United States
Environmental Prelection
Municipal £nv IronrnantaJ
Laboratory
C»nc inridti OH 4S288
Ci'A, I»HO. 2 7H 133
ami Development
Dry-Weather
Deposition and
Flushing for
Combined Sewer
Overflow Pollution
Control
-------
EPA-600/2-79-133
August 1979
DRY-WEATHER DEPOSITION AND FLUSHING
FOR
COMBINED SEWER OVERFLOW POLLUTION CONTROL
by
William C. Pisano
Gerald L. Aronson
Celso S. Queiroz
Environmental Design & Planning, Inc.
Cambridge, Massachusetts 02139
Frederic C. Blanc
James C. O'Shaughnessy
Northeastern University
Department of Civil Engineering
Boston, Massachusetts 02115
Grant No. R804578
Project Officers
Richard Field
Richard P. Traver
Storm and Combined Sewer Section
Wastewater Research Division
Municipal Environmental Research Laboratory (Cincinnati)
Edison, New Jersey 08817 -
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
.DISCLAIMER
This report has been reviewed by the Municipal Environmental
Research Laboratory, U.S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessarily reflect
the views and policies of the U.S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or
recommendation for use.
ii
-------
FOREWORD
The Environmental Protection Agency was created because of
increasing public and government concern about the dangers of pollution to
the health and welfare of the American people. Noxious air, foul water, and
spoiled land are tragic testimony to the deterioration of our natural
environment. The complexity of that environment and the interplay between
its components require a concentrated and integrated attack on the problem.
Research and development is that necessary first step in problem
solution and it involves defining the problem, measuring its impact, and
searching for solutions. The Municipal Environmental Research Laboratory
developes new and improved technology and systems for the prevention,
treatment, and management of wastewater and solid and hazardous waste
pollutant discharges from municipal and community sources, for the preserva-
tion and treatment for public drinking water supplies and to minimize the
adverse economic, social, health, and aesthetic effects of pollution. This
publication is one of the products of that research, a most vital communica-
tions link between the researcher and the user community.
The deleterious effects of storm sewer discharges and combined
sewer overflows upon the nation's waterways have become of increasing
concern in recent times. Efforts to alleviate the problem depend in part
upon the development of integrated technologies involving non-structural
best management practices with structural storage and treatment concepts.
This report presents the summary results of a two year field-
oriented data collection effort aimed at assessing the feasibility and
effectiveness of flushing small diameter combined sewer laterals. Manual
methods using a flush tanker were used to effectively remove pollutants
that deposit during dry weather periods. These deposits containing sub-
stantial organic, nutrient and heavy metal pollutants would otherwise be
suspended during wet weather periods and overflow into our nation's water-
ways. The world's first automated sewer flushing module was designed,
fabricated, installed and successfully operated yielding comparable
pollutant removal effectiveness as manual flushing.
Francis T. Mayo, Director
Municipal Environmental Research
Laboratory
-------
ABSTRACT
This report summarizes the results of a two year study aimed at
addressing the feasibility, cost-effectiveness and ease of application of
upstream solids control as an integral part of overall combined sewer manage-
ment. The project was functionally divided into four major phases. The
first three phases were intensive field engineering investigations, while the
fourth phase was relegated to data reduction and desk-top analytical efforts.
In the first phase of field work, four test segments on different
streets in the Boston sewerage system were field flushed over an extended
period using different flushing methods. External sources of fresh water,
as well as sewage, were used. The experiments were aimed at quantifying the
effectiveness of flushing deposition accumulations from a single pipe seg-
ment on a routine basis, as well as roughly estimating deposition character-
istics within collection system laterals. Removals of 75 to 90% for grit,
organic and nutrient contaminants can be expected for single manhole to man-
hole small diameter combined and separated sewer laterals. All flushing
methods yielded comparable flushing pollutant removals. The most effective
flushing method was an application of about 50 cubic feet (1.42 cubic meters)
of water, injected at discharge rates exceeding 0.5 cfs (14.4 liters per
second).
The second phase of field work was concerned with the problem of
flushing a long flat stretch of combined sewer laterals. Flushes were in-
jected into the uppermost manhole and pollutant levels in the flush wave
passing three downstream manholes were monitored. Work was divided into two
subphases. Initially, pollutant removals over the three segments were deter-
mined for different flushing conditions established in the first manhole,
providing insights into flushing effectiveness over three segments of pipe.
The results of these experiments indicated that a single flush at the upper
end of the street was reasonably effective in removing most of the deposited
load along the 675-foot (206 m.) stretch of 12-inch (30.5 cm.) combined
sewer lateral. Next, settleability tests were performed for the purpose of
crudely extrapolating how far beyond the flushing monitoring manholes would
the materials be carried. The experiments showed that heavier grit fractions
would quickly resettle, leaving the lighter solid fractions in suspension.
Roughly 20 to 30% of suspended solids and about half of the BOD and nutrient
loads would remain in suspension after 30 minutes of settling time. Analysis
of the heavy metals results from the settleability experiments indicated that
about 20 to 40% of the heavy metals would not settle within two hours of
settling.
In the final phase of field operation, an automatic sewer flushing
module was designed, fabricated, installed and operated on a single segment
iv
-------
for an extended period. Flushed pollutant loads were determined for seven
flushing events, and are comparable to removals noted in the first phase of
work, where flushing was accomplished by manual means using a flush truck.
The purpose of this work was to begin to develop operational experience using
automated flushing equipment. The state-of-the-art with respect to opera-
tional automated flushing methods, equipment and sensing interfaces has not
been fully demonstrated at this point in time. The effort in this study is
viewed as a pilot prototype investigation.
In the fourth phase, various predictive deposition loading and
flushing criteria were generated from the large data base developed during
the field programs. These formalisms allow for scanning of large-scale
sewer systems to identify problem pipes with respect to deposition. The
refined tools will allow for comparative analysis of upstream solids control
vs selected structural options to compare program efficiencies.
This work was submitted in fulfillment of Grant No. R804578 by
Northeastern University, under the joint sponsorship of the U.S. Environmen-
tal Protection Agency and the Division of Water Pollution Control, State
of Massachusetts. This report covers the period July, 1976 to February, 1979,
and work was completed February, 1979.
-------
CONTENTS
Disclaimer ii
Foreword iii
Abstract iv
Figures xi
Tables xvi
Abbreviations and Symbols xxii
Acknowledgement xxv
1. Introduction 1
1.1 Conceptual Overview of Project 4
1.2 Synopsis of Work Completed 7
1.3 Report Format 9
2. Conclusions -. 11
A. General Overview 11
B. Technical Flushing Removal Conclusions 12
G. Equipment & Methodological Conclusion 15
3. Recommendations 17
4. Details of Test Segments 23
4.1 Foreword 23
4.2 Selection of Test Segments 23
4.3 Description of Test Segments 25
4.4 Pre-Cleaning Test Segments 29
5. Field Procedures and Equipment Details 34
5.1 Foreword 34
5.2 Manual Flushing Approaches 34
5.3 Details of Flush Truck and Ancillary Equipment 37
5.4 First Phase Flushing Procedures 40
5.5 Second Phase Flushing Procedures 49
5.5.1 Serial Flushing - Pollutant Removals 49
5.5.2 Serial Flushing - Settleability Analyses 52
5.6 Automated Sewer Flushing Module 54
5.6.1 Design Details of the Automated Sewer
Flushing Module 55
5.6.2 Operational Details of the Automated
Flushing Module 58
5.7 Flow Gaging Methodologies 58
5.7.1 Dry and Wet Weather Flow Gaging 58
5.7.2 Steady State Flush Wave Calibration Procedures. . . 59
5.7.3 Dye Injection Calibration 60
6. Laboratory Analyses and Procedures 64
6.1 Foreword 64
6.2 Discrete Flush Wave Samples 64
VII
-------
CONTENTS (continued)
6.3 Heavy Metals Analysis of First Phase Flow
Composited Flushing Samples 64
6.4 Analysis of Solids Scrapings 66
6.4.1 Wet Sieving Techniques 66
6.4.2 Dry Sieving Techniques 66
6.5 Settleability Experiments 67
6.5.1 Imhoff Cone Testing 67
6.5.2 Settling Column Procedures . . 68
6.6 Analytical Methods 70
6.6.1 Pretreatment of Liquid Samples for Heavy
Metals Determinations 71
6.6.2 Pretreatment of Sediment and Scraping Samples ... 71
6.6.3 Heavy Metals Determination 71
7. Computational Methods 72
7.1 Introduction 72
7.2 Raw Data Handling 73
7.3 Missing Data Fill-In Procedure 73
7.3.1 Estimation of Missing Data - First Phase 75
7.3.2 Estimation of Missing Data - Second Phase 75
7.3.3 Estimation of Missing Data - Third Phase 75
7.4 Stage Discharge (H-Q) Relationships 76
7.4.1 Definition of the Flush Input Volumes 76
7.4.2 Discharge Estimates: Manning's Equation
With Pipe Slope 80
7.4.3 Discharge Estimates: Fitted Slope to Calibra-
tion Points Using Least Squares 83
7.4.4 Discharge Estimates: Loop-Rating Curve 88
7.4.4.1 Overview of the General Methodology ... 91
7.4.4.2 General Overview Details of
Optimization Model 91
7.4-4.3 Basic Concept of Looping State -
Discharge Curve 94
7.4.4.4 Preliminary Looping Stage/Discharge
Formulations 96
7.4.4.5 Preoptimization 99
7.4.4.6 Final Form of the Stage-Discharge
Function Fj}j 100
7.4.4.7 Comparison of Computer Versus
Measure Flush Wave Rates 108
7.5 Masses of Pollutants Removed by the Flush Wave 110
8. Single Segment Flushing Results 116
8.1 Foreword 116
8.2 Typical First Phase Flushing Results 116
8.3 Solids, Organics and Nutrient Flush Removal Results . . . 119
8.4 Heavy Metals Flush Removals 152
8.5 Comparison of Flush Methods ..... 160
8.6 Phase I Sediment Characteristics 176
vm
-------
CONTENTS (continued)
9. Serial Flushing Results 187
g.l Foreword . 187
9.2 Preliminary Flush Wave Hydraulic Experiments 187
9.3 Pollutant Removal Results 188
9.4 Discussion of Sediment Characteristics 194
10. Settling Characteristics of Materials Transported
by Flush Waves 196
10.1 Foreword 196
10.2 Assessment of Initial Concentration Data 197
10.3 Presentation of Results 199
10.3.1 Heavy Metals Analysis 206
10.3.2 Correlation Analysis 211
11. Automated Sewer Flushing and Related Topics 216
11.1 Foreword 216
11.2 Results of Automated Sewer Flushing Module 216
. 11.3 Storm Event Monitoring . 221
11.4 Background Sewage Characteristics 224
12. Predictive Tools 226
12.1 Foreword 226
12.2 Downstream Transport of Solids in Suspension 226
12.2.1 Overview of Methodology 226
12.2.2 Definition of the Flush Wave Forward Velocity . . 229
12.2.3 Overflow Rates or Settling Velocities 230
12.2.4 Percent Solids Remaining Versus
Settling Velocity Curve 230
12.2.5 Final Results 231
12.2.6 Verification 234
12.3 Downstream Transport of Organics and Nutrients 236
12.3.1 Verification 238
12.4 Simplified Sewer System Deposition 239
12.4.1 General Concepts 239
12.4.2 Single Segment Deposition Model 241
12.4.3 Multi-Segment Models 241
12.5 Verification of the Deposition Model 245
12.5.1 Introduction 245
12.5.2 Model Input Data 245
12.5.3 Verification'of Results 247
13. Development of Generalized Predictive Deposition Models .... 251
13.1 Introduction 251
13.1.1 Foreword 252
13.1.2 Data Information Sources 252
13.2 Executive Overview of Methodology 252
13.3 General Methodology - Detailed Overview 253
13.3.1 Discussion of Model Variables 253
13.4 Design of Experiment 257
13.4.1 Description of Three Sewer Systems 257
13.4.2 Range of Flows 259
13.4.3 Age and Maintenance Conditions 261
IX
-------
CONTENTS (continued)
13.5 Data Preparation for the Regression Analysis 261
13.5.1 Deposition Model Results 263
13.5.2 Areas and Total Pipe Lengths 264
13.5.3 Distribution of Pipe Slopes 264
13.5.4 Average Pipe Diameter 273
13.5.5 Distribution of Solids Deposited-
by Pipe Length 275
13.5.6 Pipe Lengths Corresponding to 80% of the
Loads Deposited - Lpp 279
13.5.7 Slope Corresponding to PLn (Spp) 279
13.5.8 Slope Corresponding to PLn./4 - (Spo/4) .... 279
13.5.9 Summary of Input Requirements for
Regression Analysis 279
13.6 Regression Analysis 280
13.6.1 Regression Method 281
13.6.2 Regression Analysis 281
13.6.3 Alternative Model Selections 282
13.6.4 Effects of Age and Maintenance 284
13.6.5 Estimation of Other Pollutants Using
TS Results 285
13.7 Model Utilization 286
, 13.7.1 Estimation of Total Pipe Length. 286
13.7.2 Estimation of Mean Pipe Slope S 287
13.7.3 Distribution of Total Solids Deposited by
Pipe Segment - Determination of PLn. 287
13.7.4 Determination of Slopes Spp and Spp/4 288
13.7.5 Formula for the Average Pipe Diameter 288
13.7.6 General Description of User's Steps 289
13.8 Test Case Application of Prediction Procedures .... 291
14. Flushing Guidelines 297
14.1 Foreword 297
14.2 Management Overview 297
14.3 Conceptual Overview for Developing a
Sewer Flushing Program 299
14.4 Sewer Flushing Costs 302
14.5 Sewer Flushing Techniques - A Discussion 304
15. Assessment of Treatment Costs with Sewer Flushing 308
15.1 Foreword 308
15.2 Dry Weather Analysis 308
15.2.1 Discussion of Pertinent Wastewater 309
15.2.2 Development of Unit Operation -
Operational Cost Models 311
15.2.3 Operational Unit Operation - Cost
Model Modifications 312
15.2.4 Estimation of Additional Treatment Costs . . . 316
15.3 Wet Weather Considerations . . . 320
16. User Details 321
16.1 Foreword 321
16.2 Application of Multi-Segment Deposition Model 321
16.3 Downstream Transport of Solids in Suspension 328
References 332
-------
FIGURES
Number Page
1 Conceptual overview of project 6
2 Proposed auto flushing concept 21
3 Map of study area 24
4 Location of flushing segments 26
5 Plan map of sewer flushing test segments: Port Norfolk
and Walnut Streets 27
6 Photographs of Port Norfolk and Walnut Streets test segments ... 28
7 Plan map of sewer flushing test segments: Shepton and
Tempieton Streets ,31
8 Photographs of Shepton and Tempieton Streets test segments .... 32
9 Representative backup and release flush methods 35
10 Representative external source flush injection methods 35
11 Mechanical piping schematic diagram of manual flushing unit ... 38
12 Photograph of flush truck 39
13 Photographs of manifold delivery system and pressurization
equipment, flush truck 41
14 Photographs of flushing nozzles and inflatable sewer
stoppers 42
15 Photographs of inflatable sewer stopper and field
engineer equipped with safety gear 44
16 Photographs of the sediment scraping operation 45
17 Photographs of flush wave injections at different feed rates
on Shepton Street 46
-------
FIGURES (continued)
Number Page
18 Photograph of flush wave injection at Walnut Street 47
19 Photographs of flush wave hand sampling scoups and flush
wave grab sampling operation, Walnut Street 48
20 Schematic of second phase field operation 50
21 Settling column flush sampler 53
22 Diagram of automated sewer flushing module 56
23 Photographs .of automated sewer flushing module 57
24 Diagram of dye injection flow calibration methodology 61
25 Photographs of dye injection equipment 63
26 Overview of sample handling during phase I 65
27 Diagram of settling column 69
28 Plots of measured versus computed flush volumes - Manning's
equation (plan & profile slope) - first phase 82
29 Comparative stage-discharge curves - Port Norfolk Street 85
30 Comparative stage-discharge curves - Shepton Street 86
31 Plots of measured versus computed flush volumes - Manning's
equation (least squares slope) - first phase 89 "
32 Typical loop-rating curve showing the progress of a flood
wave 90
33 'Simplified flow chart of the optimization process 92
34 Derivation of the loop-rating curve 95
35 Typical loop rating curve, Port Norfolk Street, 2/10/77 102
36 Typical loop rating curve, Shepton Street, 10/22/76 103
37 Typical loop rating curve, Templeton Street, 10/01/76 104
38 Typical loop rating curve, Walnut Street, 10/12/76 105
39 Plots of measured versus computed flush volumes - search
optimization (first phase) . . 106
-------
FIGURES (continued)
Number Page
40 Plots of measured versus computed flush volumes - search
optimization (second phase) 107
41 Comparison of flush wave hydrographs by dye injection
measurements and hydraulic computations, using
optimization procedure 109
42 Photographs of high rate domestic waste deposition 117
43 Photographs of low rate domestic waste deposition 118
44 Plots of flush wave solids and organic pollutant concentrations
on Shepton Street, 8/18/76 120
45 Plots of flush wave nutrient pollutant concentrations 121
46 Flush wave TSS and VSS concentration plots - Port Norfolk
Street (11/1/76) and Walnut Street (11/6/76) 122
47 Flush wave TSS and VSS concentration plots - Shepton Street
(9/13/76) and Templeton Street (9/13/76) 123
48 Flush wave pollutant mass versus time plots - Port Norfolk
Street 11/1/76 124
49 Mass removal/rainfall plots - Port Norfolk Street test
segment 135
50 Mass removal/rainfall plots - Shepton Street test segment. . . . 137
51 Mass removal/rainfall plots - Templeton Street test segment. . . 139
52 Mass removal/rainfall plots - Walnut Street test segment .... 141
53 Comparative flush rate efficiency plots - Shepton and
Templeton Streets 172
54 Comparative flush rate efficiency plots - Walnut and
Port Norfolk Streets 173
55 Cumulative mass versus .volume of flush - Shepton Street
flushes (35 cf) - 174
56 Cumulative mass versus volume of flush - Shepton Street
experiments (50 cf) 175
57 Typical phase II results, plots of flush wave concentrations
- Port Norfolk Street 3/3/77 189
58 Plot of TSS remaining vs settling velocity 204
xiii
-------
FIGURES (continued)
Number Page
59 Plot of COD remaining vs settling velocity 205
60 Plot of TSS concentration vs settling velocity. , . 207
61 Plot of TSS concentration vs settling velocity 208
62 Plot of copper concentration vs settling velocity 212
63 Plot of nickel concentration vs settling velocity 213
64 Flush wave concentrations for automated sewer flushing -
Shepton Street 10/3/77 217
65 Flush wave mass removals -
Shepton Street 10/3/77 219
66 Flush volume details of third phase program - Shepton Street . . 220
67 Shepton Street - typical storm hydrograph/pollutograph 6/7/77 . . 223
68 Overview of procedure for estimation of flushed solids
transported downstream . . . . , , . . . 228
69 Plot of percent suspended solids remaining versus
settling velocity 232
70 Percent TSS, BOD and TKN remaining in suspension versus
downstream distance from flushed manhole 233
71 Schematic of collection system 244
72 General methodology of the study 254
73 Collection system pipe slope variables. , 256
74 Representation of sediment beds and pipe slope for two
age and maintenance conditions 262
75 Distribution of pipe slopes per collection system 267
76 Distribution of pipe slopes per collection system 268
77 Complementary distribution of pipe slopes
Basin 49 - Dorchester , 269
78 Complementary distribution of pipe slopes
Basin 73 - Fitchburg 270
xiv
-------
FIGURES (continued)
Number Page
79 Histogram of collection system pipe slopes 271
80 Distribution of solids deposited by pipe lengths - WRNDB system . 276
81 Distribution of solids deposited by pipe lengths -
Dorchester system 277
82 Cumulative distribution of solids deposited vs pipe length. . . . 278
83 User steps to determine total solids deposited - TS 290
84 Diagram of pertinent information system for basin 70 -
Dorchester 292
85 Sewer flushing concepts 298
86 Conceptual overview for developing a sewer flushing program . . . 300
87 Determination of the cut-off slope for a percentage of
mass deposited 302
88 Flowsheets of primary trickling filter & conventional
activated sludge treatment plants 310
89 Sewage strength versus design flow 314
90 Treatment plant utilization versus utilization factor 317
91 Schematic of collection system - Walnut Street 322
92 Schematic of second phase field operation 329
xv
-------
TABLES
Number Page
1 Outline of Report 9
2 Description of Flushing Segment Characteristics Before Experiments 30
3 Analytical Parameters Measured for the Different Column and
Imhoff Cone Tests, Second Phase Program 67
4 Sample of Data Cards for the Flush of 10/4/76 at Walnut Street ... 74
5 Characteristics of the Tail of the Flush Wave and Flush Wave
Velocity 79
6 Estimated Ranges of Potential Biases to the Metered Flush
Volumes (cf) 80
7 Comparative Objective Functional Values for Alternative Flow
Computational Approaches 98
8 Overview of Estimated Segment Pipe Slopes 101
9 Comparison of Measured Versus Optimized Flush Wave Hydraulic
Characteristics, Dye Injection Experiments 108
10 Analytical Results, Walnut Street 8/30/76 ' 112
11 Pollutant Masses Flushed, Walnut Street 8/30/76 - 113
12 Cumulative Pollutant Masses Flushed, Walnut Street 8/30/76 114
13 Cumulative Percentiles of Pollutant Masses Flushed, Walnut
Street 8/30/76 115
14 Phase I - Field Flushing Pollutant Removals - Port Norfolk Street. . 125
15 Phase I - Field Flushing Pollutant Removals - Shepton Street .... 126
16 Phase I - Field Flushing Pollutant Removals - Tempieton Street ... 127
17 Phase I - Field Flushing Pollutant Removals - Walnut Street 128
18 Summary of First Phase Flushing Event Characteristics. 130
xvi
-------
TABLES (continued)
Number Page
19 Statistical Summary of Phase I Sewer Flushing Results
Mean and Standard Deviation of Total Mass Removals - Port
Norfolk Street 131
20 Statistical Summary of Phase I Sewer Flushing Results
Mean and Standard Deviation of Total Mass Removals -
Shepton Street 132
21 Statistical Summary of Phase I Sewer Flushing Results
Mean and Standard Deviation of Total Mass Removals -
Templeton Street 133
22 Statistical Summary of Phase I Sewer Flushing Results
Mean and Standard Deviation of Total Mass Removals -
Walnut Street 134
23 Statistical Summary of Phase I Sewer Flushing Results
Mean and Standard Deviation of Mass Removals Normalized by
Antecedent Days Between Flushes - Port Norfolk Street
Test Segment 143
24 Statistical Summary of Phase I Sewer Flushing Results
Mean and Standard Deviation of Mass Removals Normalized by
Antecedent Days Between Flushes - Shepton Street Test Segment. . 145
25 Statistical Summary of Phase I Sewer Flushing Results
Mean and Standard Deviation of Mass Removals Normalized by
Antecendent Days Between Flushes - Templeton Street Test
Segment 147
26 Statistical Summary of Phase I Sewer Flushing Results
Mean and Standard Deviation of Mass Removals Normalized by
Antecedent Days Between Flushes - Walnut Street Test Segment . . 149
27 Summary of Average Phase I Flushing Pollutant Removals
Normalized by Antecedent Days Between Flushes and by
Estimated Tributary Population 153
28 Phase I Field Flushing Heavy Metals Removals Per Unit Mass
of Solids Flushes - Port Norfolk Street Test Segment 155
29 Phase I Field Flushing Heavy Metals Removals Per Unit Mass
of Solids Flushed - Shepton Street Test Segment 156
30 Phase I Field Flushing Heavy Metals Removals Per Unit Mass
of Solids Flushes - Templeton Street Test Segment 157
31 Phase I Field Flushing Heavy Metals Removals Per Unit Mass
of Solids Flushes - Walnut Street Test Segment 158
xvii
-------
, L TABLES (continued)
Number Page
32 Summary of Phase I Heavy Metals Mass Loadings Per Unit
Mass of Solids Flushed 159
33 Statistical Summary of Phase I Sewer Flushing Heavy Metals
Results - Mean and Standard Deviation of Total Mass
Removals (Flush Solids Fraction) - Port Norfolk Street
Test Segment 162
34 Statistical Summary of Phase I Sewer Flushing Heavy Metals
Results - Mean and Standard Deviation of Total Mass
Removals (Flush Solids Fraction) - Shepton Street Test Segment . 163
35 Statistical Summary of Phase I Sewer Flushing Heavy Metals
Results - Mean and Standard Diviation of Total Mass
Removals (Flush Solids Fraction) - Tempieton Street
Test Segment 164
36 Statistical Summary of Phase I Sewer Flushing Heavy Metals
Results - Mean and Standard Deviation of Total Mass
Removals (Flush Solids Fraction) - Walnut Street Test Segment. . 165
37 Statistical Summary of Phase I Sewer Flushing Heavy Metals
Results - Mean and Standard Deviation of Mass Removals
Normalized by Antecedent Days Between Flushes (Flush Solids
Fraction) - Port Norfolk Street Test Segment 166
38 Statistical Summary of Phase I Sewer Flushing Heavy Metals
Results - Mean and Standard Deviation of Mass Removals
Normalized by Antecedent Days Between Flushes (Flush Solids
Fraction) - Shepton Street Test Segment 167
39 Statistical Summary of Phase I Sewer Flushing Heavy Metals
Results - Mean and Standard Deviation of Mass Removals
Normalized by Antecedent Days Between Flushes (Flush Solids
Fraction) - Tempi eton Street Test Segment 168
40 Statistical Summary of Phase I Sewer Flushing Heavy Metals
Results - Mean and Standard Deviation of Mass Removals
Normalized by Antecedent Days Between Flushes (Flush Solids
Fraction) - Walnut Street Test Segment 169
41 Summary of Average Phase I Heavy Metal Mass Removals Normalized
by Antecedent Days Between Flushes and by Estimated
Tributary Population 17'0
42 Visual Observations of Deposition Characteristics
First Phase Flushes - Port Norfolk Street 177
43 Visual Observations of Deposition Characteristics
First Phase Flushes - Shepton Street 179
xvi i i
-------
TABLES (continued)
Numbe
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
r
Visual Observations of Deposition Characteristics
First Phase Flushes - Tempi etoh Street
Visual Observations of Deposition Characteristics
First Phase Flushes - Walnut Street
Representative Sediment Scraping Analysis -
First Phase Flushing Program
Results of Upstream Scrapings Prior to Flushing Phase I
Phase II - Field Flushing Pollutant Removals - Port
Nqrfolk 'Street
Phase II - Field Flushing Program - Percentages of Total Mass
Transported Per Flush at Each Sampling Manhole
Average Percentages of Pollutant Loads Removed Per Flush
for Each Pipe Segment
Initial Concentration and Percent Volatile of Initial
Concentration for All Flushes and Manholes
Column Test Results 8/25, Flush 1 /Manhole 1
Column Test Results 9/7, Flush 1 /Manhole 1
Imhoff Cone Test Results 8/25
Comparison of Heavy Metals Concentrations from Settling
Column Tests
Settling Column Results Correlation Matrix
Phase III - Automated Sewer Flushing Pollutant Removal Results . .
Summaries of Background Sewage Characteristics
Average TSS Mass and Percent Removal - Second Phase Program . . .
Comparison of Measured Versus Estimated Average TSS Mass
Removals - Port Norfolk Street (Second Phase)
Average BOD Mass and Percent Removal By Flush and By Manhole,
Second Phase Serial Flushing Program . .»
Comparison of Measured Versus Estimated Average BOD Mass
Removals - Port Norfolk Street (Second Phase)
Deposition Analysis of Idealized System
Page
181
183
1
185
186
190
192
194
198
200
201
203
209
214
218
225
235
236
238
239
243
XIX
-------
TABLES (continued)
Number ' Page
64 Estimates of Solids Waste Rates in Lb/Capita/Day 246
65 Ratios of 20-Minute Peak to Average Daily Flows 247
66 Comparison of Deposition Model Predictions of Daily TSS
Accumulation with Estimates from the Flushing Experiments . . • 250
67 Sewer Density 258
68 Population Density 260
69 Per Capita Waste Rates for Various Population Densities
and Infiltration Rates : 260
70 Per Capita Values Relative to the Density of 45 Persons/Acre. ... 260
71 Total Pipe Lengths and Areas of the Basins 265
72 Slopes Corresponding to the Intervals Shown on Figure 79 272
73 Formulas for Equivalent Circular Diameters Used in Computing
the Basin Average Diameter 273
74 Summary Data on Derived Lengths, Slopes and Pipe Diameters 274
75 Ranges, Means, and Standard Deviations of the Independent
Variables Used in the Regression 280
76 Average Values of the Ratios of Computed Loads in Deposited
Pipes Over Clean Pipes 285
77 Average Pounds of PolTutants Per Pound of TS Per Antecedent Days 286
78 Distribution of Pipe Slopes for Basin 70 293
79 Comparison of Estimated Daily Solids Deposited For Basin 70
Using Different Procedures . , 296
80 Estimated Costs of Sewer Flushing Methods 305
81 Assumptions Used in Preliminary Design of Primary, Trickling
Filter and Activated Sludge Plants 313
82 Parameters for Operation and Maintenance Cost Models -
Primary, Trickling Filter and Conventional Activated
Sludge Plants 315
xx
-------
TABLES (continued)
Number Page
83 Factors Used to Compute Applied Loadings Per Unit Process 316
84 Summary of Additional Treatment Operation and Maintenance
Costs For Primary, Trickling Filter and Activated
Sludge Plants 319
85 Example of Application of Deposition Model 323
86 Comparison of Measured vs Estimated Average TSS Mass Removals -
Port Norfolk Street (Second Phase) 330
xxi
-------
LIST OF ABBREVIATIONS AND SYMBOLS
ABBREVIATIONS
CDF
cf
cfs
cm
DMH
EMT
fpm
fps
ft
ft2
gpad
gpcd
gpd
gpd/sf
ha
ID
kg
km
1
1/m
Ib/cap/day
Ib/day
log
m
mg
mg/1
mi
ml
mm
NPDES
ppb
psf
psi
psig
cumulative distribution
function
cubic feet
cubic feet per second
centimeter
downstream manhole
electrical metal tubing
feet per minute
feet per second
feet
square feet
gall on per acre per day
gallons per capita
per day
gallons per day
gallons per day per
square foot
hectare
inner diameter
kilograms
kilometer
liter
liters per minute
pounds per capita per
day
pounds per day
logarithm
meter
milligram
milligram per liter
mile
mi Hi liter
millimeter
National Pollution
Discharge Elimination
System
part per billion
pounds per square foot
pounds per square inch
pounds per square inch-
gauge
pvc
S.D.
WRNDB
SYMBOLS
ai
A
A
BOD
BOD
COD
IF
Di
SETT
F(hi/D)
FC
Polyvinyl chloride
standard deviation
sewerage system within the
area covering portions of
West Roxbury, Dedham,
Newton and Brook!ine in
metropolitan Boston.
microgram
- Flow cross-sectional area
(ft2)
- Flow cross-sectional area,
ith measurement
- Area of collection system
(acres)
- Flow cross-sectional area
(ft2) . 2
- Flow surface area (ft )
- Set of flushes performed
at one pipe segment dur-
ing a given phase
- Biochemical Oxygen Demand
(5 day)
- Fraction of BOD removed by
settling
- Chemical Oxygen Demand
- Mean pipe diameter of a
collection system, (in)
- Pipe diameter of sewer
segment i, (in)
- Base of the natural loga-
rithms
- Euclidean n dimensional
space
- Function of flow depth i
relative to diameter, D
- Fecal coliform
- Complex stage/discharge
function
xxn
-------
max
L
L1
LPD
-PM
m
n
ni
N
NH-,
OF? )
OP
OR
P
P
P(a)
PL
PL
D
Indicates the cumulative
probability of a value s of
the pipe slopes
Indicates complementary
cumulative probability
distribution
Water depth (ft)
jth flush wave depth,
flush i
Background flow depth
Maximum depth of flow
during flush
Length downstream from
point of flush (ft)
Total length of the
collection system (ft)
Length of sewer segment i
Estimated length of pipe
over which 80% of the total
loads deposit in the
collection system
Estimated length of pipe
over which the percentage
PM of the total loads
deposit in the collection
system
Summation index
Manning's roughness
coefficient
Manning's roughness
coefficient at full pipe
Manning's roughness
coefficient for ith
measurement
Nitrogen
Ammonia
Objective functional
Ortho Phosphate ?
Overflow rate (gpd/ft )
Particle size (mm)
Percent TSS remaining in
suspension
Indicates the probability
of a
Percentage of pipe length
corresponding to a per-
centage of PM of the loads
depositing in the collec-
tion system
Percentage of pipe length
corresponding to 80% of
PMD/4
pp
q
QAV
QMAX
ri
R
s
s
S
So
S
*
S
Si
^6
DP
S"p[
S
PD/4
TC
TKN
TKN
SETT"
the loads depositing in the
collection system
- One fourth of PLo
- Any given percentage of the
solids deposited in a
collection system
- Population in service area
- Discharge per capita, includ-
ing infiltration, (gpcd)
- Flow rate (cfs)
- Flow rate corresponding to
jth sample of ith flush
- Average daily dry weather
flow, (cfs)
- Peak daily dry weather flow
(cfs)
- Hydraulic radius (ft)
- Hydraulic radius of ith
measurement
- Multiple regression coeffi-
cient on the regression
analysis
- Portion of the total variar
tion about the mean (pre-
dicted by the regression
equation) which is explained
by the regression
- Mean pipe slope of the
collection system
- A particular value of pipe
slope
- Energy slope
- Pipe slope
- Estimated energy slope
- Starting value of slope, S,
in computations
- Slope of sewer segment i
- Mean ground slope
- Slope corresponding to P,
in the CDF of the pipe
slopes
- Average of pipe slopes below
SpD in the CDF
- One fourth of SpD
- Slope corresponding to PL in
the CDF of the pipe slopes
- Total coliform
- Total Kjeldhal Nitrogen
Fraction of TKN removed by
settling
LD
xxiii
-------
TP
TS
TS
a-b
TSS
TSS
V
V,
SETT
V
ETT
Vml
VSS
X
Zi
ZSi
a
5
P
P<
- Total Phosphorous
- Indicates the total mass of
solids that deposit in
collection system (Ib/day)
- Indicates the total mass of
solids that deposit in the
collection system, assum-
ing pipe bottom sediment
varying from a to b (inches)
- Total Suspended Solids
- Fraction of TSS removed by
settling
- Average foreword velocity
(fps)
- Forward wave velocity at
distance L (fps)
- Settling velocity (fps)
- Estimated flush volume i
- Measured input volume of
flush i
- Volatile Suspended Solids
- Major dimension of non-
circular pipe
- Minor dimensions of non-
circular pipe
- Percentage daily solids
deposition rate in pipe
segment i
- Objective functional value
at kth iteration
- Amount of daily dry weather
sewage solids input along
pipe segment i
- Slope of regression line
- Pre-set objective function-
al tolerance for terminating
computation
- Specific weight of water
- Specific weight of solids
- Fluid 'shear stress
- Critical shear stress
- jth time interval
NOTATION
- Summation
- For all values
- Contained within the set
CHEMICAL SYMBOLS
Cd
Cr
Cu
Hg
Ni
Pb
Zn
Cadmium
'Chromium
Copper
Mercury
Nickel
Lead
Zinc
xxiv
-------
ACKNOWLEDGMENTS
The authors would like to commend so many of the staff of EDP and
the four graduate students funded by this grant who worked so hard to make
this program a success. This study, through the dilligent efforts of many,
collected the largest data base of its kind available, through conditions of
severe weather, workload and stress.
We would like to particularly express appreciation to members of the
EDP field crew, including Paul Soper, Paul Williams, Karlson Greene and David
Walsh, for their many hours of difficult times endured. Appreciation is also
expressed to members of the EDP analysis team, particularly David Watson for
his many hours of toil in data assembly. Certainly vital to the study, and
to be commended for their efforts, are the four graduate students, Conrad
Desrosiers, Robert Noonan, Kevin McBrien and Anthony Inniss, for their many
hours spent collecting and analyzing the thousands of samples taken as part
of this effort.
Finally, last but certainly not least, we would like to express our
sincerest appreciation to Nora W. Scanlan for her patience in typing the
bulk of this report.
Northeastern University was the recipient of the combined grants
from the U.S. Environmental Protection Agency and the Division of Water
Pollution control, State of Massachusetts. Drs. Frederic C. Blanc, James
C. O'Shaughnessy and Prof. Richard J. Scranton were the project officers
for Northeastern University. Environmental Design S Planning, Inc. was the
major engineering subcontractor to NU for the project. Dr. William C. Pisano,
EDP, was the overall project director. Mr. Gerald L. Aronson, EDP, super-
vised all field work, including all equipment design and fabrication. Mr.
Celso Queiroz, EDP, supervised all data processing and computations for the
project. Drs. Frederic C. Blanc and James C. O'Shaughnessy supervised all
analytical laboratory efforts. The final report was prepared by EDP.
PROJECT OFFICERS/SPONSORS
Richard Field, Chief Thomas McMahon, Director
Richard P. Traver, Staff Engineer Water Resources Commission
Storm and Combined Sewer Section Massachusetts Division of Water
Wastewater Research Laboratory Pollution Control
Municipal Environmental Research Laboratory Westboro, Massachusetts
U.S. Environmental Protection Agency
Anthony Casazza, Commissioner Northeastern University
City of Boston Public Works Department Department of Civil Engineering
Boston, Massachusetts Boston, Massachusetts
xxv
-------
SECTION 1
INTRODUCTION
Control of combined and storm sewer overflow is a problem of increas-
ing importance in the field of water quality management. The control of com-
bined sewer overflows employing structural measures such as sewer separation,
storage and treatment have been used for a number of major cities in the Unit-
ed States. Nationwide application of these techniques for the control of com-
bined sewer overflows would require expenditures critically taxing present and
forseeable future resource allocations. New strategies are needed to reduce
these costs to tolerable limits. Non-structural controls such as sewer sys-
tem upgrading and active maintenance, improved catchbasin operation, street
sweeping and sewer flushing are upstream collection system management practices
that collectively can reduce total combined sewer pollutant loadings and ac-
cordingly the costs of downstream structural controls.
From a national standpoint, the costs of implementing controls on
storm and combined sewer overflow sources of pollution are estimated to be
$102.7 bill ion.(1) Of this total, -66.5 billion would be required to control
stormwater runoff, and $36.1 billion to control combined sewer overflows.Quite
obviously, amounts of this magnitude cannot feasibly be raised. Another consi-
deration in this area of concern is that structurally-oriented solutions for
controlling combined sewer overflows generally imply periods of disruption in
major urban areas. As such, the real social costs of applying these approaches
may even exceed the current survey estimates that include only direct construc-
tion costs. The question then arises whether this is a viable tradeoff for
the attendant water quality improvements.
The answer to this dilemma is not clear. The estimated monies will
probably never be available and a more realistic view would be to spend limit-
ed dollars in maximizing the potential of existing capital outlays. The NPDES
permit program recognized that the first and most logical step in fully uti-
lizing our nation's municipal pollution control expenditures is to maximize
existing treatment plant performance via nonstructural options such as in-
creases in O&M dollars and chemical addition. (2) Similar management con-
cepts as applied to sewer collection systems may provide large savings in con-
trolling wet-weather pollution from combined sewer overflows. Sewer flushing
is a potential low-cost, non-structural control alternative which should not
be viewed as a substitutable alternative for structural control. Sewer flush-
ing can significantly reduce overall costs when integrated together with other
upstream management practices and downstream structural options as required
where necessary.
-------
The deposition of sewage solids during dry weather in combined sew-
er systems has long been recognized as a major contributor to first-flush
phenomena occurring during wet weather run-off periods. Another manifestation
of first flush, in addition to the scouring of materials already deposited in
the lines, is the first flush of loose solid particles on the urban ground
surface that are transported into the sewerage system. These particulates may
settle out in the system and be available for flushing during periods of larg-
er flows. The magnitude of these combined loadings during runoff periods has
been estimated to range up to 30 percent of the total daily dry weather sewage
loadings.
Studies in Buffalo, N.Y., have shown that 20 to 30 % of the
annual collection of domestic wastewater solids settle and eventually are dis-
charged during storms(3). Other studies have indicated similar rates of domestic
wastewater solids deposition (4, 5, 6 and 7). As a result, a large residual
sanitary pollution load over and above that normally carried is discharged
over a relatively short interval of time, often resulting in what is known as
a "first flush" phenomenon. This can produce shock loadings detrimental to
receiving water characteristics.
One of the underlying reasons for considerable sewage solids depo-
sition is the combined sewer hydraulic design. Combined sewers are sized to
convey many times the anticipated peak dry weather sewage flow. Combined
sewer laterals can carry up to 1000 times the expected background sewage flow.
Ratios of peak to average dry weather flow usually range from Z to 10 for
interceptor sewers. The oversized combined sewer segments possess substan-
tial sedimentation"potential during' dry weather periods. Dry weather flow
velocities are typically inadequate to maintain settleable solids in suspen-
sion, and substantial solids tend to accumulate in the pipes. During rain-
storms, the accumulated solids may resuspend and, because of the limited hy-
draulic capacity of the interceptor, overflow to receiving waters. Suspended
solids concentrations of several thousand parts per million are not uncommon
for combined sewer overflows.
Some simple calculations illustrate the potential impact of overflow
on the receiving waters. If twenty-five percent of the daily pollution load-
ing accumulates in the collection system, an intense rainstorm lasting two
hours after four days of antecedent dry weather may wash the equivalent of a
full day's flow of raw sewage overboard to the receiving waters. The average
antecedent dry period between storm events is about four days for many areas
of the United States, especially along the eastern seaboard. Furthermore,
one day's equivalent of raw sewage is discharged within a two-hour period,
or at twelve times the rate at which raw sewage is entering the collection
system. The shock pollution loading potential of combined sewer overflow can
be substantial. It is clear that the sewage treatment plant simply never sees
a substantial portion of the polluting materials entering a combined sewer
collection system. Furthermore, the rate at which the accumulated pollutional
loads are discharged to receiving waters can represent shock loadings several
orders of magnitude greater than the rate at which raw sewage is being gener-
ated by the community.
The concept of sewer flushing is to either scour and transport de-
posited pollutants to the sewage treatment facility during dry weather and/or
-------
to displace solids deposited in the upper reaches of large collection systems
closer to the system outlet. The idea is either to reduce depositing pollutants
that may be resuspended and overflow during wet events and/or to decrease the
time of concentration of the solids transport within the collection system.
During wet weather events these accumulated loads may then be more quickly dis-
placed to the treatment headworks before overflows occur or would be more ef-
ficiently captured by wet weather storage facilities.
Flushing of sewer lines, although widely used around the turn of the
century as a maintenance practice, is still in its infancy in regard to being
viewed as a viable pollution control alternative for combined sewer systems.
The concept of sewer flushing is a controversial issue since it involves gen-
erally low capital first-cost investments but high operational and maintenance
costs. Federal funding mechanisms for sewerage conveyance and treatment facil-
ities presently favor high first-cost programs with low operational costs.
Federal funding does not cover operational costs. Moreover, the notion of in-
creasing the municipal commitment for greater manpower investments runs counter
to the historical trend toward decreasing public works' budgets in the area of
sewerage system management. Another situation compounding the problem is that
municipalities in many cases have one department involved with sewage col-
lection and another for treatment and sewer flushing is collection system ori-
ented. In another vein, little applied research has been performed to develop
and quantify criteria for estimating deposition loadings and for flushing sew-
ers. These.criteria are necessary to quantify the need for and. the extent of po-
tential sewer flushing management programs. As a consequence, planners were
heretofore reluctant to investigate flushing as a pollution control alternative
in the context of overall combined sewer management. Recently, however, Con-
gress has mandated that all alternative forms of wet weather pollution control
with emphasis on the non-structural Best Management Practices (BMP) such as
sewer flushing and street sweeping, be thoroughly considered in any new com-
bined sewer management facility planning effort (8).
The tasks of identifying those portions of a sewerage system where
deposition may occur and developing control policies to eliminate these con-
ditions are indeed non-trivial. There has been simply no quantitative in-
formation available on the locations of depositing materials, their character-
istics, or the hydraulic conditions minimally necessary to dislodge and trans-
port them. The literature is sparse in this regard. Inaba (9) reported that
deposition in a combined sewer district in Tokyo was limited to pipe diameters
less than 21 inches (0.53 m), and that greater volumes of depositions were
found in smaller diameter pipes. In 1898 Ogden (10) reported on flushing ex-
periments conducted in Ithaca, New York.
The Storm and Combined Sewer Research Program, Federal Water Pol-
lution Control Administration (now U.S. Environmental Protection Agency), ini-
tiated research efforts in 1966 through a contract with the FMC Corporation,
Santa Clara, California, to demonstrate the feasibility of reducing pollution
from combined sewer overflows by means of periodic flushing during dry weather.
It was contemplated at that time to have three phases of work in this area.
The first phase included a study of the overall flushing concept, small-scale
hydraulic modeling, and design and development of cost estimates for construc-
ting test equipment. The results of this work appeared in a final report en-
titled "Feasibility of a Periodic Flushing System for Combined Sewer
-------
Cleansing" (11), and set the stage for the second phase, which allowed the
effectiveness of flushing under various conditions be be determined.
This second phase (another contract to FMC) was completed in 1972
and the work is described in a final report entitled "A Flushing System for
Combined Sewer Cleansing" (12). This work produced a flushing evaluation faci-
lity at FMC, consisting of 12 and 18 inch diameter test sewers about 1600 feet
(488 m) long, supported above ground, thus allowing slope adjustments, and in-
cluding holding tanks at three points along the test sewers for the flushing
experiments. This phase of work developed limited experience in periodic flush-
ing of simulated combined sewer laterals within a limited size range (12 and
18 inches diameters). The report recommended that further studies (the third
phase) be made for flushing of larger sizes of pipe, of wave sequencing, and
of solids buildup over longer time periods. It was also suggested that a
demonstration in an operating combined sewer system be performed.
In 1974 a combined sewer management study aimed at assessing alter-
native strategies for the abatement of combined sewer overflows discharging to
portions of Boston Harbor was completed (4,13). As part of the research work
conducted during that study, a number of theoretical formalisms for prediction
of dry weather deposition and flushing criteria for sewers were developed. The
development of the deposition analysis techniques stemmed from critical shear
stress considerations. The theoretical formalisms developed were roughly
checked in the field using visual inspection techniques to assess solids ac-
cumulation. The results of that anlysis although admittedly crude, were en-
couraging. This model was used to analyze deposition problem segments within
a service area of 3000 acres (1215 ha) entailing roughly 0.5 million lineal
feet (152 km) of sewer. Roughly 3000 manhole to manhole segments were ana-
lyzed for deposition loadings and it was determined that roughly 17% of the
segments contained about 75% of the estimated daily dry weather sewage depo-
sitions. It turned out that most of these segments were small diameter com-
bined sewer laterals. Flushing criteria were empirically developed using data
generated by FMC (11,12) and then were used to estimate flushing volumes.
The research results reported in this document can be considered as
the envisioned third phase of the 'two FMC studies. Much of the theoretical
work of the aforementioned Boston study was used as the starting point for
this research.
1.1 Conceptual Overview of Project
The solids control demonstration/research program was developed to
address many of the issues relating to the feasibility, cost-effectiveness,
and ease of application of upstream solids' control program as an integral
part of overall combined sewer management. Basically, there are five funda-
mental issues that must be answered before widespread acceptance of upstream
solids control may be considered. The issues include: 1) what are the best
flushing methods to use for a given situation; 2) what is the expected pol-
lutant removal efficiency associated with the various methods; 3) what are
the costs associated with such programs; 4) how do you screen large systems
to identify problem pipes with respect to deposition; and 5) what are the
-------
effects on stormwater runoff of such a strategy as applied to combined sewer
systems.
Research Objectives
1. Test the feasibility of applying various solids control tech-
niques as a method of deposition control in combined and sani-
tary sewer lines on test segments in the Boston sewer system.
2. Monitor deposition rates on a number of test segments.
3. Monitor pollutant removals including solids, organics and nutri-
ents associated with the various solids control techniques.
4. Assess pollution oriented characteristics of both the flushed
and remaining materials versus maintenance problems of grit,
sand and gravel accumulations.
5. Recommend most favorable solids control techniques for opera-
tional testing by both automated and manual means.
6. Develop, test and evaluate automated control system in a field
operational testing program.
7. Develop, test and evaluate manual sewer flushing techniques uti-
lizing specially equipped water tankers in a field operational
testing program.
8. Assess the operational feasibility and performance of flushing
both long and short upstream collection segments.
9. Assess the effects of stormwater washoff on the characterization
of combined sewer solids.
10. Refine existing deposition model and flushing criteria.
11. Develop user guideline for solids control program as an inte-
gral part of sewer management schemes.
Figure 1 is an overview schematic of the effort. The program was
broken into three distinct components: 1) a field feasibility analysis of
various manual flushing techniques to test the feasibility of applying various
techniques to single manhole to manhole sewer segments; 2) an operational
testing program to assess serial flushing effectiveness using manual methods
over a long combined sewer lateral and to assess effectiveness of an automated
flushing; and 3) a detailed data analysis and costing phase to develop a
reasonable deposition model and flushing criteria, and analyze the concept of
upstream solids control as an integral part of combined sewer abatement
schemes.
The feasibility analysis was aimed at answering the question of
what are typical deposition rates in sewerage collection systems, what are
-------
Feasibility Analysis
Operational Testing
Detailed Data Analysis
and Costing
CTV
In situ comparative
analysis of flushing
methods
Characterization of
deposition rates,
flushing efficiency
and pollutant reduction
for each method
Characterization of
particle distribution
of flushed materials
and remaining sediments
First step in charac-
terization of the effects
of stormwater flow on
deposition characteristics
In situ operational
analysis of most favor-
able techniques
Analysis of multi-
variate (network) case
Continued characteriza-
tion of the effects of
stormwater flow on
deposition characteristics
Refinement of exist-
ing deposition model
and flushing criteria
Analysis of expected pollu-
tant reductions
Development of predictive
procedures for estimating
deposition loadings
Development of procedures
for establishing flushing
program
Development of program costs
and user guidelines
FIGURE 1: CONCEPTUAL OVERVIEW OF PROJECT
-------
the best flushing techniques to use, and what pollutant reductions can be
reasonably expected as well as supplying a wealth of data for the refinement
of the existing deposition model and flushing criteria. In the operational
testing subprogram the more promising strategies developed in the feasibility
analysis were utilized in a field program aimed at continuing data development
as well as testing the operational feasibility of such a program by both man-
ual and automated means.
From the large data base developed during the two field programs,
procedures for estimating collection system deposition leadings and flushing
criteria were generated. These formalisms allow for scanning of large-scale
sewer systems to identify problem pipes with respect to deposition.
1.2 Synopsis of Work Completed
The work performed can be functionally divided into four major pha-
ses. The first three phases were intensive field engineering investigations
while the fourth phase was relegated to data reduction and desk-top analytical
efforts.
In the first phase of field work, four small diameter (12 and 15
inch, or 0.31 and 0.39 m) test segments on different streets in the Boston sew-
erage system were field flushed over an extended period using different flushing
methods. External sources of fresh water, as well as sewage, were used. The
experiments were aimed at quantifying the effectiveness of flushing deposi-
tion accumulations from a single pipe segment on a routine basis, as well as
roughly estimating deposition characteristics within collection system laterals.
The second phase of field work was concerned with the problem of
flushing a long flat stretch of a 12-inch sewer lateral. The street contains
five manholes and is roughly 1000 feet (305 m) in length. Flushes were injected
into the uppermost manhole and pollutant levels in the flush wave passing three
downstream manholes were monitored. Work was divided into two subphases. Ini-
tially, pollutant removals over the three segments were determined for dif-
ferent flushing conditions established in the first manhole. These results-
provided insights into flushing effectiveness over three segments of pipe.
Next, settleability tests were performed on samples taken from flushes con-
ducted in a similar manner for the purpose of crudely extrapolating how far
beyond the flushing monitoring manholes would the materials be carried.
In the final phase of field operation, an automatic sewer flushing
module was designed, fabricated, installed and operated on a single segment
for an extended period. The purpose of this work was to begin to develop
operational experience using automated flushing equipment. The effort in this
study should be viewed as a pilot prototype investigation.
In the fourth phase, various predictive deposition loading and
flushing criteria were generated from the large data base developed during the
field programs. These formalisms allow for scanning of large-scale sewer sys-
tems to identify problem pipes with respect to deposition. Simplified desk-
top procedures were also prepared for assessing the magnitude of deposition
loadings within combined sewer collection systems .and for quickly establishing
-------
the extent of flushing programs. The refined tools permit comparative analy-
sis of upstream solids control vs selected structural options to compare pro-
gram efficiencies.
l
Accomplishments
I. First Phase Field Program
86 Flushing experiments with samples analyzed;
- 5600 Analytical determinations (solid, organic, nutrient
and bacterial levels) of flush wave samples;
150 Physical analyses of sediment scrapings;
600 Heavy metals determinations of flush wave samples
and numerous flow measurements;
II. Second Phase Field Programs
36 Serial flushing experiments with samples analyzed
for pollutant levels;
10 Serial flushing experiments - flush wave characteris-
tics noted;
10 Flush wave discharge experiments using florimetric
methods;
18 Settleability experiments;
- 5000 Analytical determinations (solids, organics, nutrients
and heavy metals) of flush wave samples;
100 Physical analyses of sediment scrapings.
III. Third Phase Field Programs
1 Automated Sewer Flushing Module designed, fabricated,
installed and operated;
,7 Flushes sampled;
3 Storm events sampled;
500 Analytical determinations of flush wave samples.
IV. Analysis Phase
- All results of flushing experiments data processed and
computer files prepared;
- Optimization procedure developed to compute flush wave
discharge from stage information;
- Pollutant masses computed for all experiments;
- Simplified procedures developed'for estimating dry weather
deposition within collection systems;
- Cost analysis of flushing program impacts on sewage
treatment; and
- General guidelines and cost information for sewer flushing.
-------
1.3 Report Format
This report contains sixteen chapters. Summary conclusions and
recommendations are presented in Chapters 2 and 3. Examples of procedures
used and developed throughout the report are given in Chapter 16. The remain-
ing chapters can be classified under three major headings, including:
a) experimental methodologies and procedures, b) reporting of experimental
results, and c) analysis and analytical desk top extension of field results
into user guidelines. Table 1 shows the three major groupings and the rele-
vant chapters.
TABLE 1
OUTLINE OF REPORT
A. Experimental Methodologies and Procedures
Chapter Title
4 Details of Test Segments
5 Field Procedures and Equipment Details
6 Analytical Laboratory Procedures
7 Data Processing and Computational Methods
B. Experimental Findin_gs_
Chapter Title
8 Single Segment Flushing Results
9 Serial Segment Flushing Results
10 Settleability Testing Results
11 Automative Flushing Results and Related Topics
C. Analysis and Synthesis
Chapter Title
12 Predictive Tools
13 Simplified Procedures For Estimating
Deposition Loadings
14 Flushing Guidelines
15 . Assessment of Treatment Costs With
Sewer Flushing
The first major grouping includes all the methodological procedures
used to gather experimental data. Pertinent information describing the selec-
tion of test segments, their characteristics and pre-cleaning procedures are
described in Chapter 4. All field flushing procedures and details of various
-------
equipment fabricated for use in the project are described in Chapter 5. Analy-
tical laboratory procedures and techniques used to determine pollutant charac-
teristics of the materials flushed during the various experimental programs
are presented in Chapter 6. Data processing details and computational proce-
dures used to compute flush wave hydraulic characteristics and estimates of
the flushed pollutant loadings are given in Chapter 7.
All experimental flushing results are presented in the second por-
tion of the report. In Chapter 8 tabulations and statistical summaries of
pollutant loadings flushed during the first phase of the field program are
given. Serial flushing results from the second phase are given in Chapter 9.
Summary findings of the settleability experiments conducted during the serial
flushing program are presented in Chapter 10. These results are useful in
extrapolating flushing effectiveness for pipe lengths longer than the test
segments. Automated module flushing results are given in Chapter 11. An
experiment is also described where flushing was immediately conducted follow-
a wet weather runoff event. Pollutant loadings transported during the storm
event were monitored along with flushed loadings permitting comparative
assessment of runoff versus flushed pollutant loadings. The computational
procedures described in Chapter 7 were used to compute the flushed pollutant
loads reported in Chapters 8 through 11, using raw field and analytical lab-
oratory data. All background sewage sampling conducted during the project is
summarized in Chapter 11.
The final portion of this report contains the analysis and synthesis
of the field data into user guidelines and tools. Chapter 12, entitled
"Predictive Tools", presents the details of an existing multi-segment predic-
tive deposition model (4) and calibration efforts using field flushing results
described in Chapters 8 and 9. This chapter also contains an empirical proce-
dure useful in extrapolating flushing effectiveness for segment lengths in ex-
cess of 1000 feet (305 m). This procedure was prepared using the settling column
findings presented in Chapter 10, and the results of the serial segment flush-
ing program described in Chapter 9. Simplified procedures for estimating
daily deposition loadings within collection systems are given in Chapter 13.
Flushing guidelines and costs are presented in Chapter 14. The last chapter
presents a comparative cost assessment of the impacts of flushing on dry
weather sewage treatment costs and a rudimentary analysis of the impact of
dry weather sewer flushing on wet weather storage and treatment combined sewer
overflow abatement programs.
10
-------
' SECTION 2
CONCLUSIONS
Conclusions derived from this investigation are divided into three
broad categories including: general overview; technical flushing removal
conclusions; and equipment and methodological conclusions.
A. General Overview
1. Sewer flushing has been shown to be an effective means for substan-
tially reducing dry weather sewage pollutant related deposition materials in
small diameter combined sewer laterals. Removals of 75 to 90% for organic and
nutrient contaminants can be expected for single manhole to manhole segments.
Removals of 65 to 75% for organic and nutrient deposits can be expected for
serial segments up to 700 feet (213 m) and roughly 35-45% removals are pro-
jected for segment lengths greater than 1000 feet (305 m).
2. Sewer flushing .is a reasonably effective means for reducing dry wea-
ther sewage grit/inorganic related deposition materials in small diameter
combined sewer laterals, with removals of about 75% for single manhole to
manhole flushing segments. Removals of 55-65% for solids can be expected for
serial segments up to 700 feet (213 m) and roughly 18 to 25% removals are pro-
jected for segments lengths greater than 1000 feet (305 m).
3. Sewer flushing is a most effective means for suspending and trans-
porting great distances heavy metals associated with light colloidal solids
particles. Approximately 20-40% of heavy metals contained within sewage sedi-
ments would be transported by flush waves at least 1000 feet (305 m) and pro-
bably much further, including cadmium, chromium, copper, lead, nickel and zinc.
4. Extensive field experience has indicated that sewer flushing by manual
means (water tank truck) is a simple reliable method of combined sewer solids
control for smaller diameter laterals and trunk sewers.
5. Recommendations 1-4 are based on extensive field flushing experience
using manual flush methods with a water tanker. An average of 300 gallons per
flush were used during the experiments, representing about 0.5% of the total
daily water-consumption in the area. Various flushing methods were investi-
gated, including different combinations of externally supplied flush volumes
and rates together with backup and release using sewage. All methods yielded
comparable flushing pollutant removals. The most effective flushing method
for 12-15 inch (0.31-0.39 m) lateral sewers was an application of about 50
cubic feet (1.42 cubic meters) of water injected at discharge rates exceeding
0.50 cfs (14.4 liters per second).
11
-------
6. Initial experiments with an automated sewer flushing module indicated
great operational promise with comparable removals to other techniques for
flushing. The module was designed to backup sewage with quick release for
establishing the flush wave and requires no external water supply.
7. Sewer flushing has an additional inestimatable benefit, in that the
"hands-on" presence of field crews continuously surveying the collection sys-
tems would encounter and note possible malfunctions that would otherwise go
undetected. Correction of these malfunctions, such as broken, clogged or
constricted pipes, or inoperative regulators and/or tide gates, could substan-
tially reduce potential overflow volumes and pollutant loads, and possibly
upstream flooding.
8. There were substantial sediment beds in the test segment sewer later-
als prior to commencing the flushing program. Except for the heaviest grit
particles, sediment deposits in the sewers were maintained at minimum levels
by flushing during the experimental period. Inspection of the laterals sev-
eral months after the flushing programs were terminated revealed that sedi-
ment layers returned to pre-project conditions. Sewer flushing is therefore
a viable means for minimizing grit accumulations, which can reduce hydraulic
capacity.
9. An analysis of the potential impact of nominal sewer flushing pro-
grams vs the costs of primary and conventional secondary treatment, including
solids handling, indicated that annual operational and maintenance costs
would rise about 3 to 6% depending on the type of treatment plant. Therefore,
sewer flushing would not significantly increase treatment costs.
B. Technical Flushing Removal Conclusions
1. During the first phase of operation 86 separate flushing experiments
were conducted during the period of August 30, 1976 to November 12, 1976.
Roughly 20 flushes on a 3-4 day interval were accomplished for each of 4 test
manhole to manhole segments. Three different methods of manual flushing were
performed. The first method consisted of backing up the upper end of the
flushing manhole with an inflatable rubber stopper with quick release. The
other two methods were gravity and pressurized dump discharge into the flush
manhole. Pollutant removals for the flushing experiments indicated that all
methods provide about the same degree of removal. The best method is an ex-
ternal source high volume/high rate flush. Average flush volume during this
experimentation period was 300 gallons (1.13 cubic meters). A minimum flush
volume of 225 gallons (0.85 cubic meters) is recommended for a single manhole
to manhole segment of a small diameter (12-15 inch, or 0.31-0.39 m) sewer
lateral. The periodic flushing removed the domestic sewage deposits that
accumulated between flushing events and maintained minimal levels of grit,
rock and debris.
2. The average flushing pollutant removal rates normalized by both ante-
cedent days between the flushing events, and tributary population during the
first phase program, were the following for separated sewer segments: COD =
(1.43x10 ) grams/capita/day(lbs/capita/day). Similar average results for the
12
-------
combined sewer flushing laterals were: COD = 22.0 (4.84xlO~2), BOD = 7.98
(1.76x10-2), TKN = 0.64 (1.41x10-3}, NH = 0.22 (4.84xlQ-4)5 TP = 0.14
(3.08x10-4), TSS = 19.03 (4.19x10"^, and VSS = 12.21 (2.69X10'2) grains/cap-
ita/day (Ibs/capita/day). The flushed removals on the combined sewer segments
were about two to four times the levels found on the separated sewer streets.
3. Sediment scrapings of sanitary deposits prior to flushing during the
first phase program ranged form .026 to .037 Ibs per linael foot (39 4 to 55 8
grams per meter).
4. Composited flush wave samples from the first phase were allowed to set-
tle for periods of approximately four hours prior to heavy metals analyses
Heavy metals analyzed included: nickel, chromium, cadmium, lead and mercury.
Heavy metals concentrations in the supernatant of settled flush wave were very
low. Heavy metals levels in the settled fraction were high. Average heavy
metal results for the combined sewer test segments were roughly twice those
of the separate sanitary sewered streets, indicating the impact of wet weather
street wash load.
5. During the second phase field flushing program, 6 serial flushing
experiments were performed on three flat combined sewer segments, 675 feet
(206 m) in length. Three flushes were accomplished per experiment and pollu-
tant masses were determined at each of three downstream sampling manholes per
flush. The last flush per experiment was always a maximal flush meant to com-
pletely remove any residual pollutants. The average results for all six ex-
periments indicated that most of the loads for all three segments were removed
during the first flush. Nearly 88% of the total BOD load transported by the
first sampling manhole was accomplished by the first flush. The removals
slightly decreased for two other sampling manholes further downstream. The
experiments indicate that a single flush at the upper end of the street was
reasonably effective in removing most of the deposited load along the 675
foot (206 m) stretch of 12 inch (0.31 m) combined sewer lateral.
6. Settleability tests were performed on composited flush wave samples
from the three sampling manholes in the second phase flushing operation. The
experiments showed definite shifts in suspended solids/settling velocity dis-
tribution from the first to the third downstream sampling manholes, indicating
that heavier grit fractions would quickly resettle leaving the lighter solid"
fractions in suspension. About 20 to 30 percent of suspended solids would
remain in suspension after 30 minutes of settling time. The fractions of
volatile solids relative to the suspended solids increased both with settling
time during the experiment and with the distance downstream from the flush-
ing manhole. Distribution of COD and BOD versus the settling time showed
the similar characteristics as the suspended solids settling behavior.
About half of the initial BOD levels would remain in suspension after 30
minutes of settling. Organic and nutrients concentrations correlated extreme-
ly well with both TSS concentrations and settling velocities (correlations
ranged from 0.6 to 0.8).
7. Analysis of the heavy metals results from the settleability experi-
ments indicated that about 20 to 40% of the heavy metals present in the com-
posited flush waves would not settle within two hours of settling. The bal-
ance of the metals were associated with heavier solids particles and rapidly
13
-------
settled. The metals associated with the lighter colloidal fractions could
easily be transported by flushing during dry weather to treatment facilities
or, alternatively, would be transported to receiving waters during overflow
periods. Sewer flushing is therefore a viable means for reducing an impor-
tant and significant source of heavy metals in overflows.
8. An automated sewer flushing module consisting of an oil-on-air hydrau-
lic gated device triggered by an automatic time clock was designed, fabricated,
installed and successfully operated on a daily basis for a 5 1/2 month test-
ing period. The device backed up sewage to predetermined levels and then,
retracting, induced a flush wave. Flushed pollutant loads were determined
for 7 flushing events, and are comparable to removals noted in an earlier
phase of work where flushing was accomplished by manual means using a flush
truck.
9. A special field measurement program was conducted to monitor pollu-
tant loads during storm events and then to immediately flush the segment fol-
lowing the end of the storm event. The results indicated that runoff from
slight to moderate rainfall events only partially removed fresh sanitary de-
posits, whereas the flushing removed significant organic loads. Significant
organic deposits were observed after the storm event and none after the flush-
ing event. Several inferences can be drawn from the data. The flush wave
provided the necessary turbulence to suspend, entrain and transport fresh
organic deposits. Flushing frequency intervals need not be determined by
return periods defined by slight to moderate rainfall events. "First flush"
runoff loads during intense storms can be the results of long-term sewage
solids deposition accumulations occurring over both dry and moderately wet
runoff periods.
10. Background sewage concentrations were measured at four upstream
sewer laterals over a two year period. Pollutant concentrations found in the
laterals were much higher and with far more variability than levels normally
encountered further downstream at treatment plants. This phenomenon has been
observed on numerous occasions in other locations.
11. Field flushing results from the second phase program and the analy-
tical results of the settling column tests were used to develop an empirical
model relating the percentage of flushed masses remaining in suspension at
downstream points as a result of upstream flushing. This model was favorably
compared with other flushing pollutant removal data from the serial flushing
experiments. The model indicates that at least 20% of flushed solids would be
transported at least 1000 feet (305 m) from a point of flush. Similar esti-
mates for organic'and nutrient flushed loadings are 45 to 50% for the same
distance. Most combined sewer laterals do not exceed this distance and may
discharge into trunk sewers with high shear stress characteristics, that
is, good solids carrying capacity. These results imply that sewer flushing of
combined sewer laterals could result in significant reductions of dry weather
deposits containing pollutant related contaminants.
12. An existing generalized procedure for estimating daily dry weather
sewage solids deposition loadings within each manhole to~manhole segment of
an entire collection system network was roughly calibrated using field flush-
14
-------
ing pollutant removal results. This procedure is therefore recommended for
application where detailed segment-by-segment estimates of deposition rates
are desired. This procedure can only be applied when the hydraulic character-
istics, such as the pipe length, size, shape and slope, are known for each
segment.
13. A simplified methodology was prepared for providing first-cut assess-
ments of the total amounts of solids (Ib/day) that deposit in a sewerage col-
lection system; and the extent of the collection system over which the depo-
sition takes place. The complex distribution-parameter dry weather sewage
deposition model described in conclusion 12 was applied to 75 separate and
combined sewer collection systems in eastern Massachusetts to generate esti-
mates of solids deposited daily per system (Ib/day). These estimated loads
were then regressed with selected variables representing the physical charac-
teristics of these collection systems including total pipe length, service
area, average collection system pipe slope, average pipe diameter and other
more complicated variables representing various points on the lower end of
the collection system pipe slope cumulative density function. Four alterna-
tive predictive single term power functions were developed from the regression
analysis. The degree of fit of the non-linear functions to the data set were
remarkably -high. The R2 values of the alternative models ranged from 0.85
for the simplest approach requiring little external data analysis and prepara-
tion, up to 0.95 for the most complex model requiring substantial external
engineering and data reduction analyses. These simplified procedures are re-
commended for general application in combined sewer management planning.
14. In addition to the general predictive procedures for estimating
solids deposition within collection systems as a function of sewer shed char-
acteristics (as described in conclusion 13), the effects of sewer system age
and maintenance on solids deposition was simulated by considering prior sedi-
ment deposits to develop multiplicative coefficients to the predictive equa-
tions mentioned in conclusion 13.
15. The first phase field flushing results were used to develop mean
ratios between other pollutants, such as BOD, COD, TKN, TP, NHs and VSS with
suspended solids. This therefore permits the use of the predictive equations
for total solids deposited (described in conclusions 13 and 14) to be used for
the estimation of other pollutants.
16. Extensive statistical analyses of sewerage system pipe slopes in this
effort revealed'that collection system pipe slopes can be represented by an
exponential probability model. Analysis of the distribution of loads deposit-
ed versus cumulative pipe length led to the development of generalized curves
as a function of collection system mean slope for estimating the total frac-
tion of collection system pipe footage over which a given percentage of the
total loads deposit. These findings can be combined to locate segments asso-
ciated with the required fractions.
C. Equipment & Methodological Conclusions
1. A specially designed water tanker equipped with two 1000-gallon
tanks mounted on a steel I-beam skid was fabricated for delivering flush
waters under controlled discharge conditions. The tanker was equipped with a
15
-------
pneumatic system to pressurize the tanks. The operation under gravity con-
ditions provided a controlled flush release of 35 to 50 cubic feet (0.99 to
1.42 cubic meters) at a rate of 0.25 to 0.50 cfs (7.2 to 14.4 liters per
second). Under pressurized conditions the same volumetric range of flush was
accomplished at a rate of 0.5 to 1.25 cfs (14.4 to 35.4 liters per second).
The water tanker was successfully used for 300 different experiments over a
1 1/2 year period.
2. An automated sewer flushing module was designed, fabricated, in-
stalled and successfully operated for an extended period. The device deve-
loped was an air-operated gate capable of backing up sewage flows to predeter-
mined levels and then suddenly retracting, inducing a flush wave. The flush-
ing gate was controlled from a master control timer capable of pre-programmed
flushing of varying sequences with flushing intervals ranging from 6 to 72
hours. The entire unit was powered by a 12 volt automobile battery. Air
supply to the system was by means of a small high pressure air cylinder re-
quiring replenishment every 150 flushes. The device worked remarkably well
with very little down time for repairs. The device is amendable to package
fabrication and installation.
3. A special dye injection procedure was developed to provide a prac-
tical and reliable method of measuring steady state and non-steady state dis-
charge for waste streams containing suspended solids levels up to 10,000 mg/1.
Non-steady state discharge levels of flush waves were determined. The proce-
dure consisted of three distinct operational units including: a) the high
pressure injection nozzle system; b) the pumping and air separation unit;
and c) a flow-through cell equipped florimeter with recording readout. The
dye injection experiments utilized Uracine dye with a special filter system
to eliminate any background interference. The procedure is recommended for
similar difficult-to-measure waste streams.
4. A specially designed settling column and procedures were developeo
to perform settleability tests on flush wave samples. A yoke-frame installa-
tion was devised to permit axial and transverse mixing of the column before
experimentation since the settling velocities for a considerable portion of
the flush wave solids were extremely rapid. U.S. Environmental Protection
Agency is currently investigating refined adaptation of this design for
settling column analyses of combined sewer overflow samples.
5. A mathematical programming procedure was developed and utilized to
determine the parameters of a non-steady state loop-rating curve approach
for estimating flush wave discharge from recorded stage levels. The approach
minimized the variance between computed and measured flush truck volumes, and
substantially reduced the error in variance between measured and computed
flush volumes in defining parameters of stage discharge curves. The methodo-
logy can be logically extended to estimate stage/discharge rating curve para-
meters for a multi-flow measurement site system.
16
-------
SECTION 3
RECOMMENDATIONS
Results of the present study indicate that sewer flushing effective-
ness dislodges and resuspends deposited pollutants in small diameter combined
sewer laterals. Most of the flushing experiments were conducted by manual
means in which a water tanker was used to deliver the flush volumes. One
automated flushing module was fabricated and tested for a five and one-half
month period.
The study yielded a massive amount of data on a subject primarily
a source of conjecture in the past. As noted in Chapter 2, many significant
conclusions.were reached as to the effectiveness of sewer flushing as a com-
bined sewer central measure, as well as indications generated as to the source
and nature of combined sewer transport and subsequent overflow during storm
events. As in many research and demonstration efforts, 'the sewer flushing
project also yielded a number of interesting and important questions. The
most significant of these questions have been coalesced into a series of
recommendations for further study. Additionally, since this sewer flushing
effort was aimed primarily at testing the potential of sewer flushing as a
combined sewer pollution central method, and since the results generated were
quite positive, a more comprehensive study aimed at overall development and
testing of an automated sewer flushing network is proposed. The following
are the recommendations of this effort.
l) The scope of this study was limited to a maximum flush length, care-
fully sampled and analyzed, of approximately 675 feet (206.1 m). Longer flush
length studies were conducted, but were limited to visual observation of flush
wave characteristics. This study did generate some "rough-cut" wave/solids
travel predictions, but.actual field investigations should be conducted to
verify their results.
2) Under any circumstance, single input flushing removes and should
successfully transport a minimum of 20% of the solids and up to 50% of BOD,
COD, nutrients and metals a distance in excess of 1000 feet (305 m).
This distance should be sufficient in most systems to transport that fraction
of the deposited load far enough to reach a trunk sewer or interception capa-
ble of retaining the materials in suspension until they reach the wastewater
treatment plant. To further improve flushing efficiency, especially in small
upstream networks or areas where upstream offline storage is considered via-
ble, booster flushing to further push solids to a downstream location might
be advantageous, this type of additional or booster flush was not tried as
part of this effort and should be investigated, especially to further define
the movement potential within a long, highly depositing sector.
17
-------
3) Flush waves were shown to remove significant masses of heavy metals
with large percentages, up to 50 %, remaining in suspension after extended
quiescent periods. Initial settling column tests conducted as part of
this study indicated that most of the heavy metals analyzed including nickel,
chromium, cadmium and lead, showed Tittle tendency to -settle in the
column in periods extending to 2 hours. Tests conducted in an earlier phase
of the project where flush samples were allowed to settle for periods of
approximately 4 hours indicated large percentage settling of heavy metals
after the time period. A definite question exists, possibly due in part to
difficulties with existing settling column procedures, especially pertaining
to light near-colloidal fraqtions. As part of an ongoing effort, EPA has
funded a study to develop an improved settling column and procedure, and test
the unit on combined sewer overflows. Due to the importance of heavy metals
as a potential health hazard and the potential for sewer flushing to remove
large percentages of the metals from the combined system between storm events,
further analysis of the settleability of flush wave entrained heavy metals
should be conducted using the new procedure if it proves successful. The
metals program should be coupled with the long-distance flushing program
mentioned as recommendation 1 to assess metals movement and subsequent over-
flow pollution source.
4) To the authors' knowledge, the automated sewer flushing module de-
signed and built for use in this project, represents a first of its kind ever
actually applted and tested in the field. Results of approximately 5 1/2
months of operation were very positive. The device as tested was quite sim-
plistic, aimed primarily toward prototype concept testing rather than rigorous
proofing of equipment. The automated module represents only one of a spectrum
of different types of simple devices that could be developed for sewer flush-
ing. Prior to widespread application of sewer flushing technology, a further
effort should be expended testing differing automated flushing approaches
in as rigorous a fashion over an extended time period.
5) Sewer flushing over a limited time period proved to be a technical
viable combined sewer abatement method. Comparative assessments made between
sewer flushing and removals due to wet weather indicated that flushing was
far more efficient than runoff from low to moderate intensity storms in moving
deposited combined sewer solids. This fact was further iterated by the conti-
nuous decrease in sediment levels in the flushing segments during the active
program. Unfortunately, the program was of quite limited duration, leaving the
long-term question of dry vs wet weather control still somewhat unresolved.
Based on the results of this study a long-term program assessing dry/wet wea-
ther flushing/abatement performance should be instituted to generate real op-
erational data on effectiveness and cost to compare to other abatement techno-
logies.
Proposal for an Overall Assessment of Long-Term
Automated Sewer Flushing
Results of the sewer flushing research project indicated that sewer
flushing by manual or automated means yielded similar pollutant' removals. Due
to the large number of 201 facilities plans being developed nationally for
combined sewer systems, and the results of several major cost effectiveness
studies recently released indicating the need for alternative, realistic
18
-------
combined sewer abatement technologies, the authors recommend that the follow-
ing study be immediately implemented. In this plan, concept of a network of
automatic flushing modules would be operationally tested for an extended
period of time. The purpose of this study would be to develop operational
experience with long-term automated flushing for making sound technical and
economical comparative assessments with other forms of combined sewer overflow
abatement technology, as well as to provide empirical evidence of long-term
collection system pollutant removal effectiveness over both dry and wet wea-
ther periods.
Flushing by manual means has been shown to be an effective, reliable
pollution control mechanism, subject to the unresolved issues previously men-
tioned. Initial testing of an automated flushing technique has been shown
to be equally as successful as manual flushing.
The widespread and successful use of process computers coupled with
the advanced state of telemetric technology would suggest that a network of
automatic flushing modules could be centrally controlled to operate in-line
storage devices and/or to trigger external water sources for inducing flush
waves.
Unfortunately, a number of implicit underlying assumptions violate
present knowledge, including: (a) perfected automated flushing systems do
exist; (b) the state-of-the-art in long-term sewer monitoring in sewer lines
is reasonably perfected allowing for totally automated control; and (c) an
automated sewer flushing system would take care of debris accumulations such
as sticks and rags commonly occurring within sewer sustems.
First of all, perfected automated flushing systems do not presently
exist. Except for the conceptual effort and well-controlled experiments of
the FMC Corporation (12) which never actually demonstrated automated flushing
modules in the field, and the limited experience with one flushing module in
this study, no real work has ever been performed of the magnitude required
for proper system evolution. Automated flow regulation of large interceptors
and trunk lines is being conducted in Seattle and Detroit using movable sluice
gates, and in Minneapolis - St. Paul employing inflatable dams (14), but no
significant work has been done in small pipes where the pollutant/mechanical
difficulties lie. The nature of deposited solid movement and flow patterns
in upstream networks is completely different than that found in major collec-
tion pipes requiring considerably different sensing and control devices.
Debris clogging and grit accumulations can change dramatically from day to
day. Maintenance of a complex automated computer controlled flushing system
would be a near impossibility for most municipalities. The current trend in
municipal waste treatment is away from complex facilities entailing compli-
cated automative equipment because of the lack of resources and skilled labor
required for successful operation.
Secondly, the state-of-the-art in automated flow monitoring of sew-
er systems is not perfected. There presently exists literally dozens of dif-
ferent kinds of automated sewer flow monitoring devices ranging from a liquid-
level sensor using floats, or moveable probes to sense the liquid surface, to
pressure tranducers and various bubble sensors to sense pressure and indirectly
19
-------
liquid depth, to continuous velocity meters using dyes, ultrasonics and strain
gages. Newly developed sonic level sensors which are non-intrusive could be
applied for the determination of bed load deposition accumulation. These de-
vices are as of yet untried in this situation, and could possibly function
with little trouble in a limited application. However, in a larger flushing
program requiring hundreds of these sensors, operational difficulties might
arise, crippling the effectiveness of the telemetry system. In essence, no
system has been properly tested, especially in small pipes. The state-of-the-
art in development of automated sewer flushing systems and control devices is
presently a considerable distance from the point of sophisticated computerized
telemetric feedback systems.
In order to address these points, as well as answer the many ques-
tions raised in the conclusions of this study, a dry/wet weather program is
envisioned integrating recommendations 1-5 with a step-by-step development
of automated flushing systems. This proposed program would emphasize testing,
evaluation and development of feasible system components and assess overall
system performance. The recommended study would demonstrate approximately
five automated dry-weather sewer flushing modules in a branch sewer network
or a system of five modules encompassing 1-3 miles (1.69-4.3 km) of sewer
length, coupled with wet-weather first-flush control by upstream off-line
storage.
In the dry-weather deposition oriented program 3-5 automated flush-
ing modules would be constructed and installed in an enclosed subsystem net-
work to allow for assessment of overall pollutant reduction efficiency. A
representative scheme for the envisioned program is depicted in Figure 2.
Types of modules could probably include a hydraulically operated gate system,
an expandable diaphragm positioned above the flow and inflating down into the
sewer, an auto-siphon type system and a jetted siphon. Various types of flow
sensing devices would be used to attain optimum results. Flushing would be
operated as a staged-sequential network controlled by a central controller
located within the network. The concept would assess long duration flushing,
long distance flushing and booster flushing as illustrated in Figure 2 where,
for example, the deposits just downstream of module 1 are first moved down to
area A, then are displaced to area B by the automatic siphon at mode 2, and
finally flushed out to the trunk sewer by the forced jet of module 3.
It is envisioned that 4-6 storm events would be monitored before
the installation of the flushing system, and approximately 10 while the sys-
tem is in operation. In addition, flushes and outputs from the system would
be regularly monitored during dry weather. The program would be split into
four seasons whereby six weeks would be monitored during each season. Oh a
monthly basis, all lines within the subsystem would be TV-inspected and re-
corded on video tape to visually assess overall results of the flushing pro-
gram.
In order to provide maximum wet weather control, an upstream off-
line storage module could be placed downstream of the flushed network to cap-
ture residual wet weather first flushes emanating during wet weather. In
this system, outputs from the various flow sensors would be transmitted to
a central controller located at the storage module or other remote site.
20
-------
I BACKUP a
RELEASE
MODULE
TYPICAL COLLECTION SUBSYSTEM
SEQUENCING OF FLUSHES
MODULE l< MOVES MATERIALS FROM AREA I TO POINT A
MODULE Z< MOVES MATERIALS FROM AREA 2 a POINT A
TO POINT B
MODULE 3> MOVES: MATERIALS FROM AREA 3 a POINT B
TO TRUNK SEWER
MODULE 4> MOVES MATERIAL FROM AREA 4 TO HILL
WHERE FLOW CARRES, TO TRUNK SEWER
ONTOUR
LINES
FIGURE 2 PROPOSED AUTO FLUSHING CONCEPT
-------
During wet weather, the flushing gates or dams would be used to concentrate
a major first flush and duct it into an adjacent off-line storage tank for
detainment until after the storm has subsided. When flow levels in the inter-
cepting line had sufficiently decreased, the storage tank would discharge
back into the system where the waste would continue to the treatment facility.
Equipment development and testing would follow a step-by-step concept utiliz-
ing simple, proven pieces of equipment as much as possible. Monitoring of
the wet weather program would include input into the storage tank as well as
overall network emissions.
The proposed program would provide a means for EPA to develop, test
and evaluate flushing as a realistic combined sewer pollution control alter-
native capable of incorporation directly into 201 facilities plans. Project
duration would total 2 years, at which time all aspects of sewer flushing, as
well as testing of sewer flushing (upstream capture), would be assessed,
operation performance evaluated and real costs generated.
22
-------
SECTION 4
DETAILS OF TEST SEGMENTS
4.1 Foreword
Discussion of the selection process used to choose the experimental
flushing segments is presented in Section 4.2. Descriptions, maps and photo-
graphs of the test segments flushed in the three phases of experimentation
are given in Section 4.3. Activities relating to the preparation of the test
segments for the flushing'experiments are described in Section 4.4.
4.2 Selection of Test Segments
The combined sewerage system servicing the high-density residential
communities of Dorchester and South Boston in the metropolitan Boston area was
investigated to select potential flushing test segments. The topographical re-
lief of the 3600 acre (1458 ha) area encompassing portions of both"communities
is moderate to hilly. This area is indicatad by.crosshatching in Figure 3.
This particular sewerage system was chosen for investigation because the area
had been thoroughly mapped and intensive field physical surveys had been ac-
complished in a prior study (4). In addition, a computerized sewer atlas had
been previously prepared, consisting of nearly 3000 manhole to manhole seg-
ments representing about 0.5 million lineal feet of sewer. This information
had been used as input for a deposition model to estimate daily dry weather
deposition loadings. A number of pipe segments with high dry weather sewage
deposition rates were identified in that study and physically surveyed. Rudi-
mentary flushing experiments were conducted during that study in an attempt to
dislodge heavy sanitary deposits. This particular sewerage system was chosen
because a wealth of detailed information existed. Other areas in Boston
would have been equally as suitable for conducting the flushing experiments,
but the cost of basic mapping inventory and acquiring site-specific knowledge
would have been prohibitive.
It was decided at the onset of the project that experimentation
would be limited to small diameter (12-18 inch)'pipe segments. This limitation
on pipe size was imposed for three reasons. First of all, the underlying
motivation of the project was to investigate the feasibility of flushing up-
stream pipe-segments as an integral component of an overall source control
management program. Secondly, prior experience with this system indicated
that most of the predicted daily dry weather deposition loadings were con-
tained within the small diameter laterals. Thirdly, project budgetary con-
straints precluded flushing large diameter sewer pipe segments.
23
-------
COLUMBUS PARK HEACWORKSI
CARSON BEACH
24
-------
Existing large scale base maps, small scale street maps, topographic
maps, and pertinent existing sewerage system maps, for the selected areas were
collected and visually analyzed to detect probable candidate sites for the
field flushing experiments. Copies of plan maps and City of Boston detailed
sewer maps (1"=79') were obtained for a number of candidate areas to verify
pipe "detail. This information in conjunction with past experience with the
sewerage system enabled selection of ten potential candidate sites.
The existing deposition model results for the area were further uti-
lized to study in detail the selected pipes with respect to deposition load-
ings. The candidate sites were then field inspected to roughly assess ade-
quacy for field flushing experiments. Assessments were made on the basis of
existing deposits, pipe characteristics, access, traffic and safety.
4.3 Description of Test Segments
After a careful review and inspection program, four streets were
selected for the flushing experiments. These streets are all in Dorchester
characterized by high density, 3-story multi-family dwellings. General loca-
tion of the segments are shown in Figure 4. Two of the test segments located
on Port Norfolk and Walnut Street are served by flat combined sewer laterals
of 12 and 15-inch circular pipe, respectively. The other two test segments on
Shepton and Templeton Streets, are serviced by separate sewer laterals of 12
and 15-inch circular pipe, respectively. There are downspouts on both streets
connected to the sanitary sewer. Although these two segments are separated,
considerable stormwater inflow occurs during storm events.
A plan map of the two combined sewer laterals on Port Norfolk and
Walnut Streets is shown in Figure 5. The map was prepared using City of Bos-
ton assessor maps and relevant detailed sewer plan and profile maps. The map
shown in Figure 5 also contains the number of residences and occupants per
dwelling. This information was gathered from recent census tract information.
Due to the residential nature of the community, the number of dwellings and
occupants were reasonably stable but did vary over the course of the study
as a result of several fires and ensuing demolition.
The combined sewer lateral test segment on Port Norfolk Street was
used for both the first and second phase experiments. Starting from the
uppermost manhole on the west end of Port Norfolk Street, the first phase ex-
periments were conducted using the sewer segment located between the third
and fourth manholes. Flushes were initiated at the third manhole and sampled
at the fourth manhole. Further details of the flushing and sampling proce-
dures are given in Chapter 5. During the second phase of work, the second
manhole on the west end of Port Norfolk Street was used as the flush injection
point and the flush waves at the next three downstream manholes were sampled.
The 15-inch combined sewer lateral test segment on Walnut Street, used only dur-
ing the first phase program, is located on the westerly side of Walnut Street.
Photographs of both segments are shown in Figure 6. The photo on Port Norfolk
Street was taken from the easterly end of Port Norfolk Street, that is, oppo-
site the direction of flow. The photo on Walnut Street was taken at the most
westerly end of Walnut Street again in the direction opposite to the direction
of flow. Street grades are flat for both streets. Characteristics of these
25
-------
L STREET BEACH
CARSON BEACH
PORT NORFOL
WALNUT
FIGURE 4 LOCATION OF FLUSHING SEGMENTS
26
-------
ro
JHS PORT NORFOLK STREET 218 TEST SEGMENT B PIPE .22.5
17.0 CEMENT PIPE
OVERFLOW
CONDUIT
INVERT
MANHOLE
SANITARY SEWER
STORM SEWER
SCALE T'lOO'
FIGURE 5
PLAN MAP OF SEWER FLUSHING TEST SEGMENTS
PORT NORFOLK a WALNUT STREETS
-------
Port Norfolk Street, Dorchester
Walnut Street, Dorchester
FIGURE 6: PHOTOGRAPHS OF PORT NORFOLK AND WALNUT STREET TEST SEGMENTS
28
-------
flushing segments are summarized in Table 2. Estimates of contributing popu-
lation cited in Table 2 include all upstream input and tributory capita along
the segment down to the sampling manhole. Slopes of the other segments along
Port Norfolk Street used in the serial flushing experiments are similar to the
first phase test segment.* The pipe diameter along Port Norfolk Street re-
mains unchanged.
The plan map of the flushing test segments on Shepton and Tempieton
Street is shown in Figure 7. The flushing test segment on Shepton Street was
used during both the first and third phases of experimentation and is located
between the second and third manholes from the intersection of Florida and
Shepton Streets. The segment on Tempieton Street was flushed during the first
phase and is situated between the second and third manholes from the inter-
section of Florida and Templeton Streets. The resident population along
Shepton St. appeared stable during the study because of the percentage of long
term residences and families. It appeared that the population along Templeton
St. may have varied considerably because of the transient nature of inhabitants.
The location of each dwelling, number of residences and inhabitants per dwell-
(from census tract information) are also shown.
Photographs of both streets are shown in Figure 8. The photos were
taken from Florida Street in a westerly direction. There is a hill crest in
the middle of both streets with the street grades flattening near the inter-
section at Florida Street. The test segments are in the foreground in both
photos. General characteristics of both streets are given in Table 2.
4.4 Pre-Cleaning Test Segments
Prior to initiation of the flushing program, visual inspections of
the test segments indicated roughly 4-6 inches of deposited sanitary wastes
mixed with long term accumulations of gravel, sand and grit. The segment on
Shepton Street contained mostly domestic waste deposits. The deposits in Tem-
pleton Street.and Port Norfolk Street contained substantial quantities of
sand and gravel. The deposits along the segment on Walnut Street contained
domestic waste deposits, sand and gravel, and considerably quantities of
grease.
It was desired to remove these sediments to clean pipe conditions
for the purpose of starting the program at zero base-line deposition condi-
tions. Intensive water jetting cleaning for over a week was tried using
both fire hydrant discharges and injections from the specially designed water
tanker described in Chapter 5. Several inches of material were removed at
the Shepton Street segment but the effort was futile at the other locations.
The City of Boston Public Works Department then provided mechanical cleaning
rodding devices in an attempt to remove these sediments. These efforts were
again of little use. Precleaning efforts loosened bricks in the manhole table
on Port Norfolk Street and large quantities of sand flowed into the segment.
* Virtual pipe slopes of all pipe segments used in this research effort were
determined by application of least squares/mathematical optimization tech-
niques to flush wave data. These procedures and results are described in
Chapter 7.
29
-------
TABLE 2: DESCRIPTION OF FLUSHING SEGMENT CHARACTERISTICS
PRIOR TO FLUSHING EXPERIMENTS
Characteristics
Pipe Shape & Size
(inches)
Service Type
Length of Flush
Segment (feet)
Sewer Map
Pipe Slope
Contributing
Population
General Sediment
Appearance
Dry Weather Flow
Appearance
Street Surface
Appearance
Port Norfolk
1 2 ci re .
Combined
*
247
.0049
94
Heavy septic sani-
tary deposits &
fine sand
Impounded very
sluggish
Good surface w/
considerable
surface trash
Shepton
12 circ.
Separated w/
connected
roof leaders
226
/
.0035
230
Fresh sanitary
deposits
Slight meandering
movement
Good surface,
clean
Templeton
15 circ.
Separated w/
connected
roof leaders
187
.0032
221
Septic sanitary
deposits & grit
Impounded slowly
moving
Poor surface
dirty
Walnut
15 circ.
Combined
136
.0048
71
Septic sani-
tary deposits
& some sand/
gravel
Impounded very
sluggish
Good surface,
clean
CJ
o
Phase I segment only
-------
co
SnuMDCfl ur n
ROOF LEADER
NUMBER OF
HOUSE STRFF
HOOSE STREET
TT-O INVERT
• MANHOLE
SANITARY SEWER
STORM SEWER
SCALE
FIGURE 7 PLAN MAPOF SEWER FLUSHING TEST SEGMENTS
SHEPTON S TEMPLETON STREETS
-------
Shepton Street, Dorchester
Tempieton Street, Dorchester
FIGURE 8: PHOTOGRAPHS OF SHEPTON AND TEMPLETON STREET FLUSHING TEST SEGMENTS
32
-------
Field crews then repaired the manholes. Finally, a professional sewer clean-
ing truck equipped with a 2000 psi water jetting nozzle cleaning device
was hired to remove these materials. The intent was to remove long-term accu-
mulations from the entire upstream pipe length as well"as from the test seg-
ment at each street. Shepton Street was thoroughly cleaned with the exception
of a few large rocks, brick fragments and pockets of gravel. Several inches
of gravel and sand remained along the other three segments. All segments were
reasonably free of residual organic deposits. The cleaning and repair opera-
tion took several weeks to accomplish.
The sediment beds were maintained at constant levels over the course of
the first phase flushing program which was conducted over a three month period
during the summer and fall of 1976. The second phase of experimentation be-
gan in the spring of 1977. Over the course of the winter the deposits again
accumulated to pre-project conditions primarily due to sand from winter de-
icing practices. The professional sewer cleaning contractor was re-hired to
clean the Port Norfolk and Shepton Street segments. The second phase program
was conducted during the spring and summer of 1977 solely on Port Norfolk
Street. The segments along the entire street were maintained nearly free of
any sand and gravel accumulations during this period as a result of the re-
peated flushing experiments. The third phase automated flushing experiments
were conducted over the summer and fall of 1977. No substantial sediment lay-
ers were noted during this period. The automatic module on Shepton Street
was inspected in the early spring of 1978 and substantial pre-project heavy
organic and grit deposits were observed. In general, sediment beds in the
test pipe segments were maintained at fairly constant levels during any
sequence of flushing operations. Once the flushing operations were terminat-
ed the deposition characteristics returned to pre-project conditions.
33
-------
SECTION 5
FIELD PROCEDURES AND EQUIPMENT DETAILS
5.1 Foreword
This chapter presents a summary of all field procedures, activities
and equipment used during the sewer flushing research study. A brief discuss-
ion of the types of flushing methods considered in the study is presented in
section 5.2. Details of the specially equipped flushing truck designed, fab-
ricated and used during the first two phases of experimentation are given in
section 5.3. Field operations procedures used in the first phase of experi-
mentation are described in section 5.4. Operational procedures used during
the second phase experiments are given in section 5.5. Equipment details of
the automatic sewer flushing module used in the third phase of work are de-
scribed in section 5.6. Details of the flush wave flow monitoring procedures
are given in section 5.7.
5.2 Manual Flushing Approaches
A number of different manual methods for inducing flush wave in the
test segments were considered in this study. Manual methods were solely used
in the first two phases of work while the third phase dealt solely with an
automated approach. Two general categories of manual methods were considered:
backup and release using sewage and external flush water injection using fresh
water.
Backup and release methods considered in the study are shown in
Figure 9. Figure 9-A represents a situation where both the upstream and
downstream side of a manhole are stoppered, the manhole surcharged with a pre-
determined volume of water and then the water rapidly discharged by releasing
the downstream stopper. Figures 9-B and 9-C represent two conditions of up-
stream backup and release; case B utilizes both the upstream pipe capacity and
a fraction of the manhole capacity, and case C utilizes the upstream pipe capa-
city only.
Representative external source flush injection methods are shown in
Figure 10. Figure 10-A depicts a gravity flush feed at a low discharge rate
while Figure 10-B depicts a flush injected into the manhole at a high rate of
discharge. The flush volume in either case can be small or large. High flush
rates induce high velocity heads and the ability to scour sediments, while
large volumetric flushes provide fluid momentum and the capacity to dilute and
transport scoured materials. One of the objectives of the first phase experi-
mentation was to determine the flushing pollutant removal effectiveness for
34
-------
A. Manhole Surcharge - Rapid Release (External Source)
B. Back and Release - Collection Pipe and Manhole (Higher Head)
C. Backup and Release - Collection Pipe
FIGURE 9 REPRESENTATIVE BACKUP AND RELEASE FLUSH METHODS
35
-------
A. Low Rate Flush (Gravity Feed) (Volume High or Low).
v- •>...»-.>*>....;-/\-->..v >..•" .*.*#/'-•*:•* •:.-
--- •••
B. High Rate Flush (Pressurized Feed) (Volume High or Low).
FIGURE 10 REPRESENTATIVE EXTERNAL SOURCE FLUSH INJECTION METHODS
36
-------
differing combinations of flush volume and flush rate. Minimization of the
quantity of flush water used may be important for communities faced with water
shortages in the future, while in other applications the issue of increased
mechanical equipment complexity and energy cost considerations may be more
important. The pre-project apriori hypothesis was that flush volume was more
important since it was believed that the high velocity head induced by high
entry rates would be rapidly dissipated by friction losses within the wave as
it proceeded downstream. The results of the first phase program described in
Chapter 8 showed high volume with high rate flushes to be the most effective.
Four combinations of flush volume and flush rate were considered
in this study. The external flush rate and volume inputs were delivered by
a specially equipped flush tank truck which was prepared for the project. De-
tails of the tanker are given in Section 5.3. The operation under gravity
conditions provided a controlled flush release of 35 to 50 cubic feet (.99
to 1.98 cubic meters) at a rate of 0.25 to 0.5 cubic feet per second (7.08to
14.16 liters per second). Under pressurized conditions the same volumetric
range of flush was accomplished at a range of 0.5 to 1.25 cubic feet per
second (14.16 to 35.4 liters per second). These ranges of flush rates and
volumes were chosen from prior experience with flushing segments in this study
area (4) and moreover were chosen to be reasonably representative of the type
of flushing operation which could be implemented by most communities. Various
sized discharge nozzles (1-3 inch) were also considered early in the first
phase experiments to maximize velocity head effects in the flush wave.
5.3 Details of Flush Truck and Ancillary Equipment
•Prior to the start of the actual flushing efforts many pieces of
specialized equipment had to be designed and fabricated for use in the study.
These in general included the manual sewer flushing module, discharge nozzles,
and sampling equipment. The major single piece of equipment used during the
first two phases of the sewer flushing program was the manual sewer flushing
module or flushing tank truck. The nature of the sewer flushing research
work necessitated a wide range of discharge capabilities to be provided by the
flushing module. Since the first phase flushing program was aimed at assessing
the preformance of various methods which in essence translated primarily to
varying flush injection rates, a wide range of rate flexibility was required.
The unit also had to be mobile and capable of conducting several flushes before
refilling.
Figure 11 is a mechanical schematic showing the plan, side and
end views of the flushing truck. Figure 12 is a photograph of the flush
truck. In essence the unit consisted of two 1000 gallon (3.7 cubic meters)
pressure tanks mounted in parallel on a steel I beam frame. Fluid transmiss-
ion was conducted through a 4 inch (10.2 cm) I.D. steel piping manifold
allowing complete separation or interconnection of tanks and discharge routes.
Final discharge was made through a 3 inch (7.6 cm) quick-action gate valve,
to allow for rapid on and off as well as good throttling characteristics.
Accurate measurement of discharge volumes and subsequently discharge rate
was accomplished using a 4 inch turbine water meter, manufactured by Mersey
Sparling Co. and supplied to the study by the City of Boston Public Works
Department. The meter provided accurate volumetric measurement of the
37
-------
Hrtrr HOIAK
M UK IW WMt
r iwmoo tonw— D«-w MM ocn »U.YI
to
00
\W fWMUM U«
OOWMO3M TMM
a m WITT KL«T w.vi
FIGURE II MECHANICAL PIPING SCHEMATIC DIAGRAM
MANUAL SEWER FLUSHING UNIT
SCALE'• l" = 5'6'
-------
FIGURE 12: PHOTOGRAPH OF FLUSH TRUCK
39
-------
discharge stream at rates from 0-1000 gpm (0-3750 1/m). Figure ISA shows a
photograph of the manifold piping and equipment at the rear of the truck. The
entire sewer flushing module as assembled was 22 feet (6.7 m) in length and
8 feet (2.4m) wide. To enable mobile operation the entire unit was attached
to a 20 foot (6.1m) flat bed truck. The tanker was equipped with a pneumatic
system to pressurize the tanks to 70 psi. Tank pressurization and control
was accomplished by a gasoline powered compressor mounted on the unit. Figure
13B shows the pressurization equipment located on top of the .tank units.
Typically, during the study the flushing truck would be filled by means of a
3-inch fire hose from a hydrant through one of the 3-inch quick-connect fill
valves located on top of each tank.
Flush injection was accomplished by means of a 3-inch fire hose
connected to a discharge nozzle mounted on a support rod. An assortment of
nozzles was assembled from 90° pvc electrical sweep ells ranging from
1-3 inches in inside diameter. Each nozzle could be readily coupled to the
discharge hose by means of a 3-inch camlock fitting. All nozzles and hoses
used in the sewer flushing study were equipped with camlock fittings to enable
rapid assembly. The pvc sweep ells provided an ideal nozzle for use during
the study in that the 90° bend is made over a large radius thereby minimizing
energy losses at the discharge. The nozzle support rod was made from a
10-foot length of 1-inch electrical tubing. The electrical tubing was out-
fitted with a welded point in one end for anchoring the nozzle during dis-
charge as well as a movable nozzle holder coupling that could be adjusted to
allow for nozzle centering on any size pipe from 8-30 inches in diameter.
Figure 14A shows a photograph of the flushing nozzles, nozzle support and
hoses. Figure 14B shows a photograph of the inflatible rubber stoppers
(children's toy called "Hippity Hop") which were lowered into place by a rope
installed in place in the sewer segment and rapidly inflated using the flush
truck pressurization system. These devices were inexpensive and worked ex-
tremely well. The entire flushing system so developed could be easily oper-
ated by two people from the street surface.
5.4 First Phase Flushing Procedures
The flushing program in this phase was concerned with only the
effects of flushing a single manhole to manhole segment. Four test segments
were flushed in this program and were described in section 4.3. Three dif-
ferent methods of manual flushing were performed. The first method consisted
of backing up the upper" end of the flushing manhole with an inflatable rubber
stopper, followed by quick release. The other two methods were gravity and
pressurized dump discharge into the flush manhole with the upper end of the
flush manhole generally blocked off. Different flush volumes were used during
the external source injection experiments. All dump discharge flushes were
performed using the water tanker. All flush volumes were measured by a water
meter which was repeatedly calibrated to ensure accurate monitoring of the
delivered flush volumes.
A total of 86 separate flushing experiments were performed during
the period of August 30, 1976 through November 12, 1976. Roughly 20 flushes
on a 3-4 day basis were accomplished for each of the four test segments. The
40
-------
A. Manifold Delivery System
B. Pressurization Equipment
FIGURE 13: PHOTOGRAPHS OF MANIFOLD DELIVERY SYSTEM AND PRESSURIZATION
EQUIPMENT, FLUSH TRUCK.
41
-------
A. Flushing Nozzles, Support Rod and Hoses
B. Inflatable Sewer Stoppers
FIGURE 14: PHOTOGRAPHS OF FLUSHING NOZZLES AND INFLATABLE SEWER STOPPERS
42
-------
method of flushing was rotated per segment per flush so that all methods were
applied to each segment over the test period.
The sequence of pertinent operations during a given flushing
experiment was the following. After the safety equipment was set-up, the
segment was then visually inspected by lamping to assess solids buildup and
debris as well as to characterize the depositing matter (fine sand and
organic matter, toilet paper, rags, small rocks and sticks). Several liquid
background samples and depth flow were taken at five minute intervals.
Next, the upper end of the flush manhole was blocked-off (in most cases) and
sediment samples over a prescribed unit length was taken in both the flush
and sampling manholes. The scraped materials were visually inspected to
assess solids characteristics, collected in a suitable container and brought
back to the laboratory. A photograph of the inflated rubber sewer stopper in
the upper end of the flush injection manhole on Port Norfolk Street is shown
in Figure 15A. Figure 15B shows a photograph of a field engineer
with safety equipment for entering the flushing manhole to take sediment
scrappings. Ventilation equipment and the flush truck are shown in the
background of the photograph in Figure 15A. Figure 16 shows two photographs
of the sediment scrapping operation. A specially designed pipe squeegee was
used to scrape the sediments of a lineal foot of sewer.
The flushing experiment was. then conducted either using backup
sewage or fresh water injection from the flush tank. Dye was injected in the
wave and at the instant of arrival, a one-liter aliquot was taken with a
specially designed hand scoup for obtaining a reasonable cross-sectional
sample of the solids within the flush wave at the downstream sampling manhole.
The device specifically excluded bed load materials. After the first sample
was taken at the first visual sighting of the wave, 8 grab samples were
taken at 10 second intervals, and then an additional 8 samples were taken at
20-second intervals. Wave heights were taken at each interval of time which
were later used to determine the instantaneous flow rate for computing mass
pollutants removed by the flushing experiment. A total of 17 flush wave
grab samples were taken during a given flushing operation totalling 4 minutes.
Figure 17A shows a photograph of a low rate flush feed at 0.25 cubic
feet per second into the flush manhole on Shepton Street. Figure 17B shows
a similar photo on Shepton Street with a high feed rate of 0.75 cubic feet
per second. The jet is continuing well into the flushing segment. Hydraulic
entry of the flush feed into the pipe segments for the Shepton and Port
Norfolk Street segments generally exhibited the patterns shown in Figure 17.
Figure 18 shows a photograph of a high rate (0.75 cfs) and large volume flush
(75 cubic feet) at the Walnut Street test segment. The flush wave in this
case has induced an extremely turbulent backwater effect in the flush manhole.
Flushes at Templeton and Walnut Streets generally followed this pattern.
Figure 19A shows a photograph of the flush wave sampling hand scoups. Figure
19B shows the flush wave sampling operation. The field engineer is taking a
grab sample with the hand scoup near the end of a sampling sequence at the
Hal nut Street segment. In the far left hand side of the photograph is the
staff gage where liquid level depths to the nearest eighth inch were recorded
at the'appropriate time intervals signaled by an observer at the top of the
43
-------
r
A. Inflatable Sewer Stopper in Place.
FIGURE 15: PHOTOGRAPHS OF INFLATABLE SEWER STOPPER AND FIELD ENGINEER EQUIPPED WITH SAFETY GEAR
B. Field Engineer Equipped with Safety
Gear.
-------
FIGURE 16: PHOTOGRAPHS OF SEDIMENT SCRAPING OPERATION
45
-------
en
A. Low Flush Rate (0.25 cfs) B. High Flush Rate (0.75 cfs)
FIGURE 17: PHOTOGRAPHS OF FLUSH WAVE INJECTIONS AT DIFFERENT FEED RATES ON SHEPTON STREET, DORCHESTER
-------
FIGURE 18: PHOTOGRAPH OF FLUSH WAVE INJECTION AT WALNUT STREET, DORCHESTER
47
-------
00
A. Flush Wave Hand Sampling Scoups
FIGURE 19: PHOTOGRAPHS OF FLUSH WAVE HAND SAMPLING SCOUPS AND GRAB SAMPLING OPERATION
B. Flush Wave Grab Sampling Operation,
Walnut Street
-------
manhole (not shown on the photograph). The pipe squeegee used to scrape pipe
sediments is shown on the right hand side of the photograph. After the
sampling sequence was completed the pipe segment was then flushed for five.
minutes at a maximal flush rate of about 1.25 cfs. The purpose of this final
operation was to flush clean any residual organic matter remaining in the
segment. The segment was assumed to be clean for the next period of solids
accumulation. The final washing operation was conducted from the onset
of the first phase program till the end of October, 1976.* The balance of
the first phase program was conducted without the final flushing operation to
ascertain in an indirect way whether there were residual pollutants remaining
from a given flushing experiment.
Flow monitoring of dry weather flow characteristics, stage/
discharge calibration efforts at the sampling manholes and the special dye
injection procedure used to directly estimate flush wave discharge are
described in Section 5.7.
5.5 Second Phase Flushing Procedures
The experimental field work in this phase was concerned with the
problem of flushing a long flat stretch of combined sewer lateral. The
street contains five manholes and is roughly 1000 feet in length. Flushes
were injected into the upper most manhole and pollutant levels in the flush
wave passing three downstream manholes were monitored, A diagram of the
second phase flushing and sampling manholes along Port Norfolk Street is
shown in Figure 20. Work was divided into two subphases. Initially,
pollutant removals over the three segments were determined for different flush-
ing conditions established in the first manhole. These flushing experiments
provided insights into flushing effectiveness over three segments of pipe.
Details of the field operation for this subphase of work are described in
Section 5.5.1. In the next subphase of work, settleability tests were
perrormed on samples taken from flushes conducted in a similar manner for
the purpose of crudely extrapolating how far beyond the flushing monitoring
manholes would the materials be carried. Sampling procedures and sample
preparation for settling column tests are presented in Section 5.5.2.
5.5.1 Serial Flushing - Pollutant Removals
The purpose of these experiments were to ascertain the pollutant
removal effectiveness over three consecutive combined sewered segments on
Port Norfolk Street by flushing the uppermost manhole using the flush truck.
These experiments were also intended to provide additional information for
assessing the flushing effectiveness of the first phase flushes and to
provide further information for determining rates of dry weather sewage
deposition.
The backup and release method of flushing was not considered in
this phase since there was no appreciable contributary population at the
upstream flushing manhole. Most of the flushes conducted during this period
were delivered at high.feed rates with high volumes since this mode of flushing
proved to be the most effective during the first phase operation. Booster
* At that point 69 flushes had been conducted or approx. 17 at each segment.
49
-------
PHASE 2 PROGRAM
en
o
C) (
\FLUSH INJECTION
PRIOR SCRAPING
(PHASE 2)
) C2) 0
^ A J
FLUSH SAMPLING
AND
GAGING
162.5'
246.9'
250.0'
FIGURE 20 SCHEMATIC OF SECOND PHASE FIELD OPERATIONS
-------
flushing along the segment was not considered in this study. This concept
entails sequencing multiple flushes along a segment such, that the resuspended
pollutants and grit do not resettle. Initial flushing experiments in this
phase indicated extremely favorable pollutant removal results using a single
upstream injection. In view of the limited project resources, ,the principal
investigators believed that replication of these initial favorable findings
was.more important from the standpoint of establishing technical credibility
of simple flush methods, than in pursuing the effectiveness of multiple
sequenced flushing operations. The experimentation period began in
December 1976 and extended through March, 1977, entailing two replicate sets
of three flushing rate/volume experiments. Each experiment consisted of
three flushes conducted within a short period of each other. The first two
flushes on a given day were the same while the final flush was maximal
volume/rate flush meant to remove any remaining pollutant load in the segments.
Different combinations of flush volumes (35 to 75 cubic feet) and delivery
rates (0.3 cfs to 1 cfs) were considered. Three crews sampled the flush wave
passing the downstream manholes. Samples were taken at the same frequency
as in the first phase of experimentation described in Section 5.4. Sediment
scrapings and sampling of background sewage levels were also accomplished
as in the first phase. Results of this flushing program are reported in
Chapter 9.
A special effort was initiated in the phase of work to develop a
dragging device meant to scrape clean any residual matter remaining in the
segments either after a flushing experiment, or alternatively, prior to
flushing so that both the efficiency of flushing and the rates of deposition
could be accurately monitored. Rates of deposition could be determined either
by primary measurement using a scraping operation on an undisturbed and
pre-cleaned pipe segment or alternatively, by summing the pollutant masses
transported by the flushing to the mass measured by post-flush scraping.
Flushing pollutant removal effectiveness could be better estimated if either
the total rates of deposition or the quantity of residual materials remaining
after flushing were known. Sediment scrapings over a unit length of pipe
(one foot) had been taken before and after flushing at both ends of the test
segment during the first phase program. There were a number of difficulties
with this approach. First of all, the segments were never cleaned to zero
base-line clear pipe conditions despite two intensive weeks of cleaning
using three different procedures. This work is described in Section 4.4.4.
Secondly, there was no way of ascertaining the longitudinal profile of
sediment in the segments. FMC (12) reported that most of the deposited
material occurred within the first quarter of the experimental test
segment length and depending on the duration of the accumulation period, the
sediment profile would progressively move downstream along the segment.
Accurate delineation of deposition characteristics along the length of the
segment was beyond the scope of this research work. The envisioned scraping
operation was viewed as a compromise in which estimates of the total
quantity in the segment would be obtained respective of the actual profile
in the segment.
A special scoup fitted with a nylon catch bag was fabricated and
powered by a low-speed winch system which was connected by cables between
consecutive manholes. The operation was tried several times and failed due
51
-------
to the presence of cast iron house laterals protruding into the combined sewer
lateral along Port Norfolk Street. The approach was terminated and the
manual sediment scraping operation of a unit foot of pipe was continued in
the second phase.
5.5.2 Serial Flushing - Settleability Analyses
The purpose of these experiments on Port Norfolk Street was to
roughly assess the transport of flushed pollutants beyond the test segments
using information derived from various settleability tests. Six different
flushing experiments were performed in the period of April, 1977 through
August, 1977 in which settleability tests were performed for samples taken
at each of the three sampling manholes. Three flushes were conducted per
experiment. Settling column tests were conducted on samples taken from
first flush and Imhoff cone tests were performed on second and third flush
samples. External source flush volumes were injected into the uppermost
manhole on Port Norfolk Street using similar volumes/rates as used in the
first half of phase two, described in Section 5.5.1.
Special equipment was fabricated to obtain "undisturbed" samples
of flush waves at the three downstream manholes on Port Norfolk Street. Four
special sampling devices were constructed to collect representative
"undisturbed" flush wave samples at various intervals during a flush wave.
A diagram of the sampling device is shown in Figure 21. Each sampling device
consists of a 14-inch section of schedule 40 pvc pipe to which snap-action
end gates were attached with an approximate 5 gallon capacity when full.
The two end gates were made of one-eighth inch circular aluminum plates
attached to sliding support rods. To operate the sampling device, the two
end gates are first raised to the open position and are held in place by
a spring loaded latch. All four open tubes were placed in the bottom of
one of the downstream sampling manholes. As the flush wave passed the
sampling point, each of the tubes weretrapidly lifted at a pre-selected
times. Rapid lifting caused immediate closure of the end gates and the
capture of a segment of the actual flush wave. The aforementioned
samples were augmented using hand dipped plastic buckets.*
Sample collection for the column tests and Imhoff Cone tests
required approximately 10 gallons of the flush wave. For an accurate
representation of the .flush wave, samples were collected at various times
after the flush wave first appeared at each of the three sampling manholes.
These individual samples were then composited into the one sample
representative of the flush wave that passed through each sampling manhole.
The flush wave was sampled at six instants as it passed through the sampling
manholes. These times varied from manhole to manhole, as the characteristics
of the flush wave changed the further downstream it progressed. The sampling
_
Length of the manholes along Port Norfolk Street allowed placement of only
four sampling devices in a manhole. Since the depth of flow attained
averaged less than half depth of the samples, the volume had to be augmented
with two additional samples.
52
-------
ADJUSTABLE
GUIDE
LATCH
PIVOT
1/8" ALUMINUM
DOOR
ye" SURGICAL TUBING SPRING
1/2" ID PIPE GUIDE
LATCH POINT
1/2" STAINLESS STEEL ROD
10" ID PVC PIPE
LOCK CLIP
FIGURE 21 SETTLING COLUMN FLUSH SAMPLER
53
-------
times used attempted to divide the wave into six parts for ease of collection
and to minimize the sampling effects on the wave. The time taken for the
sample collection started when the wave first appeared at the upstream end
of the manhole. Sampling times were determined from inspection of the first
phase flushes and are shown below for each sampling manhole.
Sampling Manhole Sampling Intervals (seconds)
Referenced to First Arrival of Flush
Wave at Manhole
1 10, 20, 30, 40, 60 and 80
2 20, 30, 40, 60, 80 and 100
3 20, 40, 60, 80, 100 and 120
The depths of the flush wave passing each manhole were measured
and recorded at 10 second intervals. A flow-composited sample was prepared
from the six grab samples for each sampling manhole per flush using flush
wave depth to pipe area considerations accounting for the presence of sedi-
ments within the 12 inch lateral. Special settling column equipment and
procedures were established in order to perform settleability tests of the
flush waves. A special yoke-frame installation was devised to permit
axial and transverse mixing of the column before settleability experimenta-
tion. This ad hoc procedure was necessary since the settling velocities for
a considerable portion of the composited flush wave solids were extremely
rapid. Gentle mixing using air agitation was initially performed but
resulted in solids bulking because of the high solids content of the flush
samples. Details of the settling column and procedures are described in
Chapter 6. Results of the settling column and Imhoff Cone testing are
presented in Chapter 10.
i
5.6 Automated Sewer Flushing Module
The phase III portion of the sewer flushing research project dealt
with the development and operation of a simplistic automated sewer flushing
module. The device developed was in essence an air operated gate capable of
backing up sewage flows to a predetermined level and then suddenly retracting,
inducing a flushing wave. This type of backup and release device was well
suited to the situation as found on Templeton and Shepton Streets in Dorches-
'ter. The sewer lines in the Tempieton/Shepton Street area were previously
described in chapter 4. Upstream of the flushing manhole is a hill allowing
development of sufficient static head for a clean discharge. The module so
developed was installed on 8/30/77 and operated on a regularly serviced basis
from 8/31 - 10/31/77. During this period the module was checked at least 3
times per week to ensure performance as well as conduct sampling runs. Auto-
mated flushes were sampled seven times during the period of 9/22 - 10/13/77
with the results and discussion of overall performance presented in chapter 11
54
-------
of this report. Module operation was continued after 10/31/77 until mid Janu-
ary 1978 on a periodic inspection basis to assess long term operation and
serviceability, as well as visual performance with respect to flushing of both
the upstream or reservoir segment and the downstream flushed segments.
The automated sewer flushing module, as all other specialized pieces
of equipment used during the sewer flushing study, was designed and fabricated
by EDP Inc. Due to project constraints and the prototype nature of the flush-
ing module, construction was kept as simplistic as possible to allow for maxi-
mum flexibility. Figures 22 and 23A and B present a schematic representation
of the sewer flushing module, as well as photographs of the installed unit
taken after months in service.
5.6.1 Design Details of the Automated•Sewer Flushing Module
The automated flushing module depicted in Figure 22 was constructed
primarily of wood with a final epoxy coating for water-proofing. The module
was designed for ease of installation and maintenance and could be readily
adapted for packaged installation. As previously noted the module consisted
of an air-on-oil cylinder operated gate controlled from a master control timer
capable of pre-programmed flushing of varying sequences, with flushing inter-
vals ranging from 6-72 hours. The entire unit was powered by a 12 volt auto-
mobile battery that was capable of 6 months operation flushing daily before
recharge. Air supply to the system was by means of a small high pressure air
cylinder requiring replenishment every 150 flushes.
The automated module was constructed in two major parts: 1) the
flushing gate assemblyjand 2) the master controller. The flushing gate assem-
bly was constructed so as to sit upon the table of the sewer within the man-
hole. In order to provide maximum flexibility with respect to backup volumes
and static head of the flush wave, the sewer was extended upward from its
vertical centerline to slightly above its crown. The support frame gate guide
for the hydraulic cylinder was then coupled with the sidewall extentions to
form the completed assembly. Control of the gate action was provided by an
EDP-designed crystal controlled clock timer mechanism which emitted a signal
to the 2-way air valve, which then activated a spring returned air-piloted
4-way hydraulic valve causing the gate to go down. The gate would then stay
in the down position until the backed-up flow reached the preset level sensor
float or was down for a predetermined amount of time. The time function acted
simply as a safety mechanism in case of sensor malfunction. This safety fea-
ture was not entirely necessary in that the design of the module was such that
even if the gate was stuck in the down position, the sewage would simply over-
flow the top of the gate.
The actual flushing gate was made of plywood set into the support
frame with a positive seal on the sides. Initially, gate sealing on the sewer
invert was provided by a closed cell foam seal ring attached to the contour
fitted gate. This sealing mechanism proved to be somewhat unreliable due to
the irregular contour of the brick sewer channel in the manhole. After a few
weeks of initial testing, a polyethylene/foam floating seal was added to the
support frame forming a band along the bottom of the sewer channel. This
55
-------
MASTER CONTROL PANEL
cn
cr>
MANHOLE
PRESSURE REGULATOR
4-WAY
HYDRAULIC VALVE
SPRING RETURNED
HYDRAULIC CYLINDER
4"X 15" STROKE
L SEALING RING
INVERT •
FIGURE 22 AUTOMATED SEWER FLUSHING MODULE
-------
A. Top view of installed module showing high-pressure cylinde",
oil reservoir, control panel and hydraulic cylinder
B. Angled view of installed module
•FIGURE 23: PHOTOGRAPHS OF AUTOMATED SEWER FLUSHING MODULE
57
-------
mechanism worked very well throughout the testing period, providing close to
100 percent stoppage under any sediment condition.
5.6.2 Operational Details of the Automated Flushing Module
The automated sewer flushing module designed for this study proved
to be an effective flushing unit requiring minimal maintenance during the
testing period. Initially, there were problems with gate sealing and manhole
leakage during the first few weeks of operation. After the gate seal was
modified the unit became practically service-free. The module battery and air
supply were never replenished during almost five months of operation. The
results of the testing program were considered very positive. Once the module
was initially set at a 24-hour interval, it was left on for the entire program.
All flushes were conducted automatically. In order to provide tracking of
the backup behind the gate and a crosscheck on gate timing, a continuous
recording liquid level sensor was installed in the well behind the gate and
operated from 8/31 - 10/31/78. The level sensor recorded both dry and wet
weather flow and gate operation. Several storms occurred during the automated
flushing operation. No problems were encountered with module performance or
with sewer backups during rainfall events.
5.7 Flow Gaging Methodologies
This section describes the various field procedures used in gener-
ating data for flow gaging of flush waves, as well as dry and wet weather. Pro-
cedures for utilization of the field generated data are presented in chapter
7. Accurate and reliable flow measurement of the unsteady state, turbulent
flush waves, proved to be one of the more difficult field tasks encountered
during the sewer flushing program. Several procedures were attempted, in-
cluding: 1) steady state calibration of the flush segments using the flushing
truck; 2) high pressure dye injection utilizing a special system developed
during the study; and 3) utilization of measurement flumes specially con-
structed to minimize backwater and upstream sedimentation.
5.7.1 Dry and Wet Weather Flow Gaging
One of the parameters of interest to the sewer flushing study was
establishment of baseline flow which could then be translated into per capita
waste rates for later computation. A continuous recording liquid level
sensor was installed in each of the sewer flushing test segments for extended
periods to monitor liquid level during both dry and wet weather. This proce-
dure was adequate for most of the segments, with the primary exception being
Port Norfolk Street. The main difficulty encountered was the relatively low
depth of flow in the order of 1/4-3/4 inch (0.6-1.9 cm) and small level varia-
tions making accurate resolution of liquid levels difficult. In order to in-
crease level variability and therefore resolution of readings, a special con-
strictive flume was constructed using 4 inch (10.3 cm) pvc pipe with extended
sides and a special polyethylene inlet section. The flume so constructed in-
duced critical flow conditions with minimal head loss due to the nature of the
inlet and outlet sections. This was particularly important to avoid biasing
upstream sedimentation rates as would the use of a Palmer Bowl us or Parshall
flume. Calibration of the flume so constructed was done with time of travel
58
-------
studies using Uracine dye, as well as direct inflow measurement using the
water meter on the flush truck as an input source. In such a manner reasonably
reliable determinations of dry weather flow rates were made possible.
For those segments with enough depth of water to allow adequate, reso-
lution of dry weather liquid levels, time of travel studies utilizing dye were
conducted to develop calibration data used in the procedures described in
chapter 7.
5.7.2 Steady State Flush Wave Calibration Procedures
Initially, attempts were made to use a Manning's equation based rating
curve as a steady state surrogate for determining the flow of the passing
flush waves. Although the procedure did not work well in approximating flush
wave flow due to the non-steady state turbulent flow characteristics present,
the methodology provided reliable dry weather flow rating curves. The fol-
lowing is a synopsis of the approach used in the field to generate calibration
data for the procedure presented in chapter 7 of this report.
On arrival at a given test segment the depth of flow and depth of
sediment, if any, were carefully noted. Substantial sediment layers of dry
weather deposition were noted at some of the measurement sites and proved to
be a problem in the curve fitting -process since they varied over the course
of the study. Velocity determinations were made by measuring the time of tra-
vel of dye injected immediately upstream from the measurement site. Dye (Roda-
mine B or Uracine) was used in all cases. The length of the segment was mea-
sured in the field at all sites. At least 3 separate time of travel measure-
ments were taken for each flow condition, and, if necessary, more were taken
until they coverged narrowly to one value which would then be used in comput-
ing the velocity at that stage.
The following procedure was used for the four sites where the water
tanker was used for calibration purposes. Upon arrival at the site, the time
of travel was recorded for the background flow. The tanks on the truck were
filled from nearby fire hydrants and pressurized while still connected to the
hydrants. Water from the truck was discharged into the upstream manhole of
the test segment and the flow rate maintained for a sufficient period to allow
stabilization of both the pressure in the tanks and the water depth in the
downstream end of the pipe. The input flow rate from the trunk was then re-
corded from the flow meter on the truck. The flow depth in the sewer was re-
corded and the dye tests (time of travel) experiments were completed. Flows
at lower depths were determined by progressively lowering the delivery rate
from the truck. Using the input flow rate from the truck and an estimate of
the background flow rate the total flow rate being routed through the pipe was
determined within a small margin of error. These flow rates were used to
check the velocity and flow determinations in the sewer. All data so gener-
ated was then'input into the procedure described in chapter 7 to develop final
rating curves for each of the four test segments.
59
-------
5.7.3 Dye Injection Calibration
One of the more interesting aspects of the field efforts conducted
during the sewer flushing study included the development and utilization of a
practical, reliable method of calibration of steady state and more importantly
non-steady state flows using dye injection. Flow calibration of the flush
waves was of vital importance to the results of the study. During flushing
experiments the concentrations of solids downstream from the injection point
ranged up to 10,000 mg/1 as suspended solids. Those conditions precluded the
use of any direct velocity measurement device. Attempts were made to generate
steady state based rating curves to allow for translation of the depth meas-
urements taken during the flush waves, but they proved quite unreliable due
to varying velocities encountered during the flush wave passage. Typically,
velocities on the front side of the wave before and up to the peak were much
greater for the same depth of flow than those encountered after the peak.
An alternative procedure using high pressure dye injection was devel-
oped. The procedure basically consisted of three distinct operational units
including: 1) the high pressure injection nozzle system; 2) the pumping and
air separation unit; and 3) a flow-through cell equipped f1uorometer with
recording readout. Figure 24 is a diagram depicting the dye injection system
used. Development of the dye injection procedure was a fairly complex pro_
cess. The utilization of fluorescent dyes in tracer studies coupled with
fluorometer readouts has been widely applied to river and stream gaging stu-
dies. Similar applications as applied to sewer systems have typically failed to
provide reliable results primarily due to significant fluorescence inter-
ference from compounds in the sewerage. This is especially true with respect
to rhodamine compounds whose fluorescence peak is similar to that of phenolic
compounds often found in sewerage systems.
The dye injection experiments conducted during this program utilized
Uracine dye with a special filter system to eliminate any background inter-
ference. This filtering system was necessary since a Turner model 111 fluoro-
meter (15) was used. Initially, samples were taken from all the sewer seg-
ments and analyzed on a Perkin Elmer Hitachi model 204 research spectrofluoro-
meter. This unit has extreme selectivity and the capacity to scan both the
exciter and analyzer wavelengths independently. Using a scanning procedure
all fluorescent peaks of the sewerage were identified for both the exciter
and analyzer. Samples of both 'Rhodamine B and Uracine dye were then subjected
to the same procedure. "The results of all scans were then compared and zero
interference peaks identified. As it turned out, Uracine dye had strong
fluorescent peaks in zero interference bands with the proper exciter and
analyzer wave lengths. This fact was verified utilizing dye samples spiked
with sewage. Once the optimum exciter and analyzer peaks were established,
a Kodak filters manual (16) was used to identify combinations of filters that
would provide the proper exciter and analyzer wavelengths on the Turner
fluorometer.
The next step of the process was the development of the high pres-
sure injection system itself. The system was constructed of a dye injection
bar 6 feet (18.9 cm) in length, with a filter and control valve on one end
and the nozzles on the other. The nozzles used on the dye bar were actually
60
-------
STRIP CHART
RECORDER
TURNER MODEL III FLUOROMETER
PRESSURE GAGE
FILL AND VENT VALVE
PRESSURIZED DYE
FEED
QUICK RELEASE
AMOUNTING FRAME
MIXING ZONE
COMPLETELY MIXED DYE
ENTRAPPED AIR
FROM-4V WATER a AIR
PUMP I
V
1/4" BLEED LINE
TO |_| WASTE TO SEWER
FLUOROMETER
AIR SEPARATOR UNIT
FIGURE 24
DYE INJECTION FLOW CALIBRATION
METHODOLOGY
-------
stainless steel hypodermic needles. Needles were used because they were an
ideal injection nozzle that can be readily varied in size, and which deliver
a fine solid jet of dye at high pressures. Figure 25 shows photographs of
the dye injection system. Figure 25B is a close-up photograph of the dye in-
jection nozzles , mounting system and guard. Experiments were conducted to
generate the ideal nozzle size/flow rate to maximize dye penetration and
minimize input flow. Ultimately, 20 gage needles with 100 psi injection pres-
sure was used, giving nozzle jet penetration of approximately 12 inches
(30.5 cm). This penetration distance allowed the dye to penetrate the flush
wave stream, hit the sewer invert and totally disperse. The result tested
in a hydraulics laboratory flume produced homogeneous dye mixing within 3
inches (7.6 cm) of the injection point. The dye injection bar was coupled
to a 5 gallon (18.75 liter) reservoir that was connected to the flush truck's
pressurization system.
Dye detection was accomplished, via a continuous flow-through system
utilizing a high pressure pump through an air separater device, through the
Turner fluorometer as shown in Figure 24. The net result was a continuous con-
centration time track of the flush wave passing the sampling manhole. Figure
25A and C show a view of the street level apparatus and an actual dye injec-
tion experiment respectively.
The dye injection procedure so developed proved to be extremely re-
liable in measuring the transient flush wave flow. Computed comparisons of
injected volumes versus volumes detected downstream match within a maximum
of 3 percent. The dye injection procedure so developed was used in the
field to generate calibration data for'the optimization procedure presented
in chapter 7.
62
-------
cr>
CO
A. Photograph of Street Level Apparatus Including
Fluorometer and Recorder, Injection Reservoir,
Recirculating Pump and Lines.
C. Actual Dye Injection Experiment.
B. Dye Injection Nozzles.
FIGURE 25. DYE INJECTION APPARATUS
-------
SECTION 6
LABORATORY ANALYSES AND PROCEDURES
6.1 Foreword
Field procedures used during the three phases of sewer flushing ex-
perimentation were described in Chapter 5. Descriptions of the various labora-
tory analyses, special experiments and procedures are given in this chapter.
Parameter coverage for the discrete flush wave samples collected during all
three phases of work are described in section 6.2. Discussion of heavy metals
analyses of first phase composited flush wave samples is given in section 6.3.
Various analyses of pipe sediment scrapings are described in section 6.4.
Settleability experiments conducted during the latter half of phase two are
described in section 6.5. Finally, a listing of all analytical procedures
used in the analysis is given in section 6.6.
6.2 Discrete Flush Have Samples
The flush wave pollutant characteristics were determined by analysis
of 17 discrete liquid samples. A background sample was taken and analyzed
for each flushing experiment. Figure 26 presents an overview of the sample
handling during the first phase operations. Each of the discrete samples for
all phases were analyzed for both Total and Volatile Suspended Solids. Analy-
ses of BOD5, COD, Ammonia, Total Kjeldahl Nitrogen, Ortho and Total Phosphate
were performed for all samples for about half of the first phase flushing ex-
periments. In addition, Total and Fecal Coliform bacterial levels were deter-
mined for all samples from the initial first phase flushing experiments.
Analyses of COD, BODc, nitrogen and phosphorus levels were determined for se-
lected samples from the second and third phase flushing experiments. Samples
analyzed for these parameters were taken at the onset of the flush wave where
peak concentrations occurred, and from the tail of the wave. These determi-
nations together with estimated levels obtained by regression with Volatile
Suspended Solids concentrations were used to characterize pollutant profiles
for these flushes. The regression procedures used to fill-in missing data
are discussed in the next chapter.
6.3 Heavy Metals Analysis of First Phase Flow. Composited Flushing Samples
During the first phase, an approximate flow-proportioned one liter
sample was made from the 17 discrete flush wave samples. Depth of flow and
cross-sectional area characteristics were used in proportioning the samples to
the one liter sample. The composite samples were allowed to settle for four
hours and a supernatant sample collected. The remaining supernatant wag care-
fully decanted and a representative sample of of the settled material collected.
64
-------
cr>
in
ss
< o
< o.-
d
to
_l O
< O
"
'SETTLE 2 I
MRS. MIN.I
SUPERNATANT
SEDIMENTS
A
FLOW PROPORTIONED
COMPOSITE SAMPLER
SCRAPINGS
FROM SEWER
AIR
METALS
ANALYSIS
DRIED SAMPLE
T
WET SIEVE
PROCEDURE
FIGURE 26 : OVERVIEW OF SAMPLE HANDLING
GRAIN
SIZE
DISTRIBUTION
VIA
SIEVE
ANALYSIS
DURING PHASE
DETERMINATION OF
PERCENT VOLATILE
OF COMPOSITE
SAMPLE
DETERMINATION OF
PERCENT VOLATILE
OF FRACTIONS
-------
The supernatant sample and the bottom settled material (sediment represented
materials which could eventually settle from the flush wave under ideal quies-
cent conditions)* are unlikely to occur in a sewer system.
Initially, all supernatant and sediment samples were analyzed for
heavy metals including Cadmium, Chromium, Copper, Iron, Lead, Mercury, Nickel
and Zinc. Shortly after commencement of the program, the results of the
supernatant metals analyses indicated concentrations in the range of less
than a part per billion. These levels were considered to be so low, especi-
ally in relation to the high concentrations found in the sediments, that sub-
sequent heavy metals analyses were only conducted on the sediment fraction.
6.4 Analysis of Solids Scrapings
Before and after flushing during the first and second phase, sedi-
ments scraped from the sewer segments were evaluated using several methods.
The primary concern of the effort was to conduct, on a representative basis,
sieve analyses of the scrapings and determine the grain size distribution of
the material as well as organic and inorganic content.
6.4.1 Wet Sieving Techniques. Initially a wet sieving technique was
employed to determine grain size distribution of the scraped materials. The
procedure was to simply mix the sample and pour it through a standard sieve
series of sieve numbers 8, 16, 30, 50, 100, 200 and pan. This range of
sieves yielding mesh openings ranging from 2.38 to 0.84 millimeters or coarse
to fine sand/coarse silt on the Massachusetts Institute of Technology (M.I.T.)
classification system. Unfortunately, levels of rag and paper meterials
present in the flushed solids were very high, causing clogging of the coarse
sieves. The procedure was discontinued after a short period of testing.
6.4.2 Dry Sieving Techniques. Two separate dry sieving techniques were
evaluated. First the scraped material was placed on pre-weighed drying
trays, and initial weight recorded. Samples were then air-dried at 68°C for
several days and dry weight recorded. A portion of the dried sample was then
saved for the heavy metals analyses. Approximately 1000 grams of dried sample
was placed in a series of preweighed sieves and shaken on a standard sieve
shaker for five minutes. The sieve series used was the same as that outlined
for the wet sieves. The percent of sample retained on such sieve and pan was
determined by weighing. The portion retained after ashing at 550°C for one
hour permitted determination of the percent volatile solids of the sample.
The other dry sieving technique used involved splitting the air-
dried sample in half. One half of the dried sample was placed through the
sieves and total weights retained on each sieve determined. The other half
of the sample was ashed at 550°C for one hour, and the ashed residue sieved
used the same procedure as previously outlined.
The latter method appeared to be more accurate since most of the
paper and cloth materials were removed prior to sieving. Since paper and
cloth residues were common to most of the scraping samples, the accuracy of
the sieve analysis on the organic portion of the sample could not be evaluated.
The technique of splitting the air dried sample and conducting two separate
*This program with its .extended settling period was aimed primarily at assess-
ing dissolved versus potentially settleable fractions.
66
-------
sieve analyses was used on the majority of samples in both the first and sec-
ond phases of the study. All sieving data generated was plotted on standard
M.I.T. classification sieve analysis paper.
6.5 Settleability Experiments
Settling characteristics of flush wave pollutants were evaluated
during the latter portion of the second phase experiments. Six different
flushing experiments were conducted in which samples were taken at each of
three monitoring manholes for three consecutive flushes. These experiments
were conducted in the Port Norfolk Street test segment. Three flow-composited
samples were prepared for each flush. Details of the experiments including
sample collection procedures and flow-compositing techniques were previously
described in section 5.5.2. Each composite sample from the first flush on a
given day was subjected to a settling column test while the samples from the
second and third flushes were settled in an Imhoff cone for one hour. Table
3 lists the analytical analyses performed on samples collected during the six
experiments.
TABLE 3: ANALYTICAL PARAMETERS MEASURED FOR THE DIFFERENT COLUMN
AND IMHOFF CONE TESTS,. SECOND PHASE PROGRAM
ANALYTICAL
PARAMETERS
TSS
VSS
COD
BOD
TKN
NH3
OP
TP
Cd
Cr
Cu
Hg
Ni
Pb
Zn
DATE
7/27/77
X
X
X
X
X
X
X
X
8/4/77
X
X
X
X
X
X
X
X
8/22/77
X
X
X
X
X
X
8/25/77
X
X
X
X
X
X
X
8/29/77
X
X
X
X
X
X
X
X
X
9/7/77
X
X
X
X
X
X
X
X
X
X
X
6.5.1. Imhoff Cone Testing. The procedure used for the Imhoff cone
tests is described in Standard Methods, Section 208F, "Settleable Matter" (16)
67
-------
One liter of well-mixed sample was poured into the Imhoff cone, while another
completely-mixed sample was analyzed for the pollutants cited in Table 3. The
sample in the Imhoff cone was allowed to settle for one hour. At the end of
one hour the volume occupied by the settled material was noted and a portion
of the supernatant was withdrawn for analysis. The fractions of pollutant
mass removals were estimated using the results of the two analytical tests.
6.5.2. Settling Column Procedures^ Three alternative methods were used
for the settling column analyses including a small column, a large column
with aeration mixing, and a large column with special gravity pre-mixing appa-
ratus.
The first settling column analysis method utilized a series of 3"
diameter, 18" deep columns with a single sampling port at mid-depth to esta-
blish removals for greater vertical velocities than those that could be
feasibly measured in a larger column. A range of vertical velocites from 6
to 0.05 fpm could be evaluated. Several attempts using the small columns in-
dicated that the small vessels were not suited for evaluating the flush wave
samples. The major problem was that often the sample collected from the sin-
gle sample port was in a zone of hindered or,.compression- settling. Flocculent
settling characteristics in the upper portion of the column were not noted.
This-method was thereafter discontinued.
The larger settling column employed during the testing program was
a 6 foot long (1.6 meters) by 6 inches (15 cm) inside diameter acrylic tube.
The 6 inch by 6 foot column size was chosen because of the volumetric con-
straint in obtaining the flush wave samples. Each test using this column
would require approximately 10 gallons (375 liters) of sample. As explained
in section 5.5.2, the sample used in the settling column testing program was
a composite formed from six individual grab samples at each sampling manhole
during the flush. Since a flow proportional composite sample-was desired a
minimum of 50 percent excess sample or 15 gallons (56.25 liters) would be
required; A 15 gallon (56.25 liter) sample was arbitrarily chosen as the
upper limit of sample that could be withdrawn from any given manhole without
disturbing the flush wave as it proceeded downstream.
The basic procedures used during the column testing program foil owed
those outlined by Zanoni and Blomquist (17). Six settleability experiments
were conducted and their dates are listed in Table 3. The first two column
tests used diffused air to uniformly suspend particulate matter throughout the
column prior to the tests. The introduction of air resulted in the floatation
of some material altering the settling characteristics of the column test. In
order to eliminate this problem the column v/as modified so that mixing could
be accomplished without the use of air. The construction of the column modi-
fication is shown in Figure 27. The last four column tests were conducted
within two hours of sample collection. No attempts were made to mechanically
mix the materials placed in the settling column prior to testing. Mechanical
mixing would break-up organic solids, thus potentially biasing the column
testing results in relation to settling within actual sewer lines. Samples
collected from the settling column experiments were analyzed for solids, or-
ganics, heavy metals, nitrogen and phosphates. The parameter coverage for
each of the experimentations is cited in.Table 3.
68
-------
3/8" I.D.
COPPER
TUBE
3/8 CLOSED
CELL FOAM
STOPPER
\.
5.75 DIA.
1/2" ACRYLIC PLATE
FIGURE 27
DIAGRAM OF SETTLING COLUMN
69
-------
During the initial column tests compressed air was used for mixing.
The air supply was introduced as the column was being filled prior to the
start of the test run. The upper ports used for sampling were approximately
one foot from the water surface to avoid the extraction of the air-entrained
solids. During the filling of the column, heavy grit and stones rapidly set-
tled out of the sample and deposited on the base of the column despite the
continuous aeration.
Two sets of column tests were performed using this type of mixing.
The samples collected on 7/27/77 were taken at 10, 20, 30, 60, 90 and 120
minutes. The results obtained from the 7/27/77 tests indicated a reduction of
approximately 65-75 percent of the TSS concentration in the first 10 minutes
of quiescent settling. In an attempt to characterize this rapid settling, the
column runs performed on the second set of samples, collected on 8/4/77 were
initiated immediately after stoppage of aeration. The other samples were
taken at 5, 10, 20, 30, 60, 90 and 120 minutes. Results of these tests are
discussed in Chapter 10.
The settling column was modified after the first two sets of settling
column tests to permit mechanical mixing of the contents. In addition to the
"flip" mixing system, the aeration stones were eliminated and larger diameter
sample ports were installed to prevent clogging due to high solids content.
The sample ports were changed to 3/8 inch from 1/4 inch, which resulted in a
shorter time period for sample withdrawal.
The column filling procedure was similar to the two previous column
tests with the exception that the initial sample was taken from the mixing
barrel. The timing of the column run began as soon as the column was returned
to the upright position.
The modified column was used for four sets of column tests conducted
on 8/22/77, 8/25/77, 8/29/77 and 9/7/77. The sampling program for the experi-
ments conducted on 8/22/77 and 8/25/77 entailed a two hour time period with
samples collected at 5, 10, 20, 301, 60, 90 and 120 minutes. In an attempt to
better define characteristics of the faster settling particles, sampling times
of 2, 4, 6, 8, 10, 20 and 30 minutes were used on the latter two sets of ex-
periments conducted on 8/29/77 and 9/7/77. The results obtained are discussed
in Chapter 10.
6.6 Analytical Methods
The procedure used for the various parameters are as described in
Standard Methods (18). These procedures were:
Suspended Solids (TSS) Section 208D "Total Non-Filtrable Residue"
Volatile Suspended Solids (VSS) Section 208D "Total Non-Filtrable Residue"
Chemical Oxygen Demand (COD) Section 508 "Oxygen Demand (Chemical)"
Bio-Chemical Oxygen Demand (BOD ) Section 507 "Oxygen Demand (Bio-Chemical)"
Total Kjeldahl Nitrogen (TKN) Section 421 "Nitrogen (Organic)"
Ammonia-Nitrogen (NHa ) Section 102-8-f "Selective Ion Electrodes .
and Probes"
Ortho-Phosphate (OP) Section 425E "Stannous Chloride Method"
70
-------
Total-Phosphate (TP) Section 425C "Persulfate Digestion"
followed by Section 425E,
"Stannous Chloride Method"
Total Coliform (TC) Section 909A "Standard Total Coliform
Membrane Filter Procedure"
Fecal Coliform (FC) Section 909C "Fecal Coliform Membrane
Filter Procedure"
Metals - Cadmium, Chromium, Copper, Iron, Lead, Nickel and Zinc concentrations
were determined from a single sample which had been pretreated fol-
lowing Standard Methods, Section 301A-VI. A separate sample and
sample handling procedure was used for mercury analysis.
6.6.1. Pretreatment of Liquid Samples for Heavy Metals Determinations
Liquid samples collected during the first phase of the study and
used for mercury analysis were not pre-digested. As a result only free mer-;
cury results were recorded. Liquid samples analyzed for mercury from all
other phases of the study followed the procedure in Section 301A-VI of Stan-
dard Methods which measured total mercury. Pretreatment of liquid samples
for other metals analysis followed Sjtandard Methods using the method outlined
in Section 301C, subsections II-5 and 6.
6.6.2. Pretreatment of Sediment and Scraping Samples
Pretreatment of solids samples followed the method outlined in
Standard Methods, Section 301C, subsection I1-6, except for the mercury sam-
ples. Sediment samples used in mercury analysis were subjected to a concen-
trated nitric acid leaching period of two hours. Throughout the leaching
period the samples were gently agitated. The nitric acid leachate was then
analyzed for mercury content. This mercury pretreatment method was followed
for the first phase samples, all other mercury samples were pretreated using
the digestion procedure in Standard Methods, Section 301A-VI.
6.6.3. Heavy Metals Determination
Mercury determinations followed Standard Methods, Section 301A-VI
using cold vapor atomic absorption. Copper, zinc, and iron determinations
were performed using flame ionization atomic absorption, Standard Methods,
Section 301A. Direct aspiration into an air-acetylene flame was used. A
carbon cup atomizer was used for determination of cadmium, chromium, lead and
nickel because of their low concentrations. A Varian A-6 atomic absorption
spectrophotometer fitted with a carbon cup atomizer was used.
71
-------
SECTION 7
COMPUTATIONAL METHODS
7.1 Introduction
This chapter describes various computational procedures used to
estimate the quantities of pollutants transported from the flushing segments.
Conversion of the non-steady state flush wave stage levels into discharge
proved to be a difficult but essential detail in converting discrete flush
wave pollutant concentrations into quantities of mass transported. The motiva-
tion for investigating the alternative flow computational procedures presented
in this chapter was to develop reasonably accurate estimates of the non-steady
state flush wave flow rates from the field depth of flew measurements.
Data processing of the analytical laboratory results of samples
taken during the flushing experiments and pertinent physical field infor-
mation are described in section 7.2. Approaches used in the estimation of
missing flush wave pollutant concentration levels are described in section
7.3. This step was necessary since total flushed mass transport estimates
were computed using discrete values of discharge and flush wave pollutant
concentrations at fixed intervals in time. Samples were not always ana-
lyzed for all analytical parameters and during several experiments an
incomplete set of samples was taken.
Alternative procedures are described in section 7.4 for estimating
instantaneous flush wave discharge values from stage level readings. Three
alternative approaches are presented: 1) the application of Manning's
equation using plan pipe slope and variable roughness coefficients;
2) utilization of Manning's equation with a virtual slope derived from
least squares fitting of that equation to steady-state field flow calibration
points; and 3) utilization of mathematical programming techniques for
determining parameters -of complex loop-rating curves. Comparison of inte-
grated flush volumes from predicted flow rates with known flush delivery
volumes indicated that the first approach grossly misestimated the actual
flow rates. The second approach substantially improved the predicted
results but neglected to account for the unsteadiness of the flow regime
because of the strict application of the steady uniform flow rating
curve. Actual field observations and measurements suggested that the
flush wave hydraulic characteristics are best described by a looping
stage-discharge curve with higher flow rates in the front of the wave
than in the back of the wave for similar flow depths. The third approach
described this phenomena in an extremely reasonable way and was therefore
selected for converting flush wave stage recordings into discharge. Computed
flush rates are compared with field measured flush rates using dye injection
procedures described in Chapter 5.
72
-------
Section 7.5 describes the procedure for estimating the flushed pollutant
masses for all phases of work.
7.2 Raw Data Handling
Subsequent to laboratory analysis of the samples collected during
each flushing experiment, relevant information were data processed for
further handling and analysis. This information consisted of: a) the
street name, date and hour of the flush; b) a brief characterization of
the flushing technique used indicating whether the flush was a pressure
flush, a gravity flush or a backup and release flush; c) whether or not
the upstream manhole was blocked during the flush; d) whether time of
travel measurements using dye has been performed; e) the truck delivery
volume, in cubic feet,used in flushing the pipe segment(s); f) .the flush
duration, that is, the time, in seconds, during which the flush volume
was introduced in the pipe; and g) the recorded time, in seconds, elapsed
between the opening of the quick-action gate valve in the truck and the
instant a sudden rise in the water level in the downstream manhole was
noted. Next, data cards for each collected sample were prepared containing:
a) an order number of the sample collected; b) the time in seconds between
the present and previous sample; c) the flow depth in the pipe at the
time the sample was collected; and d) concentrations, in mg/1, of COD, BOD,
TKN, NH3, TP, OP, TSS, VSS, and Total Coliform and Fecal Coliform bacteria,
in colonies/100 ml. The last data card per flushing experiment contained:
a) pipe diameter; b) the average of the sediment depths at the downstream
manhole measured prior to and after the flush; c) the pipe slope; and
d) the Manning's roughness coefficient believed to be appropriate for
the particular flush. Information contained on this last card was subse-
quently reassessed and improved through an optimization technique described
in Section 7.4. Typically, cards were prepared for each flush. A sample of
the data ca.rds showing the results of the 10/04/76 experimental flush for
the Walnut Street test segment is shown in Table 4.
Once all the data cards for one particular phase of the project
were punched and edited, disk files were created for further editing,
processing, and analysis.
7.3 Missing Data Fill-in Procedure
In general, TSS and VSS were determined for all samples collected
during all three phases of the field program. Determination of COD, BOD
TKN, NHs, TP and P04 were only conducted for selected flushes. Further-
more analytical determinations were not always performed for all 18
samples collected during each flush. As a consequence there were gaps in
the data requiring filling-in before the flushed pollutant masses could
be computed. The number and time distribution of the missing data varied
during the three phases of experimentation, requiring different criteria
for estimating missing flush wave pollutant concentration points.
73
-------
TABLE 4. SAMPLE OF DATA CARDS FOR THE FLUSH OF 10/04/76 AT WALNUT STREET
4
WALNUT STREET
GRAVITY FLU5H
NO 0 35
A
* 0
** i
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
*** 15.
B
0-
o.
10-
20.
30.
40.
50-
6Q.
70.
80.
100-
120-
140.
160.
180.
200.
220.
240.
.5 0
.0 71
C
E>.5
5.75
6.
6.25
6.5
6.75
6.5
6.25
6.25
6.25
6.25
6.
6.
6.
5.75
5.5
5.5
WITH
.5 54
D
920.
840.
2120.
2^80.
2120.
208.
1630.
2720.
2240.
1630.
184-0.
1200.
•00248 0
10 04 76 ll:05 AM
3 INCH NOZZLE AND HOSE* UPSTREAM MH NOT BLOCKED
.0
E
960.
1290.
1140.
84-0.
720.
720.
795.
750.
740.
577.
585.
315.
.015
F
83.
95.
100.
95.
127.
102.
83.
81.
71.
S3.
45.
7.
G-
35.
30.
31.
32.
3b.
31.
26.
24.
23.
iO.
7.
3.
H
28.2
29.4
29.0
30.2
34.2
28.2
25.0
20.4
20.0
13.6
11.6
5.2
I
22.0
20.7
24.1
27.2
22.3
22.6
17.1
14.5
14.0
(.6
5.2
3.0
J .
22Q2.
5551.
3698.
3155.
2584.
1*43.
1*47.
1883.
1509.
1^19.
1477.
1252.
Ib54.
1466-
1627.
*22.
664.
K
1847.
3855.
27*2.
2347.
18/6.
1328.
1271.
1266.
961.
9b4.
9*8.
853.
1101.
1045.
11^4.
663.
4*3.
L
14000000.
49000000.
41000000.
1450QOOO.
looonooo.
11000000.
13500000.
14600000.
7400000.
7200000.
4000000.
M
7000000.
7QOOOOO.
9500000.
1100000.
1800000.
1100000.
500000.
270000.
320000.
KEY: A - Sample number, where (*) implies background sample and (**) implies sample taken
at first occurrence of wave.
B - Time interval of sampling.
C - Depth of flush wave (inch).
D through M - COD, BOD, TKN, NH3, TP, OP, TSS, VSS, TC and FC, respectively.
*** - Pipe diameter (inches), average sediment depth (inch), pipe slope and
estimate of Manning's coefficient.
-------
7.3.1 Estimation of Missing Data - First Phase
Out of the 83 flushing experiments 44 were analysed for pollutants
other than TSS and VSS and, of these, 19 were analysed for only BOD. In
general, sample numbers 1 through 9, 12, 15 and 18 were anlyzed for
pollutants other than TSS and VSS. The approach was to monitor over the
initial peak concentration period and then at the tail of the wave. In
a few cases gaps existed between the first nine samples and/or in the
later part of the flush wave. Linear interpolation of the missing values
between two known values was used to estimate the gaps in the first phase
flush data. Interpolation of missing data was only used for computing the
pollutant masses removed, so that the original data files remained unaltered.
7.3.2 Estimation of Missing Data - Second Phase
In the second phase 18 serial flushing experiments were performed
on Port Norfolk Street, with samples collected at three successive manholes
downstream from the manhole where the flush wave was introduced. Eighteen
sample sets were collected at the first and second downstream manholes and
15 sample sets were taken for the third downstream manhole. TSS and VSS
determinations were performed for all samples collected at each manhole
during this phase of work. Samples depicting the frontal portion of the
flush wave were analyzed for COD for all flushes. For several flushes
samples from the frontal portion of the flush wave were analyzed for BOD and
for a few flushes,TKN determinations were performed in a similar manner.
Typically, 5 to 6 samples out of the first 9 collected were analyzed for COD
and BOD. The number of TKN analytical determinations were too few to justify
presentation of results in terms of the masses of TKN removed by the flushes.
Only missing COD and BOD values were estimated using the following regression
relationships:
COD (rag/a) n 7QI-
Port Norfolk St. COD = 6,018 VSSu'/yb (R = 0.92) (!)
BOD (mg/A) , n R11
Port Norfolk St. BOD = 1.907 VSSu
-------
each flush. TSS and VSS determinations were performed on all samples
for all seven flushes. COD analyses were also performed for all flushes
on samples collected from the front part of the flush wave and a few
toward the end of the wave. Although the missing COD concentrations
could be estimated by interpolation, the time gaps between two successive
values at the tail, for half of the flushes, were large (150 seconds
out of 230 seconds for the entire flush) precluding the use of simple
linear interpolation. Regression of the missing COD data on VSS was
again utilized. The regression equation used to complete the missing
COD values is:
COD = 3.1995 VSS0'9141 (R = 0.91) (3)
This relationship was derived from flush data at Shepton Street.
7.4 Stage Discharge (H-Q) Relationships
In chapter 5 procedures for determining flush wave stage levels
were described. These recorded flow depths required translation into
flow rates in order to convert pollutant concentrations associated with the
discrete estimates of flow rate into mass rates transported from the flushed
pipe segment. Several alternative procedures for converting stage into
discharge were investigated. The performance of each procedure was evaluated
in terms of how well the estimated flow rates, when integrated over the
flush duration, could reproduce known input flush volumes for each flush.
The procedures considered are the following:
1. Application of Manning's equation assuming uniform flow, with
variable roughness coefficient, n, and pipe slope computed using manhole
elevations and segment length;
2. Application of Manning's, equation assuming uniform flow, with
variable n and a slope derived from a least square fit of the slope to field
determinated steady state stage-discharge calibration points; and finally,
3. Establishment of stage discharge relationships through a
mathematical optimization procedure that accounts for the unsteady nature
of the flow regime of the flush wave while minimizing error between computed
and measured flush volumes.
7.4.1 Definition of the Flush Input Volumes
Since the flush volumes were used to gage the relative
predictive precision of the two initial approaches and directly in the
third procedure for establishing the stage discharge relationships, it was
therefore important to accurately determine delivered flush volumes. Input
flush volumes were measured by a 4 inch turbine water meter located on the
flush truck which was repeatedly calibrated during the project using a large
vessel of known volume. For flushes where the upstream manhole was not
blocked, the additional volume due to the upstream flow over the duration
of the flush was added'to the metered flush volume to yield the total flush
76
-------
volume. Background flow contributions were endogenously estimated in the
flush flow rate computations which are discussed in Section 7.4.4.
Except in the cases of unblocked manholes, metered flush truck volumes were
finally used as the total quantity of flush water. This conclusion was
reached after small and extraneous flow contributions and metering errors
were considered in a sensitivity analysis to be negligible.
These considerations tended to either raise or lower the metered
truck volumes. Additive sources of extraneous flow included hydrant leakage
into gutters, catchbasins and eventually into the combined sewer test
segments for Port Norfolk and Walnut Streets; ground water infiltration along
the flushed segment; and, house connections along the flushed pipe segments.
Factors tending to counter balance these additive effects were slight positive
water meter bias and the fact that, for most of the flushes, the last flush
wave stage level was on the average 20% higher than the initial background
depth, indicating that the flush wave had not completely passed the sampling
manhole by the time the last smaple was collected. Rough order of magnitude
estimates for each factor and source and their effect on the metered flush
volumes follow. .
Hydrant Leakage. Hydrant leakage did not affect the flushes
at Templeton and Shepton Streets since the sewers are separated. Leakage
at Walnut Street was negligible, but could be as high as 2 cf at Port
Norfolk Street, particularly during the high pressure flushes. This
estimate resulted from computations involving gutter water depths,
gutter shape and slope and duration of the time the hydrant was leaking.
Ground Wjter Infiltration. This contribution was small for
the first phase flushes that were conducted from late August to mid-November.
Infiltration was higher for the second phase program covering the period
from January to the end of March. In the third phase of work, occurring
between late September to mid-October, the infiltration was again low. It
was estimated that ground water infiltration would range up to 1 cf during
the 4 minutes of sampling. In terms of aerial infiltration rates, this
figure corresponds to 2300 gallons/acre/day at Port Norfolk Street and 4000
gallons/acre/day at Walnut Street.
House Connections Along Flushed Segment. This contribution
was the hardest to assess in terms of generalized average values because of
the random nature of household inflows over the short periods of flushing.
Using estimated population along each experimental segment and estimates of
the number of per day uses of major household fixtures per capita and assuming
that the number of occurances of such uses in a short interval of time is
Poisson distributed (and independent of time of day), it was estimated that
there is more than 95% chance that the number of discharges in a four-minute
period is less than or equal to the values shown below:
MAXIMUM NUMBER OF ESTIMATED HOUSEHOLD DISCHARGES IN A 4 MINUTE PERIOD
STREET
Templeton
Shepton
Port Norfolk
Walnut
NO. OF DISCHARGES
3
4
2
77
-------
Estimates of dry weather sewage contributions for a four minute interval
using an average per capita waste rate of 85 gpd (4) are given below:
AVERAGE WASTE VOLUMES IN A 4 MINUTE PERIOD
STREET
Tempi eton
Shepton
Port Norfolk
Walnut
*
Contributions
POPULATION*
50
70
33
14
alonq the pipe
AV6. VOL.
IN 4 MIN. (
1.58
2.21
1.04
0.44
segment flushed.
CF)
Using the expected maximum number of discharges and assuming a conservative
figure of 1 cf per discharge, t.he probability is small that, if the contribu-
tions did occur during the flushing period, the estimated sewage volumes
would be approximately 1, 2, 3 and 4 cf for Walnut, Port Norfolk,
Tempieton and Shepton Streets, respectively.
Water Meter Inaccuracies. Meter calibration was conducted by
filling a vessel of known volume at different indicated rates. At the
onset of phase one, the errors in the flow-rate range of 0.5 to
1.0 cfs were found to be random and within 2.0%. Near the end of the
first phase flushing experiments, it was determined that between the flow
rates of 0.5 and 1.0 cfs, the range of most of the flushes, the meter, on
the average, tended to overestimate the delivered volumes by about 6% at
0.5 cfs to a maximum of 7.5% at 1.0 cfs. The average error at 0.1 cfs was
of about 2%, 3.5% at 0.2 cfs, and 4.5% at 0.3 cfs. Later testing of a new
replacement meter indicated errors averaging around 5% for flow rates
between 0.5 and 1.0 cfs.- Using the above results, it is estimated that the
metered flush volumes during first and second phase programs were over-
estimated by about 3.5% to 7.5%. This bias amounts to a minimum absolute
value of 1.23 cf for a 35 cf flush, and up to 3.75 cf for a 50 cf flush.
Wave Tail Considerations. It was mentioned earlier in this
section that the last stage reading recorded during the flushing experiments
was higher than the initial background depth for most flushes during the
first and second phases. For the first phase flushes, the final depth was
on the average 20% higher than the initial background depth of flow. This
percentage was lower for the second phase flushes.
These observations are summarized in Table 5. Columns 3, 4 and 5
of Table 5 indicate, by phase and by street, the number of flushes in which
the last depth of flow was respectively higher, equal to and smaller than
the initial background depth of flow. The mean and standard deviation, in
inches, of the differences between the last and first depth, computed for
flushes for which the last flow depth exceeded the initial depth are given
in Column 6. These results show that the average difference did not exceed
1 inch. The residual depth implies continued flush" wave movement. What
fraction of the original flush volume still remained in the flushed sewer
pipe segment is extremely difficult to establish. It is believed that any
78
-------
residual volume still remaining in the flushed pipe at the time of the last
measurement would be roughly several cubic feet. The following discussion
is provided to support this conclusion.
TABLE 5. CHARACTERISTICS OF THE TAIL OF THE FLUSH
WAVE AND FLUSH WAVE VELOCITY
(1)
Phase
1st
2nd
(2)
Street
Tempi eton
Shepton
Port Norfolk
Walnut
Port Norfolk
1 DMH**
Port Norfolk
2 DMH
Port Norfolk
3 DMH
(3)
Number
H18 > Hl*
17
20
17
17
11
15
14
(4)
of Flushes
H18 = Hl
4
2
2
4
6
3
-
(5)
where:
H18 < Hl
_
-
-
-
1
-
1
(6)
Mean/S.D. of
H,0 Above H,
18 (inches)1
0.94/0.42
0.84/0.58
0.63/0.43
0.82/0.25
0.35/0.17
0.77/0.29
0.96/0.65
(7)
Mean/S.D. of
Wave Velocity
(fps ) 1
1.92/0.52
1.77/0.46
1.88/0.45
1.54/0.31
1.85/0.32
-
-
H
**
= background depth.
= last depth measured.
1 DMH = first downstream manhole.
The last column in Table 5 presents minimal estimates of the mean
and standard deviation of wave velocities computed over all flushes for a
particular street using the pipe segment length and the recorded time elapsed
between the start of the flush and the time taken for the peak of the wave to
reach the first downstream manhole. This procedure yields low wave velocity
estimates because the travel times are referenced from the start of injection.
An approximate estimate of the location of the wave peak downstream of the
sampling manhole can be computed at the time of the last measurement using
'these velocities and the time between the passage of the peak of the wave
and the last depth measurement taken.
The time intervals between the passage of the wave peak and the
last depth measurement were, for the great majority of the flushes, between
190 to 200 seconds. On the basis of the wave velocities indicated in Table 5.
it is estimated that the flush wave peak should be between 300 ft. (Walnut
Street) to 385 ft. (Tempieton Street) downstream from the sampling manhole at
the time the last depth of flow was measured. These estimated distances
represent more than one, sometimes two pipe segments downstream from the
sampling manhole. These calculations indicate that by the time the last depth
of flow was measured most, if not all, of the flush wave would have passed.
Additional evidence is provided by the dye injection flow monitoring
experiments described in Section 5.7. Continuous recording of dye concentrations
79
-------
for a number of typical flush waves showed a considerable velocity decrease
at the end of the wave. These observations indicate that the flush flow rate
at the time of the last depth of flow measurement was slightly greater than
background flow rates. These results together with the earlier crude time of
passage calculations suggest that a nominal 5% estimate of the metered
volume be assumed as the residual volume in the upstream pipe segment at the
time the last sample was collected.
Assessment of the factors affecting the metered flush volumes
are summarized in Table 6. The first three factors represent volumes that
are additive to the nominal metered volumes while the last two factors
represent negative corrections to the metered flush volumes. Subtotals
of the ranges of both the positive and negative corrections are also
presented and indicate that the nominal metered values should be slightly
above the volume that could be computed from the depth measurements. In
other words, the metered volumes are likely to slightly overestimate the
volume of water discharged into the sampling manhole, that is, the
estimated subtotal values differ by only a few cubic feet on each street.
TABLE 6. ESTIMATED RANGES OF POTENTIAL BIASES TO THE METERED
FLUSH VOLUMES (cf)
Street
1. Hydrant leakage
2. Ground water
Infiltration
3. Household discharge
Subtotal
4. Meter bias
5. Wave tail
considerations
Subtotal
Tempi eton
+(0-1)
+(0-3)
+(0-4)
-(1.23-3.75)
-(0-2.5)
-(1.23-6.25)
Shepton
+(0-1)
+(0-4)
+(0-5)
-(1.23-3.75)
-(0-2.5)
-(1.23-6.25)
Port Norfolk
+(0-2)
+(0-1)
+(0-2)
+(0-5)
-(1.23-3.75)
-(0-2.5)
-(1.23-6.25)
Walnut
+(0-1)
+(0-1)
+(0-2)
-(1.23-3.75)
-(0-2.5)
-(1.23-6.25)
It was assumed that all these factors would tend to balance, withstanding
•the apparent slight bias. The uncorrected metered flush volumes were
taken as the yardsticksin judging the suitability of a particular stage-
discharge relationship. A stage-discharge relationship was deemed desirable
when the computed flow rates, integrated over the sampling period, yielded
a volume approximating the metered flush volume for that flush.
7.4.2 Discharge Estimates: Manning's Equation with Pipe Slope
The use of Manning's equation to compute discharge in sewers from
80
-------
flow depth measurements is of universal practice in environmental engineering.
In foot-second units, Manning's equation is given by
, .Q 2/3 cl/2
Q = 1-49 r 7 S a
where: Q = flow rate (cfs);
R = hydraulic radius (ft);
S = energy slope (ft/ft);
a = flow area (ft*); and
n = roughness coefficient.
Application of Manning's equation usually assumes steady uniform
flow by taking S .= S , the pipe slope. It is also a common practice to
assign a constant vaTue for the roughness coefficient n according to pipe
material, age and state of maintenance. Values of 0.013 and 0.015 are
often used in sewer system computations.
As a first approximation Manning's equation was used for the first
phase flushes, with two refinements:
1. The average of the sediment depths measured prior to and after
the flush was taken into account since sediment depth changes the pipe shape,
the hydraulic radius, and the area of flow; and,
2. The Manning's roughness coefficient n was assumed variable with
the flow depth.
The physical slopes of the test pipe segments, determined from
construction drawings, at the four streets are:
Street SI ope(ft/ft)
Tempieton 0.00322
Shepton 0.00345
Port Norfolk 0.00486
Walnut 0.00481
In all four cases the values of n varied from 0.015, at full pipe, to
a maximum of 0.01935 at a depth around 1/3 of the pipe diameter.
Manning's equation was applied to the 87 flushes of the first
phase to compute the flow rates from the flow depths measured in each flush.
The total volume per flush was computed by a discrete integration of the flow
rates over the flush duration. The computed flush volumes are plotted in
Figure 28. versus the measured input flush volumes for each test segment.
.Except for the Shepton Street flushes and one flush in Tempieton Street
the plotted points fall all on one side of the 45° equivalence line. The
rating curves would therefore substantially misestimate the flush volumes.
Direct application of pipe slopes in Manning's equation for computing flow
rates would have greatly misestimated flush pollutant masses.
81
-------
160-
o
I
u
120-
80.-
3
CO
2
40.-
NUMBER OF FLUSHES PLOTTED'16
« • •«
I60r
u.
o
I
llj
120.-
80.-
u
a:
CO
<
u
40. 80. 120 160
COMPUTED FLUSH VOLUME - CF '
A) TEMPLETON STREET
5=0.00322
NUMBER OF FLUSHES PLOTTED" 18
160.-
u.
o
I 120.-
Ul
80-
CO
u.
111
IT
CO
<
Ul
NUMBER OF FLUSHES PLOTTED* 17
40.
120.
(60r
U.
o
120r-
ui
1
80H
CO
40.-
160
NUMBER OF FLUSHES PLOTTED'20
0 40 80 120 160
COMPUTED FLUSH VOLUME - CF
B) SHEPTON STREET
S'0.00345
0 40. 80. 120. 160.
COMPUTED FLUSH VOLUME - CF
COMPUTED FLUSH VOLUME - CF
C) PORT NORFOLK STREET D) WALNUT STREET
5=0.00486 5'0.00481
Obs.: Several flushes may coincide in one plotted point
FIGURE 28 PLOTS OF MEASURED VERSUS COMPUTED FLUSH VOLUMES
MANNING'S EQUATION (PLAN a PROFILE SLOPE)-FIRST PHASE
82
-------
7.4.3 Discharge Estimates: Fitted Slope to Calibration Points Using
Least Squares
One improvement to the straightforward application of Manning's
equation is to adjust either the slope S and/or the roughness coefficient n.
The second alternative procedure for establishing the required stage-discharge
relationships was to utilize field stage-discharge calibration results in a
least squares approach to determine an adjusted slope S. Steady state stage-
flow measurements were conducted using the flush truck and are described in
Chapter 5. The roughness coefficient, n, was assumed variable in this analy-
sis.
Extensive laboratory and field experiments (19, 20, 21 & 22) indicate
that the value of n varies with the depth of flow in the pipe. A smooth
curve relating the ratios n/nf and h/D, where nf is the value of n at full
pipe depth was used (22). The value of n at full pipe^n-p) and its
variation with depth was assumed known and the slope, S, was taken as a
surrogate for all the uncertainties with respect to S itself and the roughness
coefficient n.
A number of the hydraulic parameters in Manning's equation can be
expressed as a function of the ratio of flow depth to pipe diameter.
Equation (4) is rewritten as:
r.2/3 s1/2 a.
Q. = 1.49 -J - — - ! (5)
where: Q. = f(h./D), flow rate (cfs);
r] = f(h]/D), hydraulic radius (ft);
S = constant, pipe slope (ft/ft);
a, = f(h.j/D), wetted area (ft*);
n^ = f(h./D), Manning's friction factor;
h. = water depth in the pipe, (ft); and
D = pipe diameter, (ft).
Assuming that the roughness coefficient is known and allowing the
slope to take up all the uncertainties with respect to the true hydraulic slope
and roughness coefficient used, the objective function of a weighted least
squares fitting is written as:
min OF(S) = Z
O I ~™ 1
alu
2/3
where: Q. = observed values of flow from the calibration and
1.49 a. r.
F(n./D)=
m = number of observed Q.; and all other variables are as defined
before.
83
-------
Setting dOF(S)/dS = 0 and solving for S the expression for the
"best-fit" value of S is given by:
S =
m
Z
1=1
1.49
m r.2/3 a^
ri ~\ |_
ni
0.5
(7)
Several other approaches could have been used to fit Manning's
equation to the data points. One intuitive approach is to optimize the
roughness coefficient, n, instead of the slope, S. The variability of n with
h/D can be expressed by an appropriate mathematical expression and the optimi-
zation of the variable n function by non-linear means would be possible. The
procedure to fit slope S, using least squares,, is much simpler and was there-
fore adopted. Least squares results using equation 7 are presented for each
test segment as follows:
Street
Temple ton
Shepton
Port Norfolk
Walnut
Number of Field
Calibration'
Points
6
9
4
6
Best-Fit
Slope
0.00115
0.00388
0.00141
0.00221*
Correlation
Coefficient
0.95
0.95
0.91
0.84
Assuming a dead zone of 3.85" at the bottom of the pipe.
An illustration of the alternative stage-discharge curves discussed
to this point for the Port Norfolk and Shepton Street test segments are
plotted in Figures 29 and 30. The field calibration points are circled.
Four alternative curves are shown in each figure representing:
1) least squares fit of the slope assuming a variable roughness
coefficient;
2) least squares fit of the slope assuming a constant roughness
coefficient n = 0.015;
3) discharge curve given by Manning's equation using plan slope
and n = 0.015; and,
4) discharge curve given by Manning's equation using the plan
pipe slope and n = 0.013.
Curves labelled 1 and 2 in Figure 30 for the Shepton Street site differ only
slightly in the range of the available calibration points. Only curve
number 1 was plotted in that range for clarity. Curves labelled 3
and 4 derive from simple application of Manning's equation described in
section 7.4.2.
84
-------
2.5-
2.0-
1.5-
o
o
LL
1.0-
0.5-
PORT NORFOLK ST.
Curve |: Best Fit §=O.OOI4l,n= variable
Curve 2! Best Fitt=O.OOI02,n=O.OI5
Curve3: Plan Slope = 0.00486 n = 0.015
Curve4: Plan Slope = 0;00486n=O.OI3
© CALIBRATION POINTS
FIGURE 29
10
12
FLOW DEPTH —INCHES
COMPARATIVE STAGE - DISCHARGE CURVES
85
-------
2?
o
2.5-
2.0-
1.5-
0.5-
0.0-
SHEPTQN ST.
Curve I: Best Fit S= 0.00388, n= variable
Curve 2: Best Fit S = 0.00243, n = 0.015
Curve 3! Plan Slope = 0.00345,n = 0.015
Curve4:Pian Slope = 0.00345,n =0.013
©CALIBRATION POINTS
FIGURE 30
10
12
FLOW DEPTH —INCHES
COMPARATIVE STAGE-DISCHARGE CURVES
86
-------
Curve number 1 represents the proposed solution, against which the
other three will be compared. Curve number 2 results from a least squares
fit of the slope to the calibration points assuming roughness coefficient at
a constant value of 0.015. Examination of curves 3 and 4 show that the
least squares fit with the variable roughness formulation is superior to the
curve with fixed roughness coefficient. Correlation coefficients computed
over the pertinent range of observed h/D values for the Port Norfolk and
Shepton Street segments support this conclusion and are given below:
Street
Port Norfolk
Shepton
Range of
Observed h/D
0.21 - 0.83
0.13 - 0.50
Variable n
(nf = 0.015)
0.91
0.95
Fixed
n = 0.015
0.57
0.93
The two examples also show that, although the two curves are not
considerably different in the lower to mid range of depth of flow, the
divergence increases at higher depths. Its implication is that the use of
variable n becomes more important at the higher flow depths. The
lack of a calibration point above the depth of 6 inches precludes a similar
conclusion for the Shepton Street segment. Both curves 1 and 2 in Figure 30
represent extrapolation for depths beyond six inches. In that range, curve
number 1 is likely to yield more precise flow estimates then curve number
2. The higher correlation coefficient for curve 1 does not warrent this
conclusion but manipulation of the least squares analysis for Port Norfolk
provides some evidence to support this assertion.
Three low to mid range depth of flow calibration points for Port
Norfolk were used in a least squares fit of Manning's equation, using both
variable (nf = 0.015) and constant roughness coefficients (n = nf = 0.015).
The resulting slopes were then used to estimate the flow rate at the highest
available calibration point, i.e., h = 10 inches. The calibration curve
using constant n underestimated the flow rate by 18% whereas the variable n
curve underestimated the same flow rate by about 7%. The correlation coeffi-
cients computed for the three calibration points are similar, 0.87 for n
variable and 0.85 for n constant. When the fourth calibration point from
the Port Norfolk Street data set, at h = 10 inches, was introduced in
the regression, the correlation coefficients in the above table changed to
0.91 and 0.57, respectively. If similar behavior can be assumed for
Shepton Street, introduction ofva higher calibration point would imply
a slight increase in the correlation coefficient for the variable n case,
and a sharp drop on the correlation coefficient for the fixed n case.
Therefore, for Shepton Street the extrapolation beyond six inches should be
more precise using the stage-discharge curve incorporating the variable
roughness formulation.
Curves 3 and 4 in Figures 29 and 30 were computed simply using
pipe slopes and fixed n values of 0.015 and -0.013 in Manning's equation.
A comparison of these curves with curve 1 in both figures illustrates the
importance of calibrating Manning's equation in any flow measurement analysis
in sewerage systems. The larger differences noted in Figure 29, for Port
Norfolk Street, are typical of the results for the Templeton and Ijalnut
Streets test segments.
87
-------
The first phase flush volumes were computed for all four streets
using least squares best-fit slope and the variable n roughness formulation.
A comparison between the computed and input volumes is presented for the four
streets in Figure 31. Comparison of Figures 28 and 31 illustrates the
improvements achieved by using a calibrated slope in Manning's equation. In
Figure 28 practically all plotted points fall to the right of the 45° line,
indicating misestimation of the flush volumes, while in Figure 31 the
plotted points are distributed in a relatively narrow band around the 45°
line.
Remaining differences between predicted and measured flush volumes
are attributed in part to the assumption of steady uniform flow which under-
lies the applications of Manning's equation. If the unsteady non-uniform
nature of the flow is accounted for, the accuracy of the predicted flush
volumes and the flow rates throughout the flushes will improve. The non-
steady state stage-discharge result will more accurately estimate the flushed
pollutant masses. A numerical procedure accounting for the flush wave
unsteadiness was developed in this study and is described in .the following
section. It should be noted that the two alternative procedures described
thus far were only used in the early stage of work, when the first phase
flush results were available. The last formulation described in the next
section was used to compute flush volumes for all three flushing phases.
7.4.4 Discharge Estimates: Loop-Rating Curve
The relationship between stage and discharge at a cross-section
of a free surface channel flow is unique only if the flow is uniform. During
the progress of a flood wave, energy slope terms neglected in Manning's
equation, other than the bed slope, SQ, cease to be negligible as compared
to S and the discharge is no longer a function of depth alone. For a given
depth of flow, h, the discharge will be greater on the rising stage of the
wave than on the falling stage, so that the stage-discharge curve will form
a closed loop which is characteristic of conditions under which the wave was
generated. Such a loop-rating curve is illustrated in Figure 32.
The solution of this unsteady non-uniform flow problem would be
extremely complex under the conditions present in the flush experiments of
this study using hydraulic theories of flow in open channels. In light of
the many uncertainties regarding the conditions of the flush, too many sim-
plifying assumptions would be necessary, casting serious doubt on the relia-
bility of an approach based on pure non-steady state hydraulic considera-:
tions. Unknowns such as the behavior of the net a short distance from the
nozzle, storage effects of the laterals and disturbances caused by flow con-
tributions along the segment, would preclude an accurate hydraulic charac-
terization. A simple numerical optimizing approach was used in which an
arbitrary but reasonable loop-rating curve, developed around Manning's
equation, was defined for each flush as to minimize the differences between
the estimated and the measured flush volumes.
88
-------
160-
120-
80r-
40r
NUMBER OF FLUSHES PLOTTED'16
40. 80. 120
COMPUTED FLUSH VOLUME
A) TEMPLETON STREET
?«0.00115
160.-
o
I IZOr-
tu
80.-
40r-
NUMBER OF FLUSHES PLOTTED'I?
160.-
o
120.-
o
80.-
V)
u.
Q 40H
u
oe
CO
Ul
160
CF
NUMBER OF FLUSHES PLOTTED'IS
0 40 80 120 160
COMPUTED FLUSH VOLUME - CF
B) SHEPTON STREET
^.0.00388
0 40. 80. 120. l(
COMPUTED FLUSH VOLUME - CF
160.-
u.
o
1 I20.H
IU
5
80-
co
40.-
CO
<
tiJ
2
0
NUMBER- OF FLUSHES PLOTTED'20
o 40. so. 120. tea
COMPUTED FLUSH VOLUME - CF
C) PORT NORFOLK STREET D) WALNUT STREET
T'0.00141 T*0.00221
Obs.: Several flushes may coincide in one plotted point
FIGURE 31 PLOTS OF MEASURED VERSUS COMPUTED FLUSH VOLUMES-
MANNING'S EQUATION (LEAST SQUARES SLOPE)- FIRST PHASE
89
-------
LU
S
<
o
CO
Q
UNIFORM FLOW RATING CURVE
STAGE
FIGURE 32 TYPICAL LOOP - RATING CURVE,
SHOWING THE PROGRESS OF A FLOOD WAVE
90
-------
7.4.4.1 Overview of the General Methodology. The methodology applied
in the computation of the flow rates and ultimately the masses transported
by all flushes performed in this study consisted of the following steps:
1. Define for each pipe segment and for each phase the coefficients
of a complex stage-discharge function that minimizes the sum of the squares
of the differences between the estimated flush volumes and their correspond-
ing measured input volumes. These computations consider all flushing
experiments conducted for a particular pipe segment. A mathematical
programming optimization package was used to determine the coefficients of
the stage-discharge function. Figure 33 outlines the approach.
2. The complex stage-discharge function containing the set of opti-
mized coefficients defined in Setp 1 was used to compute the flow rates,
mass rates, and ultimately the total flushed masses removed by each flush
at a particular site and phase.
7.4.4.2 General Overview Details of Optimization Mode j_. The overall
optimization approach to find the coefficients of the stage-discharge function
will be first described, leaving the details of the stage-discharge function
to be described in section 7.4.4.4. Assume for the moment a .general function
F-J^J relating stage to discharge, and define:
B = set of flushes performed at one pipe segment in one flush phase;
Vm.. = measured input volume of flush i, i
h- . = jth flow depth reading during flush i (j=l, 18, V flushes);
' s J
Q. . = flow rate corresponding to depth h. ., given by
I > J I 'J
F(h. ,, h. ., hmav , (Vm., V.)) = F. . = complex function
1,1 I »J IllaX • • II I ,J
I »J
relating stage to discharge for each flush and defined by a set of parameters;
h. i = initial flow depth of flush i;
i » '
h = maximum flow depth of flush i;
18
Ve. = T (Q. . + Q. -,1)/2 • At., estimated flush volume i;
1 i^ T sj I sj"*"! J
At. = time interval between depth readings h. . and h.
j i »j i »
(j = 1, 18, V flushes)
The objective function to be optimized is given by
2
minimize Z = I (Vm.. - Ve1 ) , V. :>B
All i
91
-------
Compute Objective Function
Valu*
FIGURE 33 - SIMPLIFIED FLOW CHART OF THE
OPTIMIZATION PROCESS
92
-------
The minimization of equation 8 was performed using a direct
search algorithm devised by Hooke and Jeeves (23) to provide step changes
on the parameters of the function Fisj until the optimum is found. Figure 33
is a simplified flow chart of the optimization process. A set of parameter
values is initially assumed for FJSJ. In the context of the optimization,
these parameters define a point in' the E^ space of the search, n being the
number of parameters to optimize. The flushes of a given phase and at a
particular pipe segment are taken one by one. Cpnsidering the first flush,
its measured flow depths are converted into flow rates through F-\tj and its
initial parameters, yielding a discharge hydrograph for the flush. This
hydrograph is then integrated by a discrete approach, as given by
18
The difference between the estimated flush volume Ve^ and the
known measured value Ve-j is computed, raised to the second power and
stored. All successive flushes are treated likewise, while accumulating the
values of ?
(Vmi - Ve.T , AA..
At the last flush of the data set, the value of the objective function is
known .
The pattern search routine then initializes a local exploration
about the initial point in E1"1 before a step is made. The parameters of FJSJ
are taken one at a time as the search variable. The first search variable'
(parameter) is given small positive and negative perturbations of a given size,
tailored to the particular variable, and the objective function is evaluated
as described before at those perturbation points. If improvement (reduction)
with respect to the base point is found at any of the perturbation points,
a temporary base will be moved to the lower perturbation point, while saving
the initial one. Another parameter of F-jn- is then taken as the search vari-
able and the same process is repeated until all parameters have been searched.
The first (base) and the last points of the local search are then connected by
a straight line and a pattern move of a given step size is made along that
line, defining a new base point around which local exploration will be resumed
again.
When a point is reached where local exploration reveals no improve-
ment, a minimum has been found or the exploration is on a ridge (inversed) in
the surface of the objective function, at a point where the ridge is turning.
The perturbation size is then reduced and the local search repeated. Reduction
of the perturbation size is continued until: 1) The perturbation sizes are
below a specified resolution of the parameters; 2) the difference between the
values of the objective function at two successive bases is smaller than a
prespecified target value £; or 3) a reduction of the objective function,
greater than £, is obtained. In the first two cases an optimum has been found
93
-------
whereas in the third case, the search proceeds as before. The whole process
is also limited by the prespecified maximum number of evaluations of the
objective function. The optimization procedure assumes a unimodal objective
function. Therefore, if the objective function is multimodal, a local rather
than the global optimum may result.
The optimization package consists of a main program, that computes
the objective function, and the search subroutine that commands the
objective function surface exploration and provides the step changes on the
parameter values.
7.4.4.3 Basic Concept of Looping Stage-Discharge Curve. The objective
is to develop for each pipe segment flushed during each phase of work, a
stage-discharge relationship that meets two requirements: 1) the flow is
higher at the front of the wave than at the back of the wave for the same
measured flow depth; and 2) the discharge hydrograph volumes should be
such that the sum of the squares of the differences between estimated and
measured input volumes is a minimum. The first requirement is important
because the pollutant concentrations tend to peak in the early part of the
flush wave, usually before the hydraulic peak. A uniform flow rating curve
would therefore underestimate the masses transported in the front part of
the wave and most likely the total transported masses. The second reauire-
ment insures that the resultant stage-discharge relationship closely
.estimate'both' flow rates and flush volumes. These two requirements were
combined in a mathematical optimization problem that defines the stage-
discharge function of each individual flush while meeting both requirements.
The conceptual details of the loop-rating curve are illustrated in
Figure 34. The procedure consists of determining at each value of the flow
depth h, indicated in the figure, a correction, iAh, which is added to h
before the uniform flow rating curve is used to compute the flow rate at
flow depth h. In the rising limb of the wave, i.e., for all depths measured
prior to the occurrance of the peak depth, the corrections Ah are all
positive. At the maximum depth the correction Ah is zero, and for all depths
following the maximum depth, i.e., in the falling limb of the wave, the
correction Ah are all negative.
For example, suppose the indicated flow depth h was measured in the
front of the wave. The uniform flow rating curve would indicate for h,a flow
corresponding to point A in the figure. By computing at h a correction term,
+ Ah, which is added to h, the same uniform flow rating curve will now
indicate at h + Ah a higher flow, corresponding to point B, which is then
associated to h. This is equivalent to knowing point C on the loop-rating
curve. Suppose now that h is a'depth measured in the tail of the wave.
If the correction term, - Ah, is computed at h, and added to its value,
the uniform flow rating curve will indicate at h - Ah the flow corresponding
to point D. Again, it is as if point E,. corresponding to h on the loop-
rating curve, were known. By this simple artifice a virtual loop-rating
curve is constructed for the particular flush. The unique feature of this
approach is that the loop-rating curve is a function of the initial back-
ground flow depth, the maximum flow depth occurring during the flush and
the sediment depth within the pipe segment. Although the uniform flow rating
94
-------
UNIFORM FLOW RATING CURVE
(BACKGROUND)
h-Ah h h+Ah
STAGE
'MAX.
FIGURE 34 DERIVATION OF THE
LOOP-RATING CURVE
95
-------
curve remains the same for all flushes at a particular site, the resulting
loop-rating curve represents a specific stage-discharge function for each
flush.
7.4.4.4. Preliminary Looping Stage/Discharge Formulations
Several alternative mathematical expressions were initially
investigated to represent the stage discharge function, F-JJ, with respect
to the uniform flow rating curve and the depth correction,' Ah. Partial
summary results are presented here primarily to document the iterative
"trial and error" development of the looping stage-discharge estimation model.
Findings will be presented in terms of the optimized values of the objective
function given by equation 8,
The optimization procedure described in 7.4.4.2 was originally
devised to define for the uniform flow rating curve a polynomial function
which, having more parameters than Manning's equation, would provide a better
adjustment of the computed and measured volumes. Representation of the
uniform rating curve by Manning's equation was also considered in this
comparative analysis. Polynomials of the third degree were used. Both the
complete polynomial given by
Q = ah3 + bh2 + ch + d, (10)
where Q = flow, in cfs,
h = flow depth, in inches; and
a,b,c,d = coefficients to be determined by optimization.
and several incomplete polynomial forms, especially forms having one root
equal to zero, were also-considered. These incomplete polynomial forms are:
Q = ah3 + bh2 (11)
and .,
Q = ah13 + ch (12)
which, for a < 0, have a concave shape much similar to Manning's equation .in
the first quadrant.
Associated with the polynomial expression for the uniform flow
rating curve, a number of alternative functions were used for the correction
term, Ah, which produces the hysterisis looping effect of the discharge
curve. These functions added one or more parameters to the pattern search
optimization. The various depth correction functions used in this analysis
are as follows:
h=hi
h=h.
(13)
(14)
96
-------
l - hmax) ]
2
Ah, = ± [eh, - e(h0 + hmax) hj+ehohmax] (17)
where all variables have been defined before, except for:
h = background flow depth, in inches;
h ° = maximum depth during the flush, in inches; and
max
e,f = coefficient to be optimized.
The optimization procedure was initiated using best-estimate values
for the various parameters in the formulation. Initial values of the coeffi-
cients, a, b, c and d were determined by computing the coefficients of a
polynomial about four stage-discharge points derived from the steady-state
discharge curves described in section 7.4.2. The polynomial fitting was
performed by a subroutine added to the optimization package. The particular
form of the desired polynomial was specified in the input data. The initial
values of the coefficients e and f were arbitrarily fixed. Especially with
the complete polynomial, certain conditions With respect to the position of
points of stationarity (dQ/dh = 0) had to be imposed to avoid degenerated
solutions. Table 7 presents partial results obtained using the steady-
state discharge curves described in section 7.4.1 and 7.4.2, various polyno-
mial optimizations and finally, optimization using Manning's equation with
Ah corrections given by equation 17. Objective function values shown under
column 3 were computed using a complete polynomial and h corrections as given
by equation 7.13, except for Walnut, where correction equation 14 was used.
The value shown under column 4 for the second phase Port Norfolk flushes, was
computed using the incomplete polynomial given by equation 11, and depth
corrections, Ah, given by equation 16. The polynomial optimization results
shown under column 5 were also computed using equation 11 for the uniform
flow rating curve, but equation 17 was used for the correction term, Ah.
Column 6 of Table 7 presents results obtained with Manning's
equation and the correction term Ah as given by equation 17. It should, be
noted that only two parameters, that is, the pipe slope and the coefficient
e in equation 17 were optimized in this case. The advantage of Manning's
equation, is that the number of search variables is reduced to the pipe
slope (roughness n assumed known) and the coefficient or coefficients of the
correction term, Ah, reducing the number of required iterations and conse-
quently the computational time.
97
-------
TABLE 7: COMPARATIVE OBJECTIVE FUNCTIONAL VALUES*
FOR ALTERNATIVE FLOW COMPUTATIONAL APPROACHES
PHASE
First
First
First
First
Second
Second
Second
STREET
Tempi eton
Shepton
P. Norfolk
Walnut
P. Norfolk-lDMH
P. Norfolk-2DMH
P. Norfolk-3DMH
(1)
34042.
1884.
62649.
57625.
-
-
—
(2)
6080.
1226.
3177.
2671.
'-
-
—
(3) (4)
._
631.
-
2561 .
994.
2883.
2027.
(5) (6)
7034.
691.
3604.
-
1398.
588. 522.
8, 1632.
*0bjective Function: ?
I (Vhi, - Ve.)"
all flushes
(1) Results using stage-discharge curves derived from application of
Manning's equation with pipe slope and fixed roughness coefficient -
see section 7.4.1.
(2) Results using stage-discharge curves derived from least squares fit
of observed field data to Manning's equation - see section 7.4,2.
(3) Polynominal optimization with Ah given by equation 13 except for
Walnut Street where equation 14 applies.
(4) Polynominal optimization with Ah given by equation 16.
(5) Polynominal optimization with Ah given by equation 17.
(6) Manning's equation with Ah given by equation 17.
Obs.: Several other results were obtained from different combinations of
polynomials and h corrections, but they included partial sets of flushes and
were therefore omitted here.
98
-------
It should be noted that the optimization procedure described in
7. 4. '4.2 is not guaranteed to yield the global optimum. If the objective
function is not unimodal, as seems to be the case here, a local optimum can be
obtained. This means that the numbers shown could be only local, rather
than global optima. Nonetheless, they reveal substantial improvements with
respect to the pipe slope and best-fit slope results. The total number of
objective functional evaluation in the polynomial optimizations presented in
Table? varied from 271 evaluations in the case of Shepton, column 3, to 889
evaluations in the case of Port Norfolk, column 5. Some of the preliminary-
optimization computer runs and all runs described in the following sections
utilized a simple preoptimization procedure that led the initial point very
close to the optimum.
7.4.4.5. Preoptimization
Attempts were made at starting the optimization procedure from a
good initial point in order to save iteration steps. The uniform flow rating
curve has the greatest influence in the value of the objective function.
Starting "with a good guess on the coefficients of that curve, considerable
reduction of iteration steps were realized. As mentioned before, when a
third degree polynomial was used, the initial coefficients of the polynomial
were obtained by fitting the polynomial through h-Q points computed by Man-
ning's equation with the best-fit slope defined for each segment. The proce-
dure will be described only for the case when the uniform flow rating curve
is expressed by Manning's equation. For the polynomial case the extension
is immediate. Given the initial input values for the slope, the coefficient
e of equation 17, and sediment depths, the flow rates and flush volumes are
computed by the loop-rating curve approach for all flushes, yielding the
estimated volumes, Ve^, y-j. A simple linear regression of the type:.
Vmi = aVe. (18)
is then performed on the values of Vm. and Ve. , described before where a is
the slope of the regression line which can be obtained by:
-jl Vm. * V6i
(19)
~
E Vef
Cl early, a equal to unity corresponds to the best starting point. Disregarding
for a moment the loop-rating curve and focusing on the uniform flow curve
only, the values of Ve. can be expressed by:
18 1.49rSai. , At.
Ve. = y - ^ - ^ - 1 (20)
n .
J=l i,J
where the j index is over the sampling times for flush i.
99
-------
Combining equations 18 and 20 results in the following expression:
,,« 1.49 r2/3. a. . At,
Vm. = a S1/2 I Y. ^ i (21)
J 1 sJ
By setting 9
S* = a S (22)
the value S* will be a better initial value for the slope. Nevertheless, if
S* is used instead of S and the same computation is repeated the value of the
new a will not be exactly 1. This situation results because in the loop-
rating curve computation the volumes, Ve-j, are not a linear function of the
slope as assumed in equation 20. By repeating the adjustment given by equa-
tion. .22. several times, a can approximate unity. In this procedure a maximum
number of four successive revisions were performed in the preoptimization
routine before the pattern search routine would be implemented.
7.4.4.6. Final Form of the Stage-Discharge Function F. .
19j
Although the results obtained with the polynomial and Manning's
equation optimization represented a substantial improvement over the uniform
flow approaches given by the pipe slopes and the best-fit slope, large
differences still remained between the estimated and measured volumes for
several flushes, for which there was no clear explanation. Plots of flow
depths for all flushes at all sites revealed cases where the measured depths
for flushes of equal input volume and discharge rate were considerably
different. This could' be explained by different sediment depths. Although
measurement of sediment depths were available for each flush, there is a con-
cern as to whether or not the measurements at the ends of the pipe are
representative of the average sediment depth along the entire segment and
throughout the flush. Whatever the true causes of these discrepancies in
flow depth might have been, it was felt that some improvement would be gained
by introducing the average sediment depth in the pipe, for each flush, as a
variable magnitude to be optimized by the procedure.
The final stage-discharge function, F- ., entailed optimizing:
I iJ
1) Manning's .n, which is applicable for .all flushes at one
particular pipe segment and in a given phase;
2) The coefficient, e, of the correction factor, Ah, equation 17,
which although applicable to all flushes at one site and phase, implies
different correction values at different flushes; and
3) The average sediment depth in the pipe segment at the time of
the flush. The measured sediment depths were used as the starting point for
the optimization.
The optimized sediment depths represent a manageable way of
lumping all the unknown and complex effects that the sediments have on the
flow regime. Comparative results using Manning's equation from column 6,
100
-------
Table 7 and the final results obtained by this procedure are as follows;
RESULTANT OBJECTIVE FUNCTIONAL VALUES
A
B
PHASE STREET
First Tempi eton
First Shepton
First Port Norfolk
First Walnut
Second Port Norfolk 1-DMH
Second Port Norfolk 2-DMH
Second Port Norfolk 3-DMH
= Manning's equation with Ah correction given
= Mannina's equation with Ah correction given
mi zed sediment depth.
A
7034
691
3604
-
1398
522
1632
by equation
by equation
B
1085
353
108
21
606
23
1008
17.
17 and opti-
The final values of the objective function reveal a considerable improvement
in all cases. It is believed that the flow rates .generated by Fi,j are
very close to their actual values. Figures 35. through 38 illustrate typical
rating curves obtained by the methodology.
Plots of estimated versus measured flush volumes, similar to
those described in sections 7.4.1 and 7.4.2, were also prepared for the first
and second phase values computed by this methodology. The first ph.ase results
are presented in Figure 39 whereas the second phase values are presented in
Figure 40. A summary overview of test segment, the plan and profile pipe
slopes, the measured pipe slopes, the least squares fitted slopes ?»nd the
final optimized slopes resulting from the pattern search procedure are shown
in Table 8. The actual pipe slopes were determined using surveying equipment.
TABLE 8: OVERVIEW OF ESTIMATED SEGMENT PIPE SLOPES
PHASE
First
First
First
First
Second
Second
Second
PLAN&PROFILE
STREET PIPE SLOPE
Tempi eton
Shepton
Port Norfolk
Walnut
Port Norfolk-lDMH
Port Norfolk-2DMH
Port Norfolk-3DMH
0.00322
0.00345
0.00486
0.00481
0.00701
0.00141
0.00400
MEASURED
PIPE SLOPE
0.0029
0.0026
0.0055
0.0010
0.0059
0.0055
0.0046
BEST-FIT
SLOPE
0.00115
0.00388
0.00141
0.00221
-
0.00141
—
FINAL
SLOPE
0.00191
0.00296
0.00185
0.00134
0.00221
0.00246
0.00242
The third phase automated flushes performed in Shepton Street were not
subjected to the optimization of this section. All of these flushes were
performed by back-up and quick release of stored flush waters. The para-
meters of the stage discharge function, Fi5j, determined from the first
phase computations, were considered applicable to compute the third phase
flush rates and volumes.
101
-------
o
ro
DEPTH
(IN.)
0.0
1.00
2.0C
3.00
ft.00
5.00
6.OP
7.00
8.
10.00
11.00
12.00
n. 2 o.i
I I I 111 I 11 I I 11 I 'I
0.6
W-J-t-*
FLOW (CFS)
0.8 1.0
1.2
1.4
w-fw
1.6
1.8
2.0
2.2
PORT NORFOLK 1ST DMH 2 10 77 11:55 AH
FI.DSH * 1 VOLDME =55.20 (CF).
SLOPE = 0.002209U SED. DEPTH = 1.87 (IN.)
-RISING LIMB
FALLING
LIMB
FIGURE 35 TYPICAL LOOP RATING CURVE PORT NORFOLK STREET, 2/10/77
-------
o
co
DEPTH
(IN.)
0.0
1.00
2.00
0.2
0.4
0.6
0.8
FLOH (CFS)
1.0
3.00 : -
4.00
5.00
6.00
7.00 ::
8.00 ::
9.00 :
10.00
11.00
12.00
| « 11 i 11 l > l | »i
1.6
1.8
2.0
2.2
*
*
*
- * -f
- » t
* +
SHEPTON STREET 10 22 76 12:00 PM
FLUSH t 16 VOLUME =35.00 (CF).
SLOPE = 0.00296J1 SED. DEPTH = 0.0 (IN.)
RISING LIMB
FALLING
LIMB
FIGURE 36 TYPICAL LOOP RATING CURVE SHEPTON STREET, 10/22/76
-------
DEPTH
(IN.)
0.0
1.CO
2.00
3.00
4.00
5. CO
6.00
7.00
8.00
9.00
10.00
11.00
12.00
13.00
14.00
15.00
FLOW (CFS)
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
*
IhMPLETON STBEiiT 10 01 76 03:40 PH
*
* FLUSH * 9 BACKUP & RELEASE FLUSH
*
» SLOPE = 0.0019105 SED. DEPTH = 0.0 (IN.)
; *
*
*
*
*
- * +
* +
* * -< RISING LIMB
* +
FALLING - * *
LIMB »- - * „ * +
* +
*
*
*
*
*
*
*
*
*
: OUT OF RANGE OF COMPUTER
PRINTER '" *•
FIGURE 37 TYPICAL LOOP RATING CURVE TEMPLETON STREET, IO/OI/76
-------
o
01
DEPTH
(IN.)
0.0
1.00
2.00
3.00
0.00
5.00
6.00
7.00
8.00
-9.00
10.00
11.00
12.00
13.00
!
14.00
!
15.00 :
FLOH {CFSJ
0.2 O.tt 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
t 1 1 1 1 1 1 1 1 | 1 I 1 1 1 1 1 1 1 | 1 1 1 1 1 1 1 1 1 | 1 1 1 1 1 1 1 1 I | 1 > 1 1 H 1 1 1 | 1 1 < 1 1 1 1 I 1 | < 1 1 1 1 1 1 1 1 | I 1 1 1 > t H 1 | 1 1 1 1 1 1 M 1 | IM 1 1 1 1 1 1 | 1 1 1 1 I 1 1 H | 1 1 1 1 1 >
WALNOT STREET 10 12 76 8:55 AM
FLtJSH t 12 VOLUME =50.00 (CF.)
SLOPE = 0.00133U9 SED. DEPTH = 3. 11 (IN.)
*
*
*
*
*
-*+
- * +
* +
- * + '
* + — ( RISING LIMB
* *
- * +
', * +
FALLING . . , +
LIMB ' - * +
_
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
FIGURE 38 TYPICAL LOOP RATING CURVE WALNUT STREET, IO/I2/76
-------
160-
o
I
Ul
120.-
O
80.-
Q
Ul
OH
-------
60-
o
I
120-
80.-
0.
COMPUTED FLUSH VOLUME - CF
Manhole No. 3
OB& SEVERAL FLUSHES MAY COINCIDE IN ONE PLOTTED POINT.
Port Norfolk St.
FIGURE 40 PLOTS OF MEASURED VERSUS COMPUTED FLUSH VOLUMES
SEARCH OPTIMIZATION - SECOND PHASE
107
-------
7.4.4.7 Comparison of Computer versus Measure Flush Have Rates
An independent verification of the methodology was provided by the
dye injection experiments described in Section 5.7. Using the flush volumes
measured for 10 experiments and their corresponding measured flow depths,
this methodology was applied to optimize the discharge function, Fjsj. The
flow rates given by Fj j compared reasonably well with the flow rates
measured by the dye injection. A comparison of measured flush wave hydraulic
characteristics using dye injection methods versus the optimized results are
presented in Table 9 for four of the experiments. The volume measured by
the meter on the flush truck is given in the top row of Table 9.
Next, the volumes computed by integrating the discharge curves generated by
the dye injection measurement procedure and the numerical optimization
methodology are given. Finally, the peak flush wave rates derived from
both procedures are given.
TABLE 9. COMPARISON OF MEASURED VERSUS OPTIMIZED FLUSH WAVE HYDRAULIC
CHARACTERISTICS, DYE INJECTION EXPERIMENTS
EXPERIMENT
1
DELIVERED FLUSH VOLUME (cf) 35.
MEASURED FLUSH VOLUME DYE
INJECTION (cf) 37.5
OPTIMIZED FLUSH VOLUME (cf) 31.5
MEASURED PEAK FLUSH RATE
DYE INJECTION (cfs) -71
ESTIMATED PEAK FLUSH RATE
OPTIMIZATION (cfs) -52
2 3
35. 35.'
34.9 38.0
33.6 35.9
.68 .57
.52 .46
4
35.
36.0
37.9
.50
.38
Plots of measured flush flow rates versus estimated discharge rates
using the optimization procedure for two of the field experiments.are shown
in Figure 41. The ad hoc dye injection field procedure developed for the
project worked remarkably well in view of the extreme difficulties in
accurately monitoring non-steady state flush wave discharge lasting about
two minutes in duration. The optimization procedure reasonably reproduced
the flush wave characteristics. The degree of adherence between the
predicted and measured flush flow rates was not perfect, but reasonably close
considering the complexities in numerically trying to reproduce the dynamics
of flush waves. Application of the loop rating curve concept estimated flush
wave characteristics that more closely resembled actual measured conditions
than did the two prior steady state rating curve procedures described in
Sections 7.4.2 and 7.4.3. The accuracy in estimating the flush wave hydraulic
profiles and subsequently, the masses of pollutant flushes over the course
of the project increased by an order of magnitude. A great deal of the
project resources were expended in both the dye injection calibrations and
in the "cut and try" development and application of the optimization flow
108
-------
0.60 -
O
ID
0.10
FIGURE 41 A COMPARISON OF FLUSH WAVE HYDROGRAPHS
BY DYE INJECTION MEASUREMENTS AND
HYDRAULIC COMPUTATIONS,USING THE
OPTIMIZATION PROCEDURE
I. DYE INJECTION
2.0PTIMIZATION PROCEDURE
i i \ I I I i i I I I ^ I I i i i I
20 40 60 80 100 120 140 160 180
TIME (SEC.)
-------
FIGURE 41B COMPARISON OF FLUSH WAVE HYDROGRAPHS
BY DYE INJECTION MEASUREMENTS AND
HYDRAULIC COMPUTATIONS.USING THE
OPTIMIZATION PROCEDURE
.90 -
80 -
70 -
.60-
V)
IL.
o
3
u.
.50-
.10 -
I. DYE INJECTION
2. OPTIMIZATION PROCEDURE
1 ^v^
,
1 1 1 1 1 1 1 1 1
20 40 60 80
TIME
1 III!
100 120
(SEC.)
k — ^_
^— *-•
140 160
no
-------
estimation methodology. These efforts and expenditures were not forseen
in the conceptual development of the study. Since the field flushing
experiments conducted in this project constituted the first major data
collection research effort of this type ever performed, these additional
efforts to more accurately estimate flushed masses were deemed necessary.
7.5 Masses of Pollutants Removed by the Flush Have
Utilizing the flow predictive methodology and equations
described in section 7.4, the next step was to compute, from the analytical
concentrations of each pollutant, at discrete time intervals, the estimated
mass rates of the pollutants carried by the flush wave. A pollutograph of
each pollutant analyzed was obtained for the duration of the sampled flush
wave. A numerical integration of the pollutographs yielded the total
estimated masses of the pollutants transported by the wave. A computer
program was prepared to perform all the computations. The runs were done
for flushes qrouped by site or pipe segment, and by phase. These groups are
shown below:
GROUPING OF FLUSHES FOR THE MASS COMPUTATION RUNS
GROUP NO.
1
2
3
4
5
6
7
8
PHASE
First
First
First
First
Second
Second
Second
Third
SITE
Tempi eton
Shepton
Port Norfolk
Walnut
Port Norfolk-! DMH
Port Norfolk-2DMH
Port Norfolk-3DMH
Shepton
NO. OF FLUSHES
21
22
19
21
18
18
15
7
The same mechanisms for computing flow rate from flow depth de-
veloped in the prior section were implemented in the program. By reading.
in for each group the appropriate adjusted parameters from the optimization
the same flow rates and flush volumes associated with the final iteration
of the optimization process were re-estimated and were used to compute the
mass rates for each flush of the group. The data files described in section
7.2, provided all the information on flow depths, times of the sampling and
pollutant concentrations necessary for the computation of the pollutant
mass rates and total flushed loadings mass. The procedures for filling-in
missing data, described in section 7.3, were also incorporated into this
program.
The complete output from the program consists of four tables
printed out for each flush of a group containing: a) analytical results,
b) mass calculation, c) cumulative masses of pollutant carried by the flush
wave, and d) cumulative percentile masses of pollutants carried by the flush
wave. A sample of these tables is. presented in this, section as Tables 10, 11,
12 and 13 respectively. These tables are well labelled and are self-
explanatory.
m
-------
TABLE 10. ANALYTICAL RESULTS - WALNUT STREET TEST SEGMENT - 8/30/76
ro
STREET: H»LNUT STREET
DATE: 08 30 76
FLUSH TECHNIQUE: GRAVITY FLUSH WITH 3 INCH NOZZLE AND HOSE
HOUR OF FLDSfl: 12:35 PS
DTI USED: NO
PRESSURE IK TANK: — PSIG
SIZE Of FLUSH: 50.0 CUBIC FSET
TIME TO COMPLETE FLUSH: 15.0 SECONDS
TIME FOE FLUSH
SAMPLE
NO.
BCKGHD
10
20
30
40
50
60
70
80
90
100
110
120
130
140
150
160
170
TIME
SEC.
0
0
10
20
30
1)0
50
60
70
80
100
120
140
160
180
200
220
240
TO HEACH
DEPTH
IN.
6.00
7.00
7.00
7.50
8.00
8.50
8.50
8.50
8.75
8.75
8.75
8.00
7.50
7.50
7.50
7.50
7.25
7.25
DOWNSTREAM: 48.0 SECONDS
COD
MG/L
440
520
1640
3120
4200
7800
2840
2840
1480
1680
BOD
M6/L
163
210
10
900
680
380
360
320
600
340
TKN
HG/L
36
44
72
113
166
155
127
67
48
31
NH3
MG/l
34
37
38
41
42
44
27
14
8
4
TOT. P
MG/L
S.6
11.6
17.6
20.0
34.0
23.6
27.2
16.0
11.6
8.6
ORTH.P
MG/L
7.4
8.8
13.0
17.6
18.0
17.4
13.4
7.6
5.6
3.6
SS
HG/L
1 14
309
900
2367
3379
5178
81 14
2692
8292
5306
7184
4771
3622
2398
2423
2648
1791
VSS
HG/L
78
175
423
1155
1625
2548
4067
1350
4300
2857
3797
2530
1927
1298
1328
1468
1040
-------
TABLE 11. POLLUTANT MASSES FLUSHED - WALNUT STREET TEST SEGMENT
8/30/76
STREET: K1LHUT STBEET
DATE: 08 30 76
PLUSH TECHNIQUE: GE4VITT FLUSH KITH 3 INCH HOZZLE AMD HOSE
HOUR OF FLUSH: 12:35 Pfl
DTE USED: NO
PBESSOBB II TAMK: — PSIS
SIZE OF FLUSH: 50.0 CUBIC FEET
TIME TO COMPLETE FLUSH: 15.0 SECONDS
TIHE FOR FLUSH TO BEACH
SAHPLE
No.
BCKGHD
1
2
3
4
5
6
7
8
9
10
11
12
13
11
15
16
17
TOTAL
TISE
SEC.
0
0
10
20
30
HO
50
60
70
80
100
120
1UQ
160
180
200
220
240
MASS »OL=
PLOW
CFS.
0.0 'to
0.170
0.170
0.252
0.320
0.370
0.370
0.370
0.388
0.388
0.388
0.203
0.1 20
0.120
0.120
0.120
0.089
0.089
50.2
DOHKSTBEAH: 48.0 SECOHDS
COD
3BAHS
1.9
25.1
52.1
116.9
215.5
327.1
383.7
440.3
659.1
1713.7
1168.8
326.6
132.8
192.8
146.6
100.5
79.7
84.8
6227.9
BOD
uEASS
0.7
10. 1
5.3
0.7
41.2
94.3
82.8
71.3
58.2
83.5
81.3
«1.«
23. 1
21.7
31.2
40.7
23.7
17.2
728.5
TKN
v»'« A.I S
0.2
2.1
2.8
5.1
8.4
11.8
14.6
17.4
17.6
34.1
31.0
14.6
6.6
4.5
3.9
3.3
2.0
i 1.6
181.6
NH3
GPA.1S
0,2
1.8
1.8
2.7
3.6
4.3
4.4
4.4
4.7
9.7
7.8
3.1
1.*
1.0
0.7
0.5
0.3
0.2
52.5
TOT. P
GfiAHS
0.0
0.6
0.7
1.3
1.7
2.1
2.8
3.6
3.2
5.2
5.6
3.1
US
1.1
0.9
0.8
0.5
0.4
35.1
OSTH.P
GHAHS
0.0
0.4
0.5
0.9
1.4
1.8
1.9
1.9
1.9
3.8
3.4
1.5
0,7
0.5
0.4
0.4
0.2
0.2
22.1
SS
GRAMS
0.5
14.9
43.4
168.6
306.0
542.8
696.7
850.6
295.7
1821.7
1165.7
826.0
323.8
245.9
162.8
164.5
133.6
90.4
7853.7
7SS
GRASS
0.3
8.4
20.4
82.3
147.2
267.1
3«6.7
426.3
148.3
944.7
627.7
436.6
171.7
130.8
88.1
90. 1
74.1
52.5
4063.4
-------
TABLE 12. CUMULATIVE POLLUTANT MASSES FLUSHED - WALNUT STREET TEST SEGMENT - 8/30/76
COD:
BOO:
TKN:
AHH:
T.PH:
O.PH:
SS:
VS:
?OL.:
***** CUMULATIVE
1.9 27.1
3935.4 5104.2
0.7
448.2
0.2
114.1
0.2
37.5
0.0
21.2
0.0
14.7
0.5
4741.0
0.3
2391.8
0.0
24.7
10.9
529.5
2.3
145. 1
1.9
45.3
0.6
26.8
0.5
18.0
15.4
5906.7
8.8
3019.5
0.8
32.2
1ASSES Or POLLUTANTS CABBI
79.2 196.0
5430.8 5623.5
16.2
570.9
5.1
159.7
3.7
48.4
1.3
29.9
1.0
19.6
58.9
6732.8
29.2
3456.1
2.3
37.8
16.9
594.0
10.2
166.3
6.5
49.8
2.6
31.4
1.9
20.3
227.5
70S6.6
111.5
3627.8
4.3
40.7
ED AHAY BI THE FLUSH HIVE
411.6 738.6
5816.3 5962.9
58.1
615.7
18.6
170.9
10.0
50.7
4.3
32.4
3.3
20.8
533.5
7302.5
258.7
3758.6
7.0
42.8
152.4
646.9
30.4
174.8
14.3
51.5
6.4
33.4
5.1
21.3
1076.3
7465.2
525.8
3846.7
10.3
44.8
1122.3
6063.4
235.2
687.7
45.1
178.0
18.7
52.0
9.2
34.2
7.0
21.6
1773.0
7629.7
872.5
3936.9
13.8
46.9
1562.6
6143. 1
306.5
711.4
62.5
180.0
23.1
52.3
12.8
34.7
8.9
21.9
2623.6
7763.3
1298.8
4011.0
17.4
48.7
2221.7
6227.9
364.7
728.5
80.1
181.6
27.8
52.5
15.9
35. 1
10.8
22.1
2919.3
7853.7
1447.1
4063.4
21.0
50.2
Each row represents cumulative pollutant mass transported in grams, past sampling manhole at each instant
of sampling, including background sewage contribution. Last row is cumulative flush volume in cubic feet.
-------
TABLE 13. CUMULATIVE PERCENTILES OF POLLUTANTS FLUSHED - WALNUT STREET TEST SEGMENT
8/30/76
COO:
BOO:
TKH:
ANN:
T.PB:
O.PH:
SS:
IS:
0.4
63.2 82.0
1.5
61.5 72.7
1.3
62.9 79.9
3.7
71. It 86.2
1.7
60.2 76.2
2.1
66.5 81.8
C.2
60.4 75.2
0.2
58.9 74.3
OF POLLOTAHTS CARBIED AHAY BY
1.3 3.1 6.6
87.2
2.2
78.4
2.8
88.0
7.1
92.1
3.7
85.1
4.5
88.8
0.7
85.7
0.7
85.1
90.3
2.3
81.5
5.6
91.6
12.3
94.8
7.3
89.3
8.7
92.0
2.9
89.9
2.7
89.3
Each row represents cumulative percentage of pollutant
each instant of sampling
, starting at
first occurrence
93.4
8.0
34.5
10.2
94.1
19.1
96.6
12. 1
92.4
14.9
94.4
6.8
93.0
6.4
92.5
mass
11.9 18.0
95.7
20.9
88.8
16.8
96.2
27.3
98.0
18.1
95.0
23. 3
96.4
13.7
95.1
12.9
94.7
transported past
of flush wave.
97.4
32.3
94.4
24.8
98.0
35.6
99.0
26.2
97.3
31.8
98.1
22.6
97.1
21.5
96.9
sampling manhole at
25.1
98.6
42.1
97.6
34.4
99. 1
44.0
99.6
36.3
98.7
40.3
99.2
33.4
98.8
32.0
98.7
35.7
100.0
50.1
100.0
44.1
100.0
53.0
100.0
45.4
100.0
49.1
100.0
37.2
100.0
35.6
100.0
-------
SECTION 8
SINGLE SEGMENT FLUSHING RESULTS
8.1 Foreword
All field flushing results conducted during the first phase of
experimentation are presented in this Chapter. Descriptions of the test
segments, field sampling procedures and equipment, analytical laboratory
techniques and computational methods used to process field and laboratory
information have all been presented in Chapter 4, 5, 6 and 7, respectively.
Preliminary flushing results conducted early in the program on Shepton Street
are presented in Section 8.2 along with typical results for flushes conducted
at other test segments over the course of the program. The solids, organics
and nutrient flushed loadings for 86 flushing experiments conducted during
the fall of 1976 are presented in Section 8.3. Statistical results per
flushing site are presented for the total flushed pollutant masses, pollutant
masses normalized by antecedent periods between flushes and pollutant masses
normalized by both antecedent periods and contributary population. Similar
results are presented in Section 8.4 for the masses of heavy metals flushed
per test segment. The relative pollutant removal effectiveness of different
flush methods considered in this phase of work are examined in Section 8.5.
Section 8.6 presents various visual observations and analytical results
pertaining to sediment characteristics encountered during the flushing
program.
8.2 Typical First Phase Flushing Results
The field flushing program was initated during the middle of
August, 1976. Pre-cleaning of segments was accomplished during the latter
part of that month. Figure 42 shows two photographs of heavily deposited
sewers in the study area. Figure 42A shows a photograph taken at the
intersection of Florida and Tempieton Street. A photograph of the Tempieton
Street sewer segment is shown in Figure 42B. These photographs of heavy
deposits were typical of the flat segments in the study area. The
photographs show that dry weather deposition in upstream collection system
laterals means rags, sticks, large globs of organic material, toiletpaper
along with fine silt, sand and rocks. A contrast of several sewer segments
on fairly steep streets in the area where light to moderate deposition was
present is shown in Figure 43.
Several flushing experments were conducted on each test segment
during the pre-cleaning period primarily to develop field procedures and
to uncover any mechanical difficulties with the flush truck. Two flushes at
the Shepton Street test segment were conducted on August 18, 1976 prior to
pre-cleaning. About 4 inches of thick black sediments and heavy sanitary
116
-------
A. Florida and Tempieton Streets B. Tempieton Street
FIGURE 42. PHOTOGRAPHS OF COLLECTION SEWERS IN THE STUDY AREA WITH HEAVY DEPOSITION RATES
-------
V+-. ***
?:^.**i
FIGURE 43. PHOTOGRAPHS OF COLLECTION SEWERS IN THE STUDY AREA WITH LOW DEPOSITION RATES
-------
deposit levels were present in the segment prior to the flushes.
The second flush was conducted ten minutes after the first flush
experiment was performed and flush wave sampling completed. The volume and
rate of both flushes were 50 cubic feet at 0.5 cfs. Plots of TSS, VSS, BOD
and COD flush wave concentrations for both flushes are shown in Figure 44.
Slmilaf* plots for TKN, TP, NHs and OP are shown in Figure 45. With the ex-
ception of TSS, the second flush wave concentrations were significantly less
in magnitude to the first flush. Peak COD concentrations shown in Figure 44
were 5100 and 750 mg/1 for the two flushes, respectively. Peak TKN concen-
trations shown in Figure 45 were 135 and 23 mg/1 for the two flushes,
respectively. These results were extremely encouraging in that the first
flush seemed to transport most of the organic and nutrient related pollutants
without pre-cleaning the segments. The mass loadings flushed for these two
experiments were not computed, since the wave heights were not recorded dur-
ing these two flushes.
Typical flush wave pollutant concentration plots for the four test
segments are shown in Figures 46 and 47. VSS peak concentrations ranged
between 6000 mg/1 for the Walnut Street flush conducted on 11/06/76 to
7600 mg/1 for the Tempieton Street flush conducted on 9/13/76. These plots
were typical of the first phase experimentation program. Background sewage
concentrations are also indicated on the plots. The TSS and VSS background
concentration levels shown in Figure 46 for Port Norfolk and Walnut Street
typified background sampling during the project. Test segment background
sewage levels determined during the project greatly exceeded nominal
concentrations found further downstream at sewage facilities. Summarized
background sampling results are given in Chapter 11.
An example showing the mass rate of pollutants transported during
a given flush is shown in Figure 48 for the experiment conducted at Port Nor-
folk Street on 11/01/76. Pollutographs of BOD, TSS and VSS are shown in the
left hand plot and pollutographs of NHs, TP and TKN are shown in the right
hand figure. The flushed mass rates peaked within one minute for the solids
and organics and within a half a minute for the nutrients. This phenomena
was typical of flushes throughout the program where the peak mass wave of
the lighter pollutant fraction occurred sooner than the heavier solids
loadings. The plots also show the importance of accurate definition of the
flush wave hydraulic characteristics. The rigorous flush wave discharge
computational analysis described in Chapter 7 was initiated since it was
apparent earlier in the project that any significant error in estimation of
the hydraulic profiles of the flush wave would result in erroneous estimation
of pollutant masses transported.
8.3 Solids, Organics and Nutrient Flush Removal Results
The pollutant mass removals, in kilograms, for the flushing
experiments conducted between August 30 through November 12, 1976 on Port
Norfolk, Shepton, Tempieton and Walnut Street test segments are presented
in Tables 14 through 17, respectively. Following the date of flush on the
left hand side of these tables, the information under column headings are
presented as follows: A) truck flush volume (cf); B) truck flush
119
-------
no
o
sooo -
4000 -
3000 •
1000 -
SHEPTON ST. a/18/76
FIRST FLUSH
COO
VSS
FLUSH VOLUME = 50 cf
FLUSH RATE ' 0.5 cfs
6000
3000 -
2000 -
80
120 ISO 200
TIME (SEC.)
240
280
320
SHEPTON ST.
8/18/76
SECOND FLUSH
COO
BOO
TSS
VSS
FLUSH VOLUME- 50cf
FLUSH RATE' 0.5cfs
First Observance of Flush
240
320.
TIME (SECJ
FIGURE 44.
PLOTS Of FLUSH WAVE SOLIDS AND ORGANIC POLLUTANT CONCENTRATIONS - SHEPTON STREET
8/18/76
-------
ro
SHEPTON ST.
a/16/76
FIRST FLUSH
TKN
TOTAL PHOSPHATE
AMMONIA
ORTHO PHOSPHATE
FLUSH VOLUME; 50 cf
FLUSH RATE' 0.5 cfs
First Observance of Flush Wave
20 -
10 -
no -
so -
ESO -
40 -
30 -
10 -
200 240 ZOO 320
SHEPTON ST.
8/18/76
SECOND FLUSH
• e TKN
a a TOTAL PHOSPHATE
« » AMMONIA
°—o ORTHO PHOSPHATE
FLUSH VOLUME' 50cf
FLUSH RATE1'0.5 cfs
• First Observance of Flush Wave
40 80
TIME (SEC.)
120 160 200
TIME (SEC.)
240 280 320
FIGURE 45. PLOTS OF FLUSH WAVE NUTRIENT POLLUTANT CONCENTRATIONS - SHEPTON STREET
8/18/76
-------
ro
ro
11000 -
10000 -
9000 -
8000 -
7000 -
6000 -
SOOO -
4000 -
3000 -
2000 •
1000
PORT NORFOLK ST.
11/01/76
GRAVITY FLUSH
VOLUME OF FLUSH' 47 cf
RATE-.42cf»
TSS
VSS
FIRST OBSERVANCE
OF FLUSH WAVE
IK500 -
10000 -
9000 -
8000 -
7000 -
60OO -
9OOO -
4000
2000 -
1000 -
-T 1-
120 160
TIME (SECJ
WALNUT ST.
11/08/76
GRAVITY FLUSH
VOLUME OF FLUSH* 50.0 of
RATE".42cf«
TSS
VSS
-i p 1 1 1 r
120 160 200
TIME (SEC.)
FIGURE 46. FLUSH WAVE TSS AND VSS CONCENTRATIONS PLOTS FOR PORT NORFOLK STREET (11/1/76)
AND WALNUT STREET (11/6/76)
-------
ro
co
IIOOO -
10000 -
9000 -
8000 -
7000 -
6000 -
500O -
4000 -
3000
2000 -
1000 -
SHEPTON ST.
9/13/76
BACK-UP 8 RELEASE
VOLUME = 49 cf
TSS
IIOOO •
loooo -
9000 -
8000 -
sooo -
4000 -
3000 -
2000 -
1000 -
-T
160
TIME (SEC.)
320
TEMPLETON ST.
9/13/76
GRAVITY FLUSH
VOLUME OF FLUSH« 35 cf»
RATE«.43cfs
TSS
80 120 160 200
TIME (SEC.)
240 260
FIGURE 47. FLUSH WAVE CONCENTRATIONS PLOTS - SHEPTON STREET (9/13/76) AND TEMPLETON STREET
(9/13/76)
-------
ro
ioo -
70
60
40
SO -
20
FIRST OBSERVANCE
' OF FLUSH WAVE
PORT NORFOLK ST.
H/OI/76
GRAVITY FLUSH
VOL=47cf RATE = .42cfs
9 » BOD
. » TSS
vss
-ifi
40
120 160 200
TIME (SEC.)
tto
PORT NORFOLK ST.
H/OI/78
GRAVITY FLUSH
VOL«47cf RATE-.4acf»
AMMONIA
PHOSPHATE'S
TKN
FIRST OBSERVANCE
OF FLUSH WAVE
240
280
TIME (SEC.)
FIGURE 48. FLUSH WAVE POLLUTANT MASS VERSUS TIME PLOTS - PORT NORFOLK STREET H/Ol/76
-------
TABLE 14. PHASE I - FIELD FLUSHING POLLUTANT REMOVALS (kg)
PORT NORFOLK STREET
TOTAL MASS
DATE
8/30/76
9/2
9/7
9/16
9/21
9/24
10/1
10/4
10/8
10/12
10/15
10/18
10/25
10/22
10/29
11/1
11/5
11/8
11/12
A
32.9
46.5
32.8
47.0
32.6
46.6
33.7
B&R***
32.9
33.4
33.1
47.0
32.7
46.6
32.9
47.0
46.6
32.9
B&R***
B
.47
.92
.48
.44
.82
.79
.19
.46
.27
.33
.42
.63
.71
.46
.42
.85
.47
C
1
3
5
9
5
3
7
3
4
4
3
3
3
4
4
3
4
3
4
D
X
X
X
X
X
X
X
E
20
2
82
41
64
72
37
F
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
G
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
COD
0.74
1.97
2.33
3.80
10.71
1.78
4.69
4.32
1.30
BOD
0.28
0.87
0.46
1.28
4.12
0.71
1.58
1.32
4.84
TKN
0.020
0.085
0.091
0.125
0.426
0.068
0.142
0.152
0.042
NH3
0.003
0.016
0.019
0.015
0.084
0.008
0.028
0.035
0.012
(kg)
TP
0.007
0.018
0.031
0.017
0.082
0.019
0.001
0.030
0.013
TSS
1.74
16.61
2.35
2-. 14
3.29
3.26
4.46
11.42
2.70
2.65
1.87
3.96
2.19
2.01
2.68
3.59
3.16
1.34
7.69
VSS
0.54
10.18
1.64
1.46
1.-58
2.12
3.69
8.72
1.79
1.96
1.51
3.32
1.61
1.55
2.22
2.82
2.91
1.06
7.08
ro
01
Legend
A - Truck Flush Volume (cf)
B - Truck Flush Rate (cfs)
C - Total Antecedent Days Between Flushes
D - Rainfall Impacted Event
E - Hours After Rainfall Event to Flush*
F - Segment Cleaned After Flush**
G - Good Flush
**
***
Number of hours between rainfall exceeding 0.15 inch/hour and flush event.
Segment flushed after experiment using maximum truck discharge for five minutes.
"Backup and release.
-------
TABLE 15. PHASE I - FIELD FLUSHING POLLUTANT REMOVALS (kg)
SHEPTON STREET
DATE
8/23/76
8/26
8/30
9/2
9/7
9/10
9/13
9/16
9/21
9/24
10/1
• 10/4
10/8
10/12
10/15
10/18
10/22
10/25
10/29
11/1
11/5
11/8
11/12
A
47.8
35.4
37.3
B&R***
20.0
32.6
B&R***
46.6
32.9
33.1
46.5
32.9
32.9
B&R***
46.6
33.0
32.9
47.2
46.6
32.7
33.0
47.1
B&R***
B
.26
.76
.65
.69
.67
.42
.37
.93
.47
.46
.71
.47
.43
.37
.72
.61
.40
.41
C
b
3
4
3
5
3
3
3
5
3
7
3
4
4
3
3
4
3
4
3
4
3
4
D E
x 23
x 4
x 66
x 86
x 45
x 60
x 75
x 35
F
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
X
X
G
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
TOTAL MASS (kq)
COD BOD TKN NHg TP
5.73 1
.
2
1
1
1
2
1
7.28 2
2
.22 .097 .028 .027
..ANALYTICAL P
.02
.07
.97
.73
.97
.62
.21
.02 .189 .047 .051
.93
.20
TSS
3
ROB
2
7
11
4
2
6
4
2
2
5
5
4
4
6
2
4
3
2
5
5
1
.91
L E
.94
.57
.54
.52
.81
.76
.34
.76
.19
.76
.16
.30
.66
.67
.52
.31
.44
.70
.07
.29
.08
VSS
2.82
M S ...
1.79
5.55
9.03
3.60
2.24
5.53
2.88
2.01
1.51
4.58
4.10
3.12
3.77
5.36
1.98
3.63
2.79
2.17
4.21
4.21
.94
ro
Legend
A - Truck Flush Volume (cf)
B - Truck Flush Rate (cfs)
C - Total Antecedent Days Between Flushes
D - Rainfall Impacted Event
E - Hours After Rainfall Event to Flush*
F - Segment Cleaned After Flush**
G - Good Flush
**
***
Number of hours between rainfall exceeding 0.15 inch/hour and flush event.
Segment flushed after experiment using maximum truck discharge for five minutes.
Backup and release.
-------
TABLE 16. PHASE I - FIELD FLUSHING POLLUTANT REMOVALS (kg)
TEMPLETON STREET
DATE
8/30/76
9/2
9/7
9/10
9/13
9/16
9/21
9/24
10/1
10/4
10/8
10/12
10/15
10/18
10/22
10/25
10/29
11/1
11/5
11/8
11/12
A
46.5
32.8
B&R***
57.4
32.9
B&R***
B&R***
33.1
B*R***
46.6
32.9
32.9
47.1
46.6
32.7
32.9
46.9
46.7
32.7
32.9
B&R***
B
.98
.50
.83
.43
.36
.87
.46
.43
.41
.76
.59
.46
.49
.75
.56
.43
C
3
5
3
3
3
5
3
7
3
4
4
3
4
3
4
3
3
4
3
4
D
X
X
X
X
X
X
X
X
E
25
5
65
87
43
62
73
34
F
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
G
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
COD
7
4
3.55
2
1
2
3
6
BOD
.15
.27
.65
.78
.75
.91
.87
.92
.47
.56
TOTAL MASS (ka)
TKN NH3 TP TSS
7.01
18.32
4.48
11.88
.097 .010 .019 3.07
8.44
7.85
3.45
13.18
7.43
3.39
33.21
4.44
12.52
2.32
3.41
6.55
9.49
8.43
2.88
VSS
2.71
11.75
3.84
9.08
2.29
6.04
4.83
2.897
9.47
6.01
2.92
26.66
3.04
9.96
1.53
2.78
4.83
6.39
6.96
2.12
ro
Legend
A - Truck Flush Volume (.cf)
B - Truck Flush Rate (cfs)
C - Total Antecedent Days Between Flushes
D - Rainfall Impacted Event
E - Hours After Rainfall Event to Flush*
F - Segment Cleaned After Flush**
G - Good Flush
*Number of hours between rainfall exceeding 0.15 inch/hour and flush event.
**Segment flushed after experiment using maximum truck discharge for five minutes.
***Backup and release.
-------
TABLE 17. PHASE I - FIELD FLUSHING POLLUTANT REMOVALS (kg)
WALNUT STREET
DATE
8/23/76
8/26
8/30
9/2
9/7
9/10
9/13
9/16
9/21
9/24
10/1
10/4
10/8
10/12
10/15
10718
10/22
10/25
10/29
11/1
11/5
11/8
11/12
A
69.8
46.6
47.1
60.6
B&R***
32.0
46.5
32.0
46.6
B&R***
33.0*
32.9
B&R***
47.0
32.8
33.0
47.3
46.5
32.9
32.9
47.0
46.6
B&R***
B
.80
.86
.41
.65
.47
.96
1.39.
.89
.39*
.46
.98
.50
.42
.35
.95
.47
.43
.42
.88
C
3
3
4
3
3
3
3
3
5
7
3
4
4
3
3
4
3
4
3
4
3
4
D
X
X
X
X
X
X
X
X
E
21
1
62
84
40
71
38
F
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
6
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
COD
7.40
6.23
2.76
9.50
6.69
5.10
a. 92
3T.45
2.59
4.51
4.21
9.12
BOD
2.11
1.79
7.28
1.02
2.03
1.21
1.81
3.42
7.68
1.24
1.22
1.65
3.34
TKN
.183
.181
.144
.245
.147
.138
.323
.526
.139
.165
.141
.235
TOTAL MASS
NH3
.088
.052
.078
.202
.040
.033
.088
.109
.062
.072
.041
.127
(kg)
TP
.uw
.035
.049
.049
.075
.043
.091
.091
.001
.045
.030
.052
TSS
5.64
1.42
7.85
6.32
2.42
4.15
6.33
2.17
13.33
4.19
9.31
29.34
1.06
2.43
7.74
3.14
1.71
2.53
4.39
6.39
4.30
VSS
2.60
.65
4.06
3.38
1.09
3.59
3.84
1.76
3.67
2.86
6.55
16.69
.58
1.84
6.99
1.91
1.29
1.88
3.17
4.23
3.57
Legend
A - Truck Flush Volume (cf)
B - Truck Flush Rate (cfs)
C - Total Antecedent Days Between Flushes
D - Rainfall Impacted Event
E - Hours After Rainfall Event to Flush*
F - Segmented Cleaned After Flush**
G - Good Flush
^Number of hours between rainfall exceeding 0.15 inch/hour and flush event.
^Segment flushed after experiment using maximum truck discharge for five minutes.
***Backup and release.
-------
rate (cubic feet per second); C) total number of antecedent days respect-
ive of rainfall occurrences between flushes; D) an indication noted by "X:
if the event was impacted by rainfall; E) Number of antecedent hours
between rainfall exceeding 0.15 inch/hour and flush event; F) an indication
noted by "X" if the post flushing of the flushing experiment was conducted;
and 6) an indication noted by "X" if the flushing experiment was completed
without any major mechanical or'procedual problem, such as inoperative water
meter, burst hoses and structural failure of manhole table. The next set of
columns present the computed pollutant masses, in kilograms, of flushed load
for the following parameters: COD, BOD, TKN, NH^, TP, TSS and VSS. These
estimates were prepared using the loop stage rating curves developed by
optimization procedures, described in Chapter 7.
Hourly rainfall information measured at the Blue Hills Observatory
located about 5 miles from the study area were collected and compared with
strip charts from the automatic liquid level sensing devices installed at
each of the test segments. These devices were operated and maintained
during most of the experimentation period. Typical results of dry weather
flow gaging are discussed in Chapter ll. It appeared upon inspection of the
strip charts that rainfall intensities less than 0.10 inches/hour at the two
combined sewer segments, that is, on Port Norfolk and Walnut Streets, would
not result in any marked increase in depth of flow while intensities of less
than 0.15 inches/hour would not result in depth of flow increases in the
two separate sewer segments on Tempieton and Shepton.Streets. The two
separate sewer segments receive clear water inflow from roof drain
connections. These empirical criteria were then used to determine whether
the antecedent period prior to flushing events were impacted by rainfall of
sufficient magnitude that would change dry weather accumulations. The number
of hours between the last rainfall occurrence and the flushing event was.
arbitrarily defined at a rainfall intensity of 0.15 inches/hour.
The rationale for terminating the post flushing cleaning operation
in November was to determine, in an indirect way, whether substantial resi-
dual materials remained after the routine experimental flushes. The nature
of these experiences would provide indirect evidence of the flushing pollutant
removal efficiencies.
In total, there were 86 flushes performed during this period.
Table 18 summarizes for each test segment the number of flushes, the
number of flushes free from operational problems, the number of good flushes
with dry antecedent periods and the number of good flushes impacted by
rainfall events occurring during periods between flushes. Table 18 shows
that 64% of all good flushes were conducted under the conditions of dry
antecedent periods and that 86% of all the flushes were operationally
acceptable.
The mean and standard deviation of total solids, organic and
nutrient flushed mass removals for Port Norfolk, Shepton, Tempieton
and Walnut Street test segments are given in Tables 19 through 22.
Typically, each table contains pollutant mass removal statistics including
the mean, standard deviation and number of experiments used in the
computations for the following flush groupings: a) all flushes; b) all
129
-------
TABLE 18. SUMMARY OF FIRST PHASE FLUSHING EVENT CHARACTERISTICS
Flush Event
Characteristics Port
Number of Flushes
Number of Good
Flushes
Number of Good
Flushes with Dry
Antecedent Periods
Number of Good
Flushes Impacted by
Rainfall Events
Norfol k
19
18
11
7
Shepton
23
21
13
8
Tempi eton
21
17
11
6
Walnut
23
18
12
6
Total
86
74
47
27
good flushes free from operational problems; c) all good flushes with dry
antecedent dry periods; and d) all flushes impacted by rainfall. Although
these results are not normalized for antecedent periods prior to flushing,
several observations can be drawn from inspection of the tables. First of
all, the coefficients of variation, that is, the standard deviation
divided by the mean, for the mass removals are all less than unity with
the exception of the rainfall impacted mass removals for the two combined
sewer segments on Port Norfolk and Walnut Streets. This implies that the
flushing removal rates and also the rates of deposition are reasonably
stable statistics which is important from a. prediction and control standpoint.
Large coefficients of variation, on the order of 2 to 3, v/ere expected.
Secondly, the ratios of VSS/TSS for the good non-rainfall impacted flushes
were remarkably consistent, ranging from 0.65 to 0.75 with an average of
0.71. Thus, nearly three quarters of the suspended solids transported were
volatile in nature.
The total flushed pollutant masses given in Tables 14 through 18
for each of the test segments were divided by the total antecedent dry
periods between flushing events given under column C of each table.
Figure 49 through 52 show the time series of flushed masses for various
pollutants normalized by the antecedent dry periods for Port Norfolk,
Shepton, Templeton and Walnut Streets, respectively. Plots of daily rain-
fall collected at the Blue Hills Observatory are presented at the top of
each figure. Plots of total phosphate are followed by TSS with VSS. The
final two plots per figure include TKN with NHs, and COD with BOD. The
remarkable feature of the plots for any given pollutant is the degree of
flushing removal consistency for a given test segment.
Statistical summarys of the pollutant mass removals normalized by
antecedent days between flushes are presented in Tables 23 through 26 for
the Port Norfolk, Shepton, Templeton and Walnut test segments, respectively.
Means, standard deviations and the number of flushes used in the computations
are presented for the seven pollutants for eight different groupings of the
130
-------
TABLE 19. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING RESULTS
MEAN AND STANDARD DEVIATION OF TOTAL MASS REMOVALS
PORT NORFOLK STREET TEST SEGMENT
POLLUTANT REMOVALS (kg/flush)
TYPE
A: ALL
FLUSHES
B. ALL GOOD *
FLUSHES
C: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
D: ALL RAINFALL
IMPACTED
FLUSHES
COD
3.51
3.03
(9)
3.51
3.03
(9)
3.22
1.50
(5)
3.89
4.59
(4)
BOD
1.60
1.58
(10)
1.00
1.58
(10)
1.12
0.41
(5)
1.39
1.83
(4)
TKN
Q.128
0.120
(9)
0..128
0..120
(9)
0.109
0.045
(5)
0.151
0.186
(.4)
NH3
0.024
0.024
(9)
0.024
0.024
(9)
0.021
0.010
(5)
0.029
0.038
(4)
TP
0.024
0.024
(9)
0.024
0.024
(9)
0.016
0.010
(5)
0.035
0.033
(4)
TSS
4.16
3.83
(19)
4.28
3.91
(18)
3.32
1.72
(ID
5.78
5.84
(7)
vss
3.04
2.66
(19)
3.13
2.71
(18)
2.72
1.66
(11)
3.76
3.93
(7)
* See definition in text, page 129.
KEY: Top Row - Mean
Middle Row - Standard Deviation
Bottom Row (in parentheses) - Number of flushes used in computations
-------
TABLE 20. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING RESULTS
MEAN AND STANDARD DEVIATION OF TOTAL MASS REMOVALS
SHEPTON STREET TEST SEGMENT
POLLUTANT REMOVALS (kg/flush)
TYPE
A: ALL
FLUSHES
B. ALL GOOD *
FLUSHES
C: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
D: ALL -RAINFALL
IMPACTED
FLUSHES
COD
6.50
1.10
(2)
6.50
1,10
(2)
6.50
1.10
(2)
-
BOD
1.63
0.62
(11)
1.63
0.62
(ID
1.60
0.69
(8)
1.73
0.45
(3)
TKN
0.143
0.065
(2)
0.143
0.065
(2)
0.143
0.065
(2)
-
NH3
' 0.0375
0.0134
(2)
0.0375
0.0134
(2)
0.0375
0.0134
(2)
-
TP
0.039
0.017
(2)
0.039
0.017
(2)
0.039
0.017
(2)
-
TSS
4.56
2.24
(22)
4.38
1.53
(20)
4.36
1.47
(12)
4.43
1.72
(8)
vss
3.54
1.79
(22)
3.39
1.27
(22)
3.47
1.27
(20)
3.28
1.35
(8)
CO
ro
* See definition in text, page 129.
KEY: Top Row - Mean
Middle Row - Standard Deviation
Bottom Row (in parentheses) - Number of flushes used in computations
-------
TABLE 21. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING RESULTS
MEAN AND STANDARD DEVIATIONS OF TOTAL MASS REMOVALS
TEMPLETON STREET TEST SEGMENT
POLLUTANT REMOVALS (kg/ flush)
TYPE
A: ALL
FLUSHES
B: ALL GOOD
FLUSHES *
C: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
D: ALL RAINFALL
IMPACTED
FLUSHES
COD
3.55
-
(D
-
BOD
3.13
2.31
00).
3.49
2.41
(8)
3 63
2.06
(5)
2.61
3.08
(4)
TKN
0.097
-
(D
_
0.097
CD
NH3
0.010
-
(1)
_
0.010
-
(1)
TP
0.019
-
(D
_
0.019
-
(D
TSS
8.59
7.16
(20)
9.30
7.82
(16)
7.62
4.03
(10)
10.32
10.54
(8)
vss
6.31
5.63
(20)
6.95
6.11
(16)
5.75
3.03
(10)
7.34
8.46
(8)
co
CO
* See definition in text, page 129.
KEY: Top Row - Mean
Middle Row - Standard Deviation
Bottom Row (in parentheses) - Number of flushes used in computations
-------
TABLE 22. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING RESULTS
MEAN AND STANDARD DEVIATION OF TOTAL MASS REMOVALS
WALNUT STREET TEST SEGMENT
POLLUTANT REMOVALS (kq /flush)
TYPE
ALL
FLUSHES
ALL GOOD *
FLUSHES
ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
ALL RAINFALL
IMPACTED
FLUSHES
COD
8.21
7.69-
(12)
8.21
7.69
(12)
5.10
2.39
(7)
12.56
10.65
(5)
BOD
2.75
2.23
(13)
2.75
2.23
(13)
1.77
0.73
(8)
4.32
2.99
(5)
TKN
0.214
0.113
(12)
0.214
0.113
(12)
0.164
0.036
(7.)
0.284
0.151
(5)
NH3
0.083
0.047
(12)
0.083
0.047
(12)
0.072
0.031
(7)
0.098
0.064
(5)
TP
0.052
0.026
(12)
0.052
0.026
(12)
0.040
0.020
.(7)
0.068
0.025
(5)
TSS
6.01
6.14
(21)
6.30
6.57
(18)
3.29
1.68
(12)
11.46
8.24
(7)
vss
3.63
3.43
(21)
3.75
3.69
(18)
2.14
1.16
(12)
6.45
4.74
(7)
co
* See definition in text, page 129.
KEY: Top Row - Mean
Middle Row - Standard Deviation
Bottom Row (in parentheses) - Number of flushes used in computations
-------
1
3.0-
DAILY 2<0~
RAINFALL
IN
INCHES |.o-
.03-
TOTAL P •°2"-
.01-
4.0-
TSS 3.0-
a
VS S 2.0-
1.0-
0.1-
TKN
a
NH, .05-
3.0-
Ca° 2-^
BOO
1.0-
ALL POLLUTANT
SCALES IN
KILOGRAMS
PER DAY
YIASS REMOVAL / RAINFALL PLOTS
PORT NORFOLK ST
POLLUTANT MASS REMOVALS NORMALIZED BY ANTECEDENT DAYS
FIGURE 49-A
T Tl n IT ITlm-i T rTI-, T
1 1 1 1 1 1 1 i iii 1 1 1 1 1 1 M 1 1 1 1 1 1 1 1 1 n 1 1 1 1 1 n \ \
n 0 fl
I M 1 1 I i i i 1 h it i i i 1 1 i i i i i 1 1 i 1 1 1 1 1 1 i i i i i 1 1 M
H-TSS
1— vss
R 1 R R H 1
IMIIII i niniiiriiLiiiiiiniiiriiriiMin
n— TKN
n H hkNH*
MIMM 1 Ml 1 1 1 1 M 1 1! 1 1 1 1 1 1 M 1 II 1 1 1 1 1 1 1 M
H^,COD
-rrB,°,D, .,,,
i rn 1 1 f n
It M W 1 8 10 W 20 25 10
AUGUST | SEPTEMBER
1
135
-------
MASS
3.0-
DAILY 2-°~
RAINFALL
IN
INCHES 1.0-
.03-
TOTAL P -°2-
.01-
4.0-
TSS 3.0-
a
VSS 2.0-
0.1-
TKN
a
NH3 .05-
3.0-
BOD
1.0-
ALL POLLUTANT
SCALES IN
KILOGRAMS
PER DAY
REMOVAL / RAINFALL PLOTS
PORT NORFOLK SI
POLLUTANT MASS REMOVALS NORMALIZED BY ANTECEDENT DAYS
JlpT
1 n fl l
l n
1 1 1 1 1 1
„
1 1
1,',,
1 1 n 1 1
j
n i n 1 1
1 6
FIGURE49-B
1
T m T ~l fl-l n TI-ITI-I fl
i 1 1 1 n 1 1 1 1 iiiniiiiiiiinniiiiiiiiiii
n n • n
l l M i l i l i l i i 1,1 l l l l l 1 1 l i i i i 1 1 i i i i i in i
L! '
IL-TSS
l—vss y |
i 1 1 1 i 1 1 1 1 a 1
n inininininiiininiininiiniilii
In
n
-TKN
H R
1 1 1 1 1 1 1 1 1 1 n i ii 1 1 n 1 1 1 1 1 1 1 1 1
COD
HI— BOD n
1 1 1 1 1 1 1 1 1 1 n 1 1 M i n i M 1 1 1 M i
10 15 20 25 SO i 1 5 10 IB
OCTOBER 1 NOVEMBER
1
136
-------
1*
3.0-
DAILY 2'°~
RAINFALL
IN
INCHES |.o-
.03-
TOTAL P -02~
.01-
4.0-
TSS 3.0-
a
VS S 2-0-
1.0-
0.1-
TKN
a
NH3 .05-
3.0-
T —
BOD
1.0-
ALL POLLUTANT
SCALES IN
KILOGRAMS
PER DAY
dASS REMOVAL / RAINFALL PLOTS
SHEPTON ST.
POLLUTANT MASS REMOVALS NORMALIZED BY ANTECEDENT DAYS
FIGURE 50-A
T 1 HT Ch T J~L T
1 1 1 1 1 1 1 lii i n 1 1 it 1 1 1 1 1 1 1 M rm 1 1 n h 1 1
n
1 1 1 1 1 1 1 1 1 MI 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 M 1 1 1 1 1 1 M 1 1 1
In g TSS
1 y 1 vss
Mil N H
MlnlnininlMninilllll
NH, a TKN
1 1 1 M ! 1 I MI 1 1 1 1 1 M 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 II
n C°D • •
N BOD 1 N H n
1 1 1 1 1 1 1 1 1 1 1 1 M 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 M 1 1 1 1 1 1
t» 26 31 1 B 10 » 20 26 50
1
AUGUST 1 SEPTEMBER
1
137
-------
MASS REMOVAL / RAINFALL PLOTS
SHEPTON ST.
3.0-1
DAILY 2-0-
RAINFALL
IN
INCHES 1.0-
.03-
TOTAL P •°2-
.01-
4.0-
TS S 3.0-
a
VSS 2.0-
'-
0.1-
TKN
ft
NH, .06-
3.0-
COD _ _
2.0-
B
BOD
1.0-
ALL POLLUTANT
SCALES IN
KILOGRAMS
PER DAY
POLLUTANT MASS REMOVALS NORMALIZED BY ANTECEDENT DAYS
FIGURE 50-B
4UnTT n-iT IfH n Ti-iTn fl
innii iifTiiiii h In 1 1 1 1 1 li i n n 1 1 Hi 1 1 1 1 1 1
it 1 1 1 1 1 1 1 1 1 1 1 1 1 n 1 1 1 1 1 1 1 1 1 1 1 1 1 M M 1 1 1 1 1 1 i 1 1 1 1 I
. 1 1 ill i 1 li 1 1 -
"iniiniiHiHiniininiiHiniininiiniiin
n_TKN
n — NHs
1 1 1 1 1 1 1 1 M 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 II 1 1 1 1 1 1 1 1
COD
— -BOD •
n 1
n i T 1 1 1 1 1 1 1 1 1 1 1 1 1 M < 1 1 1
1 t 10 16 K) tt Mil 9 tO W
OCTOBER 1 NOVEMBER
1
138
-------
MASS REMOVAL / RAINFALL PLOTS
TEMPLETON ST.
3.0-
DAILY
RA1NFALL
IN
INCHES
|.o-
TOTAL P
TSS
a
VSS
.03-
-02
.01-
4.0
3.0-
2.0
1.0-
0.1-
TKN
a
NH .05-
Ok
BOD
ALL POLLUTANT
SCALES IN
KILOGRAMS
PER DAY
3.0-
2.0-
1.0-
POLLUTANT MASS REMOVALS NORMALIZED BY ANTECEDENT DAYS
FIGURE .51 -A
Ff
-n
Hi rrn Ti rTh
u
iiiiiii i i n 1 1 1 1
n i n i n'n in
1 1 1 1 1
COD—
n BOD-U TI
111111111111 n i M 111111
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 M
I* 28 311 6 10 !• 20
1
AUGUST I SEPTEMBER
1
1 1 1 1 1 1 1
28 30
139
-------
MASS REMOVAL/ RAINFALL PLOTS
TEMPLETON ST.
3.0-j
DAILY 2-°-
RAINFALL
IN
INCHES 1.0-
.03-
TOTAL P -02~
.01-
4.0-
TSS 3.0-
a
vss 2.0-
1.0-
0.1-
TKN
a
NH3 .05-
3.0-
2.0—
BOD
1.0-
ALL POLLUTANT
SCALES IN
KILOGRAMS
PER DAY
POLLUTANT MASS REMOVALS NORMALIZED BY ANTECEDENT DAYS
FIGURE 51 -B
"
R_ r-,T T n-, T 1 FH n TnTn 11
n n 1 1 iirhini iiiniiiiihinniiiiiiiiiii
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 it 1 1 1 1 1 1 1 1 1 1 1 i M i n
n y
n U I
DID n y 1 1 D— Tss
..iiinilili
"mil™ ininin n I n 1 1 n I n i I n i n 1 1 1 i 1 1 1 11
1 1 1 1 1 1 1 1 1 1 1 1 M 1 1 { 1 1 1 1 1 1 1 1 1 1 1 1 II 1 1 1 1 M 1 1 1 1 1 1 1
..III 1
lililin lilimii nllilinliiilHiiiiiniiiii
1 B 10 IS 20 2B SO i 1 S 10 IB
OCTOBER 1 NOVEMBER
1
140
-------
DAILY
3.0-1
2.0-
MASS REMOVAL / RAINFALL PLOTS
WALNUT ST.
RAINFALL
IN
INCHES |.o-
TOTAL P
TSS
a
VSS
.03-
.02-
.01-
4.0-
3.0-
2.0-
1.0-
0.1-
TKN
a
NH, 05-
COD
a
BOD
3.0-
2.0-
1.0-
ALL POLLUTANT
SCALES IN
KILOGRAMS
PER DAY
POLLUTANT MASS REMOVALS NORMALIZED BY ANTECEDENT DAYS
FIGURE 52-A
I PI 11111 n n 11 r Hi rm M HTi ITI
Til I I I I II 1 I I I I I I I I I I I I I I I I I I I I I I I M
~TSS
IT
fl-
T I I II I I I I I Mi I I I I I I I I I I I I I I I I
H^TKN
^-NH,
I I I I I I 1 I I
II I I I I I I I I
I
-COD
.BOD
I I I I I I
19 20 SI I B
AUGUST I
I I I I I I I I I I I I I I I I I I I I I I I
IO IS 20 25 30
SEPTEMBER
141
-------
f
3.0-
DAILY 2-°~
RAINFALL
IN
INCHES 1.0-
.03-
TOTAL P -02~
.01-
4.0-
TSS 3.0-
a
VSS 2.0-
,.0-
0.1-
TKN
a
NH, .05-
3.0-
T «*-
BOD
1.0-
ALL POLLUTANT
SCALES IN
KILOGRAMS
PER DAY
dASS REMOVAL / RAINFALL PLOTS
WALNUT ST.
POLLUTANT MASS REMOVALS NORMALIZED BY ANTECEDENT DAYS
FIGURE 52-B
n-i M 11 n
i n 1 1 i m MI iiiniiiiii inn MINIMI
1 fl n fl
1 1 111 i i 1 1 n i MI 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 M i
Ii y
= = i 1 i . i 1 1 i::::;
MI ininin n i n M n i n 1 1 n i n 1 1 H 1 1 1 1 1
In r-, — TKN
H
111 i ii mi 1 1 1 n 1 1 1 1 1 n t M 1 1 n 1 1 1 1 1 1 1 1 1
COD
, 1 11 fl 1 BOD
i i y u i
niMin i i iiniiiiiniiiiiriiiiiiniiiiiMM
IS 10 15 20 Z5 30 1 5 10 IB
OCTOBER NOVEMBER
142
-------
TABLE 23. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING RESULTS
MEAN AND STANDARD DEVIATION OF MASS REMOVALS NORMALIZED
BY ANTECEDENT DAYS BETWEEN FLUSHES - PORT NORFOLK STREET TEST SEGMENT
POLLUTANT REMOVALS (kg /day/ flush)
TYPE
A: ALL
FLUSHES
B: ALL GOOD *
FLUSHES
C: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
D: ALL RAINFALL
IMPACTED
FLUSHES
E: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES UP
TO 11/1
COD
1.07
1.04
(9)
1.07
1.04
(9)
0.88
0.58
(5)
1.30
1.52
(4)
p. 83
0.64
(3)
BOD
0.46
0.46
(10)
0.46
0.46
(.10)
0.45
0.40
(6)
0.48
0.60
(4)
0.33
0.18
(4)
TKN
0.038
0.041
(9)
0.038
0.041
(9)
0.029
0.018
(5)
0.049
0.062
(4)
0.033
0.018
(4)
NH3
0.008
0.008
(9)
0.008
0.008
(9)
0.006
0.004
(5)
0.009
0.012
(4)
0.007
0.005
(4)
TP
0.007
0.008
(9)
0.007
0.008
(9)
0.004
0.004
(5)
0.011
0.011
(4)
0.004
0.004
(4)
' TSS
1.25
1.31
(19)
1.30
1.33
(18)
0.90
0.44
(ID
1.94
1.97
(7)
0.84
0.31
(8)
VSS
0.88
0.88
(19)
0.92
0.89
(18)
0.73
0.42
(11)
1.21
1.33
(7)
0.65
0.26
(8)
co
(continued;
* See definition in text, page 129.
-------
TABLE 23. Cont. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING RESULTS
MEAN AND STANDARD DEVIATION OF MASS REMOVALS NORMALIZED
BY ANTECEDENT DAYS BETWEEN FLUSHES - PORT NORFOLK STREET TEST SEGMENT
POLLUTANT REMOVALS (kg /day/flush)
TYPE
F:** ALL GOOD
NON- RAINFALL
IMPACTED
FLUSHES (FLUSH
RATE COMPARISON)
F:*** ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES (FLUSH
RATE COMPARISON)
G: ALL GOOD
B&R FLUSHES
COD
0.88
0.58
(5)
.
-
3.57
-
(1)
BOD
0.30
0.17
(5)
-
1.29
0.12
(2)
TKN
0.029
0.018
(5)
-
0.142
-
(1)
NH3
0.006
0.004
(5)
-
0.028
-
(1)
TP
0.004
0.004
(5)
-
0.027
-
(1)
TSS
0.77
0.35
(7)
0.87
0.19
(3)
2.86
1.33
(2)
vss
0.62
0.30
(7)
0.66 '
0.11
(3)
2.34
0.80
(2)
Excluding B&R flushes, all good non-rainfall impacted experiments divided into two sets using
median of truck delivery rates. Experiments where flush rate is less than or equals median
flush rate. (.19 - .48 cfs).
hli
-Excluding B&R flushes, all good non-rainfall impacted experiments divided into two sets using
median of truck delivery rates. Experiments where flush rate exceeds median flush rate (.63 -.85 cfs)
KEY: Top Row - Mean
Middle Row - Standard Deviation
Bottom Row (in parentheses) - Number of flushes used in computations
-------
TABLE 24. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING RESULTS
MEAN AND STANDARD DEVIATION OF MASS REMOVALS NORMALIZED
BY ANTECEDENT DAYS BETWEEN FLUSHES - SHEPTON STREET TEST SEGMENT
POLLUTANT REMOVALS (kg/day/flush)
TYPE
A: ALL
FLUSHES
B: ALL GOOD*
FLUSHES
C: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
D: ALL RAINFALL
IMPACTED
FLUSHES
E: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES UP
TO 11/1
COD
1.79
0.91
(2)
1.79
0.91
(2)
1.79
0.91
(2)
-
1.79
0.91
(2)
BOD
0.48
0.22
(11)
0.48
0.22
(ID
0.48
0.24
(8)
0.47
0.19
(3)
0.47
0.26
(7)
TKN
0.041
0.031
(2)
0.041
0.031
(2)
0.041
0.031
(2)
.
0.041
0.031
(2)
NH3
0.011
0.007
(2)
o.on
0.007
(2)
0.011
0.007
(2)
-
0.011
0.007
(2)
TP
o.on
0.008
(2)
0.011
0.008
(2)
0.011
0.008
(2)
-
o.on
0.008
(2)
TSS
1.29
0.65
(22)
1.29
0.60
(20)
1.32
0.59
(12)
1.25
0.65
(8)
1.28
0.63
(10)
VSS
1.00
0.53
(22)
1.00
0.49
(20)
1.05
0.50
(12)
0.93
0.51
(8)
1.02
0.53
(10)
(continued)
* See definition in text, page 129.
-------
TABLE 24. Cont. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING RESULTS
MEAN AND STANDARD DEVIATION OF MASS REMOVALS NORMALIZED
BY ANTECEDENT DAYS BETWEEN FLUSHES - SHEPTON STREET TEST SEGMENT
TYPE
F:** ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES (FLUSH
RATE COMPARISON)
F:*** ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES (FLUSH
RATE COMPARISON)
G: ALL GOOD
B&R FLUSHES
POLLUTANT REMOVALS (kg/ day/ flush)
COD
1.79
0.91
(2)
-
-
BOD
0.42
0.22
(4)
0.53
0.29
(4)
0.67
(1)
TKN
0.041
0.031
(2)
-
-
NH3
0.011
0.007
(2)
-
-
TP
0.011
0.008
(2)
-
-
TSS
1.40
0.54
(6)
1.23
0.68
(6)
1.80
1.02
(2)
vss
i.n
0.46
(6)
0.99
0.57
(6)
1.32
0.76
(2)
en
***
Excluding B&R flushes, all good non-rainfall impacted experiments divided into two sets using
median of truck delivery rates. Experiments where flush rate is less than or equals median
flush rate. -(.26 - -47 cfs)
Excluding B&R flushes, all good non-rainfall impacted experiments divided into two sets using
median of truck delivery rates. Experiments where flush rate exceeds median flush rate (.67 - .93 cfs)
KEY: "Top Row - Mean
Middle Row - Standard Deviation
Bottom Row (in parentheses) - Number of flushes used in computations
-------
TABLE 25. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING RESULTS
MEAN AND STANDARD DEVIATION OF MASS REMOVALS NORMALIZED
BY ANTECEDENT DAYS BETWEEN FLUSHES - TEMPLETON STREET TEST SEGMENT
POLLUTANT REMOVALS (kg/ day/ flush)
TYPE
A: ALL
FLUSHES
B. ALL GOOD*
FLUSHES
C: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
D: ALL RAINFALL
IMPACTED
FLUSHES
E: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES UP
TO 11/1
COD
1.18
—
(D
-
-
1.18
-
(D
-
BOD
0.95
0.70
(10)
1.05
0.74
(8)
1.04
0.52
(5)
0.85
1.04
(4)
0.90
0.57
(3)
TKN
0.032
^
(D
-
-
0.032
-
(D
-
NH3
0.003
_
(D
-
-
0.003
-
(D
-
TP
0.006
M
(D
-
-
0.006
-
(D
-
TSS
2.41
1.96
(19)
2.56
2.08
(16)
2.09
1.04
(10)
3.01
2.99
(7)
2.23
1.10
(8)
vss
1.79
1.49
(19)
1.91
1.58
(16)
1.57
0.79
(10)
2.22
2.29
(7)
1.66
0.82
(8)
(continued)
* See definition in text, page 129.
-------
TABLE 25. Cont. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING RESULTS
MEAN AND STANDARD DEVIATION OF MASS REMOVALS NORMALIZED
BY ANTECEDENT DAYS BETWEEN FLUSHES - TEMPLETON STREET TEST SEGMENT
TYPE
R ** ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES (FLUSH
RATE COMPARISON)
F:*** ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES (FLUSH
RATE COMPARISON)
G: ALL GOOD
B&R FLUSHES
POLLUTANT REMOVALS (kg/ day/ flush)
COD
-
-
-
BOD
0.64
0.47
(2)
1.15
0.39
(2)
1.64
0)
TKN
-
-
-
NH3
-
-
-
TP
-
-
-
TSS •
1.33
0.53
(5)
3.09
0.76
(4)
1.73
0.22
(2)
vss
1.00
0.37
(5)
2.35
0.55
(4)
1.16
0.27
(2)
• ,j^>
00
Excluding B&R flushes, all good non-rainfall impacted experiments divided into two sets using
median of truck delivery rates. Experiments where flush rate is less than or equals median
flush rate. (.36 - .49 cfs).
***
Excluding B&R flushes, all good non-rainfall impacted experiments divided into two sets using
median of truck delivery rates. Experiments where flush rate exceeds median flush rate (.56 - .83 cfs)
KEY: Top Row - Mean
Middle Row - Standard Deviation
Bottom Row (in parentheses) - Number of flushes used in computations
-------
TABLE 26. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING RESULTS
MEAN AND STANDARD DEVIATION OF MASS REMOVALS NORMALIZED
BY ANTECEDENT DAYS BETWEEN FLUSHES - WALNUT STREET TEST SEGMENT
POLLUTANT REMOVALS (kg/ day/ flush)
TYPE
A: ALL
FLUSHES
B: ALL GOOD *
FLUSHES
C: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
D: ALL RAINFALL
IMPACTED
FLUSHES
E: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES UP
TO 10/29
COD
2.32
1.95
(12)
2.32
1.95
(12)
1.56
0.87
(7)
3.38
2.64
(5)
1.30
0.72
(5)
BOD
0.78
0.56
(13)
0.78
0.56
(13)
0.55
0.27
(8)
1.16
0.72
(5)
0.45
0.17
(6)
TKN
0.062
0.032
(12)
0.062
0.032
(12)
0.051
0.018
(7)
0.079
0.042
(5) _J
0.046
0.068
(5)
NH3
0.025
0.017
(12)
0.025
0.017
(12)
0.023
0.012
(7)
0.029
0.023
(5)
0.021
0.010
(5)
TP
0.015
0.008
(12)
0.015
0.008
(12)
0.012
0.007
(7)
0.019
0.008
(5)
0.012
0.008
(5)
TSS
1.66
1.52
(21)
1.72
1.62
(18)
0.99
0.56
(12)
3.03
1.95
(7)
0.86
0.50
(9)
vss
1.02
0.88
(21)
1.04
0.95
(18)
0.64
0.38
(12)
1.75
1.17
(7)
0.54
0.33
(9)
(continued)
* See definition in text, page 129.
-------
TABLE 26. Cont. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING RESULTS
MEAN AND STANDARD DEVIATION OF MASS REMOVALS NORMALIZED
BY ANTECEDENT DAYS BETHEEN FLUSHES - WALNUT STREET TEST SEGMENT
POLLUTANT REMOVALS (kg /day/ flush)
TYPE
F:*?f ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES (FLUSH
RATE COMPARISON)
F:*** ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES (FLUSH
RATE COMPARISON)
G: ALL GOOD
B&R FLUSHES
COD
1.00
0.36
(3)
2.33
0.78
(3)
0.09
™"
(1)
BOD
0.41
0.15
(3)
0.71
0.30
(4)
0.34
"•
(1)
TKN
0.038
0.016
(3)
0.065
0.012
(3)
0.048
™
(1)
NH3
0.013
0.008
(3)
0.032
0.010
(3)
0.026
~
(1)
TP
0.006
0.005
(3)
0.018
0.003
(3)
0.016
~
(1)
TSS
0.79
0.37
(7)
1.38
0.76
(4)
0.81
(1)
VSS
0.59
0.34
(7)
0.78
0.50
(4)
0.36
(1)
en
o
**
Excluding B&R flushes, all good non-rainfall impacted experiments divided into two sets using
median of truck delivery rates. Experiments where flush rate is less than or equals median
flush rate. (.39 - .50 cfs).
Excluding B&R flushes, all good non-rainfall impacted experiments divided into two sets using
median of truck delivery rates. Experiments where flush rate exceeds median flush rate (.80 - .95 cfs)
KEY: Top Row - Mean
Middle Row - Standard Deviation
Bottom Row (in parentheses) - Number of-flushes used in computations
-------
flush removal data. Statistics for each of the groupings in these tables
are presented as follows: A) all flushes; B) all- good operational flushes;
C) all good non-rainfall impacted flushes; D) all rainfall impacted flushes;
E) all good non-rainfall impacted flushes with post cleaning performed;
F) all good non-rainfall impacted flushes where the experimental flush rates
are first less than the median flush rate of the experiments and then those
flushes where the flush rates exceed the median flush rate; and G) all
backup and release flushes. Comparison of groupings C and D show the
impact of rainfall within the periods between flushes. Comparison of
groupings C and E show the net effects of post cleaning the segments, or
equivalent!,/, the gross overall degree of flushing effectiveness assuming
that the post cleaning operation removes all residual pollutants after
flushing. Close agreement between these two groupings would suggest high
flushing removal efficiency. The comparison within groupings F roughly show
the impact of increased flush rate on pollutant removals. Partitioning the
sample set on the basis of the median flush rate was arbitrary.
Several general observations can be drawn from inspection of the
normalized pollutant removal statistics shown in Tables 23 through 26.
First of all, the coefficients of variation are all between 0.5 to 1.0
indicating that the variation of removal about the average is small. Second-
ly, comparison of the categories C and D statistics for the combined sewer
laterals on Port Norfolk and Walnut Streets show that the flushed loadings
for events impacted by rainfall events exceed the non-rainfall impacted
flushed masses by 50 to 100 %.
Similar comparison on the separate sewer streets shows mixed
results. The normalized rainfall impacted loadings for Shepton Street
generally decreased in comparison to the non-rainfall impacted events. The
results reversed for Templeton Street in that the normalized TSS and VSS wet
weather impacted flush loadings increased. One explanation for these
results is the following. Both streets receive clear water inflow from
roof drains connected into the sewers. The segments on both streets are
flat but are at the foot of a hill. Materials could either be washed into
the segment, thereby increasing the deposition loading or, washed out of
the segment, depending on the relative intensity of prior storm events.
Visual inspections of both segments during wet weather indicated that the
flow in Templeton Street was sluggish and under slight backwater conditions
from the main trunk sewer on Florida Street while the flow in Shepton Street
discharged freely. Materials would "wash-out" during storm events at
Shepton Street and settle more rapidly at Templeton Street. There was no
backwater effect at Templeton Street during dry weather conditions.
The third general observation of groupings C and E in Tables 23
through 26 is that the post flush cleaning operation with the exception of
Templeton Street reduced the average pollutant removals suggesting that the
flushing experiments were in general extremely effective. The issue of
flushing effectiveness is addressed more rigoriously in the phase two serial
flushing program described in Chapter 9.
The fourth observation of the data is that the average flushing
removal rates were greater for the higher flush rates. An analysis of
151
-------
flushing method effectiveness is presented in Section 8.5. The final
observation of the data is that the flushed loadings for the backup and
release experiments were comparable to the other removal rates. The backup
and release flush experiments at Shepton Street were extremely favorable and
were much higher than the average removals for good non-rainfall impacted
experiments. This result' led to the placement of the automatic sewer
flushing module on this segment. Operational results of the automatic
module is described in Chapter 11.
The average normalized flushed pollutant loadings presented in
Tables 23 through 26 for each segment were again normalized by estimates of
the total upstream tributary population including population contributions
along the segment. Two estimates of population for the Port Norfolk Street
segment were used. The census information indicated a population of 94
people while dry weather flow results described in Chapter 11 indicated
that a population estimate of 61 people would be more reasonable. The aver-
age normalized results per segment by antecedent period and population are
given in Table 27 for all good flushes and for all good operational non-
rainfall impacted events. The results for the two separated and the two
combined sewer streets were averaged and are presented in the last two rows
of Table 27. The average flushed loads for the combined sewer streets are
generally two to three times the loads for the separate streets.
8.4 Heavy Metals 'Flush Removals
Heavy metals were determined for the solids fraction of settled
flush wave composites including: cadmium, chromium, copper, lead, nickel,
zinc and mercury. Sample handling procedures for the composited flush wave
samples are discussed in Chapter 6. Preliminary heavy metal analyses for
both the supernatant and settled solids fractions showed that the heavy
metals within the flush wave supernatant were extremely low, on the order
of less than half a part per billion range. Flush wave composites were
settled for four hours under ideal quiescent conditions. The purpose of the
heavy metals tests conducted in this phase of work was aimed at assessing
dissolved and settleable fractions under ideal conditions. The heavy metals
analyses performed as part of the settling column experiments in the
second phase program portray a more realistic assessment of heavy metals
settleability characteristics. Those experiments showed that roughly half
of the flushed heavy metals would rapidly settle therefore, associated with
heavier solids particles and that the remaining fraction would not settle
within the time period of the settling column test and therefore would be
transported to a treatment facility. Results of those experiments are
presented and discussed in Chapter 10.
Results of the heavy metals analyses for the settled solids
fraction of the flush wave composites are presented in Tables 28 through 31
for each of the test segments. Heavy metal results are presented in terms
of micrograms per kilogram of total dry suspended solids. Table 32
summarizes the results given in Tables 28 through 31 in terms of the minimum,
average and maximum flushed metals per solid rates for each test segment.
High zinc and copper levels may be attributable to the copper and^brass
piping used in many of the older residences in the area. Statistical
152
-------
TABLE 27 . SUMMARY OF AVERAGE PHASE I FLUSHING POLLUTANT REMOVALS NORMALIZED
BY ANTECEDENT DAYS BETWEEN FLUSHES AND BY
ESTIMATED TRIBUTARY POPULATION
SHEPTON STREET (Separated)
(Estimated Tributary
TYPE
ALL
GOOD
FLUSHES
ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
TEMPLETON STREET (Se
(Estimated Tributary
ALL
GOOD
FLUSHES
ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
Population
COD
7.78
(2)
7.78
(2)
Darated)
Population
-
-
WALNUT STREET (Combined)
(Estimated Tributary
ALL
GOOD
FLUSHES
ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
Population
32.64
(12)
21.97
(7)
= 230)*
BOD
2.10
(11)
2.08
(8)
= 221)*
4.75
(8)
4.71
(5)
= 71)*
11.02
(13)
7.72
(8)
POLLUTANT
TKN
0.178
(2)
0.178
(2)
-
-
0.873
(12)
0.718
(7)
REMOVALS
NH3
0.048
(2)
0.048
(2)
-
-
0.352
(12)
0.324
(7)
(grams/ day/capita/f lush)
TP
0.048
(2)
0.048
(2)
-
-
0.211
(12)
0.169
(7)
TSS
5.60
(20)
5.72
(12)
11.58
(16)
9.46
(10)
24.23
(18)
13.90
(12)
VSS
4.35
(20)
4.57
(12)
8.64
(16)
7.10
(10)
14.65
(18)
8.97
(12)
en
oo
(continued)
-------
TABLE 27. Cont. SUMMARY OF AVERAGE PHASE I FLUSHING POLLUTANT REMOVALS NORMALIZED
BY ANTECEDENT DAYS BETWEEN FLUSHES AND BY
ESTIMATED TRIBUTARY POPULATION
PORT NORFOLK STREET (Combine
(Estimated Tributary Populat
TYPE
ALL
GOOD
FLUSHES
POP. =94
ALL
GOOD
FLUSHES
POP. =61
ALL GOOD
NON-RAINFALL
IMPACTED FLUSHES
POP. =94
ALL GOOD
NON-RAINFALL
IMPACTED FLUSHES
POP. =61
SUMMARY
SEPARATED
STREETS ALL
GOOD FLUSHES '
(AVERAGE)
COMBINED
STREETS ALL
GOOD FLUSHES
(AVERAGE)**
COD
11.38
(9)
17.54
(9)
9.30
(5)
14.34
(5)
7.78
22.01 -
25.09
POLLUTANT REMOVALS (grams/day/capita/flush)
•d)
ion = 94*; Estimated Population using DWF consideratio
BOD
4.94
(10)
7.61
(10)
4.819
(6)
7.43
(6)
3.43
7.98-
9.32
TKN
0.404
(9)
0.623
(9)
0.308
(5)
0.475
(5)
0.178
0.639-
0.748
NH3
0.079
(9)
0.123
(9)
0.064
(5)
0.098
(5)
0.048
0.215-
0.238
TP
0.074
(9)
0.115
(9)
0.043
(5)
0.067
(5)
0.048
0.143-
0.163
is = 61)
TSS
13.83
(18)
21.3
(18)
9.57
(11)
14.75
(11)
8.59
19.03-
22.77
VSS
9.77
(18)
15.06
(18)
7.79
(11)
12.00
(11)
6.50
12.21-
14.86
en
Population from census information.
Range reflects various tributary population
estimates for Port Norfolk Street.
-------
TABLE 28. PHASE I FIELD FLUSHING HEAVY METALS REMOVALS PER UNIT MASS OF
SOLIDS FLUSHED - PORT NORFOLK STREET TEST SEGMENT
DATE
8/30/76
9/2
9/7
9/16
9/21
9/24
10/1
10/4
10/8
10/12
10/15
10/18
10/25
10/22
10/29
11/1
11/5
11/8
11/12
A
32.9
46.5
32.8
47.0
32.6
46.6
33.7
B&R***
32.9
33.4
33.1
47.0
32.7
46.6
32.9
47.0
46.6
32.9
B&R***
B
.47
.92
.48
.44
.82
.79
.19
.46
.27
.33
.42
.63
.71
.46
.42
.85
.47
C D
1 x
3 x
5
9
5 x
3
7
3 x
4 x
4 x
3
3
3
4 x
4
3
4
3
4
E
20
2
82
41
64
72
37
F
x
X
X
X
X
X
X
X
X
X
X
X
X
X
X
G
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
MASS (yq
Cd Cr
6.00 5.5
6.28 34.8
-
6.73
4.44 41.7
2.09 40.2
7.72 9.7
8.31 60.1
2.72 37.0
2.26 39.7
8.17 54.8
_
-
6.17
2.37 2.9
2.69 33.7
3.78 31.5
6.97 31.7
OF METALS/kq TOTAL
Cu
1517
1093
-
334
375
322
807
134
308
255
411
-
-
252
47
203
366
412
Pb
42.4
137.0
-
244.9
269.7
138.9
8.3
559.0
454.0
60.4
624.2
-
-
209.7
43.3
126.9
275.6
338.0
SUSPENDED SOLIDS)
Ni
28.8
28.6
-
48.9
54.7
12.3
27.8
5.8
21.0
12.7
44.6
-
-
22.2
5.3
42.4
28.4
19.9
Zi
1261
1155
-
1510
1293
689
1533
2042
282
892
1180
-
_
855
129
817
948
1172
Hq
10.70
-
-
-
3.16
1.08
-
3.50
0.50
-
1.71
-
-
1.72
2.11
2.08
-
-
en
en
Legend
A - Truck Flush Volume (cf)
B - Truck Flush Rate (cfs)
C - Total Antecedent Days Between Flushes
D - Rainfall Impacted Event
E - Hours After Rainfall Event to Flush*
F - Segment Cleaned After Flush**
G - Good Flush
Number of hours between rainfall exceeding 0.15 inch/hour and flush event.
**Segment flushed after experiment using maximum truck discharge for five minutes.
***Backup and release.
-------
TABLE 29: PHASE I FIELD FLUSHING HEAVY METALS REMOVALS PER UNIT MASS OF
SOLIDS FLUSHED - SHEPTON STREET TEST SEGMENT
DATE
8/23/76
8/26
8/30
9/2
9/7
9/10
9/13
9/16
9/21
9/24
10/1
10/4
10/8
10/12
10/15
10/18
10/22-
10/25
10/29
11/1
11/5
11/8
11/12
A
47.8
35.4
37.3
B&R***
20.0
32.6
B&R***
46.6
32.9
33.1
46.5
32.9
32.9
B&R***
46.6
33.0
32.9
47.2
46.6
32.7
33.0
47.1
B&R***
B C
.26 5
.76 3
.65 4
3
5
.69 3
3
.67 3
.42 5
.37 3
.93 7
.47 3
.46 4
4
.71 3
.47 3
.43 4
.37 3
.72 4
.61 3
.40 4
.41 3
4
D
X
X
X
X
X
X
X
X
E
23
4
66
86
45
60
75
35
F
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
G
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
MASS
Cd
1.71
-
7.57
4.48
4.08
5.34
-
4.10
3.88
2.72
6.01
2.25
3.72
4.85
2.84
5.56
-
3.93
5.80
2.50
(yq OF METALS/kq TOTAL
Cr
33.4
-
56.8
37.4
31.9
40.1
-
27.5
34.7
18.3
52.6
29.2
31.0
53.0
15.4
52.8
-
32.7
13.8
23.4
Cu
97
-
353
221
539
348
-
214
242
252
196
188
280
216
175
216
-
239
357
161
Pb
61.7
-
298.4
232.2
98.3
570.7
-
44.4
129.0
269.0
524.2
45.0
300.8
235.5
15.9
161.7
-
190.7
70.4
121.4
SUSPENDED SOLIDS)
Ni
6.9
-
25.5
26.9
19.7
3260.0
-
23.1
12.0
21.0
37.8
11.9
16.7
10.9
13.1
20.0
-
20.6
26.1
11.2
Zi
20
-
1310
1210
1058
1306
-
13
722
770
938
549
1074
911
407
899
-
907
550
975
H9
-
-
-
-
3.65
-
- •
3.47
-
-
4.13
-
3.90
4.40
3.50
2.48
-
1.67
1.70
1.70
01
Legend
A - Truck Flush Volume (cf)
B - Truck Flush Rate (cfs)
C - Total Antecedent Days Between Flushes
D - Rainfall Impacted Event
E - Hours After Rainfall Event to Flush*
F - Segment Cleaned After Flush**
G - Good Flush
Number of hours between rainfall exceeding 0.15 inch/hour and flush event.
**
Segment flushed after experiment using maximum truck discharge for five minutes.
Backup and release.
-------
TABLE 30. PHASE I FIELD FLUSHING HEAVY METALS REMOVALS PER UNIT MASS OF
SOLIDS FLUSHED - TEMPLETON STREET TEST SEGMENT
DATE
9/2
9/7
9/10
9/13
9/16
9/21
9/24
10/1
10/4
10/8
10/12
10/15
10/18
10/22
10/25
10/29
11/1
11/5
11/8
11/12
.A
32.8
B&R***
57.4
32.9
B&R***
B&R***
33.1
B&R***
46.6
32.9
32.9
47.1
46.6
32.7
32.9
46.9
46.7
32.7
32.9
B&R***
B
.50
.83
.43
.36
.87
.46
.43
.41
.76
.59
.46
.49
.75
.56
.43
C
3
5
3
3
3
5
3
7
3
4
4
3
4
3
4
3
3
4
3
4
D
X
X
X
X
X
X
X
E
5
65
87
43
62
73
• 34
F
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
G
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
MASS
Cd
4.11
6.72
5.77
8.63
8.93
,7.65
3.12
7.51
5.05
2.56
2.04
8.40
3.56
3.13
-
3.75
7.90
3.02
4.14
18.90
(yg OF METALS/kq TOTAL
O Cu
29.5 1565
246
45.5 318
47.9 385
51.0 1172
35.8 370
19.5 120
40.2 354
98.4 643
33.0 259
30.8 233
58.1 295
35.6 823
43.0 211
-
39.1 254
14.1 141
25.2 339
31.0 274
35.7 195
Pb
169.0
128.0
233.6
372.5
52.1
684.2
75.7
598.9
315.0
238.5
280.0
410.2
62.9
68.3
-
242.8
548.1
244.5
355.2
277 .4
SUSPENDED SOLIDS)
Ni
21.6
20.4
28.1
25.9
46.4
34.4
14.0
55.8
28.4
18.1
8.9
25.2
18.8
22.5
-
22.5
45.7
22.7
12.4
28.6
Zi
1572
31
2230
mo
1438
1280
361
1116
Hq
2,62
-
-
-
-
-
5.24
-
352 13.1
845 1
525
607
854
975
-
578
314
786
718
601
4.0
-
3.75
-
2.62
-
1.71
2.67
-
-
-
en
•-J
Legend
A - Truck Flush Volume (cf)
B - Truck Flush Rate (cfs)
C - Total Antecedent Days Between Flushes
D - Rainfall Impacted Event
E - Hours After Rainfall Event to Flush*
F - Segment Cleaned After Flush**
G - Good Flush
**Number of hours between rainfall exceeding 0.15 inch/hour and flush event.
Segment flushed after experiment using maximum truck discharge for five minutes.
Backup and release.
-------
TABLE 31 . PHASE I FIELD FLUSHING HEAVY METALS REMOVALS PER UNIT MASS OF
SOLIDS FLUSHED - WALNUT STREET TEST SEGMENT
DATE
8/23/76
8/26
8/30
9/2
9/7
9/10
9/13
9/16
9/21
9/24
10/1
10/4
10/8
10/12
10/15
10/18
10/22
10/25
10/29
11/1
11/5
11/8
11/12
A
69.8
46.6
47.1
60.6
B&R***
32.0
46.5
32.0
46.6
B&R***
33.0
32.9
B&R***
47.0
32.8
33.0
47.3
46.5
32.9
32.9
47.0
46.6
B&R***
B C
.80 3
.86 3
.41 4
.65 3
3
.47 3
.96 3
1.39 3
.89 5
.39* 7
.46 3
4
.98 4
.50 3
.42 3
.35 4
.95 3
.47 4
.43 3
.42 4
.88 3
4
D
X
X
X
X
X
X
X
X
E
21
1
62
84
40
71
38
F
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
G
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
MASS
Cd
5.82
9.08
7.58
7.70
8.16
9.03
5.81
5.67
8.16
4.40
7.63
5.80
5.03
4.67
3.94
8.56
7.20
-
(vfl
Cr
30.3
69.9
79.6
52.5
27.7
98.9
52.8
56.9
56.6
14.7
35.6
45.3
26.6
87.5
20.7
45.9
32.7
-
OF METALS/ kq TOTAL
Cu
1883
2670
1063
456
1722
863
395
801
343
87
83
284
74
151
180
437
249
-
Pb
394.0
711.0
246.0
186.9
874.0
226.0
615.5
50.6
176.0
99.7
144.0
387.4
108.0
70.9
191.6
296.5
285.9
-
SUSPENDED SOLIDS)
Ni
34.8
62.4
76.2
50.4
63.0
20.3
47.3
35.5
48.7
9.9
35.3
39.9
45.2
15.7
70.7
84.9
58.5
-
Zi
1295
2262
2145
1540
1629
1680
1710
904 1
1360
371
1419
1130 1
1105
10009 1
1750
1550
993
-
Hg
5.29
-
-
-
6.33
8.99
5.44
8.8
3.19
5.0
-
8.89
-
1.44
5.29
3.83
3.33
-
01
00
Legend
A - Truck Flush Volume (cf)
B - Truck Flush Rate (cfs)
C - Total Antecedent Days Between Flushes
D - Rainfall Impacted Event
E - Hours After Rainfall Event to Flush*
F - Segment Cleaned After Flush**
G - Good Flush
*Number of hours between rainfall exceeding 0.15 inch/hour and flush event.
**Segment flushed after experiment using maxium truck discharge for five minutes.
***Backup and release.
-------
TABLE 32. SUMMARY OF PHASE I HEAVY METALS MASS LOADINGS
PER UNIT MASS OF SOLIDS FLUSHED
HEAVY METALS RATES (yq/kg OF SOLIDS)
SHEPTON
CADMIUM
CHROMIUM
COPPER
LEAD
NICKEL
ZINC
MERCURY
Win.
1.7
13.8
175.0
15.9
10.9
13.3
1.7
Avg.
4.2
34.4
252.6
198.2
18.9*
810.0
3.1
Max.
7.5
56.8
539.0
570.7
3260.0
1360.0
3.65
Min.
2.0
14.1
120.0
52.1
12.4
31.0
1.7
TEMPLETON
Avq.
6.0
39.6
431.4
281.9
26.3
857.5
5.7
Max.
18.9
98.4
1565.0
684.2
55.8
1577.0
14.0
PORT NORFOLK
Min.
2.1
2.9
47.0
8.3
5.3
129.0
0.5
Avg.
5.1
32.6
455.7
235.5
26.9
1050.5
2.9
Max.
8.3
60.1
1517.0
624.2
54.7
1533.0
10.7
Min.
3.9
14.7
73.6
50.6
9.9
371.0
3.2
WALNUT
Avq.
6.7
49.1
690.6
297.9
47.0
1932.5
8.0
Max.
9.1
98.9
2670.0
615.0
84.9
2262.0
18.9
en
to
Excludes high nickel concentration flush - 9/21 - (3260 mg/kg)
-------
summaries of the total heavy metals mass removals, in micrograms per flush,
are presented in Tables 33 through 36 for each of the test segments. The
mean, standard deviation and number of flushing experiments used in the
computations are given for each pollutant. The data set is again partitioned
into four groupings including all flushes, all good flushes, all good
non-rainfall impacted flushes and rainfall impacted flushes. These results
were computed using the solids mass removals given in Tables 19 through 22
and the heavy metals solids concentration given in Tables 28 through 31
for each test segment. The coefficients of variation range from about 0.5 to
1.5 which are low considering the unknown origin of the pollutant sources.
Heavy metals results normalized by antecedent periods between
flushes for each pollutant and data groups are given in Tables 37 through 40
for each test segment. 'Inspection of the results shows that the heavy
metals loadings for the combined sewer streets are generally much higher
than for the separated sewer segments. Table 41 presents a summary of the
heavy metals removals both normalized by antecedent days between flushes and
estimates of the tributary population. Again these results show significant
differences between the combined and separated sewer segments. The results
are otherwise scattered to make any other general observations.
8.5 Comparison of Flush Methods^
One aim of the first phase flushing operation was to assess pollu-
tant removal effectiveness of different flushing methods for small diameter
laterals. Plots are presented in Figure 53 depicting flushed TSS and VSS
loadings normalized by antecedent periods between experiments versus flush
discharge rate for all good operational flushes for the Shepton and Tempieton
Streets test segments. Similar plots for Port Norfolk and Walnut Streets are
shown in Figure 54. Flush volumes are grouped into 35 and 50 cubic feet
categories and differences between nonrainfall and rainfall impacted flushes*
are also noted. Finally, results for backup and release flushes are also
noted in the plots for both nonrainfall and rainfall impacted flushes.
Visual inspection of the plots indicates that the TSS and VSS re-
movals are roughly invariant of the method used but tend to increase with
both the higher flush rates and the higher flush volumes. This observation
was derived in the following manner. The TSS and VSS flushed masses normal-
ized by antecedent periods for all good nonrainfall impacted flushes shown
in Figures 53 and 54 were compared with their respective averages for each
street. Backup and release flushes were excluded from the comparison. The
flushing experiments were arbitrarily divided into a four-way categorization
by flush volume and flush rate, that is, flush rates less than 0.5 cfs and
flush rates exceeding 0.5 cfs, and flush volumes approximately equal to 35
and 50 cubic feet, respectively. These categories are shown below and are
labeled as low volume/low rate, low volume/high rate, high volume/low rate,
and high volume/high rate. The total number of experiment:; where the normal-
ized TSS and VSS flushed masses exceeded the good nonrainfall impacted
TSS and VSS averages per segment over all four segments were tallied and
are noted below. The total number of events used in this comparison is 41
flushes with about an eqijal division between the four categories. Inspection
* See definition on page 129.
160
-------
of the table below shows that the greater preponderance of flush masses ex-
ceeding the mean TSS and VSS mass removals per street occurs within the high
volume/high rate flush category.This implies that the experiments with flushing
volumes of about 50 cubic feet and rates exceeding 0.5 cfs were more effective
in comparison to the average removals than were, for example, low volume/low
rate flushes. This flush category typified both high flush energy ...and momen-
tum. Another observation that can be drawn from the table is that pollutant
removals exceeded the average more frequently for the higher flush volume ex-
periments (independent of rate) than for the higher flush rate experiments
(independent of volume). The optimal condition, however, is for both high
rates and high volumes.
NUMBER OF GOOD NONRAINFALL IMPACTED TSS AMD VSS
NORMALIZED FLUSH MASSES EXCEEDING AVERAGE REMOVALS PER STREET
FOR DIFFERENT FLUSHING CONDITIONS
Low Volume/Low Rate
TSS: 2
VSS: 4
High Volume/Low Rate
TSS: 6
VSS: 6 .
Low Volume/High Rate
TSS: 2
VSS: 2
High Volume/High Rate
TSS: 10 •
VSS: 10
Comparison of the backup and release flushes for the four test seg-
ments indicate that the flush at Port Norfolk exceeds the average dry weather
TSS and VSS mass removals, whereas the flushed pollutant masses for experi-
ments on Walnut and Tempieton Streets are less than the average. Backup and
release flushes on Shepton Street were all conducted following a rainfall
event. The average TSS and VSS removals for the three backup and release
flushes on Shepton Street exceed the averages for the rainfall impacted
events. These results agree well with field experiemce since backup and re-
lease flushes were difficult to accomplish on Templeton and Walnut Streets.
The above discussion deals with the analysis of the optimal flush
method with respect to flush pollutant removal efficiency. All methods re-
sulted in comparable removal effectiveness. Another issue worth investigating
is over-flushing. Figure 55 and 56 show plots of cumulative VSS mass removals
versus the cumulative fraction of flush volume passing the sampling manhole
at Shepton Street for delivered flush truck volumes of 35 and 50 cubic feet,
respectively. Flushing rates used in the experiments are indicated on the two
figures. Both plots show that ninety percent mass removal occur at roughly
the "knee-of-the-curve" points. The average percent volume corresponding to
the "knee-of-the-curve" point is roughly 80% in Figure 55 for the 35 cubic loot
flushes and 70% for the 50 cubic foot flushes in Figure 56. This means that
ninety percent of the load would be moved or transported by flush volumes of
about 30 cubic feet. Flush volumes of 30 cubic feet are recommended for single
segment flushing.
161
-------
TABLE 33. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING HEAVY METALS RESULTS
MEAN AND STANDARD DEVIATION OF TOTAL MASS REMOVALS (FLUSH SOLIDS FRACTION)
PORT NORFOLK STREET TEST SEGMENT
TYPE
A: ALL
FLUSHES
B: ALL GOOD
FLUSHES
C: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
D: ALL RAINFALL
IMPACTED
FLUSHES
HEAVY METAL REMOVALS (yg/flush)
CADMIUM
27.14
30.01
(15)
27.14
30.01
(15)
16.73
15.25
(10)
47.94
42.61
(5)
CHROMIUM
119.5
66.7
(13)
119.5
66.7
(13)
121.2
79.2
(9)
115.7
31.9
(4)
COPPER
3278.
6464.
(15)
3278.
6464.
(15)
1247.
979.
(10)
7341.
10637.
(5)
LEAD
854.8
789.6
(15)
854.8
789.6
(15)
849.2
916.9
(10)
866.1
539.1
(5) -
NICKEL
128.9
127.5
(15)
128.9
127.5
(15)
90.5
63.3
(10)
205.7
191.9
(5)
ZINC
5443.
.6039.
(15)
5443.
6039.
(15)
3196.
2403.
(10)
9939.
8763.
(5)
MERCURY
25.51
57.28
(9)
25.51
57.28
(9)
6.78
2.09
(6)
62.96
99.79
(3)
ro
KEY: Top Row - Mean
Middle Row - Standard Deviation
Bottom Row (in parentheses) - Number of flushes used in computations
-------
TABLE 34. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING HEAVY METALS RESULTS
MEAN AND STANDARD DEVIATION OF TOTAL MASS REMOVALS (FLUSH SOLIDS FRACTION)
SHEPTON STREET TEST SEGMENT
TYPE
A: ALL
FLUSHES
B: ALL GOOD
FLUSHES
C: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
D: ALL RAINFALL
IMPACTED
FLUSHES
HEAVY METAL REMOVALS (yg/ flush)
CADMIUM
18.50
8.60
(17)
18.25
8.42
(15)
17.62
9.52
(10)
15.50
5.30
(6)
CHROMIUM
151.4
71.5
(17)
150.2
70.9
(15)
142.4
79.1
(10)
146.7
67.5
(6)
COPPER
1186.
805.
(17)
968.
557.
(15)
1031.
557.
(10)
846.
389.
(6)
LEAD
882.6
759.4
(17)
912.5
779.3
(15)
622.6
604.2
(10)
707.2
757.8
(6)
NICKEL
145.4
256.9
(17)
150.8
264.6
(15)
76.5
36 ."9
(10)
191.3
379.4
(6)
ZINC
3551.
2170.
(17)
3321.
2026.
(15)
2945.
2254.
(10)
2598.
1989.
(6)
MERCURY
12.50
7.20
(17)
11.38
6.08
(11)
10.70
6.34
(9)
10.80
3.00
(6)
en
co
KEY: Top Row - Mean
Middle Row - Standard Deviation
Bottom Row (in parentheses) - Number of flushes used in computations
-------
TABLE 35. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING HEAVY METALS RESULTS
MEAN AND STANDARD DEVIATION OF TOTAL MASS REMOVALS (FLUSH SOLIDS FRACTION)
TEMPLETON STREET TEST SEGMENT
TYPE
A: ALL
FLUSHES
B: ALL GOOD
FLUSHES
C: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
D: ALL RAINFALL
IMPACTED
FLUSHES
HEAVY METAL REMOVALS (yg/ flush)
CADMIUM
42.18
27.39
(18)
43.57
29.18
(15)
44.12
30.58
(9)
40.43
27.86
(7)
CHROMIUM
346.6
263.7
(18)
354.3
275.8
(15)
281.0
181.8
(9)
419.0
357.2
(7)
COPPER
4920.
6888.
(18)
4837.
7172.
(15)
3013.
3077.
(9)
6661.
10038.
(7)
LEAD
5657.4
9489.0
(18)
6306.1
9952.6
(15)
8169.4
12508.1
(9)
3173.0
3210.0
(7)
NICKEL
228.9
176.7
(18)
237.1
181.3
(15)
252.5
209.7
(9)
194.8
124.0
(7)
ZINC
8873.
8357.
(17)
8770.
8939.
(15)
7921.
7803.
(9)
9631.
9383.
(7)
MERCURY
33.81
28.18
(8)
33.81
28.18
(8)
17.82
5.04
(4)
49.80
32.42
(4)
en
KEY: Top Row - Mean
Middle Row - Standard Deviation
Bottom Row (in parentheses) - Number of flushes used in computations
-------
TABLE 36. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING HEAVY METALS RESULTS
MEAN AND STANDARD DEVIATION OF TOTAL MASS REMOVALS (FLUSH SOLIDS FRACTION)
WALNUT STREET TEST SEGMENT
HEAVY METAL REMOVALS (yg/ flush)
TYPE
A: ALL
FLUSHES
B: ALL GOOD
FLUSHES
C: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
D: ALL RAINFALL
IMPACTED
FLUSHES
CADMIUM
41.90
52.00
(18)
29.30
17.00
(14)
21.52
13.44
(10)
81.98
77.29
(6)
CHROMIUM
291.6
364.7
(18)
194.5
129.6
(14)
142.9
76.8
(10)
566.0
551.7
(6)
COPPER
4265.
5328.
(18)
3300.
4514.
(14)
1086.
925.
(10)
10380.
5244.
(6)
LEAD
1826.0
1785.9
(18)
1469.6
1576.0
(14)
847.9
812.9
(10)
3625.6
1864.8
(6)
NICKEL
282.70
313.80
(18)
221.17
127.64
(14)
178.27
124.09
(10)
522.85
446.73
(6)
ZINC
8963.
8709.
(18)
7027.
3998.
(14)
6074.
4161.
(10)
15307.
11166.
(6)
MERCURY
51.30
56.30
(12)
50.24
59.71
(10)
16.52
5.87
(6)
99.37
54.54
(5)
en
KEY: Top Row - Mean
Middle Row - Standard Deviation
Bottom Row (in parentheses) - Number of flushes used in computations
-------
TABLE 37. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING HEAVY METALS RESULTS
MEAN AND STANDARD DEVIATION OF MASS REMOVALS NORMALIZED BY
ANTECEDENT DAYS BETWEEN FLUSHES (FLUSH SOLIDS FRACTION) - PORT NORFOLK STREET TEST SEGMENT
TYPE
A: ALL
FLUSHES
B: ALL GOOD
FLUSHES
C: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
D: ALL RAINFALL
IMPACTED
FLUSHES
HEAVY METAL REMOVALS (yg/day/flush )
CADMIUM
8.00
9.75
(15)
8.00
9.75
(15)
4.54
3.96
(10)
14.90
13.51
(5)
CHROMIUM
32.4
18.1
(13)
32.4
18.1
(13)
32.1
21.4
(9)
33.1
6.0
(4)
COPPER
1010.
2102.
(15)
1010.
2102.
(15)
316.
223.
(10)
2397.
3204.
(5)
LEAD
232.1
224.7
(15)
232.1
224.7
(15)
237.3
261.8
(10)
221.8
119.2
(5)
NICKEL
37.82
42.12
(15)
37.86
42.12
(15)
25.24
19.62
(10)
63.10
59.97
(5)
ZINC
1929.
2187.
(15)
1929.
2187.
(15)
1355.
1532.
(10)
3076.
2772.
(5)
MERCURY
8.28
18.02
(9)
8.28
18.08
(9)
2.08
0.76
(6)
20.69
27.36
(3)
KEY: Top Row - Mean
Middle Row - Standard Deviation
Bottom Row (in parentheses) - Number of flushes used in computations
-------
TABLE 38. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING HEAVY METALS RESULTS
MEAN AND STANDARD DEVIATION OF MASS REMOVALS NORMALIZED BY
ANTECEDENT DAYS BETWEEN FLUSHES (FLUSH SOLIDS FRACTION) - SHEPTON STREET TEST SEGMENT
TYPE
A. ALL
FLUSHES
B: ALL GOOD
FLUSHES
C: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
D: ALL RAINFALL
IMPACTED
FLUSHES
HEAVY METAL REMOVALS (yg/day/flush)
CADMIUM
5.41
3.01
(17)
5.70
2.84
(16)
5.84
3.73
(9)
4.80
1.47
(7)
CHROMIUM
44.3
23.8
(17)
46.7
122.5
(16)
45.9
24.1
(9)
47.7
20.3
(7)
COPPER
354.
272.
(17)
374.,
268.
(16)
455.
325.
(9)
270.
98.
(7)
LEAD
244.7
196.8
(17)
258.0
195.3
(16)
216.8
216.3
(9)
310.9
152.6
(7)
NICKEL
61.04
114.62
(17)
64.67
117.37
(16)
74.26
145.32
(9)
52.33
71.50
(7)
ZINC
1108.
667.
(16)
1165.
650.
(15)
1393.
737.
(8)
904.
417.
(7)
MERCURY
3.71
2.56
(12)
4.01
2.46
(11)
4.09
2.71
(9)
3.61
(2)
en
KEY: Top Row - Mean
Middle Row - Standard Deviation
Bottom Row (in parentheses) - Number of flushes used in computation
-------
TABLE 39. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING HEAVY METALS RESULTS
MEAN AND STANDARD DEVIATION OF MASS REMOVALS NORMALIZED BY
ANTECEDENT DAYS BETWEEN FLUSHES (FLUSH SOLIDS FRACTION) - TEMPLETON STREET TEST SEGMENT
HEAVY METAL REMOVALS (yg/ day/flush)
TYPE
A: ALL
FLUSHES
B: ALL GOOD
FLUSHES
C: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
D: ALL RAINFALL
IMPACTED
FLUSHES
CADMIUM
12.20
8.34
(18)
11.98
8.31
(15)
11.96
7.69
(9)
11.54
9.14
(7)
CHROMIUM
90.4
65.5
(17)
89.6
67.7
(15)
76.3
48.1
(9)
100.9
86.4
(7)
COPPER
1305.
2247.
(18)
1305.
2388.
(15)
717.
782.
(9)
1932.
3414.
(7)
LEAD
651.3
619.6
(18)
738.8
643.5
(15)
673.7
515.4
(9)
771.4
732.4
(7)
NICKEL
60.00
45.24
(18)
60.30
43.87
(15)
64.69
45.63
(9)
49.84
38.71
(7)
ZINC
2500.
2784.
(18)
2488.
2925.
(15)
2168.
2597.
(9)
2707.
3079.
(7)
MERCURY
8.45
5.02
(8)
8.45
5.02
(8)
5.94
1.99
(4)
10.92
5.36
(4)
en
oo
KEY: Top Row - Mean
Middle Row - Standard Deviation
Bottom Row (in parentheses) - Number of flushes used in computations
-------
TABLE 40. STATISTICAL SUMMARY OF PHASE I SEWER FLUSHING HEAVY METALS RESULTS
MEAN AND STANDARD DEVIATION OF MASS REMOVALS NORMALIZED BY
ANTECEDENT DAYS BETWEEN FLUSHES (FLUSH SOLIDS FRACTION) - WALNUT STREET TEST SEGMENT
TYPE
A: ALL
FLUSHES
B: ALL GOOD
FLUSHES
C: ALL GOOD
NON-RAINFALL
IMPACTED
FLUSHES
D: ALL RAINFALL
IMPACTED
FLUSHES
HEAVY METAL REMOVALS (yg/day/ flush)
CADMIUM
11.45
11.69
(18)
11.57
12.59
(15)
6.23
4.33
(10)
21.58
15.69
(6)
CHROMIUM
81.8
93.3
(18)
80.1
101.0
(15)
40.5
22.8
(10)
157.4
135.0
(6)
COPPER
1260.
1627.
(18)
1067.
1318.
(15)
312.
290.
(10)
3084.
1687.
(6)
LEAD
512.7
534.2
(18)
468.7
516.1
(15)
221.6
176.9
(10)
1052.2
608.9
(6)
NICKEL
80.41
80.77
(18)
84.68
85.05
(15)
52.30
37.98
(10)
126.74
119.96
(6)
ZINC
2505.
2201.
(18)
2522.
2300.
(15)
1689.
1083.
(10)
4285.
2931.
(6)
MERCURY
14.53
17.12
(12)
15.26
17.76
(11)
4.28
1.77
(6)
28.40
19.65
(9)
en
to
KEY: Top Row - Mean
Middle Row - Standard Deviation
Bottom Row (in parentheses) - Number of flushes used in computations
-------
TABLE 41. SUMMARY OF AVERAGE PHASE I HEAVY METAL MASS REMOVALS NORMALIZED BY
ANTECEDENT DAYS BETWEEN FLUSHES AND
BY ESTIMATED TRIBUTARY POPULATION
SHEPTON STREET (Separated)
(Estimated
TYPE
ALL
GOOD
FLUSHES
Tributary Population
CADMIUM
24.82
ALL GOOD
NON-RAINFALL 9t- on
IMPACTED "••"
FLUSHES
TEMPLETON
(Estimated
ALL
GOOD
FLUSHES
STREET (Separated)
Tributary Population
55.21
ALL GOOD
NON-RAINFALL ,. ,,
IMPACTED Wtl*
FLUSHES
WALNUT STREET (Combined)
(Estimated
ALL
GOOD
FLUSHES
Tributary Population
161.34
ALL GOOD
NON-RAINFALL on 14
IMPACTED °y>1^
FLUSHES
HEAVY METAL MASS REMOVALS (nanograms/day/capita/flush)
= 230.)*
CHROMIUM COPPER LEAD NICKEL ZINC
203.3 1627. 1121.8 281.16 5065.
199.8 1979. 943.0 322.89 6058.
= 221.)*
409.2 5908. 2947.4 271.49 11315.
345.5 3245. 3048.4 292.71 9813.
= 71.)*
1152.6 17757. 7221.7 1132.58 35287.
570.6 4402. 3121.7 736.62 23789.
MERCURY
17.42
17.80
38.24
26.89
204.63
60.27
•-J
o
(continued)
-------
TABLE 41. Cont. SUMMARY OF AVERAGE PHASE I HEAVY METAL MASS REMOVALS NORMALIZED BY
ANTECEDENT DAYS BETWEEN FLUSHES AND
BY ESTIMATED TRIBUTARY POPULATION
PORT NORFOLK STREET (Combined
(Estimated Tributary Populati
TYPE
ALL
GOOD
FLUSHES
POP. =94
ALL
GOOD
FLUSHES
POP. =61
ALL GOOD
NON-RAINFALL
IMPACTED FLUSHES
POP. =94
ALL GOOD
NON-RAINFALL
IMPACTED FLUSHES
POP. =61
SUMMARY
SEPARATED
STREETS ALL
GOOD FLUSHES
(AVERAGE)
COMBINED
STREETS ALL
GOOD FLUSHES
(AVERAGE)**
CADMIUM
85.07
131.09
48.33
74.48
40.02
123.21-
146.22
HEAVY METAL MASS REMOVALS (nanograms/day/capita/flush)
)
on = 94*; Estimated Population using DWF considerations =61)
CHROMIUM
345.4
532.3
342.2
527.3
306.3
749.0-
842.5
COPPER
10750.
16566.
3370.
5194.
3767.
14254.-
17162
LEAD
2469.6
3805.7
2524.7
3890.5
2034.4
4845.7-
5513.7
NICKEL
402.79
620.69
268.54
413.82
276.33
767.69-
876.64
ZINC
20523.
31625.
14422.
22224.
8190
27905.-
33456.
MERCURY
88.12
135.79
22.15
34.13
27.83
146.38-
170.21
**
\
Population from census information.
Range reflects various tributary population estimates for Port Norfolk Street.
-------
TSS
VSS
*
9
s
u
cc
wO.5-
4.0-,
Q
LU
,ao-
co
<
S
1.0-
0
TSS
(830) -
0
02
0.4
0.6
0.8
..6-
1.2-
0.4-
0.2 04 0.6 0.8 1.0
DISCHARGE RATEKCFS)
©
©
0.2 0.4 0.6 0.8 1.0
SHEPTON ST.
(6.6 6} A VSS
po^_
L-4.0
NOTE SCALE
CHANGE
®
1.5-
1.0-
0.5-
1.0
r-4.0
-3.0
-2.0
&©
Q6
0.8
Q-50 CF FLUSH VOL.
&-35 CF FLUSH VOL.
TEMPLETON ST.
• RAIN IMPACTED
0 DRY WEATHER
02. 0.4 0.6 0.8 1.0
DISCHARGE RATE = (C F S )
BACKUP 8 RELEASE
DRY WEATHER
RAIN IMPACTED
MASS REMOVAL NORMALIZED BY ANTECEDENT DAYS BETWEEN FLUSHES
FIGURE 53
COMPARATIVE FLUSH EFFICIENCY PLOTS
SHEPTON AND TEMPLETON STREETS
172
-------
3.0-
2.5-
^
-g 2.0-
0>
~" 1.5-
Q
LU
i i.o-
tu
oc
s °-5~
S
0-
TSS
A •<7-33) 3.0-1
2.5-
m ® • 2.0-
• • ©
1.5-
f^
0 0 1.0-
^(V
^ £& ° °'5"
0_
H
vss ,,^
A
•
O
0
•^
0\1/
d^ * o
A
% 0
1 1 1 1 1
0 02 0:4 0.6 0.8 1.0 0 02 0.4 0.6 0.8 1.0
DISCHARGE RATE (CFS)
WALNUT ST.
TSS VSS
3.0^
">; 25-
i
—• 2.0-
g w-
o
2
UJ 1.0-
(£
| 0.5-
0-
3.0-
2.5-
90-
A
1.5-
©
©
Q
« 1.0-
r\ f?\
A\ V-'
^ ' _ • A 0.5-
I 1 I 1 1
*
©
©0
***% ^* A
i i i 1 " 1
0 0.2 0.4 0.6 0.8 1.0 0 02 0.4 0.6 0.8 1.0
DISCHARGE RATE (CFS)
PORT NORFOLK ST. BACKUP 8 RELEASE
0 — 50. CF FLUSH VOL. • RAIN IMPACTED DRY WEATHER
&-35. CF FLUSH VOL. © DRY WEATHER RAIN IMPACTED
MASS REMOVAL NORMALIZED BY ANTECEDENT DAYS BETWEEN FLUSHES
FIGURE 54 COMPARATIVE FLUSH EFFICIENCY PLOTS
WALNUT AND PORT NORFOLK STREETS
173
-------
SHEPTON STREET
35 CUBIC FEET FLUSHES
100 -,
CO
CO
Q
UJ
LU
CC
CO
CO
o
LL
O
vP
8TO -
60 -
40 -
20 -
9/24-.37cfs
20
ll/5-.40cfi
9/IO-.69cfs
10/18-.47 cfs
I/|-.6lcfs
40 60
% VOLUME
80
I
100
FIGURE 55
CUMULATIVE MASS VERSUS VOLUME
OF FLUSH SHEPTON STREET FLUSHES
(35 cf)
174
-------
SHEPTON STREET
50 CUBIC FEET FLUSHES
IO/l5-.7ICfs
IO/!-.93cfs
l!/8-.41cfs
100
% VOLUME
FIGURE 56 CUMULATIVE MASS VERSUS VOLUME
OF FLUSH SHEPTON STREET
EXPERIMENTS (50cf)
175
-------
8.6 Phase I Sediment Characteristics
Sediment characteristics within each of the test segments were ob-
served before and after each flushing experiment. Visual observations of
sewer segment sediment characteristics for each street are given in Tables 42
through 45. Field observations are often noted for both the upstream flush-
ing and downstream sampling manholes. Visual observations given in Tables
42 through 45 indicate that in general the organic sanitary deposits were re-
moved by the flushing, and that sand, grit and gravel remained. The amount of
organic deposits found on any given day before flushing was fairly constant
over the three month flushing period. The quantity of inorganic sediments
decreased with time during the flushing program. This phenomena was also
observed during the second and third phase programs.
Representative sediment scraping analyses are shown in Table 46 for
each of the test segments. After the name of the location and date, the per-
cent volatile content of scrapings taken upstream prior to flushing and down-
stream after flushing are given in Table 46. Sieve analyses of total dry
solids and of the volatile fraction for downstream sediment scraping samples
remaining after flushing are then given. Solids handling procedures and
analysis techniques are discussed in Chapter 6. Difficulties encountered
performing sieve analyses on sediment scrapings taken prior to flushing are
also described in Chapter 6.
The volatile content of upstream scrapings taken prior to flushing
showed a high fraction of organic matter in the sediments.- The volatile
content of downstream sediments scraped subsequent to the flushing operation
were low. The sieve analyses of those samples indicated that roughly 80% by
mass of the dry solids sieved were greater than 0.2 mm or greater than fine
sand. The dry weight sieve analyses for Shepton and Port Norfolk Streets
scrapings indicated that most of the material was in the range of grit, that
is, between 0.2 mm to 2 mm. The dry weight sieve analyses for Templeton and
Walnut Streets scrapings indicated that a sizeable fraction of the material,
ranging from 18% for 11/12/76 sample on Walnut up to 48% for the 10/15/76
sample on Templeton Street, exceeded the 2 mm particle size range. Sieve
analyses of the volatile fraction of the scraping samples indicated that lit-
tle of the organic material remaining was smaller than fine sand. Little
fine organic material remained after flushing.
Dry solids determinations for a number of scrapings taken over a
unit length of pipe at the upstream flushing manhole prior to flushing are
presented in Table 47. These scrapings provide a good indication of dry
weather deposition rates since the upper end of the segment was always cleaned
bare by the flushing experiment. Accumulated materials during the period to
next flush would therefore start from a zero base-line condition. The total
dry solids weights per lineal foot of scrapings were normalized by the ante-
cedent period between flushes, and are also shown in Table 47. Averages are
presented for each street. Scrapings were not performed on Walnut Street,
due to the "soupy" deposits in suspension in the segment. These primary esti-
mates of deposition rates are used in Chapter 12 in a comparative analysis
of predicted versus measured deposition loading rates.
176
-------
TABLE 42: VISUAL OBSERVATIONS OF DEPOSITION CHARACTERISTICS
FIRST PHASE FLUSHES - PORT NORFOLK STREET
Date
Prior Flush
Post Flush
8/18/76
8/30
9/2
9/7
9/16
9/21 *
9/24
10/1
10/4
10/8
10/12
10/15
10/18
10/22
Heavy sanitary deposition
Heavy gravel and organic
sediment
Heavy sand
Heavy sand buildup, light
sanitary deposits
Light-medium sanitary
deposition
Light sanitary deposits
Heavy sanitary deposition
Heavy sanitary suspension
Heavy sanitary suspension
Moderate sanitary solids
Moderate sanitary suspension,
no si It-sand
Heavy sanitary suspension,
light deposits
Heavy sanitary suspension
Upstream: heavy sanitary
suspension
Downstream: heavy sanitary
with sand deposits
Partial removal
Slight removal
Reduced
Cleaned sanitary depo-
sits, sand remained
Clean
Clean
Cleaned sanitary
deposition
Clean
Clean
Upstream: some sanitary
deposits remaining
Downstream: grit
Clean
Clean
Removal of suspension/
deposition of silt
remained
Removed sanitary sus-
pension & deposition,
=2" sand remained
(continued)
177
-------
TABLE 42: VISUAL OBSERVATIONS OF DEPOSITION CHARACTERISTICS
FIRST PHASE FLUSHES - PORT NORFOLK STREET (Cont'd)
Date
Prior Flush
Post Flush
10/25
10/29
11/1
11/5
11/8
11/12
Upstream: heavy sanitary sus-
pension
Downstream: 4" sediment, very
fine sand
Upstream: heavy sanitary depo-
sition
Downstream: 3" sediment, fine
sand
Heavy sanitary suspension & sedi-
ment, some sand
Upstream: heavy sanitary sus-
pension
Upstream: heavy sanitary sus-
pension
Downstream: medium sanitary
deposition & grit
Upstream: moderate sanitary sus-
pension
Downstream: light sanitary sus-
pension & 2" of grit
Removed sanitary deposi-
tion, 1" sand removed
Sanitary deposits removed,
sand remained
Removed sanitary solids,
sand left but moved a
little
Removed sanitary suspen-
sion; sand remained
Upstream: clean
Downstream: heavy sani-
tary in suspension
Most of sanitary suspen-
sion removed, grit
remained
178
-------
TABLE 43: VISUAL OBSERVATIONS OF DEPOSITION CHARACTERISTICS
FIRST PHASE FLUSHING PROGRAM - SHEPTON STREET
Date
Prior Flush
Post Flush
8/23/76
8/26
8/30
9/2
9/7
9/10
9/13
9/16
9/21
9/24
10/1
10/4
10/8
10/12
10/15
10/18
Heavy sanitary deposits
2"-3" sanitary deposition
prior to flush
Medium sanitary deposit
High sanitary deposition -
difficult to observe
Medium sanitary deposits
Medium-heavy sanitary deposits
Heavy sanitary deposits
Moderate sanitary deposits
Heavy sanitary deposits
Heavy sanitary deposition
(mostly toilet paper)
Heavy sanitary suspension
Moderate sanitary suspension
Upstream: heavy sanitary sus-
pension
Downstream: fairly clean
Heavy sanitary deposition
Heavy sanitary deposition
Heavy sanitary suspension &
deposition
Reduced
Clean
Clean
Difficult to observe
Reduced
Considerable reduction
Clean
Floating debris in flush
Clean
Removed sanitary deposits
but heavy particles
remain
Removal of suspension for-
mation of toilet paper
scum
Removal of suspension
Upstream: clean
Downstream: light gritty
deposition
Clean
Clean
Clean
(continued)
179
-------
TABLE 43: VISUAL OBSERVATIONS OF DEPOSITION CHARACTERISTICS
FIRST PHASE FLUSHING PROGRAM - SHEPTON STREET (Cont'd)
Date
Prior Flush
Post Flush
10/22
10/25
10/29
11/1
11/5
11/8
11/12
Very heavy sanitary suspension.
Approx. 1" sediment down-
stream.
Light sanitary suspension
Heavy sanitary suspension
Moderate sanitary suspension
& light deposition
Heavy sanitary suspension
& light deposition
Upstream:heavy sanitary sus-
pension
Downstream: Medium-light sani-
tary suspension
Upstream: heavy sanitary
deposits
Downstream: light grit
Removed suspension & part
deposition, very silty
Removed
Suspension removed, light
gritty deposits near
downstream manhole
Clean
Clean
Still heavy sanitary sus-
pension remaining
Upstream: heavy sanitary
suspension
Downstream: light sani-
tary suspension
180
-------
TABLE 44: VISUAL OBSERVATIONS OF DEPOSITION CHARACTERISTICS
FIRST PHASE FLUSHING PROGRAM - TEMPLETON STREET
Date
Prior Flush
Post Flush
8/30/76
9/2
9/7
9/10
9/13
9/16
9/21
9/24
10/1
10/4
10/8
10/12
10/15
10/18
10/22
4" sanitary deposit
High sanitary deposition
Heavy sanitary deposits & silt
Heavy sanitary deposits
Heavy sanitary deposits
Heavy sanitary deposits/gravel
Some change
Reduced
Reduced
Clean
Clean
Sanitary deposition
removed, not gravel
Heavy sanitary deposits & silt Cleaned, silt remained
Heavy sanitary deposition
{50% toilet paper)
Heavy sanitary suspension
& silt
Heavy sanitary suspension &
sand deposits
Upstream: heavy deposition of
solids & toilet paper
Downstream: moderate deposition
of sanitary & gritty solids
Heavy sanitary suspension,
little sand/silt
Heavy sanitary suspension,
light deposits
Heavy sanitary suspension up-
stream
Heavy sanitary, sand, silt
downstream
Very heavy sanitary suspension.
Approx. 1" sediment down-
stream
Heavy deposition remained
in line
Partial removal
Removal of sanitary sus-
pension
Upstream: clean
Downstream: some toilet
paper & small rocks left
Removal of. suspension
Removal of suspension
Removal of sanitary sus-
pension
Removed suspension & part
deposition, very silty
(continued)
181
-------
TABLE 44: VISUAL OBSERVATIONS OF DEPOSITION CHARACTERISTICS
FIRST PHASE FLUSHING PROGRAM - TEMPLETON STREET (Cont'd)
Date
Prior Flush
Post Flush
10/25
10/29
11/1
11/5
11/8
11/12
Heavy sanitary suspension,
-2" sand downstream
Heavy sanitary suspension
upstream
Heavy suspension & deposition
downstream
Heavy sanitary suspension & 1'
in downstream
Upstream: heavy sanitay sus-
pension
Downstream: heavy sanitary sus-
pension & light sediment
Upstream: heavy sanitary sus-
pension
Downstream: -2" sanitary & gritty
deposits
Upstream: heavy sanitary depo-
sition
Downstream: grit & sanitary
deposition
Removed sanitary suspen-
sion, sand remained
Suspension removed, gritty
deposition remained down-
stream
Upstream: clean
Downstream: left sand &
some sediment
Sanitary suspension re-
moved, sediment left
Removed most of suspension
& deposition
Deposition removed and a
suspension formed in up-
stream & downstream
manholes
182
-------
TABLE 45: VISUAL OBSERVATIONS OF DEPOSITION CHARACTERISTICS
FIRST PHASE FLUSHING PROGRAM - WALNUT STREET
Date
Prior Flush
Post Flush
8/23/76
8/26
8/30
9/2
9/7
9/10
9/13
9/15
9/21
9/24
10/1
10/4
10/8
10/12
10/15
10/18
Heavy sanitary deposits
Heavy sanitary deposits
No sign of deposits
4" sanitary deposition
Heavy sanitary deposits &
sand
Mid-heavy sanitary deposits
& sand
Heavy deposits
Difficult to observe
Light deposits
Light deposits 4"-6" gravel
Heavy deposition
Heavy sanitary suspension &
heavy silt
Sanitary suspension
Heavy sanitary deposits
Moderate silt/sand/gravel &
light sanitary deposition
Heavy sanitary suspension,
light deposits
Heavy sanitary suspension,
light silt & gravel
Reduced
Cleaned
No change
Reduced
Cleaned sanitary deposits
Sand remained
Cleaned sanitary deposits
Sand remained
Difficult to observe
Cleaned
Cleaned deposits, not
gravel
Heavy deposition remained
Removal of suspension
Removal of suspension
Heavy sanitary deposits
Sanitary removed
Removed
Removal of sanitary sus-
pension & some silt
(continued)
183
-------
TABLE 45: VISUAL OBSERVATIONS OF DEPOSITION CHARACTERISTICS
FIRST PHASE FLUSHING PROGRAM - WALNUT STREET (Con't)
Date
Prior Flush
Post Flush
10/22
10/25
10/29
11/1
11/5
11/8
11/12
Heavy sanitary suspension &
deposition
Upstream: no deposition
Downstream: 1" mud & medium
sanitary deposits
Clean upstream
Heavy sanitary suspension &
1/2" mud in downstream
manhole
Upstream: light black suspen-
sion
Downstream: moderate sanitary
suspension
Medium sanitary suspension,
no deposition
Upstream: heavy sanitary sus-
pension
Downstream: medium sanitary
deposits & grit
Upstream: fairly clean
Downstream: 1" of grit
Removed, some silt re-
mained =1"
Removal of sanitary depo-
sition
Clean
Clean
Clean
Unchanged
Upstream: light sanitary
suspension
Downstream: light grit
184
-------
TABLE 46: REPRESENTATIVE SEDIMENT SCRAPING ANALYSES
FIRST PHASE FLUSHING PROGRAM
00
01
%
LOCATION DATE
Shepton 10/15/76
Port 10/18
Norfolk
10/22
10/25
11/1
11/8
Templeton 10/15
10/18
11/5
Walnut 10/15
10/18
10/25
11/12
VOLATILE
UPSTREAM
SCRAPING
PRIOR TO
FLUSH
89.1
84.0
57.3
88.0
80.9
79.1
72.3
% VOLATILE
DOWNSTREAM
SCRAPING
AFTER
.FLUSH
2.2
2.0
1.1
1.1
3.2
4.3
5.6
6.2
9.7
5.4
3.2
4.2
6.2
DOWNSTREAM SCRAPING DRY SIEVE RESULTS
% WEIGHT OF MATERIAL BETWEEN STATED
PARTICLE SIZE INTERVALS:
( >2mm 2-.6mn .6-. 2mm .2-. 06mm <.
4
4
2
12
9
14
48
35
40
28
20
44
18
12 64
17 61
33 61
39 45
61 27
72 13
25 18
29 25
26 23
20 30
24 35
30 17
24 38
19
17
4
4
3
1
8
10
10
20
19
9
20
*
06mm )
1
1
0
0
0
0
1
1
1
2
2
0
0
DOWNSTREAM SCRAPING VOLATILE
% WEIGHT OF METERIAL BETWEEN
PARTICLE SIZE INTERVALS:
( >2mm 2-. 6mm .6-. 2mm
16 34
5 27
4 36
17 38
25 55
20 73
41 47
39 47
42 26
33 26
36 34
28 29
38
55
57
43
19
6
9
10
20
27
22
31
SIEVE RESULTS
STATED
.2-. 06mm <.06mn )
11
13
3
2
1
1
2
3
10
12
8
12
1
0
0
0
0
0
1
1
2
2
0
0
M.I.T. Classification Scheme Used (see Section 6.4.1):
>2mm: larger than coarse sand
2-.6mm: coarse sand
,6-.2mm: medium sand
.2-.06mm: fine sand - coarse silt
<.06mm: smaller than coarse silt
-------
TABLE 47:
RESULTS OF UPSTREAM SCRAPINGS PRIOR TO FLUSHING
PHASE I
JATE
9/24
10/8
10/15
10/18
10/25
10/29
9/24
10/8
10/12
10/15
10/18
10/22
10/25
10/29
ANTECEDENT
3
4
3
3
3
4
3
4
4
3
3
4
3
4
DAYS DRY GRAMS/ FOOT
SHEPTON STREET
55
45
15
77
15
20
TEMPLETON STREET
134
100
40
59
50
25
6
18
NORMALIZED FOR
ANTECEDENT DAYS
18.3
11.3
5.0
25.6
5.9
5.9
44.6
25.0
10.0
19.7
16.7
6.3
2.0
4.5
AVERAGE
GRAM/ FT/DAY
12.0
16.1
PORT NORFOLK STREET
10/8
10/12
10/15
10/18
10/22
10/25
10/29
4
4
3
3
4
3
4
30
65
33
67
79
53
54
7.3
16.3
11.0
22.3
19.8
17.7
13.5
15.4
186
-------
SECTION 9
SERIAL FLUSHING RESULTS
gj Foreword
Flushed pollutant removal results for experiments conducted during
the first half of the second phase flushing program on Port Norfolk Street
during the winter/spring of 1977 are presented in this Chapter. Settleability
experiments performed on flush samples gathered during the latter portion
of Phase two in the summer of 1977 are discussed in Chapter 10. Six
experiments consisting of three flushes per experiment were conducted.
Details of the field procedures are given in Chapter 5. In Section 9.2
preliminary flushing experiments are described that were conducted to
assess flush wave hydraulic characteristics for the multiple flush segment
experimentation. Flushed solids and organic loadings for the six experiments
are described in Section 9.3. Characteristics of sediments noted during
this period of flushing are presented in Section 9.4 as well as for the
period of sampling for settleability testing.
9.2 Preliminary Flush Have Hydraulic Experiments
The first phase flushing program entailed a single manhole to
manhole flushing operation. It appeared from that program that those
experiments conducted using high flush volume (50 cf) together with high
injection rates, that is, in excess of 0.5 cfs, were the most effective
in removing pollutants from the segments. The hydraulic characteristics
of flush waves travelling in excess of 276 feet were not known at that point
in time. A series of hydraulic experiments were conducted during December
of 1976 on Port Norfolk Street to assess flush wave characteristics for the
envisioned serial flushing program. Flushes were injected into the upstream
injection manhole and wave heights as a function of time were noted at the
three downstream manholes. Various flush volumes and rates were investigated
and it appeared that flush rates in excess of 0.45 cfs would be adequate to
ensure a noticeable wave roughly 675 feet from the point of injection. In
addition, the larger volume flushes (50 cf) maintained slightly better wave
form at the furtherest downstream manhole, presumably from.increased
momentum. Flush waves with water depth of 4 to 5 inches were noted at the
end of the street, roughly 1000 feet downstream for flush volumes of 50
cubic feet injected at rates exceeding one cfs. These experiments together
with first phase experience suggested that flush volumes of around 50 cubic
feet injected at rates of 0.4 cfs or better would be adequate for the serial
flushing program aimed at assessing multiple segment flushing effectiveness.
187
-------
9.3 Pollutant Removal Results
Six flushing experiments on Port Norfolk Street were conducted
during the winter and spring period of 1977. Three flushes were conducted
per experiment with the flush injected into an upstream manhole and flush
wave samples noted at three downstream manholes. Typical flush wave
concentration levels for this phase of work are shown in Figure 57 for the
experiments conducted on March 3, 1977. Nine plots of TSS and VSS concentra-
tions versus time of flush passage at each manhole are presented. Each row
of plots present flush wave concentration as a function of time for the three
sampling manholes per flush. The top three plots refer to the first flush
while the last three plots are associated with the third flush. Computed
flushed loadings of TSS and VSS in kilograms, are indicated on each plot.
These loadings were estimated using flush wave discharge,rates computed using
the loop stage rating curves described in Chapter 7 and the measured flush
wave concentrations at each manhole. The scale change in flush wave TSS and
VSS concentration between the first and subsequent flushes should be noted.
The peak flush wave concentrations for the first flush ranged from 2500 and
1800 mg/1 for TSS and VSS, respectively, at the first manhole, up to 7200
and 6000 mg/1 at the second and third manholes. The peak TSS and VSS
concentrations noted at all three sampling manholes for the second flush are
all less than 800 mg/1. With the exception of the first sample at the first
sampling manhole for the third flush, the peak TSS and VSS concentrations
are again less than 800 mg/1. Similar plots of BOD, COD and TKN showed the
same dramatic concentration levels for the initial flush and much lower
levels for the second two flushes.
The flushed pollutant masses, in kilograms, for the 18 flushes
conducted during this phase of work are given in Table 48. After the date
of the flushing experiments in Table 48, the following information under the
labelled columns are: a) the flush number; b) the order of the sampling
manhole, that is, in a downstream sequence; c) the delivered flush volume,
in cubic feet; and d) the flush rate, in cubic feet per second. The next
set of columns present the estimated flushed masses for the following
pollutants: COD, BOD, TKN, TSS and VSS. The antecedent periods between
experiments were 11, 35, 21, 5 and 9 days, respectively for the experiments
taken in chronlogical order. These flushed pollutant masses were then
normalized by antecedent periods between experiments and sums of pollutant
masses transported by each sampling manhole computed. Percentages of mass
transported per flush relative to the total mass transported during an
experiment for each of the pollutants were then computed. These percentages
for the six flushing experiments are presented in Table 49. Information in
Table 49 under the labelled headings is as follows: a) flush number; b) order
of sampling manhole, that is, in a downstream sequence; c) delivered flush
volume in cubic feet; and d) flush rate, in cubic feet per second. The next
set of columns presents the percentages of pollutant mass transported per
flush. For example, the relative percentages of BOD mass transported past
the first sampling manhole for the experiment on 1/04/77 are 88, 13 and 12
percent, respectively. Inspection of the results indicate that high removal
or transport effectiveness is accomplished with the first flush and the
effectiveness decreases as the distance .from the point of flush injection
increases. The results appear to be invariant to both the volume of
188
-------
PLOTS OF FLUSH WAVE CONCENTRATIONS
PORT NORFOLK STREET, 3/3/77
8000-
,7000-
MG/L
6000H
MANHOLE NO. I
ALL MASS IN KILOGRAMS
TS S = .75
VSS =.497
MANHOLE NO. 2
MANHOLE NO. 3
FLUSH I
50 CF
AT 0.52
CFS
= 2.47
= 2.0
TSS
MG/L
FLUSH 2
50 CF
AT 0.5
CFS
1600-
1400-
'l20O-
1000-
800-
600-
400-
200-
TSS=.26
VSS = .26
VSS
TSS
I I 1 T
TSS =.46
I I
I I
= 35
= .18
FLUSH NO.I
METHOD' GRAVITY
RATEi.82 CFS
VOLUME'SO CF
..BACKGROUND
3TSS, VSS
VSS
( I I I I I I I I I I I
= 1.32
= .73
FLUSH NO. 2
METHOOi GRAVITY
RATE',50 CFS
VOLUME- 50 Cf
back-
Background
TSSaVSS
=.39
= .195
140
\ I I I I I I
0 20 40 60 80 100
I11IPI1 I I
FLUSH N0.3
METHOD'PRESSURE
RATE'1.18 CFS.
I
| VOLUMES BO CF
= 1.17
=.74
0 20 40 60 80 100
TIME (SEC) TIME (SEC)
FIGURE 57. TYPICAL PHASE 2 RESULTS
\ I \ ] I I I I I
140 0 20 40 60 80 120
I
160
TIME (SEC)
189
-------
TABLE 48. PHASE II - FIELD FLUSHING POLLUTANT REMOVALS (kg)
PORT NORFOLK STREET
TOTAL MASS (kq/flush)
DATE
1/04/77
2/10/77
3/03/77
3/08/77
3/17/77
A
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
B
1
2
3*
1
2
3*
1
2
3*
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
C
50.0
50.0
50.0
55.2
50.0
50.0
50.0
50.0
50.0
50.0
50.0
50.0
3b.O
35.0
50.0
D
0.64
0.65
0.65
1.10
0.70
0.98
0.52
0.50
1.18
0.40
0.50
1.0
0.51
0.48
0.85
COD
5.03
8.30
1.57
1.51
0.52
0.82
3.50
8.07
12.15
1.38
2.61
5.40
0.92
1.72
3.21
1.T9
3.03
8.83
0.45
0.46
3.50
0.48
0.59
1.33
2.74
4.24
7.36
0.90
1.52
2.08
0.45
0.79
0.91
3.31
3.53
14.88
0.59
0.97
3.05
0.34
0.48
1.43
BOD
0.93
1.57
0.25
0.14
0.20
0.93
1.03
1.64
0.37
0.40
2.45
0.06
0.13
0.46
0.14
0.13
0.45
TKN TSS
4.78
7.75
0.54
0.90
0.42
0.72
4.39
3.33
7.09
1.17
2.03
5.20
0.71
1.57
2.45
0.033 0.75
0.079 2.47
0.231 8.85
0.027 0.26
0.048 0.35
0.041 1.32
0.012 0.46
0.014 0.39
0.018 1.17
1.56
3.28
6.04
0.31
0.53
1.14
0.43
0.61
0.93
3.67
3.68
10.38
0.36
.0.43
1.55
0.23
0.35
1.19
VSS
3.72
5.93
0.28
0.44
0.24
0.34
3.07
2.71
3.72
0.70
1.32
3.45
0.37
0.78
1.37
0.58
2.00
6.91
0.10
0.18
0.73
0.24
0.20
0.74
1.20
2.61
4.89
0.14
0.21
0.70
0.17
0.27
0.51
3.25
3.22
10.64
0.28
0.31
1.17
0.16
0.20
0.73
(continued)
190
-------
TABLE 48 (CONT.).
PHASE II FIELD FLUSHING POLLUTANT REMOVALS
PORT NORFOLK STREET
TOTAL MASS (kq /flush)
DATE
3/24/77
A
1
2
3
B
1
2
3
1
2
3
1
2
3
C
35.5
35.0
50.0
D
0.47
0.44
0.94
COD BOD
2.70
2.84
5.21
0.50
0.96
2.57
0.70
1.24
1.84
TKN TSS
1.97
3.68
4.58
0.23
0.51
1.99
0.27
0.83
1.65
VSS
1.61
3.27
3.84
0.15
0.34
1.40
0.16
0.44
1.08
Legend
A - Flush Number
B - Downstream Sampling Manhole
C - Flush Volume (cf)
D - Flush Rate (cfs)
*
Samples not taken at third downstream manhole.
191
-------
TABLE 49. PHASE II - FIELD FLUSHING PROGRAM PERCENTAGES OF TOTAL MASS
TRANSPORTED PER FLUSH AT EACH SAMPLING MANHOLE
PERCENTAGES OF MASS
DATE
1/04/77
2/10/77
3/03/77
3/08/77
3/17/77
A
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
B
1
2
3
1
2
3
*
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
•1
2
3
1
2
3
1
2
3
C
50.0
50.0
50.0
55.2
50.0
50.0
50.0
50.0
50.0
50.0
50.0
50.0
35.0
35.0
50.0
D
0.64
0.65
0.65
1.10
0.70
0.98
0.52
0.50
1.18
0.40
0.50
1.0
0.51
0.48
0.85
COD
71
22
7
78
14
8
60
24
16
65
21
14
59
26
15
56
22
22
74
11
15
64
26
10
67
22
11
65
23
12
71
20
9
78
14
8
71
20
9
77
16
7
BOD
88
13
12
77
13
10
**
**
**
**
**
**
**
**
**
63
11
26
62
19
19
72
14
14
FLUSH
TKN
46
37
17
56
34
10
80
14
6
TRANSPORTED PER
TSS
83
9
8
83
9
8
70
19
11
48
29
23
48
35
17
51
18
31
77
11
12
78
12
10
68
13
19
74
12
14
74
14
12
86
8
6
82
10
8
79
12
9
VSS
88
6
6
88
6
6
74
17
9
56
27
17
43
40
17
59
12
29
84
7
9
82
9
9
79
9
12
82
10
8
80
12
8
88
8
4
87
8
5
85
9
6
(continued)
192
-------
TABLE 49 (CONT.). PHASE II - FIELD FLUSHING PORT NORFOLK STREET. PERCENTAGES
OF TOTAL MASS TRANSPORTED PER FLUSH AT EACH SAMPLING MANHOLE
PERCENTAGES OF MASS TRANSPORTED PER
FLUSH
DATE A
3/24/77 1
2
3
B C D
1 35.5 .47
2 35.0 .44
3 50.0 .94
1
2
3
1
2
3
COD BOD
69
13
18
56
19
25
54
27
19
TKN TSS
80
9
11
73
10
17
56
24
20
VSS
84
8
8
81
8
11
61
22
17
Legend
A - Sampling Manhole
B - Flush Number
C - Flush Volume (cf)
D - Flush Rate (cfs)
Samples were not taken at third sampling manhole.
BOD determined only for first sampling manhole.
193
-------
flush used and the flush rate.
The fractional removal percentages in Table 49 were averaged and
are summarized in Table 50 , showing the average percentage per flush of the
total load removed for each of the three segments downstream of the flush
injection manhole. These averages were computed using the loads computed
per manhole for the six sets of flushing experiments. The results indicate
that most of the loads for all three segments were removed during the first
flush. For example, 81.7% of the volatile suspended solids load was removed
from the first flush. The second and third flushes removed an additional
18.3% of the total. No appreciable gain is achieved by repeated flushing.
Furthermore, the experiments indicate that a single flush at the upper end of
the street was reasonably effective in removing most of the deposited load
along the 675 feet stretch of 12-inch combined sewer lateral.
TABLE 50 . AVERAGE PERCENTAGES OF POLLUTANT LOADS REMOVED PER FLUSH FOR EACH
PIPE SEGMENT
TSS
First
Flush
Flush
Flush
Second
Flush
Flush
Fl ush
Third
Flush
Flush
Flush
Sampling Manhole
1
2
3
Sampling Manhole
1
2
3
Sampling Manhole
1
2
3
76
12
11
72
14
13
66
20
13
.1
.7
.2
.4
.2
.4
.5
.2
.3
VSS
81.
10.
8.
79.
11.
8.
71.
17.
10.
7
1
2
5
6
9
6
8
6
COD
67
19
12
68
18
12
65
22
11
.7
.8
.5
.7
.4
.9
.7
.5
.8
BOD
87.
2.
10.
81.
10.
8.
81.
9.
9.
3
7
2
8
8
2
0
9.4 Discussion of Sediment Characteristics
Sediment within the three combined sewer segments on Port Norfolk
Street was mostly sand, grit and with some septic organic sanitary waste
deposits. The level of depositions had substantially increased during the
winter snow period when the first phase operation terminated (11/12/76) and
the preliminary hydraulic experiments described in Section 9.2 began (12/20/76)
Considerable sand from de-icing operations over the winter had washed into
the segments. Visual observations taken before and after the flushing
operations indicated that the sanitary deposits were generally washed away
leaving sand and grit accumulations. Volatile solids were determined for
32 pre and post flushing strappings and ranged from 1.4 to 50.2% with an
average of about 5.1%. Little difference in volatile content was noted in
pre/post flushing, presumably the result of a sand and silt layer. After
the end of the snow period sand, grit and gravel layers were maintained at
constant levels.
194
-------
Sediment levels were also noted along the Port Norfolk Street
test segments during the spring and summer of 1977. During this period the
latter portion of the second phase program was conducted in which flush wave
samples were taken for the settleability analyses described in Chapter 10.
In addition, the special dye-injection experiments meant to verify the loop
stage discharge methodology described in Chapter 7 were conducted during this
period. In total, roughly 50 flushes were conducted over a four month
period. The grit and sand accumulations were gradually reduced to minimal
layers toward the end of the summer of 1977.
195
-------
SECTION 10
SETTLEABILITY TESTING RESULTS
10.1 Foreword
To characterize the settleability of the solids in the flush waves
a series of flush waves were analyzed as the waves traveled from manhole to
manhole in a three segment series. Samples were taken from each of the three
manholes in the Port Norfolk Street sewer section described in Chapter 5,
using the sampling procedures described in Chapter 5. Samples taken at
each location for each flush were then composited based on the hydraulics of
the wave and analyzed according to the procedures described in Chapter 6.
The settling column testing subphase of the second phase flushing
program, phase IIB, was conducted during the period of July 27 - September 7,
1977. During that period a total of 18 flushes were sampled at each of the
three manholes described in Chapter 4 of this report. As indicated in
Chapter 5, on each flushing day three successive flush waves were injected
along the sewer line and were sampled at each of the three downstream sampling
locations. The locations were the same as those used during the serial flush-
ing, phase IIA program, that had been completed earlier in the year. Special-
ized sampling techniques using special devices were used to ensure the col-
lection of "undisturbed" flush wave samples as described in Chapter 5.
In addition to the settling column tests, Imhoff cone tests were
conducted to determine settleability and general character of the
supernatants of the flush samples. Procedures used for conducting the
settling-column and Imhoff cone tests were described in Chapter 6.
All samples taken from the settling columns for settleability
analysis were analyzed to determine pollutant concentrations associated with
the various settling velocities. Pollutants analyzed included, TSS, VSS,
BODr, COD, ammonia nitrogen, TKN, orthophosphate, total phosphate cadmium,
chromium, copper, lead, zinc, and nickel.
Attempts were made to analyze mercury levels in the various settling
column samples, but the concentrations proved to be so low, on the order of
1.0 ppb, as to make the reliability and meaning of the determination unsuita-
ble for further use.
196
-------
Classical settling theory considers that the sedimentation may be
one of four types: Type I - Discrete Particle Settling, Type II - Flocculent
Settling, Type III - Zone or Hindered Settling, and Type IV - Compression
Settling. A more detailed description of the four types of sedimentation may
be found in Metcalf and Eddy (24). The nature of the particles in suspension
determines whether Type I - Discrete Particle Settling or Type II - Flocculent
Settling will occur for suspensions with a low concentration of solids. As
the concentration of solids in suspension becomes progressively higher
Type III - Hindered Settling and Type IV - Compression Settling will be the
case for either discrete or flocculant particles. The suspension produced
by sewer flushing activities contains a mixture of discrete and flocculant
particles. Most of the larger particles in suspension such as grit settle
as discrete particles uninfluenced by surrounding flocculent particles.
Smaller particles generally undergo flocculent settling.. Thus in this
study, laboratory analyses were carried out for both types of settling.
10.2 Assessment of Initial Concentration Data
As previously indicated the composite flush wave samples were
analyzed for a number of parameters. Table 51 presents a summary of the
initial concentration or concentration of the composited flush wave samples
for all three manholes during the 18 column testing flushes on each of the
six dates. Data is presented for the primary parameters of the analysis, TSS
and percent volatile. Careful analysis of the data presented on the table
leads to several conclusions which agree favorably with the results of the
phase IIA program,as well as with the settling column and Imhoff cone test
results presented later in this chapter. Although the induced turbulent
energy of the flush wave decreases as the wave proceeds downstream along
the three sewer segments, the concentration of solids and percent volatile
tend to increase sharply. This phenomenon is indicative of the cumulative
effects of the flush wave scouring a progressively increasing deposited load
along the pipe. Although the scour or grit removal energy dissipates sharply
along segment 1-2 as indicated by the large change in percent volatile
generally exhibited, the pollutant removal still remains high. This fact is
most important in assessing the performance of sewer flushing as a
pollution control measure.
The second distinct trend indicated in the data is the relatively
high removal efficiency of the first flush. Again the results are
consistent with those found in phase IIA of the program. After the first
flush there is generally a large drop off in TSS concentration especially
with respect to manholes 2 and 3. Similarly the percent volatile or
percent organic level of the samples also decreases indicating that the
readily available surface pollutants have been carried away in segments 1 and
2 by the first flush and to a large degree in segment 3.
Since the bulk of the deposit was scoured during the first flush
as indicated in Table 51,.sett!ing column testing was only performed on the
first flush of each flushing day. Imhoff cone tests were performed on all
flushes.
197
-------
TABLE 51. INITIAL CONCENTRATION AND PERCENT VOLATILE OF INITIAL
CONCENTRATION FOR ALL FLUSHES AND MANHOLES
DATE
7/27
8/4
8/22
8/25
8/29
9/7
Average
FLUSH
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
MANHOLE
INITIAL
TSS mg/£
1029
501
226
505
134
52
528
no
103
221
58
40
126
77
-
1130
102
57
589
'164
96
#1
%VOL.
59
40
52
58
54
52
73
51
45
68
57
55
60
51
-
76
60
53
65.7
52.2
51.4
MANHOLE
INITIAL
TSS mg/A
1856
1285
475
1610
596
369
3316
665
717
2226
438
118
2669
-
200
5927
763
315
2934
749
366
#2
% VOL.
68
68
51
74
61
44
84
68
70
84
71
60
80
-
64
80
63
49
78.3
66.2
56.3
MANHOLE
INITIAL
TSS mg/A
3958
478
2593
2214
285
264
7055
482
748
1880
579
312
2254
812
423
5542
613
783
3817
542
854
#3
% VOL.
85
54
41
82
52
55
86
95
70
92
80
67
82
74
62
83
89
57
85.0
74.0
58.7
* Initial Concentration of Settling Column Experiments
198
-------
10.3 Presentation of Results
The composite flush sample suspensions are a mixture of particles
which exhibit discrete settling behavior and particles which exhibit floccu-
lent settling behavior. Total suspended solids concentrations in the
samples were found to vary between 500 and 8,000 mg/1. Utilization
of a large sedimentation column is a good method for determining the
settling characteristics of such suspensions. Column test results in this
study were graphically analyzed. Several graphical approaches were used. The
first were plots of percent of particles removed versus the settling velo-
city, V. This type of plot is extremely useful in evaluating settling char-
acteristics of pollutants as associated to various settling velocities. A
second approach plotting concentration versus settling velocity, V, was used
to differentiate characteristics of one manhole versus another.
Assessment of sewer flushing related settling column data requires
a different philosophical viewpoint than that commonly used with this type of
information. Typically, settleability testing is done to design clarifiers
or sedimentation tanks where high removals are the aim. On the other hand,
sewer flushing is aimed at resuspending deposited solids and keeping them in
suspension until they reach a treatment or removal point. This difference is
most important in reviewing the data generated, in that poor settleability
characteristics equate with increased overall flushing efficiency.
Tables 52 and 53 are the data summaries of typical settling column
tests for flush 1, manhole 1 for 8/25 and 9/7, respectively. From this data
it is possible to construct curves of percentage removal of a specific pollu-
tant versus measured settling velocity. Careful analysis of the data pre-
sented shows that sewer flushing is far more efficient with respect to long-
term pollutant removal than grit and other readily settleable inorganics.
Data presented in Tables 52 and 53 illustrates this fact by indicating that
for the particular sample, removals for TSS averaged roughly 80 % for
30-minute quiescent periods, where BOD, TKN, Total Phosphorous averaged from
50-55%. This is most significant when the converse or percent remaining is
viewed. In this case, after 30 minutes of settling only 20 % of the
solids remain in suspension, while between 45-50 % of the significant
organic pollutants remain. Extended settling periods, up to 120 minutes as
illustrated in Table 52, showed little additional change beyond the 30-minute
increment. Surprisingly, sewer flushing COD removal was much lower than
other pollutants,with approximately 80 % removal, or 20 % remaining after
30 minutes.
The marked difference between COD and BOD is most probably due to
the COD being primarily representative of grease and other long-term decom-
posing matter complexed in the heavy sediments. This type of sediment primarily
composed of'materials that are difficult to move, would probably not be moved
by storm events, and on a pollution control basis, is of lesser significance
than the more readily available BOD and nutrients.
Although the corresponding removals vary from flush to flush and
manhole to manhole, the general pattern remains the same. As this pattern
199
-------
TABLE 52. COLUMN TEST RESULTS 8/25 FL1 - MH1
SAMPLE
NO.
D13
Fl
F2
F3
F4
F5
F6
F7
F8
F9
F10
Fll
FT 2
F13
F14
FT 5
F16
F17
FT 8
TIME PORT
MIN. NO.
Back*** -
5 9
3
1
10 7
3
1
20 5
3
1
30 3
1
60 5
3
1
90 3
1
120 3
1
DEPTH
FT.
-
0.71
2.98
4.98
0.93
2.64
4.64
1.05
2.34
4.34
2.06
4.06
0.56
1.85
3.85
1.61
3.61
1.41
3.41
S.O.R.*
gpd/sf
-
1555
6437
10757
1004
2851
5011
567
1264
2344
742
1462
100
333
693
193
433
126
307
VERT. VEL.
fps x 103
-
2.40
9.93
16.60
1.55
4.40
7.73
0.89
1.95
3.62
1.14
2.26
0.16
0.51
1.07
0.30
0.67
0.20
0.47
mg/A
353
102
no
145
96
91
79
64
71
79
52
64
59
52
91
60
133
48
56
COD ORTHO-P TOTAL-P
% Rem** ing/* % Rem*,* rng/n. % Ren**
-
71.1
68.8
58.9
72.8
74.2
77.6
81.9
79.9
77.6
85.3
81.9
83.3
85.3
74.2
83.0
62.3
86.4
84.1
1.58
0.36
0.47
0.53
0.44
0.40
0.47
0.41
0.24
0.27
0.32
0.31
0.47
0.35
0.32
0.27
0.28
0.13
0.14
-
77.2
70.3
66.4
72.2
74.7
70.3
74.1
84.8
82.9
79.7
80.4
70.3
77.8
79.7
82.9
82.3
91.8
91.1
1.85
0.97
1.17
1.26
1.02
0.99
0.99
0.99
1.00
2.45
1.26
0.71
0.93
0.85
1.03
0.88
0.80
0.71
0.71
-
47.6
36.8
31.9
44.9
46.5
46.5
46.5
45.9
34.2 +
31.9
61.6
49.7
54.1
44.3
52.4
56.8
61.6
61.6
TSS
mg/x, % Rem**
221
51
68
90
61
49
50
52
47
47
43
46
43
32
40
27
26
29
22
-
76.9
69.2
59.3
72.4
77.8
77.4
76.5
78.7
78.7
80.5
79.3
80.5
85.5
81.9
87.8
88.2
86.9
90.9
VSS
mg/x. % Rem*
151
39
47
54
42
35
36
36
33
39
34
31
33
24
32
20
20
27
24
_
74.2
68.9
64.2
72.2
76.2
76.2
76.2
78.1
74.2
77.5
79.5
78.1
84.1
78.8
86.6
86.6
82.V
84.1
ro
o
o
* S.O.R.
** % Rem.
Surface Overflow Rate
Percent Removal
*** Back - Initial Concentration
+ Greater than initial concentration
-------
TABLE 53. COLUMN TEST RESULTS 9/7 FL1 - MH 1
SAMPLE TIME
NO. MIN.
PI Back***
P3 2
P4
P5
P6 4
P7
P8
P9 6
P10
PI 3
P14 8
P15
PI 6
P17 10
PI 8
PI 9
P20 20
P21
P22 30
P23
PORT
NO.
-
9
3
1
8
3
1
7
3
1
5
3
1
5
3
1
3
1
3
1
DEPTH
FT.
-
0.71
2.98
4.98
0.73
2.71
4.71
0.72
2.43
4.43
0.87
2.16
4.16
0.56
1.85
3.85
1.56
3.56
1.35
3.35
S.O.R.*
gpd/sf
-
3825
16092
26892
1958
7317
12717
1290
4374
7974
1175
2916
5616
605
1998
4158
842
1922
486
1206
VERT. VEL. COD BOD TKN
fps x 103' ing/* % Rem.** mg/a % Rem.** nig/* % Rem** TSS
-
5.92
24.83
41.50
3.04
11.29
19.63
2.00
6.75
12.31
1.81
4.50
8.67
0.93
3.08
6.42
1.30
2.97-
0.75
1.86
1682
311
1008
1492
692
621
803
545
477
523
477
508
470
485
470
447
333
439
394
374
-
81.5
40.1
11.3
58.9
63.1
52.3
67.6
71.6
68.9
71.6
69.8
72.1
71.2
72.1
73.4
80.3
73.9
76.6
77.8
345
160
308
435
204
216
294
198
210
150
174
164
153
168
157
192
162
153
153
153
-
-
10.7
26.1 +
40.9
37.4
14.8
42.8
39.1
56.5
49.6
52.5
55.7
51.3
54.5
44.3
53.0
55.7
55.7
55.7
42.6
26.3
29.1
42.0
26.6
25.2
32.2
24.6
23.8
16.8
27.2
22.7
22.4
22.7
21.8
25.2
21.3
21.8
20.2
20.7
-
38.3
31.7
1.4
37.6
40.8
24.4
42.3
44.1
60.6
36.2
46.7
47.4
46.7
48.8
40.8
50.0
48.8
52.6
51.4
1130
378
631
1021
296
297
511
237
236
320
190
198
192
192
200
204
155
182
156
153
SUSPENDED SOLIDS
J^4- •A'-A-
% Rem* VSS % Rem*
-
66.5
44.2
9.6
73.8
73.7
54.8
79.0
79.1
71.7
83.2
82.5
83.0
83.0
82.3
81.9
86.3
83.9
86.2
86.5
859
290
478
744
231
232
351
196
176
199
179
159
154
160
170
173
135
156
135
129
_
66.2
44.4
13.4
73.1
73.0
59.1
77.2
79.5
76.8
79.2
81.5
82.1
81.4
80.2
79.9
84.3
81.8
84.3
85.0
IV)
o
* S.O.R. - Surface Overflow Rate
** % Rem. - Percent Removal
*** Back - Initial Concentration
+ Greater than initial concentration
-------
became evident the initial sampling time intervals for the column tests were
decreased in an attempt to better define the shape of the curve, as evidenced
in the sampling times presented in the two tables. The sampling intervals
per each experiment are given in Chapter 6.
Due to the difficulty in mixing the initial sample suspension, very
short initial sampling times (less than 60 seconds) could not produce meaning-
ful data. In all cases, the particles with settling velocities associated with
grit and sand settled too quickly* for distribution analyses to be carried out.
It is also evident from the data in Tables 52 and 53 that a certain
portion of the particles and associated pollutants will have settling velo-
cities so low that this fraction will be transported long distances** by the
flush wave. By examining the sampling results for the 30-minute time inter-
val which represent settling velocities of .002 fps or less, it becomes evi-
dent that 15 to 30 % of the total suspended solids and volatile sus-
pended solids will remain in suspension.*** In terms of organic materials or
nutrients associated with the solids in suspension, the samples ranged from
20% to 50%. Such percentages are representative of the fraction of the flush
which will at minimum be carried to a treatment plant. The Imhoff cone data
typified by Table 54 provides an additional example of the material which will
remain in suspension after an hour of quiescent settling. Table 54 shows the
removal percentages after a one-hour settling period for all flushes and all
manholes sampled on August 25, 1977. The percentage removals for total sus-
pended solids ranged from 65 to 85 %, while the removals for volatile
suspended solids ranged from 59 to 87% on the date. In the case of such data
the first flush samples represent the bulk of the sewer deposit flushed. On
that basis it is apparent that 85% of the TSS and 80% of the VSS are removed
by the one-hour quiescent settling or that 15% TSS and 20% VSS will definitely
remain in suspension.
Figures 58 and 59 present the results of flushes conducted on 8/25
and 9/7 for TSS and COD respectively. Figures 58 and 59 provide interesting
insights into many factors of the settling column tests. First of all, the
comparison of data from the short time frame sampling on 9/7 with the long
time frame sampling on 8/25 shows amazing consistency. Although not
presented, the comparison of the results of the test conducted on 8/22 and
8/25 yield similar consistency. The second important item to note is the
relatively rapid settling velocities of a large percentage of the flushed pol-
lutants, especially solids reaching a plateau at roughly 18% remaining.
Comparison of Figures 58"and 59 reiterates the prior discussion pertaining
to tesser removals or lower degrees of settleability of the organic pollutants.
Although not shown, a similar trend was found to be present for all other
pollutants with percent remaining ranging up to 50% -for BOD and nutrients.
* Settled out in 30-60 seconds
** A minimum of 1550 feet (457.2 m)
*** A more formal analysis of transport is presented in Chapter 12.
202
-------
TABLE 54. IMHOFF CONE TEST RESULTS 8/25
FL-MH
1-1
1-2
1-3
2-1
2-2
2-3
3-1
3-2
3-3
SAMPLE
. D13
**D14 ,
D15
: Die
Ml
M2
Dll
D12
D7
"D8
D3
D4
D9
DID
D5
D6
Dl
D2
COD
rng/A %Rem*
353
-25V 92.9
4131
822 80.1
4723
1320 72.1
129
64 50.4
749
210 72.0
1166
317 72.8
'64
37 42.2
313
68 78.7
498
191 61.6
OP
mgA
1.58
0.23
10.55
8.21
9.88
10.35
0.92
0.77
2.32
1.30
2.59
1.95
0.32
0.24
0.80
0.32
0.76
0.75
TP
% Rem* mg/2 % Rem*
1.85
85.4 0.74 60^
25.61
22.2 14.78 . 42.3
24.53
18.03 26.5
0.64
16.3 0.46 28.1
2.02
44.0 0.92 54.5
2.91
42.7 1.75 39.9
3.02
25.0 1.64 45.7
5.18
60.0 3.08 40.5
1.57
1.3 1.46 7.0
TSS
mg/s, % Rem*
221
32 85.5
2226
332 85.1
1880
374 80.1
59
18 67.2
438
42 90.4
579
108 81.3
40
13 67.5
118
41 65.3
312
68 78.2
VSS
mg/x. % Rem*
151
29 80.8
1876
325 82.7
1724
318 81.6
33
.16 51.5
309
39 87.4
462
99 78.6
22
n 50.0
71
34 52.1
208
60 71.2
ml
Settled
8.5
150.0
165.0
2.0
25.0
39.0
0.7
4.5
16.0
Settleable
Matter
mg/A
189
1894
1506
41
396
471
27
77
244
Cone, of
Deposited
Solids mgA
22,235
12,627
9,127
20,500
15,840
12,077
lf
38,571_
17,111
15,250
o
co
* % Rem - Percent Removal ** Typical: sample (D13) is initial and sample (D14) is final.
-------
FLUSH 1 MH 1 • 9/7 TSS
MH 2 A
MH 3 •
100-
90-
PERCENT
REMAINING
FEET/SEC.
1 t
1 2
80-
70-
60-
50-
4O-
30-
20-
10-
(V
MH 1 0 8/25 TSS
MH 2 A
MH 3 a A •
•'.
A •
a ©
* a'
s ®"-'a • ' $
0 a3^«^°C*0*A
A - %
v t i i ttitiit i i i iiitii r I'lii'iiii
1 2 34681 2 34 681 2 34 681
io~4 io"3 io"8
r i i i i 1 1 1 i i i i i i 1 1 1 i i i i i i 1 1 1 }
34681 234681 234681 2
IO'L
VERTICAL SETTLING VELOCITY
FIGURE 58. PLOT OF TSS REMAINING VS SETTLING VELOCITY
204
-------
FLUSH 1 MH I • 9/7 COD
MH 2 A
MH 3 •
100-
90-
80-
70-
PERCENT eo.
REMAINING
50-
40-
30-
20-
10-
0-
FEET/SEC. ,
MH 1 0 8/25 COD
MH 2 A
MH 3 Q ^
*
*
• A ©
\ * f I*
jfj C|^^ A.° ...
*
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
234681 234681 23468
O"4 IO"3 IO"2
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I I 1 1
1 234681 234681 234681 23
-3 -2 -1 1
VERTICAL SETTLING VELOCITY
FIGURE 59. PLOT OF COD REMAINING VS SETTLING VELOCITY
205
-------
Figures 60 and 61 are plots of actual concentration vs settling
velocity for 8/22, 9/7 and 8/29, respectively, for TSS. The plots are parti-
cularly useful in comparing removals from the 3 manholes. What is clearly
indicated is the tendency for the flush wave to scour progressively lighter
materials as it flows downstream. It should be remembered that as the flush
wave progressed further down the line from one manhole to another, the wave
energy decreased. The result of this was that heavier particles dropped out
of suspension while the lighter particles remained.
10.3.1 Heavy Metals Analysis
An extensive program was conducted as part of the column settleabi-
lity testing program to assess the placement of heavy metals in relation to
particle settling velocities. As the problem of heavy metals discharges and
subsequent accumulations inv bottom deposits and aquatic life is addressed,
it becomes increasingly important to attempt to quantify the sources of such
metals. One possible source of significant metals discharges are combined
sewer overflows. Data generated in the first phase of this study, which are
reported in Chapter 8, indicated that the largest percentage of heavy metals
in the combined sewer system are located in fractions, which tend to settle
after extended periods. The settling column effort was aimed at assessing
to which fractions the metals were attached. Analyses were conducted for a
total of six metals, including: copper, zinc, chromium, cadmium, lead and
nickel.
Table 55 is a comparison of the heavy metals data generated per
manhole per flush. The table represents a summary of each flush, indicating
initial concentration, concentration and corresponding percent remaining of
the first sample of each settling column test, and the average concentration
and corresponding percent remaining of the metals after reaching a maximum
removal plateau. Analysis of the data presented in the table clearly indi-
cates several facts. First of all, metals exhibited a very high removal rate
for the initial few minutes of the test, as shown by the rapid drop in con-
centration between the initial and first sample of each column test. Second,
no significant settling then occurs after the initial period, as indicated
by the average plateau concentrations and percent remaining. This is very
significant in terms of metals accumulation and movement. The metals exhi-
bited relatively rapid settling characteristics, amounting to 50% or more in
metals concentration between the background and the first settling sample,
for both the short and long interval sampling tests, and thereafter generally
had a very slight positive tendency toward settling. The metals located in
'deposited sewer solids are split into two distinct fractions: roughly 50%
that are 'entrained on the heavy grit and sand, and which are not readily
transported; and the remaining 50% which are entrained in extremely light
near-colloidal fractions with settling velocities so low as to negate any
further removal once suspended.
The significance of this finding is especially clear when the remo-
vals due to storm events are assessed as discussed in Chapter 11. Low to mo-
derate intensity storms, which are of highest incidence in the northeast, tend
to move only light fractions, which is to say that these storm, although
206
-------
10,000-1
FLUSH I
1000-
to
IOOH
10"
MH I
MH 2
MH 3
9/7 TSS
0
3 0
©00
0
00
0
0
0
MH I O
MH2 A
MH 3 0
8/22 TSS
-4
10
r^ i i i i i i i
34681
-3
K>
2 34
FEET/SEC.
I III!
6 8 I
i n i i i i i i-i
2 34 681
1 I II I I I II
2 34 681
1 I I 1 1 II11
2 34 681
10
-4
1I I I l I 111
2 34 681
2 Id"1
CENTIMETER/SEC.
\ I I I M I 11
2 34 681
\ I I
234
10
FIGURE 60. PLOT OF TSS CONCENTRATION VS SETTLING VELOCITY
207
-------
1000-
_J
to
10-
FLUSH I MH I
MH 2
MH 3
8/29 TSS •
A A
.« * *
"
ft
a a
a
0 „ ^, 0©
0 © 0 MH I ©
0 MH2 A
MH3 0 8/25 TSS
i i T i <-i in r i n i i i i i 11 i i r i i m
I 2 3 4 ' 6 8 I 234681 23468
IO"4 I03 10~2
FEET/SEC.
1 I I I I I I ITT ' ] [ I IITTTI I I I I I I I 11 I I I I I I I I I I I
I 234681 234681 234681 234681 23
IO'4 10 I0~2 |0"' ID*
CENTIMETER/SEC.
VERTICAL SETTLING VELOCITY
FIGURE 61. PLOT OF TSS CONCENTRATION VS SETTLING VELOCITY
208
-------
TABLE 55'. COMPARISON OF HEAVY METALS CONCENTRATIONS FROM SETTLING COLUMN TESTS
DATE
7/27
7/27
7/27
8/4
8/4
8/4
8/22
8/22
8/22
8/25
8/25
8/25
8/29
8/29
8/29
9/7
9.7
9/7
H.H.
*
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
CADMIUM
BK**
.0180
.0090
.0860
.0063
.0086
.0050
-
-
-
-
-
-
.0020
.0065
.0006
.0047
.0062
.0065
1st
SAMP.
.0050
.0039
.0023
,0026
.0088
-
-
'
-
-
-
-
.0022
.0058
.0021
.0026
.0050
.0042
%
REM.*
27/7
43.3
2.7
41.3
-
-
-
-
-
-
-
-
-
89.2
-
55.3
80.6
64.6
AV6.
.0046
.0022
.0020
.0044
.0050
.0042
-
-
-
-
-
-
.0028
.0028
.0030
.0023
.0032
.0038
% REM.
AV6.
25.6
24.4
2.3
69.8
58.1
84.0
-
-
-
-
-
-
-
43.1
-
48.9
51.6
58.5
COPPER
BK
.6000
.9827
.9400
.7857
.6550
.4750
.2250
.7850
5.8500
.2750
.6250
.5550
.1800
.5850
.4100
.4091
6.1000
5.0000
1st
SAMP.
.1800
.5078
.1938
.1210
.6250
.5851
.4300
.3850
.1900
.1150
.2950
-
.1544
.8500
.7500
.2354
.8000
.1900
%
REM.
30.0
51.7
20.6
15.4
95.4
-
-
49.0
3.2
41.8
47.2
-
85.8
-
-
57.0
13.0
AVG.
.1600
.3015
.1837
.2731
.3288
.3368
.3576
.1403
2.76
.2673
.3197
.7363
.4386
.3928
1.600
.2644
2.009
3.632
% REM.
AVG.
26.7
30.7
19.5
34.8
50.2
70.9
-
17.9
47.2
97.2
51.2
-
-
67.1
-
64.6
32.9
24.2
CHROMIUM
BK
-
-
-
-
-
-
-
-
-
-
-
-
.028
.120
.006
.019
-
.034
1st
SAMP.
-
-
-
-
-
-
-
-
-
-
-
-
.063
.021
.011
.026
.013
.030
%
REM.
-
-
-
-
-
-
-
-
-
-
-
-
-
17.5
-
-
-
88.3
AVG.
-
-
-
-
-
-
-
-
-
-
-
-
.033
.038
.028
.017
.018
.024
% REM.
AVG.
-
-
-
-
-
-
-
-
-
-
-
-
-
31.7
-
89.5
-
70.6
O
IQ
- Indicates negative removals
* % Remaining
** BK - Background (initial) concentrations (mg/1)
(continued)
-------
TABLE 55 (Cont'd). COMPARISON OF HEAVY METALS CONCENTRATIONS FROM SETTLING COLUMN TESTS
7/27
7/27
7/27
8/4
8/4
8/4
8/22
8/22
8/22
8/25
8/25
8/25
8/29
8/29
8/29
9/7
9/7
9/7
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
ZINC
BK**
2.0800
2.0755
2.3520
1.2751
1.8600
1.8600
.1400
.4040
.7080
.7040
1.6400
.2480
.5200
2.0000
.7200
1.3900
4.8000
5.2000
1st
SAMP.
.4080
.9188
.6500
.4516
1.6800
1 . 2000
.0640
.1320
.2200
.3680
.6400
-
.5588
1.8400
1.3320
.5246
.2600
1.5200
* *
REH.
19.6
44.2
27.6
35.4
90.3
64.5
44.4
32.6
31.1
52.2
39
-
-
92.0
-
37.7
5.4
28.7
AVG.
.8176
.5738
.6323
.9895
1.1700
.9693
.0883
.3564
.1992
.5125
.5857
.7738
.7190
2.0500
1.2900
.5961
1.2700
1.8500
% REM.
AVG.
39.3
27.6
26.9
77.6
62.9
52.1
61.3
88.2
28.1
72.8
35.7
-
-
-
-
42.9
26.5
35.0
LEAD
BK
-
-
-
-
-
-
-
-
-
-
-
-
.0057
.0107
.0003
.0010
.0012
.0016
1st
SAMP.
-
-
-
-
-
-
-
-
-
-
-
-
.0027
.0089
.0030
.0033
.0080
.0038
%
REM.
-
-
-
-
-
-
-
-
-
-
-
-
47.4
83.2
-
-
-
-
AVG.
-
-
-
-
-
-
-
-
-
-
-
-
.002
.035
.002
.002
.006
.003
% REM.
AVG.
-
-
-
-
-
-
-
-
-
-
-
-
35.
-
-
-
-
-
NICKEL
BK
-
-
-
-
-
-
-
-
-
-
-
-
.026
.125
-
.085
.092
.102
1st
SAMP.
-
-
-
-
-
-
-
-
-
-
-
-
.038
.117
.025
.052
.058
.111
%
REM.
-
-
-
-
-
-
-
-
-
-
-
-
-
93.6
-
61.1
63.1
-
AVG.
-
-
-
-
-
-
-
-
-
-
-
-
.035
.054
.056
1.120
.040
.066
* REM.
AVG.
-
-
-
-
-
-
-
-
-
-
-
-
-
43.2
-
-
43.5
64.7
o
- Indicates negative removals
* % Remaining
** BK - Background (initial) concentrations (mg/1)
-------
'they tend to move little of the total accumulated solids load, will wash
out significant masses of heavy metals. Conversely, the data indicates that
'sewer flushing is very effective with respect to metals removal, and there-
fore minimizes.storm entrainment and potential metals in overflows.
Figures 62 and 63 are plots of settling velocity vs concentration
for copper (8/22) and nickel (8/29), respectively. The figures clearly show
'the wide scatter of the metals data which allowed only for assessment of set-
tling trends, but again show the very high percent remaining plateau effect
of the metals. Of all the metals, copper and zinc were the only metals to
exhibit any real settling tendency as exhibited in Figure 62. Figure 63 is
quite representative of the other metals where no real settling tendency
existed.
The results of the metals testing program definitely showed that a
significant portion of the heavy metals were associated with large particles
with very high settling velocities, as shown by the rather large drop in
J,concentration from the background to sample. The remaining fraction of fifty
.percent or more tended not to settle at any significant rate, and would be
carried downstream for long distances.
10.3.2 Correlation Analysis
A correlation analysis was conducted between the various monitored
"pollutant parameters and settling velocity, TSS and VSS, in a linear and
'logarithmic fashion for various combinations of flushes. The results are
shown in Table 56. High correlations were shown to exist between all para-
meters except metals. Significant correlations existed between cadmium, cop-
per and zinc, with settling velocity, TSS and VSS. The relationship between
,these metals and settling velocity tends to reiterate the slightly positive
-settling velocity exhibited. The correlation analysis reiterated all of the
prior findings.
211
-------
BACKGROUND CONCENTRATIONS
MHI O .225 mg/1 '"'
MH2 A .785mg/l
MH3Q 5.85;mg/l
10 n ' - „
8/22 COPPER
1.0-
5 .
3 c
O -
A
C3
I E
m Q A
o u
o °
A
A
i i i T i i i i i i i i nrrni i i r T^ r m
I 234681 234681 23468,
IO'4 10-3 FEET/SEC. I0'2 IO'1
I i i i i 11 M i i i i i i 1111 I I i i i m i I I i Mini I I
I 234681 234681 234681 234681. 23
IO'4 10-3 I0'2 lO'l - I01
CENTIMETERS/SEC. , *
VERTICAL SETTLING VELOCITY
FIGURE 62 PLOT OF Cu CONCENTRATION VS SETTLING
VELOCITY • •:.
212
-------
BACKGROUND CONCENTRATIONS
MH I O 0.026 mg/1
MH2A 0.125 mg/l
MH3Q
8/22 COPPER
1.0 H
.Oh
D
O
A
'-' L-J ©
13
ID o
i f i i i i i i i i i i r i i ri i r T i i ri i ii
I 234681 234681 23468
IO"4 10-3 FEET/SEC. I0'2 |QH
I I 1 i i i 1111 I I i i i 1111 I | | I I 111 I I I I | I I III T~l
I 234681 234681 234681 234681 23
IQ-4 IQ-3 10-2 lO'1 I01
CENTIMETERS/SEC.
VERTICAL SETTLING VELOCITY
FIGURE 63 PLOT OF Ni CONCENTRATION VS SETTLING
VELOCITY
213
-------
TABLE 56. SETTLING COLUMN RESULTS CORRELATION MATRIX
COLUMN TESTS 8/22, 8/25, 8/29, 9/7+
IND.
VARIABLE
VEL-LOG
VEL-LINEAR
TSS-LOG
TSS-LINEAR
VSS-LOG
VSS-LINEAR
IND.
VARIABLE
VEL-LOG
VEL-LINEAR
TSS-LOG
TSS-LINEAR
VSS-LOG
VSS-LINEAR
IND.
VARIABLE
VEL-LOG
VEL-LINEAR
TSS-LOG
TSS-LINEAR
VSS-LOG
VSS-LINEAR
MANHOLE
1
DEPENDENT
COD
.34
.46
.91
.94
.95
.96
BOD TKN TSS
.68
.88
.85
.87
.86
.86
.51
.75
.65
.58
.67
.57
.54
.61
1
1
.98
.98
VSS Cd
.44 .
.55
.97 .
.98 .
1
1
MANHOLE
29
15
34
18
31
16
2
DEPENDENT
COD
.59
.43
.90
.90
.89
.90
BOD
.61
.•74
.68
.84
.68
.83
TKN
.73
.65
.71
.79
.70
.78
TSS
.62
.50
1
1
.98
.98
VSS Cd
.60 .
.50
.98 .
.98 .
1
1
MANHOLE
41
09
28
14
31
14
3
DEPENDENT
COD
.39
.22
.78
.54
.78
.53
BOD
.29*+
.43
.79*+
.85*+
.70*+
.84*+
TKN
.59*+
.61*+
.67*+
.76*+
.60*+
.74*+
TSS
-.56
.83
1
1
.98
.98
VSS Cd
.56 .
.83 .
.98 .
.98 .
T
1
37
36
32
36
27
34
VARIABLE
Cr
.31
.27
-.08
-.07
-.12
-.08
VARIABLE
Cr
.56
.39
.38
.35
.38
.35
VARIABLE
Cr
.36
.23
.18
.10
.17
.10
Cu
.34
.11
.31
.07
.29
.06
Cu
.34
.13
.51
.95
.53
.44
Cu
.36
.59
.54
.61
.54
.60
Pb
.31
.12
.26
.03
.27
.02
Pb
.37
.39
.28
.14
.28
.14
Pb
.38
.27
.04
.05
.01
.05
Zn
.38
.21
.02
.21
-.06
.19
Zn
.49
.62
.41
.46
.42
.45
Zn
.54
.51
.25
.42
.25
.41
Ni
.20
-.03
.14
-.03
.10
-.05
Ni
.01
-.10
.17
-.09
.18
-.09
Ni
-.27
-.20
-.14
-.05
-.15
-.06
+ - Date set includes both short and long interval experiments.
*
Less than ten observations.
214
-------
TABLE 56. SETTLING COLUMN RESULTS CORRELATION MATRIX
COLUMN TESTS 8/29 and 9.7 (Cont'd) ++
MANHOLE 1
IND.
VARIABLE
VEL-LOG
VEL-LINEAR
TSS-LOG
TSS-LINEAR
VSS-LOG
VSS-LINEAR
DEPENDENT
COD
.22
.42
.92
.93
.94
.95
BOD
.68
.88
.85
.87
.86
.86
TKN
.51
.75
.65
.58
.67
.57
TSS
.40
.57
1
1
.97
.97
VSS
.31
.52
.97
.97
1
1
Cd
.29
.15
.34
.18
.31
.16
VARIABLE
Cr
.31
.27
-.08
-.07 -
-.12
-.08 -
Cu
.30
.05
.15
.01
.14
.02
Pb
.31
.12
.26
.03
.27
.02
Zn
.22
.10
.19
.20
.16
.19
Ni
.20
-.03
.14
-.03
.10
-.05
MANHOLE 2
IND.
VARIABLE
VEL-LOG
VEL-LINEAR
TSS-LOG
TSS-LINEAR
VSS-LOG
VSS-LINEAR
DEPENDENT
COD
.64
.50
.84
.88
.84
.88
BOD
.61
.74
.68
,84
.68
.83
TKN
.73
.65
.71
.79
.70
.78
TSS
.70
.61
1
1
.96
.97
VSS
.69
.61
.96
.97
1
1
Cd
.65
.47
.50
.60
.52
.60
VARIABLE
Cr
.56
.39
.38
.35
.38
.35
Cu
.48
.11
.54
.50
.56
.50
Pb
.37
.39
.28
.14
.28
.14
Zn
.67
.60
.59
.82
.59
.81
Ni
.01
-.10
.17
-.09
.18
-.09
MANHOLE 3
IND.
VARIABLE
VEL-LOG
VEL-LINEAR
TSS-LOG
FSS-LINEAR
VSS-LOG
VSS-LINEAR
DEPENDENT
COD
.42
.61
.79
.70
.77
.69
BOD
.29
.43
.74
.85
.70
.84
TKN
.59
.61
.67
.76
.60
.74
TSS
.62
.89
1
1
.94
.96
VSS
.60
.88
.94
.96
1
1
Cd
.37
.36
.32
.36
.27
.34
VARIABLE
Cr
.36
.23
.18
.10
.17
.10
Cu
.32
.57
.54
.71
.52
.70
Pb
.38
.27
.04
.05
.01
.05
Zn
.33
.40
.43
.63
.42
.62
Ni
-.27
-.20
-.14
-.05
-.15
-.06
++ - Date set includes only long interval experiments.
215
-------
SECTION 11
AUTOMATED SEWER FLUSHING AND RELATED TOPICS
11,1 Foreword
The results of the third phase field flushing program are described
in this chapter. Several ancillary topics are also presented, covering joint
wet weather sampling followed by flushing, and summary results of all back-
ground sewage sampling conducted during the project. Operational results for
the automated flushing module installed on Shepton Street during the fall of
1977 are described in section 11.2. In the summer of 1977 the pollutant
levels of several storm events were monitored on Shepton Street, using auto-
matic samplers. Flushing experiments.were immediately conducted on the seg-
ments to compare the relative fraction of material transported from the seg-
ment by the storm event and by sewer flushing. These results are described
in section 11.3. The results of all the background sewage flow and quality
sampling activities conducted during the three phases of field operations
are summarized in section 11.4.
11.2 Results of Automated Sewer Flushing Module
The third phase of the sewer flushing research project dealt with
the development and operation of a simplistic automated sewer flushing module.
The device developed was an air-operated gate capable of backing up sewage
flows to a predetermined level and then suddenly retracting, inducing a flush-
ing wave. Mechanical and operational details of the device are presented in
Chapter 5. This type of backup and release device was well suited to the
situation as found on Tempieton and Shepton Streets in Dorchester. The sewer
lines in the Tempieton/Shepton Street area were previously described in Chap-
ter 4. Upstream of the flushing manhole is a hill allowing development of
sufficient static head for a clean discharge. The module so developed was
installed on 8/30/77 and operated on a daily basis from 8/31 - 10/31/77.
During this period the module was checked at least 3 times per week to ensure
performance as well as conduct sampling runs. Automated flushes were sampled
seven times during the period of 9/22 - 10/13/77. Module operation was con-
tinued after 10/31/77 until mid January 1978 on a periodic inspection basis
to assess long-term operation and serviceability, as well as visual perform-
ance with respect to flushing of both the upstream or reservoir segment and
the downstream flushed segments.
Typical flush wave pollutant concentrations are shown in Figure 64
for the automated flushing module sampling experiment conducted on 10/3/77.
The plots of COD, TSS and VSS shown in the figure are for samples collected
216
-------
ro
7500-
6000-
TSS-
4500-
3000-
vss-
1500-
COD —
SHEPTON ST.
10/3/77
COD
vss
BG 0 10 30 50 70 90 110 130 150 170 190 2tO 230 250
I
TIME
FIRST OBSERVANCE OF FLUSH WAVE
-BACKGROUND LEVELS
FIGURE 64 FLUSH WAVE CONCENTRATIONS FOR AUTOMATED SEWER FLUSHING
-------
when the module released stored sewage, thereby inducing a flush wave. The
high TSS and VSS background sewage levels for this experiment were atypical
of sewage characteristics generally found on Shepton Street. Time plots of
mass transported for COD, TSS and VSS are shown in Figure 65 for the 10/3/77
flush. Total masses removals of 15.76, 9.02 and 6.48 kilograms for COD, TSS
and VSS are also indicated in Figure 65. Flush wave concentrations for the
third phase experiments were similar to the first phase flushing experiments
conducted at this location.
Flush volume details of the third phase flushing module operation
are shown in Figure 66. A profile sketch of the two pipe segments upstream of
the flushing module location is given in Figure 66. The first phase flushing
manhole is 226 feet upstream of the module. The flushing module used in the
third phase operation is located in the first phase sampling manhole. Pro-
files of backed-up sewage are depicted starting from 10 inches and ranging
up to 16 inches of sewage referenced to the invert within the flushing module
manhole. It should be noted from Figure 66 that the profile of backed-up
sewage extends up to the first phase flushing manhole for a depth of about
10 inches. Depth of sewage backup greater than this height would extend be-
yond the first phase flushing manhole and well up into the second upstream
segment. Accordingly, any flushing experiment where sewage backup is greater
than 10 inches would actually be flushing portions of the second upstream
segment from the flushing module. This distinction is important in inter-
preting the pollutant flushing removals.
A table of estimated volumes versus depth of sewage is also pre-
sented in Figure 66. Two estimates of volume are given for a given level of
backup. The first estimates,-labeled A in Figure 66, assume that only pipe
segment volume is used whereas the second estimates, labeled B, account for
manhole and house connection laterals volumes. These estimates were graphi-
cally determined from profile plots of the segments. Actual volume occupied
during any experiment is not known accurately but probably fall somewhere in
the range cited in the table.
The flushing mass removal results for the sewer sampling events are
presented in Table 57. After the date of the experiment the depth of sewage
backed up in the flushing manhole is given followed by estimates of the total
mass removals of COD, TSS and VSS. The total mass removals are also the mass
removals normalized by antecedent days between flushes since the module was
programmed to flush daily.
TABLE 57: PHASE III - AUTOMATED SEklER FLUSHING POLLUTANT REMOVAL RESULTS
DATE
9/26/77
10/3/77
10/6/77
10/7/77
10/11/78
10/13/78
Depth (inches) of Backup
at Flushing Module
15.0
16.0
9.5
9.0
11.9
14-0
Total
COD
11.30
15.74
1.85
1.17
2.33
4.08
Mass Removals
TSS
9.75
9.02
1.22
.86
1.53
2.62
(kg/Flush)
VSS
6.93
6.48
.89
.53
1.09
1.72
218
-------
275 -i
ro
10
SHEPTON ST.
TOTAL MASS REMOVALS
COD = 15.76 kg
TSS = 9.02 kg
VSS= 6.48 kg
10/3/77
0 10 30
I I I I I I I I I I I I I I i 1 I I I i 1-vss
50 70 90 110 130 150 170 190 210 230 250
TIME ,Sec.
FIRST OBSERVANCE OF FLUSH WAVE
I— BACKGROUND LEVELS
FIGURE 65 FLUSH WAVE MASS REMOVALS
-------
ro
ro
o
50
Ul
UJ
u.
UJ
UJ
46 -
DEPTH OF BACKUP
(INCHES)
e
10
12
M
16
VOLUME
A(Cf) B
54
82
. 108
151
194
67
100
.. 132
180
230
A = PIPE VOLUME
B= PIPE VOLUME PLUS MANHOLE VOLUME PLUS LATERAL VOLUME
16"
15"
14
BACKUP LEVELS FOR
AUTOMATED FLUSHING MODULE
THIHD PHASE
FLUSHING
MODULE SITE
T
BACKED-UP
SEWAGE
47 -
46
400
I
300
• -- 200
PIPE LENGTH (FEET)
FIGURE 66 FLUSH VOLUME DETAILS OF THIRD PHASE PROGRAM
i
100
-------
The pollutant removal results for the 10/6/77 and 10/7/77 experi-
ments approximate the removal rates during the first phase flushing program
on Shepton Street. Average normalized mass removals of COD, TSS and VSS for
all good experiments conducted in the first phase are 1.79, 1.29 and 1.0
kg/day/flush. The depth of sewage backed up during these two experiments was
less than 10 inches. These two experiments were flushing the same pipe length
as in the first phase program. The depths of back.ed-up sewage exceeded 10 inches
for experiments conducted on 9/26/77, 10/3/77, 10/11/77 and 10/13/77. For
the experiment conducted on 10/13/77 the sewage profile extended 175 feet up
Into the second segment, or 400 feet from the flushing module. Inspection of
the mass removal results for these dates in Table 57 indicate that the flushed
loadings were substantially higher for these experiments in comparison to the
two flushing experiments on 10/3/77 and 10/6/77. The last time the second
segment upstream from the flushing module was flushed was during the pre-
cleaning work conducted in September of 1976, in preparation for the first
phase program.
The mass removal results shown in Table 57 indicate that the auto-
mated flushing removal effectiveness was comparable to first phase flushing
results. Sediment levels in both upstream segments were noted during the
flushing events and during the maintenance checks. No accumulation of sani-
tary deposits was present in the pipes, indicating that the flushing waves
adequately scoured and entrained any materials that deposited during the
backup period. All pipe segments including the downstream sampling manhole
were free of deposits during this period.
The module was left in operation at the end of the sampling phase
until the middle of January, 1978 when the site was inspected. Review of
the chart from the liquid level sensory devices, which recorded the depth of
sewage backed up during a flushing operation, indicated that the module con-
tinued to perform daily unattended until the end of the year. The car bat-
teries that provided electrical power to the device were inoperative. Tempo-
rary batteries were installed and the module went through a normal sequence.
The module was again checked in late April of 1978. Visual inspection of the
sediments indicated that sanitary deposition, sand and grit had again accumu-
lated to pre-project levels.
11.3 Storm Event Monitoring
During the summer of 1977 an automated sampling was installed in
the sampling manhole on Shepton Street for the purpose of monitoring the im-
pact of stormwater flows on sewage deposits. The Shepton Street test seg-
ment is separated but receives substantial clear water inflow from roof
drains connected into the sewer. The idea of the experiments was to monitor
the pollutant masses transported during a storm, and then immediately flush
the segment to determine residual deposited masses. Visual wet weather flow
in small diameter pipe segments over the course of the project indicated re-
latively smooth flowline turbulent flow conditions. Flush waves at com-
parable depths of flow were always extremely turbulent with many random erra-
tic flow eddies. It was of interest to determine just what effect storm-
water conditions would have on suspending and transporting deposited organic
221
-------
pollutants. The concept of the experiment was simple. In practice," the
logistics of field flushing just after a rainstorm was almost insurmountable.
The first rainstorm monitored occurred on May 5, 1977. Very little
rainfall was recorded at the Blue Hill Observatory. Heavy rainfall was noted
at the site for about 1 1/2 hours during the storm event. The automatic sam-
pler went into sequence and withdrew 24 1-liter samples at 10-minute inter-
vals. Concentrations of TSS, VSS and BOD ranged from 150 to 250 mg/1 over
the four-hour period. BOD concentrations correlated with VSS levels. The
liquid level sensing device was inoperative so that depths of flow were not
recorded for the entire event. An average flow of .05 cfs was estimated for
the storm event. The amount of BOD transported during the storm event, less
the base level dry weather contribution, was roughly estimated at 3.53 kg.
Heavy sanitary deposits were noted in the segment. The segment was immediate-
ly flushed within 2 hours using 50 cubic feet of v/ater at a rate of about 0.5
cfs. Peak VSS and BOD levels during the flush event equaled 8000 and 1800
mg/1. The flushed masses of VSS and BOD were estimated to be 10.86 and 6.16
kg, respectively. The segment was flushed clean by the experiment.
A second storm event was monitored on June 7, 1976 on Shepton Street
The storm began at 2 a.m. and stopped at 10 a.m. Rainfall intensities of
0.17 and 0.14 inches per hour for the first two hours of the storm were recorded
at the Blue Hill Observatory, and thereafter varied from trace to 0.10 inches
per hour at 9 a.m. The automatic sampling began its sequence at 2:05 a.m.
and continued for four hours, taking samples at 10-minute intervals. The
liquid level sensing device monitored flow levels throughout the storm event.
Figure 67 shows the storm hydrograph and the VSS pollutograph for the first
four hours of the event. The peak VSS concentration monitored during the
storm event was 1077 mg/1. The total mass under the measured VSS polluto-
graph is 3.48 kg, and the estimated base-line dry weather contribution is
0.53 kg. Although the rainfall information shows only significant rainfall
falling within the last hour of the storm, the hydrograph during the final
5 hours shows several minor broad peaks. A flushing experiment was conducted
at 10 a.m. on that morning. Peak flush wave VSS concentrations reached 8100
mg/1 and the estimated flushed mass is 7.37 kg. The segment contained sani-
tary deposits prior to flushing and was clean after the flushing event.
Although the results of these two storms are meager, several sug-
gestive inferences can be drawn from the data. First of all, sanitary depo-
sits are not easily dislodged and transported during storm events. Extreme
"first flush" phenomena may be the results of an extreme storm event flush-
ing the accumulated deposits of a long sequence of dry and alternatively light
rainfall periods. Secondly, flushing does cause the required turbulence to
to suspend, entrain and transport sanitary deposits. Thirdly, the time
interval required for flushing can be, on the average, longer than the aver-
age return period for storm events, since minor to moderate intensity storms
may not flush clean the pioe segments.
222
-------
FIGURE 67
SHEPTON STREET- TYPICAL STORM
HYDROGRAPH/ POLLUT06RAPH
6/7/77
(0 FS ) VSS MASS REMOVAL .
FLOW (nfl) Kg/sec.
i.o
ro
ro
co
J09
.08-
.07-
.06-
JOB-
.04
.03-
.02-
.01-
0-
.80-
ASSUMED BASE FLOW
= 0.01 CFS
I \
I \
I \
"\
STORM BEGAN
AT 2'OB
TIME (MINUTES)
\ /
T 1 1 1 1 1 1 1 1 r
320 340 360 380 400 420 440 460 480 600
STORM ENDED
AT lO'OO AM
-------
11.4 Background Sewage Characteristics
Summary statistics of background sewage characteristics in mg/1 for the
four test segments are given in Table 58. The mean and standard deviation
shown for each pollutant was computed using data collected over the entire
field flushing program. The results are generally atypical of sewage strength
levels normally encountered at treatment plants for strictly domestic sewage.
Both the mean levels and coefficients of variation are markedly higher, es-
pecially for Port Norfolk and Walnut_Streets, than comparable treatment plant
levels. Great care was taken in sample collection to ensure that sediments
were not scraped into sample bottles.
Dry weather sewage flows were measured at different times of the
year over the course of the project. Procedure measurement details are given
in Chapter 5. Measurements were primarily performed at Shepton Street and
Port Norfolk Street test segments since discharge levels at Walnut and
Tempieton Streets were extremely sluggish and difficult to measure. Instant-
aneous flow rate determinations at Shepton Street ranged from 5800 to 14,200
gpd with an average of 12,900 gpd. The per capita waste rate for the
Shepton Street test segment is 56 gpcd using the average discharge and the
census population of 230 people. Dry weather sewage flow gaged at the
first phase sampling manhole at Port Norfolk Street ranged from 3000 to 4000
gpd with an average value of 3600 gpd. A special flume described in
Chapter 5 was installed in the first phase flush injection manhole. The
flume was flow calibrated and then used to determine dry weather flow levels.
Discharge rates at this location averaged about 0.0025 cfs corresponding to
a per capita waste rate of 53 gpcd. The estimated population tributary to
the first phase manhole down to the sampling- manhole was adjusted downward
from 94 to 60 people ba'sed on an assumed rate of 60 gpcd for the entire
street. Per capita waste rates of 60 gpcd were assumed for the Templeton and
Walnut Street segments based on the flow measurements at Shepton and Walnut
Streets. The per capita waste rates are low but agree with previous
measurement work performed in the Dorchester area (4).
224
-------
TABLE 58. SUMMARIES OF BACKGROUND SEWAGE CHARACTERISTICS
BACKGROUND CONCENTRATIONS (mg/1)
POLLU-
TANT
COD
BOD
TKN
NH3
TP
OP
SS
VSS
PORT NORFOLK
Mean/Std.Dev.*
2065/2066
774/714
117/103
29/17
17/11.2
8/5.2
1116/1829
1332/1813
No.
Samples
21
14
12
9
9
9
50
35
WALNUT
Mean/Std.Dev.*
2314/4238
1069/1468
91/69
43/15.2
20/13.2
3.5/7.5
1792/2879
1523/2074
No.
Samples
17
13
12
12
12
12
27
26
TEMPLETON
Mean/Std.Dev*
546/ -
407/273
17/ -
87 -
4.8/ -
1.87 :
1112/1323
1021/1226
No.
Samples
1
10
1
1
1
1
26
20
SHEPTON
Mean/Std.Dev*
405/ -
256/49
547 -
33/ -
11.57 -
10.0/ -
392/971
265/204
No.
Samples
2
5
1
1
1
1
36
12
ro
en
*Std. Dev. - Standard deviation
-------
SECTION 12
PREDICTIVE TOOLS
12.1 Foreword
In this chapter two predictive approaches are presented that are
useful in the general assessment of potential sewer flushing programs. The
first model deals with estimating the transported fraction of flushed pollut-
ant loading as a function of distance downstream from the point of flush.
This approach was empirically developed from the second phase field flushing
program data. A comparison 'is presented of results using the prediction proce-
dure with other second phase data. The second model discussed is a generalized
procedure for estimating daily dry weather sewage solids deposition loadings
within each manhole to manhole segment of an entire collection system net-
work. A comparison of predicted pollutant deposition loadings for the phase
one test segments with phase one field flushing results is also presented.
Estimates of flushed solids transported downstream using the first
of the two aforementioned modes are presented in Section 12.2. This procedure
was expanded to provide estimates of organics and nutrients transported down-
stream by flushing and these results are presented in Section 12.3. Details of
a generalized sewer system deposition model are given in Section 12.4. Verific-
ation of the approach using field flushing results are given in Section 12.5.
12.2 Downstream Transport of Solids in Suspension
12.2.1 Overview of Methodology
Solids suspended by a flush wave will tend to redeposit some
distance downstream from the point of suspension due to the action of gravity.
The distance downstream at which particles of a given size will settle is a
function of the particle characteristics, concentrations and the residual
energy and shear force of the flush wave or flow at those distances.
The question as to how far the suspended materials will go is
fundamental in establishing the spatial frequency of flushing to guarantee
that a specified fraction of the flushed materials will ultimately reach a
treatment plant. The objective here is to provide a crude and preliminary
answer to that question. A more accurate analysis of the flush wave effect-
iveness in carrying the scoured solids long distances downstream would
require characterization of the settleability characteristics of the
solids picked up by the wave and the development of a predictive tool for
the propagation of the wave downstream.
226
-------
Knowledge of the evolution of the wave or flow in space and time, as-
sociated with some criteria of shear stress, would probably provide the best
assessment of its downstream carrying capacity. Nevertheless, it should be
recognized that regardless of the level of effort and sophistication used,the
answers that can be derived from such an approach are still based on ideal
conditions and crude approximations to the problem. This operational
conclusion was reached due to the impossibility of actually predicting the
behavior of the wave and the materials in suspension under the effects of
manhole drops, sharp flow turns, and the downstream interference of household
connections and contributions from merging portions of the collection system.
Ackers and Harrison (25), Mitchel (26)., and Martin and deFazio (27)
dealt successfully with the specific problem of the attenuation of flood
waves in straight, circular pipes, using numerical methods for the solution
of the continuity and momentum differential equations involved. Sonnen (28)
modelled the transport of sediments in sewers by considering three
distinctive transport mechanisms recognized in the literature, namely: bed-
load, suspended load and washload transport. Each of the processes was
modelled separately, with sediment mass balances performed to guarantee that
the predictions do not exceed the amounts of materials available for each
type of transport. Although the application of these concepts to the problem
of sediment transport in collection system laterals is encouraged, many
questions regarding the whole methodology still remain unresolved and require
considerably more research. It is believed that unless the hydraulic
problem is solved with greater accuracy, especially in the case of the small
volume flushes occurring over short time intervals considered in this study,
sophisticated approaches to the sediment transport process itself are not
justified.
A much simpler approach employed here used estimates of forward
velocities together with discrete settling and indirectly accounting for
flocculance and hindrance. Figure 68 depicts the methodology used and the
following steps summarize the computations performed:
1. The travel times of flush waves to the three downstream manholes
measured during the Port Norfolk second phase flushes were used to define an
exponentially decaying function*relating velocity with downstream distance
travelled. Assuming a straight pipe of uniform slope and neglecting down-
stream flow contributions a residual wave forward velocity, VL, was estimated
at a downstream distance, L, as follows:
VL = ea+bL (23)
2. An average velocity over the distance, L, was defined as the
average of the velocities V0, at L = 0, and VL computed at L, as follows:
V = 0.5 (VQ + VL) = 0.5 ea(HebL) (24)
3. With the estimated average forward velocity V, an iterative
procedure (method of secants) was used to solve Manning's equation for the
depth of flow h;
* Empirical function with parameters determined by least squares. .
227
-------
ro
r-o
CO
SETTLING
COLUMN
TEST
2ND PHASE A
2ND PHASE B
FLUSHES
NOTE
Units: V,VL,V0, VSETT-fps
L- ft
L - ft
"*V SETT." •"
SETTLING "V"
FIGURE 68. OVERVIEW OF PROCEDURE FOR ESTIMATION OF FLUSHED SOLIDS TRANSPORTED DOWNSTREAM
-------
4. Assuming the flow depth to be uniform and equal to h from 0 to
L, the geometry of the pipe and the flow depth, h, defined the cross-sectional
flow area, Ac (ft^);
5. The flow surface area, AS (ft2), was computed by:
„ = Volume _ Ac ' L .
As Depth h '
6. The overflow rate defining the cut-off settling velocity of
the particle that just passed length L in suspension is given by:
A • V
OR =u Flow _ c _
U>K- uSett. Surface Area AS (26>
Equation 26 can be rewritten as:
Vsett. ' H* (2?)
which could have been derived directly by equating the time of fall of the
particle to the bottom of the pipe with its time of horizontal travel to the
distance L;
7. The settling column test results from all three downstream
manholes in Port Norfolk Street were used to define a curve relating the
percentage of solids remaining in suspension with settling velocity and also
accounting for flocculent and hindered settling; and
8. Finally, the fraction of the suspended solids that remain in
suspension throughout length, L, can be defined using the settling velocity
from equation 27 in the percent remaining versus settling velocity curve.
12.2.2 Definition of the Flush Have Forward Velocity
The travel times of the flush wave, measured at three downstream
manholes from the flush injection manhole, in Port Norfolk Street, were used
to define an exponentially decaying function relating the flush wave velocity
with downstream pipe length. The travel times used in each flush and at each
manhole were those c.orresp_pnding to the arrival of the peak of the flush wave,
identified by the highest depth measured. Whenever the maximum depth was mea-
sured at more than one successive time step, indicating a flat peak, the aver-
age of the times to the first and last highest depth measurements was used. A
diagram of the flushing and sampling manholes along Port Norfolk Street is
presented in Figure 20, Chapter 5.
The total distances traveled by the wave to the successive down-
stream manholes are approximately equal to:
First Downstream Manhole (DMH) 162*ft.
Second " " 409 ft.
Third " " 659 ft.
229
-------
These distances, combined with the travel times defined above, provided
estimates of the flush wave velocities. There were a total of 101 velocity
determinations for all three downstream manholes from the second phase field
program.
A linear regression of the logarithm of the velocities on the
corresponding manhole distances, for all 101 points, yielded a correlation
coefficient, about equal to -0.4. The mean and standard deviation of the
velocities at each manhole were then computed and any velocity measurement at
each manhole falling outside its corresponding range of the computed mean
±2 standard deviations was excluded in a new regression run. An improved
correlation coefficient was obtained. The resulting equation for velocity,
V, , at a downstream distance L is:
v = eO.789-0.000634L (28)
where e = 2.71828 (R = -0.63)
A total of 10 velocity points were eliminated from the original set leaving
91 points, almost equally distributed among the three manholes, for the
regression. It should be noted that the data used in the regression
encompassed different flush volumes and injection rates, which explains in
part the correlation coefficient obtained.
A partition of the velocity data by flush volumes and injection
rates could have been adopted. The percentages of solids remaining in
suspension at downstream lengths>L, could have been associated with ranges of
flush volumes and injection rates. However, it was felt that the precision
of the whole approach would not warrant such refinements.
12.2.3 Overflow Rates or Settling Velocities
For a fixed downstream distance L, the wave average velocity,V, can
be estimated using equations 23 and 28 . Computation of the overflow rate
at the distance,L,requires that a flow depth be associated to that velocity.
Manning's equation was used to determine depth of flow. The slope used
in Manning's equation derived from the results of the loop-rating curve
optimization described in Chapter 7. Although the three pipe segments
flushed in Port Norfolk Street have different slopes, the optimized slope of
the intermediate pipe segment was assumed constant throughout the range of
valid values of L. Programming logic was prepared to solve Manning's
equation for h given the velocity V, by a trial and error approach of drawing
successive secants to the velocity versus depth curve. Precise convergence
was obtained in a few iteration steps.
Using this flow depth together with the distance L and correspond-
ing forward velocity V in equation 27 the overflow rate or settling
velocity can be computed.
12.2.4 Percent Solids Remaining versus Settling Velocity Curve
The computed values of percent solids remaining in suspension and
230
-------
their associated vertical velocities, in fps, derived from the settling
columns tests described in Chapter 10, were plotted and are shown in Figure
69 for all three downstream manholes. A smooth average curve was then drawn
through those points labelled, "average settling column curve." This curve was
constructed in the following manner. The settling velocities were broken into
small ranges and all remaining fractions falling within a given range were
averaged out and the average values were plotted against the mid-range value
of the settling velocity. The "average settling column curve" was then drawn
through those averaged points with a high degree of adherence. This curve
was considered to be representative of the ideal settling column.
Another curve was- defined to approximate the turbulent settling
conditions occurring in the sewer lines, labelled "conservative turbulent
settling curve," Five plotted points falling exactly on or very near the
average settling column curve, and well spaced in the range of 1 x 10-4 to
3 x 10-3 fps, were used to construct that curve. The settling column data
referring to those points were then reviewed by: a) doubling the settling
or detention time while keeping the percentage remaining unchanged; and
b) assuming the overflow rate to be only 65% of the revised overflow rate.
A smooth curve was drawn connecting the plotted revised settling velocities
corresponding to the five selected points as indicated above. The criteria
used in drawing the revised curve is usually adopted in the design of
settling basins. Conditions in sewer lines are certainly more turbulent
implying that the resulting curve should be considered conservative in regard
to the fractions remaining in suspension.
The turbulent settling curve was fitted by a polynomial in th
range of settling velocities of 1 x 10~4 to 1 x 10~2 fps. This polynomi
the
polynomial
is given by:
P = 424.xl05Vsett - 960.xl03V^ett + 123.xl02V9ett + 10.8 (29)
where P = Percent TSS remaining in suspension
= Settling velocity, in fps.
It should be noted that a fraction of particles, i, with settling velocities
less than that given by equation 27 defined as V., will also settle along
the distance, L. In the design of settling tanks it is usually assumed that
the fraction is given by Vj/V$ett. for a particular i. That fact was
neglected here under the assumption that the conservative turbulent settling
curve already accounts for this fraction.
12.2.5 Final Results
The procedure described above was used to compute final percentages
of solids remaining for various downstream distances, L. A plot of the
computed values is shown in Figure 70 indicating the percentage of TSS
remaining in suspension, P, as a function of downstream
distances, L, in feet. The results indicate that the
231
-------
o
z
lOO-i
90-
80-
70-
u 6°H
cc
CO
a
o
to
a
ui
o
UJ
o_
co
to
LU
O
a:
UJ
a.
50-
40-
30-
20-
10-
SETTLING COLUMN EXPERIMENTS (9/76 8/25)
FLUSH 1 MH 1 ®
MH 2 A
MH 3 H
CONSERVATIVE
TURBULENT SETTLING-
CURVE
AVE.
SETTLING
© COLUMN
CURVE
FEET/SEC
10
-4
I I I I I I I I
34 681
io":
t I I i i i i i
34 681
10
-2
I I i i i i 11
34 681
\ I I I I I I I
34 681
\
2
\ I t I I I I I
34 681
I
2
10
-3
-1
CENTIMETER/SEC
10" 10
VERTICAL SETTLING VELOCITY
\ \ II I I It
34 681
101
FIGURE 69. PLOT OF PERCENT SUSPENDED SOLIDS REMAINING VERSUS
SETTLING VELOCITY
232
-------
10
100
500
1000
"2000
3000
DOWNSTREAM DISTANCE-FEET (L)
FIGURE 70 PERCENT TSS, BOD AND TKN REMAINING IN SUSPENSION
VERSUS DOWNSTREAM DISTANCE FROM MANHOLE FLUSHED.
PORT NORFOLK DATA
233
-------
percentages of solids suspended from the first segment remaining in suspension
at the successive downstream manholes in Port Norfolk are:
First Downstream Manhole (DMH) 68%
Second " " 34%
Third " " 24% .
The computations also indicate that roughly 12% of the solids suspended from
the first segment would not resettle again. This estimate agrees favorably
with the Imhoff cone results for the first flush at the first sampling man-
home described in Chapter 10. The percentage of solids remaining in
suspension is 16.9% which was computed by averaging out the results from all
Imhoff cone tests performed on samples taken from the first flush for this
location.
The smooth curve drawn in Figure 70 is mathematically expressed by:
P(%) = - 1.33 (InL)3 + 33.86 (InL)2 - 288.57 (InL) + 834.32 (30)
(160 <.L £3000)
where: P = percentage of solids remaining in suspension;
In = indicates Napierian logarithms; and
L = downstream distances, in feet
The percentages given by equation 30. should apply only to
deposits over a single segment downstream from the flushing manhole. A
convolution of equation 30 would be necessary to account for masses
removed from more than one pipe segment.
Examples of convolution computations over multiple segments involv-
ing the application of equation 30 are given in the next section and in
section 12.3.1, Furthermore, detailed examples are also presented in
Chapter 16 of this report.
12.2.6 Verification
The methodology just described was verified using average TSS
values measured in the Port Norfolk second phase serial flushing program,
described in Chapter 9. The average masses removed by the first, second
and third flushes of the day in each of the manholes were computed, and are
presented in Table 59. This table has been previously prepared in Chapter 9
but is repeated for ease in reference. The percent removals at each manhole,
by flush, were also computed and are indicated in Table 59 assuming that
after the third successive flush all deposits had been flushed out.
If the assumption is made that the first pipe segment was actually
clean after the third flush, then the percentages indicated in Table 59
for manhole 1 can be taken as average percentages of the total mass deposited
in the pipe which were removed by each successive flush. The same interpreta-
tion cannot be applied for manholes number 2 and 3, because it is unclear
how much of the measured masses originated in the pipe segment immediately
234
-------
TABLE 59. AVERAGE TSS MASS AND PERCENT REMOVAL BY FLUSH AND BY
MANHOLE - SECOND PHASE-SERIAL FLUSHES
Manhole Number
Flush
No.
1
2
3
1
Mass(kq)
2.85
0.48
0.42
%
76.05
12.74
11.21
2
Mass(kq)
4.03
0.79
0.75
%
72.39
14.20
13.41
3
Mass(kq)
7.39
2.24
1.48
01
h
66.52
20.17
13.31
Totals
3.75
100.00
5.57
100.00
11.11
100.00
upstream from the manhole, and how much came as suspended load from upstream
segments.
Using the first segment as a reference, the settling velocity
methodology just described was applied successively to the three segments
and the average total TSS masses transported past each manhole were
estimated and compared to the average measured values for the first flush.
The following computations were performed in terms of the average values
given in Table 59. The average removal for flush 1, manhole 1 is 76%. The
settling velocity criteria estimated that percentage as 67.5%, implying
that the total mass in the first pipe segment, before the flushes, was
actually 4.22 kg, as opposed to 3.75 kg. This means that after the third
flush there was still 0.47 kg of TSS remaining in the pipe segment. In
order to estimate, from the percentage removals, the actual masses flushed
out of each pipe, it is necessary to have.an estimate of the total masses
deposited in each pipe segment prior to the flush. Starting with the value
of 4.22 kg in the first segment, it is a reasonable approximation to
assume that the deposits in the other pipe segments are inversely prooortion-
al to their slopes. Total deposit estimates of 6.09 kg in the second down-
stream segment and 7.40 kg in the third using that assumption. These values
are indicated in Table 60. .
Percent removals from the successive downstream segments were
computed using exactly the same procedure described for the removals from
the first segment, only that the overflow rates or settling velocities were
computed from equation 27 modified to:
V
Sett.
_ h • V
L - L.
(31)
where: L = 0 ft for the materials removed from the first segment;
L = 162 ft for the materials moved from the second downstream
segment;
L = 409 ft for the materials moved from the third downstream
segment;
and,all other variables have exactly the same meaning as
defined for equation 27 .
235
-------
The percent removals computed by the methodology using equation 31
instead of 27. are presented in Table 60. Direct use of equation "30 in
the convolution computations would have slightly overestimated the percent-
ages indicated for segments 2 and 3 because the average wave velocities
corresponding to those segments would have been slightly overestimated. The
percentages indicated in Table 60 can be applied to the appropriate total
masses deposited in the pipe segments to yield the estimated total masses
flushed out from each manhole. Assuming that all the deposits in each
segment are at some point in time suspended by the flush wave to redeposit
again (reasonable assumption for a few pipe segments), and assuming independ-
ence in the process of scouring and deposition of the materials from each
pipe segment. The resulting values are shown in Table 60 as Item 3 -
estimated mass flushed out. Comparison of the estimated and measured average
values shown in the last two lines of Table 60 indicate a satisfactory
agreement for the masses flushed out from manholes 2 and 3. The value for
manhole 1 was fixed as a starting point.
The proceeding result supports the credibility of the crude
approach used in this analysis to define TSS masses remaining in suspension
as a function of downstream distance from the point of the flush injection.
It also encouraged using the TSS formulation as a basis for assessing flush -
wave efficiency in transporting organics, typified by BOD, and nutrients,
typified by TKN.
TABLE 60. COMPARISON OF MEASURED VERSUS ESTIMATED AVERAGE
TSS MASS REMOVALS - PORT NORFOLK ST. (SECOND PHASE)
Manhole Number
1.
2.
3.
4.
Item 1-
Estimated mass deposited (kg) 4.22
% remaining in suspension
from:
Segment 1 67.50
Segment 2
Segment 3
Estimated mass flushed out
(kg) 2.85
Measured mass flushed out
(kg) 2.85
2
6.09
34.20
46.80
-
4.29
4.03
3
7.40
24.20
27.90
41.20
*
5.77
7.39
*0.242 x 4.22 + 0.279 x 6.09 + 0.412 x 7.40 = 5.77.
12.3 Downstream Transport of Orqanics and Nutrients
i
As described in Chapter 10, the samples collected from the settling
column tests were analyzed for VSS, COD, BOD, TKN, NHg, OP and TP besides
TSS. The settling column determinations of BOD and TKN were used to
estimate flushing efficiency in transporting suspended organic matter and
236
-------
nutrients downstream.
The following steps were performed:
a) All settling column analytical results including results of
samples from all flushes and all sampling ports were used to define regress-
ion equations relating the fractions of BOD and TKN removed from suspension
to the fractions of TSS removed from suspension. The regression equations
computed in this step are given by:
o m
BODSett. = 0>861 x TSSSett (R = °'74) (32)
™Sett. = °'601 x TSSSet?5 (R = °-80> (33)
Where BOD$ett. » TK%ett. and TSS$ett. above refer to the fractions
of BOD, TKN and TSS removed during the settling analysis. The correlation
coefficients of 0.74 and 0.80 for BOD and TKN indicate that fractions of
settleable BOD and TKN can be reasonably estimated on the basis of the
fractions of TSS removed. The TSS fraction settled values used in the BOD
regression covered the range of 0.44 to 0.96, with the great majority of the
values falling above 0.7. In the TKN regression the TSS values ranged from
0.10 to 0.96, again with the great majority of the values above 0.7.
b) Equation 30 was used with the above regression equations to
determine the percentages of BOD and TKN remaining in suspension as a
function of downstream distance from the point of the flush injection. Given
equations 30 , 31 and 32 the following calculations were performed:
. For a fixed, L, equation 27 was used to estimate the percent
of TSS remaining in suspension, P;
. The value (1. - P/100) represents the fraction of TSS removed
from suspension, TSS<- t- ;
= (I- - P/100) was used in equations 32 and 33 to
estimate BODsett. and TKNSett. > tne fractions of -BOD and f KN
that redeposited; and,
. The values (1. - BODsett.) and (1- ~ TKNSett.) represent the
estimates of the fractions of BOD and TKN sti'll remaining in
suspension at the downstream distance, L.
Plots of BOD and TKN remaining as a function of downstream distance, L,
computed for several values of L, are shown in Figure 70. The results indicate
that the percentages of BOD and TKN suspended from the first segment remaining
in suspension at the successive downstream manholes (DMH) are approximately
as follows:
BOD TKN
First DMH 90% 9Q%
Second DMH 62% 69%
Third DMH 50% 60%
237
-------
The computations also indicate that roughly 33% of the BOD and 50% of TKN
suspended from the first manhole would not resettle again.
A brief assessment of the settling column results for VSS, COD,
NH3 and TP indicates that, if a similar approach used for BOD and TKN was
applied, their percent remaining curves would plot on Figure 70 relative
to the TSS, BOD and TKN curves, as follows: a) VSS - slightly above the
TSS curve; b) ,COD - between the TSS and BOD curves; c) NHs - well above
the TKN curve; and, d) TP - slightly above the TKN curve.
Notwithstanding the limitations of this procedure, a more
comprehensive predictive tool could be developed using regression or
normographs to give approximate estimates of the fractions remaining in
suspension as a function of flush volume and rate, and pipe size, slope and
length. The generalized procedure could be developed if it were repeated
for a series of different pipe sizes and slopes and if the flushes were
separated by volume and rate.
12.3.1 Verification
The average BOD mass removals from the second phase (serial) flush-
ing program are shown in Table 61, with the average percent removals by flush
and by manhole.
TABLE 61. AVERAGE BOD MASS AND PERCENT REMOVAL BY FLUSH AND BY MANHOLE
SECOND PHASE SERIAL FLUSHING PROGRAM
Flush
No.
1
2
3
Totals
Mass (kg)
0.743
0.058
0.137
0.938
1
%
79.2
6.2
14.60
100.0
Manhole
Mass(kq)
1.001
0.189
0.164
1.354
Number
2
%
74.0
13.9
12.1
100.0
Mass(kq)
2.045
0.456
0.452
2.953
3
%
69.3
15.4
15.3
100.0
Using the curve in Figure 70 for BOD remaining in suspension and employing the
same logic as given in Section 12.2.6, Table 62 was prepared. The first item
presents the estimated BOD mass deposits per segment. The second item gives
the percentages of BOD remaining in suspension per downstream manhole from
each segment. The last two items present estimates of mass flushed out and
average measured mass removals, respectively.
Comparison of the estimated and predicted BOD masses flushed out,
that is, rows 3 and 4 of Table 62 show that, despite the limitations on
both the estimated and measured values, the data does not seem to
contradict the simplified approach used in defining the flush wave efficiency
in transporting the flushed BOD loadings downstream.
238
-------
TABLE 62. COMPARISON OF MEASURED VERSUS ESTIMATED AVERAGE BOD MASS
REMOVALS - PORT NORFOLK ST. (SECOND PHASE) SERIAL FLUSHING
Manhole Number
Item
1.
2.
3.
4.
Estimated mass deposit (kg)
% remaining in suspension
from:
Segment 1
Segment 2
Segment 3
Estimated mass flushed out
(kg)
Measured mass flushed out
(kg)
1 2
0.818 1.180
90.9 62.8
78.4
-
0.743 1.439
0.743 1.001
3
1.434
50.4
57.4
78.0
2.208
2.045
12.4 Simplified Sewer System Deposition Model
Details of a procedure for obtaining estimates of the amount
of daily dry weather deposition loadings within each manhole to manhole
segment of a sewer collection system are provided in this section. A much
simpler procedure is described in Chapter 13. A number of crude "approxima-
tions and simplifications are used in this procedure and therefore, the
results are not purported to be a substitute for those provided by more
rigorous approaches (28,29).It is intended to provide estimates for only
dry weather conditions and has no provisions for considering transient wet
weather phenomena. No distinction is made between bedload, suspended load
and washload deposition and resuspension characteristics. The major
simplifying assumption of the model is that the amount of deposition remain-
ing in any segment over the course of a day is computed as the residual
loadings not washed or moved downstream during peak dry weather flow
conditions. No detailed accounting is made of the temporal pattern of
diurnal deposition, resuspension and transport phenomena. The general outline
of the approach is given here but further details are available (4).
12.4.1 General Concepts
A well designed sewerage system should not only convey flows but
should also minimize the deposition of sewage solids during dry weather
conditions. There are in use an ample number of suitable empirical and
theoretical equations for flow design but no uniform criteria have been
established to prevent solids deposition.
The approach commonly used to prevent deposition is the method of
minimum permissible velocity. However, the use of average velocity
consideration is not necessarily the most robust criterion to use for a
wide range of typical operating conditions.
239
-------
A more fundamental approach is the method of fluid shear stress, T,
given by equation 34 .
T = prs
(34)
where p = specific weight of water
r = hydraulic radius, and
s = energy slope
Yao (30) reviewed experimental results dealing with fluid shear
stress measurements and concluded that the average boundary shear stress
computed by equation 34 will approximate the actual local boundary shear
stress within the possible region of deposition, provided that the flow
depth is equal to or greater than one-third of the sewer diameter.*
Yao also concluded that a shear stress of .02 to .04 psf is adequate
for self-cleaning for removal of particles in the range of 0.2mm to l.mm in
sanitary sewers, while a shear stress of .06 to .08 pfs is necessary to dis-
lodge and transport particles of relatively larger sizes for self-cleaning of
.combined sewer systems.. ...In addition* ..this work .?.howed that the .present prac-
tice of using a constant minimum velocity for all sewer sizes tends to under-
design larger sewers and over-design smaller sewers.
Deposition Mechanisms. Shield's classic results are commonly used
to predict solids deposition in sewerage systems. Shield's results, however,
relate to bedload movement and specifically to uniform particles moving on
the surface of the bed. In simple terms, there are two primary mechanisms
involved in the transport of sewage particles: bedload transport and
suspension.
The first to use bedload transport considerations to predict
deposition in sewers was Camp (31). Assuming a particle specific gravity
of 2.65, Shield's relationship (32) for bedload transport for large shear
Reynolds numbers** is given by:
TC/(PS-P)P = -06 => TC = .02p (35)
where p = particle diameter; (mm); PS = specific particle weight, and
T = critical wall shear stress (psf)
\*
The second transport mechanism is suspension. Hughmark (33)
correlated 14 sets of data on slurry transport and Raths (34) conducted
experiments on sand sediment in sewers. In order to prevent deposition of
sand particles (specific gravity = 2.65), a critical wall shear stress must
be maintained or exceeded. The results of their experiments can be
summarized by the following relationship:
The actual or local boundary shear stress varies considerably, with the
maximum occurring around the center line of the channel and the minimum near
the water surface.
**
Shield's constant equals 0.06 for this flow condition.
240
-------
T
c
= -021 p173 (36)
The smaller particles (less than 0.05 mm) of Hughmark's data
closely agree with the above functional form. A reasonable first order
approximation is to assume that both mechanisms transport heterogenous
materials through sewer systems. The geometric average of equations 35
and 36 can be used to predict transport requirements. The equation
relating the critical wall shear stress, TC, necessary to move a particle of
given diameter, p, is the following:
TC = .02 p2/3 (37)
12.4.2 Single Segment Deposition Model
Equation 37 and sewage particle size distributions were used to
predict the quantity of suspended solids deposited from dry-weather flow
over a single length of pipe. The results computed from equation 38 with
two particle distributions (29,35) and the experimental results from the FMC
study (.12) were fitted by a simple single term power function given by
equations 36 and 37 :
-1.2
Z = 40 (—} for T > .004 psf (38)
Z = 40 for T < .004 psf (39)
where Z is the percentage of the suspended solids in the dry weather
sanitary flow that is deposited if the wall shear is less than T.
The shear stress, T, would be computed for maximum daily dry
weather flow conditions. Maximum daily peak flow, QMWJ can be computed from
average dry weather flow, Q,,y, using:
(40)
where PP is the contributing population in 1000's and a and b are determined
from analysis of flow measurements.
12.4.3 Multi-Segment Models
In considering a series of sewer pipes having low values of fluid
tractive shear, that is, characterized by low slopes or low flows (or both),
the condition can arise where solids from an upstream reach can successively
deposit in downstream pipes. The relative amounts deposited in any section
would depend on the shear stress during peak flow in that link and also on
the amounts deposited upstream. A general procedure is desired to predict
the total cumulative load in any section from all upstream sources.
241
-------
The procedure used is the following:
1. Segment the collection system into a network of "m" links
where each link may be a section of pipe between manholes
or several sections combined into a single section (similar
hydraulic characteristics);
2. Establish, for all links, a list of all downstream sections
that convey waste from the given link;
3. Compute cumulative upstream population at end of each link;
4. Compute average daily dry weather flow for each link using
the cumulative population from step 3 and an average per
capita waste rate;
5. Compute maximum daily dry-weather flow for each link using
equation 40 ;
6. Compute shear stress for each link associated with the maximum
daily flow, using equation 37 for the appropriate pipe shape;
7. Compute the dry-weather suspended solids deposition rates,
Zj(i = !,...,m) from the shear stresses calculated in step 6,
using equations 38 and 39 ;
8.
Compute the suspended solids load ZL-j (i = l,...m) developed
along each link using population per link length and daily
solids generated per capita;
9. Starting at the uppermost link, i, compute the amount of input
material that will deposit, that is Z-f x ZL^;
10. Search the list of downstream links for the deposition rate,
Zj, greater than the rate at the link where the load is
initially generated, and compute the amount deposited as the
jth link from the ith component input load using (Z.-Z.) x ZL.;
J ' *
11. Continue searching the list of downstream links for a deposition
rate Z|< greater than Zj and compute the deposition at the kth
link from the ith component using (Zk - Zj) x ZL^;
12. Set Z|< = Zj and repeat steps 10 and 11 until the complete list
of downstream links is completed;
13. Start with-the next uppermost link in the system and repeat
steps 9 through 12 while maintaining a running sum of all the
deposited loads in each link from previous iterations; and
14. Sequentially proceed downstream until all components are
completed.
242
-------
In other words, a fraction of the load generated in an upstream
section may deposit in that section (if the shear stress is sufficiently low)
and more of that load may deposit in downstream sections only if the shear
stress falls below levels experienced upstream.
The present model is coded to assume any collection system geometry
with the one rule that only three segments can be considered at a given
manhole. The model is coded to compute shear stress for circular,
ovoid, rectangular and horseshoe shaped cross-sections with or without
preset sediment~beds.
An idealized example using the schematic in Figure 71 illustrates
this procedure. Assume that the shear stress developed during peak dry-
weather flow in links 1, 4, 5, 8 and 11 results in. deposition rates of
10, 5, 5, 15, and 20 percent, respectively. Assume that the shear in all
other links, i.e., 2, 3, 6, 7, 9 and 10, is sufficiently high to preclude
any localized deposition. The dry weather load developed along each of the
11 links is, say, 100 units of dwf solids.
Table 63 shows contributions from all upstream links on each
downstream link and the total deposition in section 4 consists of loads from
TABLE 63. DEPOSITION ANALYSIS OF IDEALIZED SYSTEM*
10
Q)
.£>
3
C
•r-
1
2
3
4
5
6
7
8
9
10
11
1
10.
-
-
0.
0.
-
-
5.
-
-
5.
2
0.
0.
5.
0.
-
-
10.
-
-
5.
3
0.
5.
0.
-
-
10.
-
-
5.
Link
4 5
5.
0. 5.
-
-
10. 10. 1
-
-
5. 5.
Numbers
6 7 8 9 10 11
0.
0. 0.
5. 15. 15'.
- - - 0.
- - - 0. 0.
5. 5. 5. 20. 20. 20.
Total Amount
Deposited
Each Link
10
0.
0.
15.
5.
0.
0.
90.
0.
0.
100.
In
The ijth in the table represents the amount deposited in link i
that originated in link j. Thus element (8,1) = 5 represents the
amount deposited in link 8 that originated in link 1.
243
-------
o
0
12"
o
12"
L3J 12"
-o
15"
0
o
15"
12"
o-
r) 12
©
18"
o-
15'
-O
o
12"
-o
30"
TRUNK SEWER
FIGURE 71 SCHEMATIC OF COLLECTION SYSTEM
244
-------
sections 2, 3, 4 but not from section 1 because the deposition rate, Z] is
greater than 1%. At link 5, the only amount deposited is from the load
developed along that link (Z-] is greater than l§: no deposits, T.^ and Z3
less that Zg but 1^ equals Z$: no deposits; Z^ less that 2.$: deposits). The
overall deposition rate for the entire system is 20 percent (220 units
deposited/1100 units total load with nearly equal loadings in links 8 and 11.
Modification to the basic existing model (4) during this
project included the introduction of the following additional options:
a) use of fitted slopes for the pipe segments in addition to the pipe slopes
computed from manhole invert elevations; b) use of actual populations by
pipe segment, in addition to average uniform figures computed on the basis
of pipe length and population/100 ft of pipe, derived from global figures for
the basin; and c) use of Manning's variable roughness coefficient n in the
flow computations.
12.5 Verification of the Deposition Model
12.5.1 Introduction
The deposition model described in the previous section was developed
to: a) identify areas of extensive collection systems subject to high
degrees of deposition; b) indicate the relative degrees of deposition among
different parts of the system; and, c) provide an indication of the order of
magnitude of daily deposition throughout the system.
Comprehensive verification of such a model would involve applying
it to one or more relatively large collection systems and verifying, by
thorough inspection or more likely by a limited sampling, how well the model
performed in providing the three levels of information outlined in the
objectives mentioned above. It should be expected that the larger the
collection system the better should be the overall performance of the model.
The present study did not include in the work program field tasks
to allow a complete and formal verification of the model. The verification
effort described here focuses only on the analyses of deposition loadings
in the four test segments considered in the first phase field program. Both
qualitative (visual) and quantitative comparative analyses will be presented
for the four segments. Familiarity with the sewer segments at all four
streets and occasional- observations of pipe segments other than those being
flushed permitted pragmatic appraisal of the results given by the model,
especially in regard to the relative degrees of deposition throughout the
system. With regard to the numerical prediction of localized daily solids
mass deposition, the first phase field flushing results were used in
conjunction with estimates of the fractions removed by repeated flushing
from the second phase program to provide verification of the numerical values
given by the model. It should be recognized that such verification is very
limited in scope by the small size of the collection systems being analyzed.
12.5.2 Model Input Data
Physical System. All necessary information pertaining to the
245
-------
physical characteristics of the four collection systems upstream from the
sampling manholes were prepared from as-built maps and field verified.
Populatlpji. The model option using population estimates by
individual pipe segments was used in this analysis. Census tract information
described in Chapter 4 of this report were used to estimate the contributing
population to each pipe segment in the four streets.
Per Capita Liquid Waste Rates. Background flow measurements were
used to estimate per capita waste rates of 56 and 53 gpcd for Shepton and
Port Norfolk Streets, respectively. It is assumed that the waste rates at the
other two streets is of the same order, that is, around 60 gpcd, since the
four streets are fairly homogeneous in terms of population activity and
income level. In running the model, per capita values ranging from 60 to 200
gpcd were used.
Per Capita Solid Haste Rates. The deposition loads predicted by
the model are a linear function of the per capita solid waste rate. In
defining an estimated average solid waste rate, to be used in the deposition
model run, the averages of all measured TSS background concentrations at the
four sites were used in conjunction with a per capita contribution of 60
gpcd.* Table 64 presents the estimates of solids waste rates in lb/capita/
day for the four streets.
TABLE 64. ESTIMATES OF SOLIDS WASTE RATES IN LB/CAPITA/DAY
Street
Templeton
Shepton
Port Norfolk
Walnut
Population
221
230
94
71
Mean TSS
Background
Concentration
1112.
393.
1116.
1792.
Average for
Mean*
Background
Flow
(cfs)
0.0205
0.0214
0.0087
0.0066
all Streets
Mean
Solid Waste
Rate
(Ib/cap/day)
0.56
0.20
0.56
0.90
0.56
At 60 gpcd.
An average solids waste rate of 0.56 is estimated from field data collected
over the course of the project at the four test segments. A value of 0.5
Ib/cap/day was considered a reasonable figure to be used in the deposition
model runs at the four streets.
Peak to Average Flow Coefficients. The model estimates daily
deposition loads as a function of the tractive shear stress associated with
maximum daily dry weather flow. The peak coefficients used in the
verification runs derived from analysis of flow records covering a period
of one week at all four sites. The liquid level continuous records were
Summaries of all background flow and sewage strength data are presented in
Chapter 11.
246
-------
noted at time intervals of 20 minutes. A 20 minute peak to average daily
flow ratio was computed for each day of the week for all four sites and are
presented in Table 65. A 20 minute interval was chosen since it represented
the smallest time interval that could be read from the dipper charts. The average
peak flow coefficients appearing at the bottom of Table 65 were used in
establishing the value of the coefficient in equation 40 .
TABLE 65. RATIOS OF 20 MINUTE PEAK TO AVERAGE DAILY FLOWS
Day of the
Week
Sunday
Monday
Tuesday
Wednesday
Thursday
Fri day
Saturday
Average
Site
Tempi eton
1.72
1.54
2.02
1.73
1.62
1.72
1.70
1.72
Shepton
1.13
1.29
1.25
1.32
1.26
1.27
1.29
1.26
Port Norfolk
1.09
1.07
1.16
1.15
1.18
1.10
1.10
1.12
Walnut
1.23
1.57
1.59
1.21
1.41
1.42
1.72
1.45
12.5.3 Verification Results
. The deposition model used an arbitrary criteria to qualitatively
rank the degree of solids deposition in pipe segments which is as follows:
Degree of Deposition TSS Deposited Daily
None
Low
Moderate
High
0 - 2%
2 - 6%
6 - 15%
> 15%
The degree of deposition in the segments modelled for each street, using a
per capita waste flow rate of 60 gpcd, are shown below. The last segment for
each street shown below was the first phase test segment.
QUALITATIVE DEGREES OF DEPOSITION IN THE SEGMENTS MODELLED
., . (Per Capita Waste
Segment ^
No.* Tempi eton Shepton
1 none
2 none
3 moderate
4 high
5
low
moderate
low
moderate
moderate
Flow Rate =60
Port Norfolk
high
moderate
high-
gpcd)
Walnut
high
high
high
high
high
Upstream-downstream order; in addition,
refers to 'the flushing segment.
in each street, the last segment no.
247
-------
Observations of upstream segments as well as the test segments in each street
at various points in time indicated that the qualitative predictions of the
model could be judged with a fair degree of subjective confidence since the
predicted results were found to be in reasonably good agreement with the
visual observations of solids deposition in the four streets.*
Verification of the quantitative results given by the deposition
model was performed using the first phase field results presented in Chapter
8. Verification was done for all four flushing sites and consisted of the
following steps:
a. Estimation of the average daily mass of solids accumulated in
the single pjpe segment of each street. The computations considered in each
street the average mass removed by the first phase flushes, and fractional
estimates of the average flushed mass relative to the total average deposited
mass in each segment;
b. Use of first phase sediment scrapings taken prior to flushing;
c. Use of the deposition model described in Section 12.4 to
predict the daily accumulations of solids in those pipe segments; and
d. Comparison of the results derived from the average flushed
masses and the sediment scrapings with those predicted by the deposition
model.
The mean TSS mass removals normalized by antecedent days between
flushes from the first phase for good flushing events are as follows:
Street TSS Mass Removal (kg/day)
Tempieton 2.56
Shepton 1.29
Port Norfolk 1.30
Walnut 1.72
The above estimates represent mean values of mass transported out of the
respective pipe segments by a single flush and were reported in Chapter 8.
Experience from the second phase serial flushes in Port Norfolk Street
indicated that the mass flushed out of the first segment by the first flush
(of three) represented about 76 % of the total solids mass accumulated
in the pipe segment in the period between flushes. Similar estimates of the
fraction flushed out for the other three sites are also necessary in order
to compute the average daily accumulation rates for all four pipe segments.
In the absence of any additional primary information measured,a similar flush-
ing effectiveness rate is assumed for the other three streets.
"Jf
Previous verification of the model in the Boston area indicated that the
model predicted none or low deposition in 52 out 55 segments where visual
observation indicated no sedimentation; none or low deposition in 21 out 25
segments where visual observation indicated low deposition; low or moderate
deposition in 9 out 12 segments where visual observation indicated moderate
deposition; and moderate or high deposition in 18 out 33 segments where
visual observation indicated heavy deposition (36).
248
-------
The numerical predictions of the deposition model can only be
verified at the pipe segments flushed in the first phase, for which numerical
estimates of average deposition loads were computed. For the pipe segments
where the deposition predictions and field flushing results are being com-
pared, the optimized slopes developed for the looping stage/discharge curves
in Chapter 7 for each flush segment were used in the deposition analysis.
For the upstream segments the plan and profile map pipe slopes were used.
The first four columns in Table 66 present the results predicted'by the depo-
sition model for per capita waste flow rates of 60, 100, 150 and 200 gpcd.
The average first phase solids removals normalized by antecedent days between
flushes, kg/day for each test segment are presented under the next column,
labelled A. Two independent estimates of measured solids deposition rates,
kg/day, are presented under columns B and C. The estimate of daily deposi-
tion given under column B is computed using the phase one flushing removal
rates shown under column A and a flushing effectiveness level of 76%.* An
indirect estimate of daily deposition along the segment is presented under
column C using the measured sediment scrapings solids measurements taken over
a one-foot section of pipe to flushing in the first phase. It is assumed
that the unit deposition rates determined from the scraping operation are
applicable over the'entire segment. Sediment scraping information was not
collected for the Walnut Street test segment.
Comparison of the deposition model predictions for the per capita
waste rate of 60 gpcd with the estimates of daily accumulation from the
flushing results shows reasonable agreement with the exception of Port Nor-
folk Street. The deposition estimates derived from the sediment scraping
operation show closer agreement with Port Norfolk Street. In sum, the cali-
bration results indicate that the deposition model should be viewed as a
crude tool useful in providing rough cut estimates of dry weather solids
deposition.
Another measure of flushing efficiency can be derived by dividing the
flushed solids removals given under column A by the measured scraping
given under column C. The average effectiveness for the three streets
where scrapings were performed is 55%.
249
-------
TABLE 66. COMPARISON OF DEPOSITION MODEL PREDICTIONS OF DAILY TSS ACCUMULATION WITH ESTIMATES OF THE
FLUSHING EXPERIMENTS
ctreet Deposition Model Predictions (kg/day)
60 gpcd 100 gpcd 150 gpcd 200 gpcd ABC
Templeton 3.49 2.60 2.09 1.79 2.54 3.34 3.01
Shepton
1.69
1.28
1.02
0.87
1.31
1.73
2.71
en
0 Port Norfolk 3.33
2.54
2.01
1.71
1.30
1.71
3.80
Walnut
2.88
2.99
2.41
2.02
1.72
2.26
LEGEND
A - Average First Phase Flushing Results (kg/day)
B - Daily Accumulation from Flushing Results (kg/day)
C - Daily Accumulation from Measured Scraping Results (kg/day)
-------
SECTION 13
DEVELOPMENT OF GENERALIZED PREDICTIVE DEPOSITION MODELS
13.1 Introduction
Deposition of sewage solids during dry weather in combined sewer
systems has long been recognized as a major contributor to "first-flush"
phenomena occurring during wet weather runoff periods. Estimation of these
loadings for a given sewer system is an extremely difficult task. Measure-
ment for extended periods is possible but extremely expensive. Some
literature information is available from experiments on build-up of sanitary
sewage solids in a pilot sewer study conducted by the FMC Corporation(ll).
Techniques presently available to estimate dry weather deposition in sewer-
age systems involve the use of computerized mathematical models, that are
both complex and expensive and requiring more effort than appropriate for
preliminary "first-cut" assessments.(4 , 28).
The objective of the analysis presented in this Chapter is to
provide planners, engineers and municipal managers with readily obtainable
technical information so that they can make intelligent informed decisions
on potential sewer flushing programs. In this Chapter a set of generalized
procedures for estimating pollutant loadings associated with dry weather
sewage solids deposition in combined sewer systems is presented. A complete
exposition of this analysis has been described in a planning document (37)
prepared earlier in this study. A summary of that analysis .is presented in
this chapter.
The predictive equations relate the total daily mass of pollutant
deposition accumulations within a collection system to physical characteris-
tics of collection systems such as per capita waste rate, service area,
total pipe length, average pipe slope, average diameter and other more
complicated parameters that derive from analysis of pipe slope characteris-
tics. Several alternative predictive models are presented reflecting
anticipated differences in the availability of data and user resources.
Pollutant parameters include TSS, VSS, BOD, COD, TKN and TP. Sewer system
age and degree of maintenance was also considered. Factors are presented
for estimating the increase in collection system deposition resulting from
improper maintenance. A users' guide has been presented to establish the
necessary data input to utilize the predictive .procedures.
251
-------
13.1.1 Foreword
An executive overview of the methodology used to develop the
predictive simplified deposition predictive models is presented in section
13.2. A more detailed elaboration of the approach is given in section 13.3.
An overview of the design of the numerical experiment is presented in section
13.4. Data input preparation for the numerical regression analyses are given
in section 13.5. Summary results of the regression analysis and alternative
model selections are presented in section 13.6. A user's application guide
is given in section 13.7. Finally, an analysis demonstrating the numerical
predictive sensitivity of the various models and approaches is presented in
section 13.8.
13.1.2 Data and Information Sources
The data and information for this analysis were derived
principally from three data sources: (a) sewer atlas physical data for
portions of West Roxbury, Dedham, Newton and Brook!ine, Massachusetts for an
infiltration/inflow study conducted by Environmental Design & Planning, Inc.
for the Metropolitan District Commission(38);(b)sewer atlas physical data for
portions of the City of Fitchburg, Massachusetts for a section of 208 combined
sewer management study conducted by Environmental Design & Planning, Inc. for
the Montachusetts Regional Planning Commission (5); and (c) sewer.atlas
physical data for portions of Dorchester and South Boston for a combined
sewer management study sponsored by the Metropolitan District Commission (4).
13.2 Executive Overview of Methodology
An empirical model relating pollutant deposition loadings to
collection system characteristics is the goal of this study. The approach is
to use least squares to fit parameters of a postulated model. The data base
used in the fitting process consists, in part, of a number of collection
system parameters developed from an extensive data analysis of the physical
details or several major sewerage collection systems in eastern Massachusetts.
These characteristics are some of the independent variables -.used in the
analysis. The data for the dependent variables are the total daily sewage
solids deposited in these collection systems for a wide variety of different
operating conditions. These quantities are estimated using an existing
exogenous model that uses extremely detailed information to compute deposition
loadings throughout an entire collection system network. An analysis of the
detailed outputs of this model together with some of the physical data of
the collection systems 'provided the remaining independent variables in the
data base. Simply stated, the dependent variable data was generated from an
exogenous predictive analysis while the independent variable data was obtain-
ed from primary collection system data and from a secondary analysis of the
exogenous simulation outputs with selected collection system data.
Results of the field flushing programs have been earlier describ-
ed in Chapters 8 and 9. Methodological details of an existing exogenous
deposition model that predicts solids deposition in all segments of an entire
sewerage collection system have been discussed in Chapter 12. In addition,
calibration efforts using the field flushing results and the aforementioned
252
-------
model were described in Chapter 12. Those results were given to justify the
application of that model to produce simulated data for the purposes of
analysis described in this Chapter.
13.3 General Methodology-Detailed Overview
The general methodology used in the study is outlined in
Figure 72. The first step is to define the general characteristics and
parameters of the conceptual model. This discussion is presented in section
13.3.1. Next, a series of experiments is designed to generate deposition
loadings using the deposition model described in Chapter 12 for a wide
range of conditions likely to be encountered in practice. The regression
equations would be valid for use over these ranges of conditions. The
design of the experiments, described in Section 13.4 consists of defining
the study areas to be used in the experiments and the hydraulic conditions
under which the numerical experiments would be performed. The suspended
solids per capita waste rate is discussed in the design of the experiments.
The next step involved the collection of all pertinent physical
data associated with the selected collection systems, such as system
configuration, pipe lengths, shapes and sizes, invert elevations, so that the
deposition model referred to in Chapter 12 could be used. This physical
data, together with the deposition model and the total loads deposited
simulated for each of the collection systems is described in section 13.5.1.
Analyses of the collection system service areas and total pipe
lengths, distributions of pipe slopes and average collection system pipe
diameters are presented in sections 13.5.2 to 13.5.4, respectively.
Deposition model results included the total Ib/day deposited in each pipe
segment of each basin and the total loads accumulated throughout each
system. Information on loads by pipe segment is then used to generate curves
for each basin, showing the accumulated percentages of the loads deposited
against the accumulated percentage of pipe lengths where deposition took
place. . This part of the work is described in section 13.5.5.
Physical data of the system together with the distribution of
loads by pipe length are then used to define the derived variables LpD and
SPD/4 wnic'1 are described in sections 13.5.6 to 13.5.8, respectively. Total
loans by basin generated by the deposition model together with primary
variables (pipe length, area, average slope, average diameter) and the
derived variables (LPD, SPQ, SDQ/A) are then used as input for the regression
analysis described in section 13.o. Results of the regression analysis were
examined and considered satisfactory and the process was complete.
13.3.1 Discussion of Model Variables
A discussion of the independent variables considered in the model
and a few descriptive details of the preliminary analyses preceding the
selection of the complete list of variables is given in this section.
The abvious and simplest of variables that can be used to
characterize a collection system are the total service area, total pipe
253
-------
ro
en
RED
VAR
FFINFH "DFRIVFD" <£--.,„ NO
IABLES
V
DATA MASSAGE TO DEFINE
ADDITIONAL "DERIVED" VARIABLES
j START j
V
CONCEPTUAL
MODEL
V
DESIGN OF EXPERIMENTS:
-DEFINE STUDY AREA
-DEFINE SURROGATE PER
CAPITA WASTE CONTRIBU-
TIONS TO COVER A WIDE
RANGE OF THE FLOW
CONDITIONS
-DEFINE PER CAPITA
SOLID WASTE CONTRIBU-
TION
A A
DEFINE % LOAD VS.
PIPE LENGTH
CURVES
A
\
ASSEMBLE . DEPOSITION
COLLECTION ANALYSIS
SYSTEM > ^
PHYSICAL
DATA
1
[ FINISH
AYES
SATISFACTORY |
A
ANALYSIS OF
RESULTS
A
PREDICTIVE
EQUATIONS
A
f
REGRESSION
ANALYSIS
A
FIGURE 72
GENERAL METHODOLOGY OF THE STUDY
-------
length, average slope and the average pipe diameter. It was believed from
the onset of this study that these variables alone would not be adequate to
explain the variability of the estimated loads from the deposition model.
Clearly, a better characterization of the collection systems was necessary.
An exploratory analysis applying the deposition model on a
number of sample collection systems revealed an interesting insight. Plots
of the cumulative percentages of total loads deposited in each basin versus
cumulative pipe lengths were prepared. A number of these curves can be
inspected from Figures 80 and 81 presented in Section 13.5.5. The curves
spread around the range of 70% to 90% of the total mass deposited suggested
the use of the pipe length corresponding to 80% of the total mass deposited
as a potential variable to include in the regression analysis.
Another set of plots of the cumulative distribution of pipe
slopes for a few basins also suggested that the mean pipe slope alone would
not be adequate to explain the effects of the pipe slopes on the variations
of the deposition loads. A better characterization of the collection system
pipe slopes could be obtained by defining various parameters at the flatter
pipe slope range. Three other pipe slope parameters -besides the mean pipe
slope were initially selected for inclusion into the regression model. These
parameters are as follows:
a) the pipe slope corresponding* to the percentage of the pipe
length where 80% of the total load of the collection system
deposits (SpD);
b) the average of the slopes in the basin below SpD (SpD) ;
c) the slope corresponding to some fraction of SpD, arbitrarily
taken as the slope corresponding to 1/4 the percentage of
pipe lengths below which 80% of the total mass deposits
(spo/4).
These slope parameters can be seen in Figure 73 .
Further analysis revealed that Spg and Spp were very strongly
correlated, so that retaining both in the regression_analysis was not necess-
ary. This finding was fortunate since the variable Spo is much more difficult
to determine than Spp. The variable SpQ was excluded from the analysis.
Finally, it is clear that the deposition process is also
strongly affected by the sewage flows in the system. Variations in popula-
tion density and the degree of infiltration affects the dry weather flow
rates. These effects are incorporated into the per capita waste rates used
in the deposition model simulations and in the regression analysis. The
summary list of variables considered in the regression analysis is the
following:
Note that the correspondence indicated in Figure 73 does not necessarily
imply that the pipe length over which 80% of the load deposits has slope
smaller than or equal to SpD at all segments.
255
-------
\
en
100-
a.
jf
a.
CL
50-
50 80 IO
% Suspended Solids Deposited, TS
FIGURE 73 COLLECTION SYSTEM PIPE SLOPE VARIABLES
-------
1. Total collection system pipe length (\L) - ft;
2. Service area of collection system (A) - acres;
3. Average collection system pipe slope (S) - ft/ft;
4. Average collection system pipe diameter (D) - inches;
5. Length of pipe corresponding to 80% of the solids deposited
in the system (LpD) - ft;
6. Slope corresponding to LpD(SpD) - ft/ft;
7. Slope corresponding to 1/4 of the percentage of pipe
length (PLn) below which 80% of the solids deposit
(SpD/4)- ft/ft; and
8. Flow rate per capita, including allowance for infiltration
(q) - gpcd.
With respect to the mathematical forms of the regression model
both linear and alternative non-linear models were initially postulated.
Non-linear fitting techniques were not needed in the analysis since the
linear models, that is, the strictly additive form and the logarithmic
multiplicative form converted in the log domain, resulted in excellent
fitting results with the R2 approaching 95%.
13.4 Design of Experiment
In this section an overview will be presented of how collection
system data from three major sewerage systems was used to design the data
base for the multivariate regression experiment. A description of the
three sewer systems whose data were assumed to represent an adequate sample
from the universe of all collection system is presented in section 13.4.1.
A discussion of the per capita flow waste rates used in the experiment is
presented in section 13.4.2. These surrogate waste rates reflect wide
variations in population density and infiltration conditions encountered in
practice. This parameter can be considered as a decision variable from a
planning standpoint. -Various sewer system age and maintenance considerations
are discussed in section 13.4.3.
13.4.1 Description of Three Sewer Systems
The physical characteristics of the three major collection
systems used in this analysis derived from three prior studies. The first
area, covering portions of West Roxbury in Boston, Dedham, Newton and
Brook!ine is strictly separated. The second area covering major portions of
Dorchester and South Boston, two neighborhoods of the Boston metropolitan
area is a mixed combined and separate area while the third basin covering
a portion of the City of Fitchburg is served by a combined sewer system. The
total pipe length, service area and pipe density for each basin are given in
Table 67. The total pipe footage for all three areas entails 196 miles
257
-------
(315.5 km) of separate and combined sewer systems encompassing a total area
of 8.9 square miles (204.5 square km).
TABLE 67. SEWER DENSITY (mi/acre)
Overall Pipe
System Length (mi)
WRNDB**
35 basins
Dorchester
(37 basins)
Fitchburg
(3 basins)
64.87
119.85
11.17
*
Weighted averages: by pipe
"kit
Sewerage system with area
Newton and Brookline in Bo
Pipe
Area Density
(acres) (mi/acre)*
2464. 0.026
2753. 0.044
485. 0.023
length: 0.036, by area: 0.034.
covering portions of West Roxbury, Dedham,
ston metropolitan area.
The land use in the first area in West Roxbury and neighboring
communities is mostly moderate to high density single and two family
dwellings with a population density ranging from 10 to 15 people/acre. The
topography is mild with several hilly portions in the area. This area was
investigated in a recent infiltration/inflow study and was subdivided into
35 distinct sewer collection subsystems.
The land use in Dorchester and South Boston is mostly high
density multi-family dwellings with population density ranging from 30 to 60
people/acre. The topography in Dorchester is moderate with a number of hilly
sections while portions of South Boston are fairly flat. There are a total
of 37 distinct sewer collection systems in this study area.
The land use in the third area in Fitchburg is mixed commercial
and high density multi-family dwellings with a small portion of-single
family homes. The population density is similar to Dorchester. The study
area is subdivided into three collection systems.
A total of 75 different sewer collection systems form the data
basis for the analysis. It is assumed that these basins collectively
represent a wide variety of different pipe slope conditions, pipe sizes and
shapes and network system configurations. Some basins serve narrow strips
of land while others are broad fanned-shape with a high hierarchial network
order. A central assumption is that the collection system characteristics
represented by the sample of 75 sewer sheds is an adequate representation of
the total universe of collection systems. This assumption is not completely
valid since, for example, extremely flat collection systems were not part of
the sample set. Future work should broaden this data base. This sample
258
-------
however is deemed reasonably complete for the purposes of this analysis.
A complete sewer atlas of manhole to manhole descriptive
physical data including pipe length, slope, shape, size and network ordering
designations was available for each of these systems. Much of this data had
been previously processed for computer application although a considerable
portion of the data had to be placed in EDP format for purpose of this study.
Roughly 6000 manhole to manhole segments incorporating all of the aforement-
ioned parameters were necessary to represent the hydraulic characterization
of the 75 sewer collection systems.
13.4.2 Range of Flows
The degree of deposition in a sewer pipe is strongly dependent
on the discharge. As flow increases through a pipe the depth-, velocity,
hydraulic radius all change resulting in higher shear stress with less
deposition. Discharge therefore is an extremely important parameter in the
analysis. The dry weather discharge in a sewer system is dependent upon the
local population density, the domestic per capita contribution, the degree of
infiltration and any industrial waste contributions.
It was envisioned that a single per capita surrogate waste
rate would be generated incorporating a wide range of population density
and infiltration conditions encountered in practice.* This variable would
embed all these variations and be used in both the deposition model to
predict daily dry weather solids deposition and in 'the regression model as
an independent variable.
Population densities ranging from 15 people/acre up to 90
people/acre were considered. Using a factor of 0.035 miles of sewer per
acre the corresponding number of people under 100 feet of sewer pipe was
computed. These factors are shown in Table 68 and are used in the
deposition model which requires as input the number of people per 100 feet
of sewer.
The dry weather per capita contribution of 85 gpcd was consider-
ed fixed in this analysis. Four different infiltration estimates of 500,
1000, 2000 and 4000 gallons per acre per day were used to cover the range
of normally encountered infiltration conditions. The adjusted per capita
waste rates incorporating the various rates of infiltration for the range
of population densities considered in the analysis are shown in Table 69
These per capita values are again adjusted to the mid-range of population
density of 45 people/acre and are given in Table 70. This last conversion
permits considering one single range of surrogate per capita flow rates
using 45 people/acre as the norm. Four different flow rates are considered
in the analysis and cover the full range of per capita waste rates for
different population densities and infiltration conditions. The per capita
waste rates used in the analysis are: 40, 110, 190 and 260 gpcd.
Industrial waste contributions were not explicitly considered. The user can
readjust the per capita waste rates used in this analysis to reflect indust-
rial contributions.
259
-------
TABLE 68. POPULATION DENSITY (PERSONS/TOO FT OF PIPE)
Density (person/acre) Persons/100 .ft of pipe*
15 8.15
30 16.30
45 24.46
60 32.61
90 ' 48.91
*
Assumes 0.035 mi/acre or 184.8 ft of pipe/acre
TABLE 69. PER CAPITA WASTE RATES FOR VARIOUS POPULATION
DENSITIES AND INFILTRATION RATES*
Density Infil
(person/acre) 500
15
30
45
60
90
*
Assumes
**
Gallons
118.3
101.7
96.1
93.3
90.6
tration Rate (3£ad>**
1000
151.7
118.3
107.2
101.7
96.1
a dry weather contribution
per acre per day.
2000
218.3
151.7
129.4
118.3
107.1
of 85 gpcd.
400C
351.6
218.3
173.9
151.7
129.4
TABLE 70. PER CAPITA VALUES RELATIVE TO
THE DENSITY OF 45 PERSONS/ACRE
Density
(person/acre) 500 1000 2000 4000
*
**
15
30
45
60
90
Minimum value.
Maximum value.
39.
67.
96.
124.
179.
1*
3
1
4
7
50
78
107
135
190
.1
.2
.2
.6
.6
72
100
129
157
212
.1
.3
.4
.7
.6
116.
144.
173.
202.
256.
2 I
3
9
3
3**
260
-------
The per capita solids waste rate of 0.5 Ib/capita/day was used in
all computations. This parameter derives from analysis of background sample
and flow monitoring results and was established in Chapter 12. All regression
results presented in section 13.6 can be linearly scaled for any other
desired per capita solids waste rate.
13.4.3 Age and Maintenance Conditions
The presence of long-term accumulations of organic matter, sand,
gravel, grit and debris in the form of sediment beds, shoals, or bars can
significantly alter the hydraulic characteristics and accordingly the degree
of deposition, particularly for lateral pipes with little dry weather
discharge. These accumulations can easily result in new well-constructed
sewer systems with sound joints and few hydraulic obstructions such as
protruding house connections. Similar deposits can occur in systems
that are rodded and frequently cleaned but either are old and/or have poor
joints and many hydraulic obstructions. Perforated manhole lids provide the
perfect opportunity for children to jam sticks into manholes that can result
in massive blockages of accumulated rags and toilet paper. The above
conditions are but a few of the possible age and maintenance problems
encountered in practice.
Three different categories of sewer system age and maintenance
were considered in this analysis. The first category of clean pipe conditions
represents good maintenance practices and well-constructed sewer systems. No
sediment beds were considered in this case.
Two cases simulating different degrees of maintenance other than
perfect clean pipe conditions were also considered. In the first case or the
intermediate maintenance category, sediment beds ranging from 1 to 3 inches
in depth were assumed for all pipes with slopes less than 0.0075. Figure 74
shows the assumed ranges of beds between pipe slopes of 0.0005 and 0.0075.
In the third category, the zero maintenance care, the sediment beds range
from 3 to 6 inches for the same range of pipe slopes. This range was
established using judgment and also based on visual inspection of numerous
combined sewer pipes in eastern Massachusetts combined sewer systems.
These three conditions were used in the deposition model analysis
to compute daily collection system deposition loadings.
13.5 Data Preparation for the Regression Analysis
Descriptions of the precursory analyses necessary for generating
the regression analyses input data are presented in this section. Details
of the analysis for generating daily estimates of total sewerage system
deposition loadings for each of the 75 collection systems are presented in
section 13.5.1. Descriptions of collection system service areas and
corresponding sewerage system pipe lengths are given in section 13.5.2. An
analysis of statistical distributions of collection system pipe slopes is
discussed in section 13.5.3. Methods for computing average pipe diameters
per collection system are given in section 13.5.4. Results of predicted
deposition loadings for the 75 collection systems considered in section 13.5.1
261
-------
g
£
o
o
W
cc
A
CASE 2 - INTERMEDIATE MAINTENANCE (A=3INCH,B=1 INCH)
CASE 3 - POOR MAINTENANCE (A=6 INCH, B=3 INCH)
.0005 .0075
COLLECTION SYSTEM PIPE SLOPE
FIGURE 74 REPRESENTATION OF SEDIMENT BEDS
AND PIPE SLOPE FOR TWO AGE AND
MAINTENANCE CONDITIONS
. 262
-------
are used to determine distributions of solids deposition as a function of
cumulative pipe length. This analysis is described in section 13.5.5.
Discussions of three other parameters, LpQ, SpQ, and Sp^ which derive from
the analysis described in section 13.5.5, are presented in sections 13.5.6
through 13.5.8, respectively.
13.5.1 Deposition Model Results
Description of the Deposition Model. The deposition model used
in this analysis to generate estimates of daily solids deposition for each
of the 75 collection systems is described in detail in section .12.4 of
Chapter 12. The model considers peak daily dry weather flow and uses a shear
stress critera to determine the limiting diameter of the solid particles
that deposit at each segment. Then, with this limiting particle size and
assuming a given distribution for the particle sizes present in the dry
weather sanitary flow, the model determines the percentage of the suspended
solids that deposit at each pipe segment. The model also has mechanisms to
account for the fact that particles of diameters up to a given size that
deposit in a given pipe segment are not available for deposition in down-
stream segments.
Some of the results given by the model are: (1) the flow
conditions at each pipe segment, including discharge, average velocity, water
depth and shear stress; (2) the loads (Ib/day) deposited at each pipe segment
and (3) the accumulated value,of the loads deposited in all upstream segments.
Input Data Required by the Deposition Model. The input data
required by the deposition model consists of:
- Segment identification (by a segment number);
- Segment upstream and downstream pipe inverts;
- Segment length;
- Pipe shape (10 shapes possible);
- Pipe sizes (diameter or height and width);
- Segment type (zero, one or two tributary segments);
- Network location designation (defined by the segment type
in conjunction with the next downstream segment number);
- Sediment depth in the segment;
- Population per 100 feet of pipe;
- Average daily waste flow contribution in gpcd;
- Peak daily to average flow peaking coefficients;
- Manning's resistance coefficient, n, and its variability with
flow depth; and
- Total solids contribution in Ib/capita/day.
Deposition Input Data Preparation. In section 13.4.1, the
description of the three different sewerage systems considered in this study
was presented. A total of 75 subsystems were used in this analysis
including 35 separate collection systems from the WRNDB sewer system, 37
collection systems from the Dorchester and South Boston combined sewer
systems; and 3 collection systems from the Fitchburg combined sewer system.
263
-------
All the necessary physical data in the form of computer cards
were available for the Dorchester and Fitchburg system from prior studies.
For the WRNDB system all the pipe elevations, lengths, shapes and sizes
were also available from a previous study, but all the segment numbering and
the additional information required to establish the system configuration had
to be generated in this study.
Other information on waste flow rates and solid matter contribu-
tion to the systems, necessary to run the model, were given in section 13.4.2.
Deposition Model Runs and Results. Three sets of runs were
performed for all 75 basins. The first set of runs were performed assuming
no previous sediment deposits present in the pipes, that is clean pipe
conditions. The second set of runs were performed in which sediment depths
ranging from 1 to 3 inches were assumed to represent moderate maintenance
conditions. The third set of runs were performed assuming sediment depths
from-3 to 6 inches intended to simulate poor maintenance.
Selected information from these runs were punched out on cards
for use in future phases of the study. The values of the loads (Ib/day)
deposited by pipe segment were used to define for each basin the accumulated
percentages of the total load versus the accumulated percentages of total
pipe lengths where deposition occurs. An overall curve for all 75 basins
was also prepared. These curves were useful in deriving several variables
used in the regression analysis. Representative curves for individual
collection systems and the overall curve are presented in section 13.5.5.
The total loads per basin were used as the observed values of the dependent
variable in the regression analysis.
13.5.2 Areas and Total Pipe Lengths
The total service area and total pipe lengths were known from
prior studies for all 75 basins and are presented in Table 71. The first
35 basins cover portions of the WRNDB sewerage system. Basins 36 through 72
cover the Dorchester and South Boston sewerage system while the last three
basins cover portions of the City of Fitchburg sewerage system. The data
on Table 71 were also used for a simple regression of total pipe length on
total area and is described in section 13.7. This regression may be useful
in extreme cases where the total pipe length is not known or cannot be
immediately determined.
13.5.3 Distribution of Pipe Slopes
The regression model proposed in section 13.4.1 included several
collection system pipe slope parameters, SPQ and Spp/4, that required
computation of the cumulative pipe slope distributions. A computer program
was prepared to compute these distributions from data on the pipe segments
upstream and downstream invert elevations and segment lengths. The program
computed the slope distribution weighing the segment slopes by their
lengths. The mean, standard deviation, coefficient of variation, coefficient
of skewness and coefficient of kurtosis of pipe slopes per collection system
were also computed. The program computed the distribution and the afore-
264
-------
TABLE 71. TOTAL PIPE LENGTHS AND AREAS OF THE BASINS
Basin No
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Pipe Length
(ft)
47180.
5945.
610.
669.
3309.
360.
1900.
2251.
650.
1160.
2158.
1410.
15610.
1551.
990.
1305.
71621.
3279.
14415.
263.
489.
1146.
4180.
27385.
12653.
3331.
14016.
738.
14997.
14540.
1374.
16988.
17131..
7728.
26981.
6245.
7735.
1750.
Area
(Acre)
230.
38.
5.
7.
19.
5.
9.
13.
3.
8.
6.
13.
84.
70.
70.
6.
641.
58.
100.
6.
8.
4.
6.
173.
90.
19.
82.
6.
120.
140.
6.
96.
100.
45.
178.
42.
33.
29.
Basin No
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
Pipe Length
(ft)
4265.
3485.
3400.
3170.
4000.
5060.
11325.
13133.
7757.
7764.
112638.
875.
5200.
4501.
11890.
5830.
2235.
7145.
5276.
3115.
10585.
8741.
35501 .
13051
5635.
11325.
33005.
9899.
5644.
14220.
4492.
35033.
134528.
69274.
31748.
12092.
14754.
Area
(Acre)
19.
17.
26.
16.
25.
25.
52.
44.
24.
42.
245.
5.
36.
27.
51.
28.
15.
34.
91.
9.
120.
65.
228.
78.
26.
54.
177.
47.
30.
57.
24.
233.
315.
360.
264.
78.
143.
265
-------
mentioned statistics for each system (WRNDB, Dorchester and Fitchburg) and
finally an overall distribution and the first four moments of all data
lumped into one data .set.
Plots of the slope distributions for a few basins are shown in
Figures 75 and 76. The concave shapes of those cumulative distributions
(CDF) without a point of inflexion, suggest an exponential distribution for
the pipe slopes. Several of the cumulative distributions were plotted on
normal, log normal and Gumbel's probability paper. .All plotted curves
resulted in very non-linear shapes, indicating that the pipe slopes do not
follow any of those distributions. Plots of the complementary CDF of the
pipe slopes on semi-logarithmic paper, nonetheless, resulted in remarkably
linear shapes shown in the illustrative cases in Figures 77 and 78.
Although no formal numerical test of goodness-of-fit was performed, this
fact indicates that, at least for the sample data used in this study, the
distribution of the pipe slopes is exponential.
The solid lines drawn on Figures 77 and 78 were plotted using
the expression of the exponential cumulative distribution function given by:
Fs = 1 - e-s/S (.41}
where F~ = P(s < S) (cumulative pipe slope distribution);
s = any given slope;
S = the mean slope computed for the basin, as indicated above
and
e = the base of the natural logarithms.
Figure 79. presents the histograms for the WRNDB, Dorchester and
Fitchburg sewerage systems and the overall histograms considering all data.
The slope values corresponding to the intervals in Figure 79 are given in
Table 72.
Two observations can be noted from these histograms:
a) the shapes of the histograms for all 3 systems are similar
with minor differences between them (the same is true for
the global histogram compared to any of the other three); and
b) they all indicate an exponentially decaying shape, character-
istics of the exponential distribution with the CDF given by
equation 1 .
A final justification for the assumption that the pipe slopes
are exponentially distributed is the following. The mean and standard
deviation of the coefficients of variation of the slope for all 75 basins
are 0.87 and 0.25, respectively. The mean value of 0.87 for the coefficients
of variation very closely approximates the theoretical value of 1.0,
266
-------
01
zi
a
CO
>
5
2
o
1.0-
.9-
.8-
.7-
.6-
.5-
.4-
,3~
.2-
.h
.01
INDIVIDUAL COLLECTION SYSTEM DESIGNATION NUMBERS
NOTING BASINS IN DORCHESTER & SOUTH BOSTON SEWERAGE SYSTEMS
COLLECTION SYSTEM PIPE SLOPE s, x 102
I.62
2.90
4.I8
5.47
6.75
8.04
FIGURE 75 CUMULATIVE DISTRIBUTION OF PIPE SLOPES PER
COLLECTION SYSTEM
9.0
-------
ffl
I
Q.
o
1.0-
•9-
B-
>r
£-
.5-
4-
.3-
.2-
INDIVIDUAL COLLECTION SYSTFM DESIGNATION NUMBERS
/NOTING BASINS IN THE WRNDB SEWERAGE SYSTEM
COLLECTION SYSTEM PIPE SLOPE s, x 10
.01
FIGURE 76
I.62
2.90
4J8
5.47
6.75
8.04
CUMULATIVE DISTRIBUTION OF PIPE SLOPES PER
COLLECTION SYSTEM
9.0
-------
1.0-
.9-
.8-
.7-
.6-
.5-
.4-
.3-
.2-
a>
U.
UJ
CL
O
UJ
Q.
U.
O
O
z .08-
LJ
5 -07-
I .06-
o
o
.05-
.04-
.03-
.oa-
.01'
BASIN 49 - DORCHESTER
O.O .012943 .028996 J04505 .061104 .077157 D9
COLLECTION SYSTEM PIPE SLOPE,s
FIGURE 77 COMPLENTARY DISTRIBUTION OF PIPE SLOPES= Gs= I'Fs
269
-------
1.0-
.9-
.8-
.7-
.6-
.5-
.4-
.3-
1 .2-
ii
w
to
CO
LU
Q.
O
CO
LU
I JH
.09-
.08-
.07-
.06-
.05-
u.
Q
CJ
cc
LU
2
til
a.
o .0.4-
o
.02-
.01'
BASIN 73 - FITCHBURG
0.0 .001384 .O0299 .04505 .061104 J077I57
COLLECTION SYSTEM PIPE SLOPE,*
JOS
FIGURE 78 COMPLEMENTARY DISTRIBUTION OF PIPE SLOPES: Gs= I'
270
-------
.300-1
ro
.zooH
.100-
WRNDB(MDC)
DORCHESTER
FITCHBURG
OVERALL
f I 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
PIPE SLOPE INTERVAL NUMBERS
FIGURE 79 HISTOGRAM OF COLLECTION SYSTEM PIPE SLOPES
-------
TABLE 72. SLOPES CORRESPONDING TO THE INTERVALS
SHOWN ON FIGURE 79
Interval
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Slope Limit
of Range
0.000100
0.003311
0.006521
0.009732
0.012943
0.016154
0.019364
0.022575
0.025786
0.028996
0.032207
0.035418
0.038629
0.041839
0.045050
0.048261
0.051471
0.054682
0.057893
0.061104
0.064314
0.067525
0.070736
0.073946
0.077157
0.080368
0.083578
0.086789
0.090000
> 0.090000
characteristic of the exponential distribution.
This finding has significant importance in the practical
application of the regression model developed in this study. This is
especially true for cases where detailed statistical analysis of pipe slopes
information is not available and only the mean value of the pipe slope is
required to completely define with reasonable accuracy the pipe slope
distribution.
The special slope parameters Spp and Spp/^ can then be estimated
through this pipe approximation. Definition of the pipe slopes distribution
using the exponential model is also important in delineating the extent and
geographic location of the deposition loads in the system. This topic will
•
272
-------
be further discussed in section 13.7.2*
13.5.4 Average Pipe Diameter
The average pipe diameter of all segments within each collection
system was computed by weighing circular diameters of each segment by its
corresponding lengths. An equivalent circular pipe diameter was first
determined for non-circular sections before the weighted average was computed.
The fraction of non-circular pipes was a small percentage of the total pipe
length. All pipes are circular in the WRNDB system. The number of non-
circular pipes represent about 5% of the total for the Dorchester system,
whereas in Fitchburg they represent less than 1% of the total pipe lengths.
The formulas used for the equivalent circular diameter for the
non-linear sections are presented in Table 73 and have been derived to yield
roughly the same hydraulic radius at low depths of flow. For rectangular and
U-shaped pipes a simple equivalence of total areas are indicated in Table 73
These forms are non-existent in the data for this study. The average pipe
diameters determined for all 75 basins are presented in Table 74 .
TABLE 73. FORMULAS FOR EQUIVALENT CIRCULAR DIAMETERS
USED IN COMPUTING THE BASIN AVERAGE DIAMETER
Pipe,
Shape
Circular
Rectangular
Ellipse
Egg
Horseshoe
Oval
Ovoid
Modified Circle
Arch
U
Dimensions
D
X,Y
X,Y
X,Y
X,Y
X,Y
X,Y
X,Y
X,Y
X,Y
Equivalent
Circular D
D
(1.273 XY)1/2
(X Y)1/2
5.51X5'38/(X+Y)4'39
(X+Y)/2.0
0.67 X
5.51X5'38/(X+Y)4'39
(X Y)1/2
Y
(1.273 X Y)1/2
It should be stressed here that in the generation of the slope data for the
regression analysis described in section 13.6,the assumption that the pipe
slopes are exponentially distributed was not used. The probability
distributions of slopes for all basins were determined from their respective
slope data.
273
-------
TABLE 74. SUMMARY DATA ON DERIVED LENGTHS, SLOPES AND PIPE DIAMETERS
.BASIN HO.
1
2
1
4
S
6
7
6
9
10
11
12
1?
14
1 5
If
17
IP
I'l
ZO
21
Z.2
23
2.4
25
26
2.7
is
23
30
3!
32
33
34
35
Jfa
37
38
39
40.
41
•V2
43
44
45
46
47
4^
49
50
51
5jt
53
54
55
56
5?
58
59
60
61
62
63
64
65
66
67
68
69
70
71
?2
73
74
^f
SOT PEpns.
LENGTH) pT|
PI.Q
15050.
2080.
43'..
450.
1502.
130.
741.
1019.
36.
560.
304.
685.
6056.
634.
405.
715.
34879.
1983.
593P.
83.
171.
648.
1S3$>.
9228.
4517.
1370.
5914,
4BG.
929.
6005.
5Z6.
79A7.
6903.
3956.
10954.
2591.
1461.
936.
1100.
1913.
IV5.
786.
2540.
16«5.
4«£°.
5279,
2893.
2647.
52.9.3.
571.
10, I
13.1
14.3
12.7
12. Z
12.6
12.!
12,5
12.9
13.1
14.3
12.7
12.0
15.6
13.3
12.4
12.4
10.7
13.6
«S«|
0.02?90l
0.036746
0.046065
0.009US
0.02196^
0. 008611
C. 0141 57
0.0152 OH
0.006400
0.017163
0.012293
0.013985
0.013979
0,019666
0.030103
0.0045.78
0. 014539
0.009127
0,0 1 73 ',4
0.01785?
0.026175
0,019257
O.D21631
0.027730
1.; 119560
0.016939
0.016298
O.3f)4415
0.025510
0.013617
0*016441
f, 016991
0. 015655
0.(« 1 73 55
O.OIH20'
0.014155
0.0346&9
0.005660
0.035361
0.005365
0.027941
0,037994
0.002575
0.029394
0.006341
0.0*4465
0.011777
0.025934
0-022566
Q.Q38857
0.035336
0.012463
.",015249
0.011539
0.005995
0.041551
0.011366
0.0 10? 69
0.'M73432
O.OOQ732
0.005012
0. 001311
0.005064
0.01A521
0.00634*
0.006970
0.008*49
0.006143
0,002892
0.004463
0.005396
0.011499
0. 01)4523
0.025736
O.Q13749
0.00754?
0.019364
0,004945
0.007763
0.00'065
O.OS6154
0.001077
0.047799
0,C07'-67
0. •> 11 3 54
0.00251*
0,007625
n. 105086
0.012943
0.>PC28'>?
0,004675
0,009732
0.006021
Q. 0^5449
O.OOZ80?.
0.002 143
0,009732
0.003311
O.QU5VJ
O.OC64J5
0-. 003106
O.OU068
C. 006571
0.05Z7S7
O.OOa?S7
0.009732
0,006521
0.(J161S4
SPD/4
0,003870
0.007320
0.02*900
0.004290
0.006230
0.009030
0.001900
0.005050
0.002200
0.004560
0,. 003430
O.OOV290 •
0.002040
0.003660
0,010600
0.000089
0.001830
0.001230
0.003430
0,006270
0,01040'!
0. 004 7 80
0.004460
0.003660
0.001 850
0,003810
0,001840
0.002130
0.001060
0.001910
0,003230
0,004070
0.002100
fl.OOZZW
O.C03469
0.000798
0,001100
0.000380
O-PQ^'i^O
O.GOI2'50
0,0073fO
0.004240
0,007100
0-. GO^HOO
0,003330 '
C.0041BO
O.OOOBM
0.006410
O.OD0344
0. 002 560
0,C02350
0,007200
0.000704
0.00*020
O.i?0217n
O.OOC788
0.0013TO
S. 001780
O.,')03930
0.003'J'jO
0,00? 020
0.000?76
0.000611
0.003800
0.001 ISO
0«fi0461Q
0.001790
O.G232BO
0.004420
0.003020
0,000772
Q.C00310
0.004270
0.004920
.",00*560
274
-------
13.5.5 Distribution of Solids Deposited by Pipe Length
The cumulative distribution of solids deposited in the collection
system versus the cumulative length of sewer pipe where the deposition
occurs is the only exogenous information that the user will have to accept in
applying this methodology. In other words, this distribution is the only
information that cannot be derived from local data and can only be modified
in a minor way by input from local conditions. These modifications will be
covered later.
Some simplifications had to be made in deriving these distribu-
tions especially considering the great number of pipe segments (around 6000)
that were numerically considered. The procedure established for each basin,
as a function of the "maximum and minimum estimated loads, a series of 200
intervals where loads of approximately equal values were accumulated,
together with their corresponding pipe lengths. Whenever the basin had less
than 200 pipe segments the number of intervals would be made equal to the
number of pipe segments of the basin. The cumulative values of loads versus
lengths were then computed for each of these intervals.
Plots of several cumulative probability functions for a few
basins are presented in Figures 80 and 81 • An overall curve computed from
all 75 basins is also presented in Figure 81. A total of 5000 intervals were
used, in computing the overall curve, yielding a smooth curve as can be seen
from Figure 81 . Plots in semi-log paper of the complementary values of the '
load probabilities, (l.-F load) (in the log axis) versus the cumulative
probability of pipe lengths, resulted in nearly straight lines, indicating
that this distribution is also exponential.
It was observed by associating the curves in Figures 80 and 81
with their respective mean slopes that the average basin slope increases
moving from the lower to the upper curves. This explains comparatively
higher percentages of loads depositing in lower percentages of pipe lengths.
Based on this observation, Figure 82 was prepared. In its preparation a
smooth curve was drawn at approximately the middle of the range of curves
from Figures 80 and 81 with mean slopes corresponding approximately to the
range limits in the legend of Figure 82. With the mean pipe slope known
for a given basin, Figure 82 can be used to estimate the percentages of pipe
lengths corresponding to given percentages of total solids deposited.
Further discussion on the applications of Figure 82 will be presented in
Chapter 14 which deals with flushing strategies.
The choice of the pipe length corresponding to 80% of the loads
deposited as an independent variable in the regression analysis discussed
in section 13.3.1, resulted solely from the observation of higher separation
of the curves around the value of 80%. Other percentages of mass deposited
could have given the same results. It should be noted that the loads
estimated by the regression equations presented in section 13.6 correspond
to 100% of the loads deposited. The length LpD corresponding to 80% of the
loads is only one regressor devised to explain the variation in the total
loads.
275
-------
PO
--J
CTl
INDIVIDUAL COLLECTION SYSTEM DESIGNATION NUMBERS
.3 .4 j* ^4 7
FRACTION PIPE LENGTH
i
1.0
FIGURE 80
DISTRIBUTION OF SOLIDS DEPOSITED
PIPE LENGTH -WRNDB SYSTEM
BY
-------
ro
FIGURE 81
INDIVIDUAL COLLECTION SYSTEM DESIGNATION NUMBERS
.5
.8
1.0
.3 .4 .5 .6 .7
FRACTION PIPE LENGTH
DISTRIBUTION OF SOLIDS DEPOSITED BY
PIPE LENGTHS - DORCHESTER SYSTEM
-------
100-1
00
RANGES OF
AVERAGE SLOPES
0.01-0.018
0.018-0.025
0.025-0.037
% PIPE LENGTH
FIGURE 82 CUMULATIVE DISTRIBUTION OF SOLIDS
DEPOSITED VS. PIPE LENGTH
-------
13.5.6 Pipe Lengths Corresponding to 80% of the Loads Deposited - L
The cumulative distribution of loads deposited versus the
cumulative lengths of pipe for all basins were prepared. The values of Lpo
to be used in the regression analysis were determined in the following manner.
The computer printouts of the distributions were scanned for the percentages
of pipe length (PLn) corresponding to 80% of the total loads deposited and
either read directly or more often they were interpolated. Those percentages
were then applied to the total pipe lengths in the basins resulting in the
LpQ values for all basins presented in Table 74.
13.5.7 Slope Corresponding to PLD(SpD)
The determination of Spp, the slope corresponding to the
percentage of pipe (PLp) where 80% of the total loads deposits is illustrated
in Figure 73. The value of PLp in the cumulative distribution of pipe
slopes for a given basin is established and the corresponding slope value
Sp[) is determined. This was performed for all basins by reading directly or
interpolating values in the tables of the cumulative distributions of the
pipe slopes of these basins. The values of Spp for all the 75 basins
considered are given in Table 74.
There is no theoretical justification for the choice of SpQ as
a regressor. As a matter of fact the average of the slopes (Spjo) below SpD
had originally been thought of as a better regressor than Spo itself, and
was the first to_be included in the regression analysis. A high correlation
between Spp and Spp was observed and Spp was included in the regression
since its_reduction of the sum of squares was slightly better than SpD. The
variable Spp was dropped from further consideration since its determination
required far more effort than that for
13.5.8 Slope Corresponding to PLD/4 - (SpD/4)
The determination of Spp/4 is illustrated in Figure 73- First
of all, multiply by 1/4 the percentage, PL^, corresponding to 80% of the
loads deposited. Next enter that value in the cumulative distribution of
pipe slopes for a given basin and determine the corresponding value, Spp/4.
This step was done by reading directly or interpolating values in the
tables of the cumulative distributions of the pipe slopes for the basins.
The values of SpD/^ for all 75 basins are presented in Table 74.
The choice of Spo/4 as an independent variable in the analysis
regression did not involve any theoretical consideration. Its choice
resulted from an "a priori" belief that representing the lower ranges of the
pipe slope distribution would be significant in the regression analysis.
It is shown in section 13.6 that this assertion is true.
13.5.9 Summary of Input Requirements for Regression Analysis
Descriptions of the data preparation for the regression analysis
has been presented in sections 13.5.1 through 13.5.8. The only remaining
279
-------
independent variable not described in those sections is the waste flow rate
contribution in gpcd. This variable, as described in section 13.3.2 was
fixed at the values of 40., 110., 190., and 260. gpcd, covering, therefore,
a wide range of flows.
To summarize, the values obtained for the independent variables
are given in Tables 71 and 74. _The tabular values of L' and A are given in
Table 71 and the values of Lpp, D, S, and SpD and Spp/4 determined for all
75 basins are given in Table 74. Summary statistics of the independent
variables used in regression analysis are given in Table 75 including the
range, the mean and the standard deviation for each variable. The results
from the deposition model are lengthy and are not presented in this report.
TABLE 75. RANGES, MEANS AND STANDARD DEVIATIONS OF THE
INDEPENDENT VARIABLES USED IN THE REGRESSION
Variable
L' (ft)
A (acre)
S (ft/ft)
D (in)
LPD
SPD (ft/ft)
spD/4(ft/ft
q (gpcd)
*
q is not a
Range
263. -134528.
3. -641
0.00238-0.0799
8.0-20.0
36. -34879
0.0011-0.0589
0.000089-0.0249
40. -260.
random variable.
Mean
13702.
76.
0.0210
11.5
4026.
0.0101
0.0037
*
Standard
Deviation
22867.
102.
0.0126
2.0
5811.
0.0093
0.0033
*
3.6 Regression Analysis
In this section analyses are presented that relate daily pollutant
deposition accumulations in sewerage collection systems with overall
collection system physical characteristics. First of all, summary results
of various regression models for estimating total suspended solids deposition
in collection systems under clean pipe conditions, that is, under the
assumption of a well-maintained system, are described in section 13.6.1
through 13.6.3. Secondly, regression relationships are presented in
section 13.6.4 that are useful for approximating deposition loadings under
conditions of existing sediment deposits, that is, mimicing the situation
where poor maintenance prevails. These relationships modify the clean-pipe
280 .
-------
results given in section 13.6.3. Finally, empirical factors for computing
daily accumulations for other pollutants are presented in section 13.6.5
using the total suspended solids estimates given by the procedures in
section 13.6.3 and 13.6.4. These factors derive from the data reduction
analyses of the field flushing data presented in Chapter 8.
Various combinations of independent variables were initially
considered including only primary variates such as total pipe length, L,
service acres, A, average pipe diameter, D, and average collection system
pipe slope, S. More complicated parameters such as LDQ, Spp and SpD/4 were
than incorporated into alternative sets of independent variables.
Multiple correlation coefficients for the untransformed data,
that is, the linear additive models, ranged from 0.814 to 0.906. Similar
analyses using the logarithmic transformed data yielded multiple correlation
coefficients ranging from 0.923 to a maximum of 0.974. This implies that
the -coefficient of determination, R2, ranged from 0.852 to 0.949. The
multiplicative models were-all superior to the linear forms explaining
roughly 95 percent of the total variability of the dependent variable. For
this reason only the multiplicative model results will be presented.
Statistical details of all the alternative regression models considered are
not presented but are cited in the earlier report describing this work (37).
13.6.1 Regression Method
The linear regression program used to empirically establish the
relationships of the total daily suspended sol ids (TS) deposition within a
sewerage collection system with the independent variables described in
previous sections, is one that operates in -a step forward manner. At each
step in the analysis the particular variable entered into the regression
equation accounts for the greatest amount of variance between it and the
dependent variable i.e., the variable with the highest partial correlation
with the dependent variable. The program is flexible to allow any
independent variable to be: (1) left free to enter the regression equation
by a criteria of the sum of squares reduction; (2) forced into the
regression equation; or (3) be kept definitely out of the regression
equation in one given selection. The procedure permits examination of
several alternative considerations of the independent variables by optional
selections of variables to be forced in and out of the regression equation
or to be simply left free to enter the equation using variance reduction
criteria.
Observation of the relative change in the standard error of
estimate was used as the stopping rule in the regression analysis. An
increase of the standard error at a given step indicates that the additional
information realized by introducing the variable is off-set by the loss in
degrees of freedom, implying that the regression equation is better off
without that particular variable.
13.6.2 Regression Analyses
The values of the independent variables used in the regression
281
-------
analysis for all 75 basins were presented in Tables 71 and 74. The range,
mean and standard deviations of the independent variables were computed
from the data given in Tables 71 and 74 and are presented in Table 75. A
total of 300 observations were used in the regression analysis since the
deposition loadings for each of the 75 basins were computed for four per
capita waste generation rates.
In this analysis various predictive models are analyzed relating
total suspended solids deposition within a collection system with the
aforementioned independent variables' under the assumption of clean pipe
conditions. These relationships are therefore applicable for situations in
which the sewer piping system is properly maintained. The effects of age
and improper maintenance on collection deposition loadings were examined and
the results are presented in section 13.6.4. The degree of increased daily
deposition resulting from improperly maintained systems was crudely simulated
using several assumed depths of bottom sediments.
Both linear additive and multiplicative models were investigated.
Untransformed observed values of the dependent and independent variables
were initially used, leading to a strictly linear regression equation. In
the second case the observed values of both the dependent and independent
variables were transformed by taking their natural logarithms, leading to a
linear equation in the logarithmic domain which can then be put into a non-
linear multiplicative form.
13.6.3 Alternative Model Selections
In this section several regression models are recommended for
user application. Alternative forms reflecting the availability of data
and/or user resources will be presented. The simplest forms require little
data and have the least predictive reliability whereas the more complicated
models, requiring greater user resources and data availability, provide
estimates with extremely high reliability.
The Elaborate Model. The highest multiple correlation coeffici-
ent, 0.974 (R^ = 0.949) was obtained using the model given by equation 42 .
TS = 0.0038 I!0'8142 SpD-°'8187 SpD/4-°'1078 q-°-5098 (R2=0.949) (42)
where TS is deposited solids loading in Ibs/day, L is total length of sewer
system in feet, Spp and Spp/4 are slope parameters defined in section 13.5.7
and 13.5.8, and q is the per capita waste rate in gpcd.
Utilization of equation 42 requires knowledge of total pipe
length, the per capita contribution and the two pipe slope parameters, SpD
and Sp[)/4- These slope parameters in turn are a function of the percentage
of pipe where 80% of the total mass (TS) deposits (PLD) and the probability
distribution of the pipe slopes. The value PLp is assumed given for the
regression analysis, derived from the extensive computer analyses performed
during this study and reported in section 13.5.5. The probability
distribution of the pipe slopes can either be derived from histograms
282
-------
computed from local pipe slope data or it can be defined with reasonable
approximation from the mean pipe slope (S) only, as described in section
13.7.2. If the pipe slopes are not available a regression relationship of
mean ground slope and mean pipe slope presented in section 13.7.2, could be
used to estimate mean pipe slope.
An Intermediate Model. A simpler model than that given by
equation 42 , including only the primary variables, is given by:
TS = 0.001303 t1'18 A'0'178 (S)"0'418 (D)0'604 q'0'51 (R2=.852) (43)
where S is average pipe slope, and D is the average equivalent diameter in
inches; and A is service area (acres).
All five independent variables present in equation 43 can be
defined with good precision in any practical application. Neither the
distribution of deposited loads versus pipe length nor the probability
distribution of the pipe slopes are required. The mean pipe slope's, again
can be correlated to the mean ground slope as described in section 13.7.2,
if no information on pipe slopes is available. A fitted equation for L' as
a function of A was derived from the data used in this study and is presented
in section 13.7.1.
If the mean pipe diameter, D, is eliminated from this regression
the loss in precision of the estimates is not significant, resulting in the
expression:
TS = 0.00389 l!1-2195 A-°-1866 ^-0.4343 q-0.51 (R2=0<848) (M)
2
The Simplest Model. The highest R value that can be obtained
with the least number of independent variables is given by the regression
equation:
TS = 0.0076 I!1'063 (S)"0'4375 q~°'51 (R2=0.845) (45)
Comments on Model Selections. The user should note that the
parameters of the regression equation estimated by the least squares
procedure are a function of the data used in their estimation. These
parameters are not known with certainty and represent only estimates of the
true parameters of the model. The estimation of the parameter is improved
as the number of data points increase and as the spread or variance of the
independent variables also increases.
It is known from regression theory that the procedure provides
the best estimates (least variance) around the means of the independent
variables used in computing the parameters of the regression equation. As
the values of the independent variables depart from the mean the uncertainty
of the estimation given by the equation increases and can become large
outside the range of values used as data in the regression. In other words,
283
-------
the uncertainty of the estimate given by the regression equation may be
large when the model is used for conditions requiring extrapolation beyond
the range of data used to develop the model.
13.6.4 Effects of Age and Maintenance
The regression equations presented in section 13.5/3 were derived
from deposition data computed under the assumption of clean pipes, with no
bottom sediments. In this section the impact of poorly maintained systems
was crudely simulated by arbitrarily assuming various levels of prior sediment
accumulation in the pipes. These sediment levels would change the bottom
cross-sectional shape of the pipe channel and hence the depth of flow, the
hydraulic radius and the shear stress characteristics.
Two cases simulating different degrees of maintenance other than
perfect clean pipe conditions were considered. In the first case, or the
intermediate maintenance category, sediment beds ranging from 1 to 3 inches
in depth were assumed for all pipes with slopes less than 0.0075. A sediment
bed of 3 inches was assumed for all pipes with slopes less than 0.0005. The
bed depths then ranged linearly starting at 3 inches for a pipe slope of
0.0005 up to one inch for a pipe slope of 0.0075. This range was established
using judgment and also based on visual inspection,of numerous combined
sewer laterals in eastern Massachusetts sewerage systems. In the second
category of maintenance, the zero maintenance case, sediment beds ranging
from 3 to 6 inches for the same range of slopes was considered.
Considering the two age and maintenance criteria mentioned here,
the deposition model was used to estimate total deposition loadings for each
of the 75 sewer systems for each of the four per capita waste generation
rates of 40, 110, 190 and 260 gpcd. Before similar regression computations
were performed on the deposition results obtained for pipes with bottom
deposits, a comparison was made of the total deposited loads computed under
the assumptions of clean and sedimented pipes.
For each basin the ratios of TS computed for sedimented pipes
with sediment beds of 1-3 inches and 3-6 inches and the TS values for clean
pipes were, calculated for all four per capita waste rates considered, i.e.,
40, 110, 190 and 260 gpcd. The resulting ratios were very stable for a
given per capita waste rate for both cases of sediment deposits. The mean
and coefficient of variation of these ratios are presented in Table 76 for
both conditions of bottom deposits.
The results shown on Table 76 suggest that the prediction of TS
in sedimented pipes could be accomplished by a simple functional multiplica-
tive correction of the results given by any of the regression equations for
clean pipes. An equation was fitted using the data of Table 76 for each of
the bed deposit conditions.
These equations are:
- For a system with deposits ranging from 1 to 3 inches;
284
-------
TS
1-3 inches
= 1.68 q-°-076 TS(clean) (R2=0.988)
(46)
TABLE 76. AVERAGE VALUES OF THE RATIOS OF
COMPUTED LOADS IN DEPOSITED PIPES OVER CLEAN PIPES
Ratios
TSl-3inches/TS(clean>
TS3-6inches/TS(clean>
The numbers in parenthesis
ratios.
Average
Per
40
1.263
(0.18)
1.312
(0.14)
indicate the
'Values of Ratios for
Capita Waste Rates
110 190
1.186 1.128
(0.14) (0.07)
1.211 1.151
(0.11) (0.09)
coefficient of variation
260
1.094
(0.12)
1.121
(0.09)
of the
where q = flow per capita, and TS(clean) = load of total solids computed
from any of the regression equations presented in section 13.6.3.
- For a system with deposits ranging from 3 to 6 inches:
TS3-6 inches = 1>79 q~°'°84 TS(clean) (R2=0.999)
(47)
The R values indicated above refer to the regression of the
ratios of TS on the values of flow per capita. The small difference found
between the two conditions of bottom deposits may well be the result of an
inappropriate accounting of these factors by the deposition model. On
the other hand it may simply have resulted from the particular combination
of pipe diameters and sediment depths used as data, which may have led to
actually small differences in flow depths above the sediment levels, and
therefore' small differences in shear stress between the cases.
13.6.5 Estimation of Other Pollutants Using TS Results
Results of the first phase field flushing experiments described in
chapter 8 were used to develop factors for estimating other pullutant deposi-
tion loadings. The normalized flushed pollutant loadings from Tables 23 thru
27 in chapter 8 were used to compute factors for estimating relative ratios of
COD, BOD, TKN, NH3, TP and VSS to TSS. Pairs of normalized flushed pollutant
loadings to solids flushed were grouped for all good non-rainfall impacted
flushes from all segments and the average ratios computed. These average re-
sults are summarized in Table 77. The mean, standard deviation and number of
flushing experiments used in the computation are shown in Table 76 for each
pollutant. The coefficients of variation for these factors with the exception
o'f TKN are all remarkably low.
285
-------
TABLE 77. AVERAGE POUNDS OF POLLUTANTS PER
POUND OF TSS* PER ANTECEDENT DAYS
COD
BOD
TKN
NH3
TP
VSS
MEAN Ib/lb TSS
1.247
0.434
0.041
0.014
0.009 .
0.745
SD
0.269
0.203
0.114
0.009
0.005
0.108
# POINTS
14
26
14
13
14
45
* All good, non-rainfall impacted flushes of all 4 streets
13.7 Model Utilization
This section summarizes the steps required by the user to
define on a preliminary basis the total amounts of solids and other pollutants
that deposit in the sewerage system. The nature of the available data and/or
the degree of resources the user can commit will define different work
programs which in turn will lead to different levels of confidence in the
results.
Several additional relationships derived from analysis of
collection system characteristics and computational methods for deriving
estimates of the various parameters in the deposition prediction models will
be given. Descriptions of the necessary user steps required to use the
predictive models will then follow.
13.7.1 Estimation of Total Pipe Length
The total pipe length of the system, L', and its corresponding
collection area, A, are generally assumed to be known. In cases where this
information is not known and where crude estimates will suffice, the total
pipe length can be estimated from the total basin area using the expressions:
L'= 168.95 A0'928 (R2=0.821)-low population density (10-20 people/acre)
L'= 239.41 A0*928 (R2=0.821)-moderate-high population density (30-60
people/acre)
(48)
These expressions were derived from the data for the 75 basins used in this
study, yielding a correlation coefficient of 0.906. Since detailed
population density estimates for each collection system in the data set were
not available, the following approach was used to account for the effect
of population density on length of sewer. The data for the 75 basins were
separated into two distinct data sets representing low population densities
(the 35 basins from the West Roxbury-Newton-Dedham-Brookline sewerage
system) and moderate to high population densities (the latter 40 basins from
the Dorchester and Fitchburg sewerage systems).- A dummy variable approach
was used in the regression analysis to obtain the two expressions given in
equation 48 . A more explicit treatment of population density would be
286
-------
possible if detailed population density estimates per collection system were
available.
13.7.2 Estimation of Mean Pipe Slope S
Pipe Slope Data is Available. In this case:
where: S = slope of segment i
1 = length of segment i
n = total number of pipe segments
Pipe Slope Data Not Available. If data on pipe slope is not
available the user will determine a mean value for the ground slope using any
procedure, such as the techniques of uniform grid sampling of random sampling.
With the mean ground slope value, determine the mean pipe slope by:
S = 0.348 (SG)0t818 (R2=0.96) (50)
where: SQ = mean ground slope (ft/ft)
The expression above resulted from a regression of mean ground versus mean
pipe slope for all 75 basins of this study. The resulting correlation
coefficient was 0.98.
This relation was derived in the following manner. Topographic
mylar overlays with ground contour at ten feet intervals were developed for
the WRNDB and the Fitchburg sewer system service areas. Similar maps
existed for the Dorchester and South Boston sewer systems. Using the
topographic maps the collection system areas were subdivided into smaller
fractions considered to have uniform slopes. The ground slopes were
computed for each of the subareas. Those ground slopes were then associated
with the percentages of the collection system area they covered. The basin
average ground slope was then computed by weighing the individual slopes by
their associate subarea percentages. Ground slopes in the smaller basins
were determined by using one or two subareas whereas the ground slopes in
the large basins were determined using 10 to 21 subareas.
13.7.3 Distribution of Total Solids Deposited by Pipe Segment -
Determination of PLp
The distribution of total solids deposited by pipe length is
presented in Figure 82. The user can determine PLg corresponding to 80% of
the solids deposited by entering the graph with "percent of mean deposited"
287
-------
equal to 80% and, from the curve that best approximates the average pipe slope
S, read on the horizontal axis, "% pipe length", the value of PI_Q. This value
will be required to determine Spp and Spp/^ as described KQ!OW.
13.7.4 Determination of Slopes Spp and Spp/4
Pipe Slope Data is Available. The collection system pipe slopes
can be directly used to define the cumulative pipe slope distribution. In
this study the pipe lengths (as integer values) associated with the observed
slopes were taken as their frequencies. The user should adopt the same
approach for compatability with the equations for TS derived in this study.
Simplified method? may also be used, with less accurate results. If, for
example, the slopes are not weighted by their lengths, this is equivalent to
assuming, that all pipe segments have the same length, which in some cases
may be a' reasonable assumption.
Pipe Slope Data Not Available. The average slope S can be
computed using either equation 49 or equation 50 to define the pipe slope
cumulative density function (CDF) as follows:
FS = l-e" (51)
If the probability Fs is known, with a fixed S the corresponding slopes can
be computed, and vice-versa. For the basins used in this study, equation
51 was verified with excellent results as described in Section 13.5.3.
To determine the slope Spp, use the value of PLp determined in
section 13.7.3, and make:
and determine s. Then:
= l-e-s/S =
i e
SPD=S
To determine the slope Spp,., make:
Fs = ! = e-s/S = 1/4PLD
and determine s. Then:
Spp/4 = s.
13.7.5 Formula for the Average Pipe Diameter
If the system contains non-circular shapes, first compute their
equivalent circular diameters using the equations for equivalent pipe
diameters presented in Table 73. Then, compute the average diameter by:
288
-------
,,
where: D^ = diameter of segment i
l.j = length of segment i
, n = total number of pipe segments.
13.7.6 General Description of User's Steps
The generalized procedure for obtaining total solids deposited
estimates is presented in Figure 83 and should be referred to in the ensuing
discussion.
The first question to be answered at the onset is whether pipe
slope data are available. If the answer is positive the user should use
equation 49 to determine the average pipe slope S; otherwise, a mean
ground slope should be determined as described in section 13.7.2 and then
equation 50 should be used to determine S.
The next question that arises is whether the total pipe length
of the collection system is known. In the negative case the total area of
the collection system is assumed known and the total pipe length is estimated
by equation 48 . The per capita waste rate including infiltration is then
established.
The selection of the equation for the estimation of the solids
deposited, TS, follows. If the simplest model is desired the user has at
this point all the elements required by equation 45 to compute TS. If the
intermediate model is chosen and the pipe diameter information is available
the mean pipe diameter is computed by equation 52 . All information is
computed to estimate TS from equation 43 . If pipe diameters are not
available then use equation 44 . If the elaborate model is chosen, the
value of P1_D must be determined using Figure 82 as described in section
13.7.3. The next variables to be determined are Spr> and SpD,4.
If no pipe slope information is available the user has no
alternative but to use the exponential approximation for the pipe slope
cumulative distribution. If the pipe data are available but the user
considers the exponential approximation satisfactory he may use equation
51 , associated with Pin and S to determine Spo and Sp^/4 as described in
section 13.7.4. Alternatively, the user can define the pipe slope CDF from the
pipe slope data and derive from it, the values of Spp and SpQ/4 as described
in section 13.7.4. Finally, the user can define the pipe slope CDF from the
pipe slope data and derive from it, the values of Spp and Spg/4 corresponding
to the percentages PL[> and 1/4P1_D, respectively. At this point all the
elements are prepared to estimate TS from equation 42 which provides the
most reliable estimate of TS.
289
-------
ro
VO
0
FROM TOTAL ARE
DETERMINE TOTA
PfPE LENGTH BY
EQUATION 1481
g
ESTIMATE TS
f " 'l45f
CHAVE PIPE\
SLOPE DATA/
I NO '
YES ^ j
\ f
DETERMINE MEAN
PIPE SLOPE s — i
FROM. EQU. 1491 |
~1 ^
=« ^as^HAVE TOTAL PIPE LENGTH'
NbYES
J1 DEFINE PER-CAPITA VALU
1 q (INCLUDES INFILTRATION
4 ^JCSS— CSIMPLEST MODEL)^ — 1
4^o
fpiNiSHV • - ESTIMATE TS £Jig— /"HAVE DATA ONV YES .„ (INTERMEDIATE MODEL)^
^—^^ FROM EQUATION \PIPE DIAMETER,/
144) 1 '
V
ESTIMATE TS ESTIMATE MEAN
1431 ' D
FROM DATA COMPUTE x_
THE CDF OF. SLOPES ~
V
ENTER PLD AND 1/4 PLD
IN SLOPE DISTRIBUTION
AND DETERMINE Son
AND SpD/4
\
ROM MEAN GROUND SLOPE
tETERMINE MEAN PIPE SLO
BY EQUATION ISO)
V ,i
A
E
1
> ENTER % MASS DEPOSITE
AND BASIN AVERAGE SLO
x- •* x
/USED GROUND VSA
[ PIPf: MEAN SLOPE 1
V RELATIONSHIP I
\ABOVE? ^/
NO
\f
1 EXPONENTIAL \ — I52 — >
V APPROXIMATION /
\I>IPE SLOPE T^/
. ESTIMATE TS f
J FROM EQUATION 1421 "
YES
V
PE
D=80X
PE 8
WINE PLD
USE a TO DEFINE THE PDF OF PIPE
SLOPES BY ASSUMING THE EXPONENTIAL
APPROXIMATION p =i_es/5
SET F8=1-» ' =PLD AND
THEN SpD=S
SET F,=1-«~s/" =1/4 PLD
DETERMINE S. THEN SpD/4
1
DETERMINES. J
..
FIGURE 83 USER STEPS TO DETERMINE TOTAL SOLIDS DEPOSITED
-------
The resulting estimate of TS is the total daily solids deposition
in the collection system of interest. If the user wishes to modify this
estimate for pipes with existing sediment beds the multiplicative equations
in section 13.6.4 should be used. If estimates of other pollutants are
desired, the deposited solids results, TS, should then be used as predictors
to compute other pollutant estimates using the equations given in section
13.6.5.
13.8 Test Case Application of Prediction Procedures
In this section an example problem illustrating the methodologies
developed in this report is presented. The test case is one of the collect-
ion systems in the Dorchester sewerage system. Figure 84 shows a diagram of
the collection system. Estimates of total daily solids deposition in this
collection system will be given for different assumptions of data availability
using the simplified procedures. These estimates will also be compared with the
results computed forthis system using the deposition model described in Chapter 12.
The test case used in basin number 70 of the Dorchester
combined sewer system. There are 190 manhole segments in this collection
system. The topography of this combined sewer collection system is fairly
hilly. The land use in the area is exclusively high density multi-family
dwellings with a population density of 30 people/acre. The area! extent and
total collection system pipe footage is about one standard deviation above
the mean of all the systems used in the regression analysis. The values of
all the independent variables for basin 70 are given in Tables 71 and 74 .
Data Requirements. The total pipe length, L', for the basin is
35,225 feet and the total service area, A, is 233 acres. The total pipe
length can also be computed using equation 48 for high population density
yielding 37,740 feet, representing an overestimation error of 7.7%.
There are three possible ways to compute the values of the pipe
slope variables required as input for the various regression equations. The
first method involves computing these parameters from the distribution
derived from actual pipe slope data. The second procedure requires knowledge
of only the mean collection system pipe slope, S, and assuming that the pipe
slopes can be represented by an exponential distribution. The third
alternative approach assumes that only ground slope information is available
and that the exponential distribution is applicable to represent collection
system pipe slopes.
The collection system pipe slope histogram representing the
distribution of 190 pipe segments in basin 70 of the Dorchester sewerage
system is given in Table 78. This table was prepared using the invert
elevations for each of the 190 segments. The pipe slope histogram is
divided into 30 intervals with the upper pipe slope limit of each interval
given. The elements under the column labeled "frequency" are the actual
pipe footages associated with each pipe slope interval. The last two
columns give the interval probability computed using the pipe footage per
interval relative to total pipe footage and the cumulative probability. The
mean and standard deviation are also given in Table 78. The average
291
-------
AREA'- 223 ACRES
PIPE FOOTAGE : 35;225
NUMBER OF SEGMENTS'- 190
POPULATION DENSITY: SO/ACRE
AVERAGE GROUND SLOPE 0.0359
FIGURE 84
DIAGRAM AND PERTINENT INFORMATION
FOR BASIN 70 - DORCHESTER
292
-------
TABLE 78. DISTRIBUTION OF PIPE SLOPES FOR BASIN 70
AVG. DIAMETER =15.6
HISTOGRAM FOR FIXED RANGES (MAX. AND MIN. ABCISSA VALUES SPECIFIED)
IRTERT1L
1
2
3
8
5
6
7
8
9,
10'
11
12
"3
18
15
16
17
•"8
19
20
21
22
23
2d
25
26
27
28
29
30
OPPES tlHTT
OP RABGE
0.000100
0.003311
0.006521
0.009732
0.0129U3
0.016158
0.019368
0.022575
0.025786
C.028996
0.032207
0.035119
0.038629
O.OIH839
0.005050
0.008261
0.051071
0.058682
0.057893
0.061101
0.060318
0.067525
0.0^0736
0.0739*6
0.077157
0.080368
0.083578
0.086789
0.090000
> C.090000
FREQUENCY
0
3680
10»77
3239
0099
1189
680
832
1229
752
U72
909
1085
371
752
287
315
U60
1528
0
560
68
1033
33
-------
equivalent circular diameter, D, is 15.6 inches and was computed using the
formulas given in Table 73, and equation 52 .
The average pipe slope, S, is 0.0194. The slope parameter
can be estimated in the following manner. Three curves are available in
Figure 82, Section 13.5.5, for relating the cumulative distribution,of solids
deposited to cumulative distribution of collection system pipe length. Curve
2 is applicable in this case since the average pipe slope is 0.0194. The
percentage of pipe associated with 80% of the total daily load
deposited is 38%.* The value of the variable, SPD, can then be determined
by entering 0.38 in the last column of Table 78 and interpolating the
corresponding value of slope in column two. The interpolated value of Spp
is 0.00629. The value of SpD/4 is obtained by enterina the same table with
0.38/4 = 0.095, yielding SpD/4 = 0.00302.
In the second procedure the mean pipe slope can be used in the ex-
ponential cumulative distribution function given by equation 78 , to compute
the values of Spn and Spny4. The value of SPD is obtained by solving equation
51 for s, having set FS ='0.38 and 5" = 0.0194. The result is Spo=s=0.00928.
The value of Sp[j/4 is found by solving equation 51 again for s, having set
FS = 0.38 = 0.095 and 5T= 0.0194. The result is SpD/4 = s = 0.00194.
4
The only information required for the third procedure is the
average ground slope, SL. The average ground slope for basin 70 was deter-
mined by graphical procedures to be 0.0359. An estimate of the mean pipe
slope, S, can be obtained using the mean ground slope in equation 50 . The
estimated value of mean pipe slope, S, using the mean ground slope, JL, is
0.02287 differing from the actual mean pipe slope by 17%. Values of ^ SpD
and SpQ/4 can now be estimated using the exponential relationship in
equation 51 , yielding SpD = 0.00985 and SpD/4 = 0.00229.
Comparison of Deposition Prediction Procedures. The detailed
manhole to manhole dry weather deposition model described in Section 12.4 of
Chapter 12 was used to estimate total solids deposited for the entire
collection system network in basin number 70. The estimated load is 169.11
Ibs/day using an average per capita waste rate of 190 gpcd. This model
requires detailed specification of the hydraulic parameters for each segment
in the system and the use of a computer. Similar estimates will be computed
from using the simplified power functions generated in this study for the
three different estimates of the pipe slope variables.
The elaborate model given in equation 43 requires specification
of L, SPD, SpD/4 and the per capita flow rate, q. This equation was solved
for the three different estimates of Spp and SpQ/4 with q = 190 gpcd. The
deposited load, TS, is 155.73 Ibs/day using the values of SpD and Sp[)/4
derived from the analysis of detailed collection system pipe slope information.
The estimated load is 118.72 Ibs/day using values of Spp and SpD/4 computed
from the mean pipe slope and the exponential cumulative function. The
estimated load is 111.11 Ibs/day using values of Spp and SpQ/4 for the third
The distribution of loads by pipe length derived for basin 70 yields
exactly 38% at 80% of the total load. The error involved in using Figure 82
is zero in this' example.
294
-------
situation where the mean ground slope is used to compute mean pipe slope for
input into the exponential function.
The intermediate model given by equation 44 requires that L,
A, S, and q be specified. Three variations are computed using this
formulation. The first estimate is determined using the measured pipe
length, L and pipe slope, S, derived from data; the second case uses the
measured pipe length and an estimate of mean pipe slope from ground slope
estimate; and the third result is computed using the estimated pipe slope S,
and an estimate of pipe length derived from equation 48 . These three
estimates are 186.89, 174.02 and 190.55 Ibs/day, respectively.
The simplest model given by equation 45 requires specification
of L', S and q. The estimated loads are 198.82 Ibs/day using exact estimates
of L' and S, 185.03 Ibs/day using exact estimates of pipe length and an
estimate of pipe slope ground from ground slope, and, 200.26 Ibs/day using
estimated values of both pipe length and mean pipe slope.
A comparison of all computed loadings for the different predict-
ive models under different assumptions of data availability is given in
Table 79. The percentage error relative to the estimate provided by the
deposition model described in Chapter 12 is also provided. These results show
that the estimated values given by all three regression equations are reason-
ably close to the value derived from the detailed collection system model.
The elaborate model consistently over estimated deposition loadings.
It is an invalid conclusion to infer from this comparison that
the simpler approach might be superior to the more elaborate one. It
should be noted that this comparative result is only for one basin and that
on the average, the elaborate model will provide consistently superior results
because the coefficient of determination RS is higher for the elaborate
model. The under estimates given by the elaborate model using approximations
for the pipe slopes are explained by the fact that, for this particular
basin, the exponential approximation over estimated by about 50% the slope,
SpQ. Nevertheless, one of the simpler models would be more appropriate in
cases where little data is available requiring many assumptions and
approximations. Moreover, the utilization of the simpler models for planning
"first-cut" purposes may be more cost effective from the standpoint of
collecting, analyzing and preparing the required data inputs.
In sum the simplified procedures given by equations 42 , 43
and 45 provide estimates of daily solids deposition using exact data for
basin 70 with a relative error of 8 to 18 % in comparison to the
estimate given by the complicated procedure described in Chapter 12. It was
shown in Chapter 12 that the complicated procedure was roughly calibrated
using actual field flush information lending credulence to~the adoption of
the simplified procedures generated in this report.
295
-------
TABLE 79. COMPARISON OF ESTIMATED DAILY SOLIDS DEPOSITED FOR
BASIN 70 USING DIFFERENT PROCEDURES
Procedure
Solids Percentage Error
Deposited Relative to Chapter
(lbs/da.y) 12 Results
Deposition Model (Chapter 12) 169.11
Elaborate Model (Eg. 42 )
Exact Data 155.73
Exponentially Dist. Slopes 118.72
SQ and Exponentially Dist. .,,, ,,
Slopes 111J1
Intermediate Model (Eg. 44 )
Exact Data 186.89
Estimated Slope S 174.02
Estimated L1 and S 190.55
Simplest Model (Eg. 45 )
Exact Data 198.82
Estimated S 185.03
Estimated Lf and S 200.26
-7.9%
-30.0%
-34.0%
+10.5%
+ 3.0%
+12.7%
+17.6%
+ 9.4%
+18.4%
296
-------
SECTION 14
FLUSHING GUIDELINES
14.1 Foreword
Overview comments on the use of sewer flushing as an integral com-
ponent of new abatement technologies for combined sewer management is pre-
sented in section 14.2. A conceptual overview dealing with the development
of a sewer flushing program is presented in section 14.3. Approximate costs
of sewer flushing using both manual and automated operation is presented in
section 14.4. Procedural details and guidelines dealing with the implemen- .
tation of a sewer flushing program are given in section 14.5.
14.2 Management Overview
The concept of controlling depositing solids in sewer lines, although
widely used around the turn of the century as a maintenance practice, is
still in its infancy in regard to being viewed as a viable pollution control
alternative for combined sewer systems. Even if shown to be effective, the
concept of sewer flushing is still a controversial issue since it involves
generally low capital first-cost investments but high operational and main-
tenance costs. Federal funding mechanisms presently favor high first-cost
programs with low operational costs. Federal funding does not cover opera-
tional costs. Moreover, the notion of increasing the municipal commitment
for greater manpower investments runs counter to the historical trend toward
decreasing public works' budgets in the area of sewerage system management.
It is the authors' opinion that sewer flushing in itself should not
be viewed as a substitutable alternative for structural control, but instead,
when integrated together with other upstream management practices and down-
stream structural options where necessary, overall costs can be significantly
reduced. The federal government's will and commitment to spending many
billion dollars over the next decade in combined sewer control must be met
with equal commitment to execute this mandate in a cost-effective manner.
Suggestive management concepts employing several state-of-the-art
combined sewer controls are illustrated in Figure 85. It is assumed that
thorough sewer system physical surveys aimed at identifying cost-effective
pollution-control malfunctions have been performed and the requirement for
additional control has been established. In short, the sewer system is
"tuned-up" and further pollutant reductions are necessary. The four addition-
al controls shown in Figure 85 are sewer flushing, upstream off-line storage,
swirl regulators and disinfection using chlorine and/or possibly chlorine
297
-------
FLAT DEPOSITING AREA
(FLUSHED)
MIXED AREA
(NO FLUSHING)
UPSTREAM
OFFLINE
STORAGE
MODULE
REGULATORS c
ro
vo
oo
CONCEPTS
Minimize upstream accumulation during dry periods using
sewer flushing where appropriate.
Detain residual "first flushes" from upstream networks
(l~5 sewer miles) with an upstream offline storage module.
Allow all residual flow to continue downstream to
wastewater treatment plant or overflow.
Install swirl regulator/treatment devices on major upstream
inlets to overflow conduits and/or in line on overflow conduit
itself to remove residual downstream pollutants.
Disinfect at or downstream of swirl storage site using ?C10o
and /or chlorine.
SOME DRY
WEATHER
ACCUMULATIONS
DEPOSITING AREA
LIMITED FLUSHING)
SWIRL
REGULATO
UPSTREAM
OFFLINE
STORAGE
MODULE
STEEP SLOPES LITTLE DEPOSITING
(NO FLUSHING)
-LITTLE DRY WEATHER ACCUMULATION
-WILL BE FIRST NETWORK
DISCHARGING TO INTERCEPTOR
foul sewer)
SWIRL REGULATOR as Treatment unit + Disinfection
(clear water)
FIGURE 85
SEWER FLUSHING CONCEPTS
• Receiving
Water Body
-------
dioxide. Sewer flushing in flat collection systems would minimize upstream
accumulations during dry weather. Upstream storage modules, optionally,
could be employed to retain dry weather flushes and residual "first-flushes."
Swirl regulators and/or other treatment devices at major inlets to overflow
conduits can be used to remove residual pollutants. Newer forms of disinfect
tion, including sequential dosing of"chlorine/chlorine dioxide mixedj).y the
swirl flow, having high' bacterial kill in extremely short detention periods,
could be then incorporated upstream of the swirl.
The upstream control concept combines flushing and off-line
storage. Pollutants removed by flushing would either go directly to the
treatment plant and/or alternatively to locations where starm flows and/or
released captured upstream stormwater would flush the loads to off-line
storage. The flushing program would be combined with an active maintenance
program where malfunctions in the collection system are quickly detected and
remedied.
The advantages of such an integrated approach would be the follow-
ing: a) reduction in size and cost of required major downstream storage/
treatment facilities; b) improved overall solids and pollutant removal;
c) more even temporal distribution of solids loadings to the wastewater
treatment plant; d) improved flow stabilization during high intensity - short
duration storms, and e) improved sewer maintenance practices with the
resulting benefit of fewer system malfunctions which can in many instances,
reduce overflow pollutant loadings.
14.3 Conceptual Overview for Developing a Sewer Flushing Program
An outline of the following steps necessary to implement a sewer
flushing program are presented in Figure 86. Brief descriptions of each
step are as follows:
Step A - Deposition Load Estimation
First, the sewerage system should be divided into manageable
collection systems, excluding major trunk and interceptor sewers. Estimates
of the daily dry weather sewage pollutant deposition must be prepared for
each collection system. A series of simplified procedures for estimating
gross levels of dry weather sewage pollutant deposits within collection
systems are avail able"and are described in Chapter 13. Date requirements
for these procedures are also given in Chapter 13. Minimal information
required to use these procedures are estimates of the service area, the
average ground slope and the overall per capita waste rate, including
infiltration. Procedures are also available in Chapter 13 for computing
pipe length from service area estimates. These estimates of collection
system deposition loadings do not include any allowances for industrial
wastes deposits. Estimates of significant industrial waste contributions
would be necessary.
299
-------
USE SIMPLIFIED PROCEDURES
IN CHAPTER 13 TO ESTIMATE
DEPOSITION LOADS
CO
o
o
DECIDE WHETHER
FLUSHING IS FEASIBLE
GATHER
PLAN 8 PROFILE MAPS
USE SLOPE CRITERIA
TO SELECT POTENTIAL
FLUSHING SEGMENTS
ALTERNATIVES
COST ANALYSIS
FIGURE 86 CONCEPTUAL OVERVIEW FOR DEVELOPING A SEWER FLUSHING PROGRAM
-------
Step B - Alternatives Cost Analysis
"Desk-top" assessments can then be performed to determine whether
further consideration is worthwhile. Two elements are necessary to roughly
appraise the viability of sewer flushing program. First, the costs of
flushing must be known and, second, the removal effectiveness must be esti-
mated. Flushing costs are given in section 14.4. A rough rule-of-thumb is
$25,000 (present worth: 20 years @ 8%) per flushing site, using either a
manual or automated means.
The extent of a flushing program can be roughly estimated in the
following, manner. Equation 50 in section 13.7.2 can be used to estimate
average collection system pipe slope using an estimate of average ground slope
in the collection system. The estimate of the average collection system pipe
slope can then be used in Figure 82, section 13.5.5 to select a curve depict-
ing the cumulative percentage of mass deposited versus the cumulative frac-
tion of pipe length. The user can select from among a family of curves pre-
pared for different average pipe slopes. The left-hand plot in Figure 87 is
an example of such a curve. The next step is to select the percentage of
mass deposited, PM, to be controlled by sewer flushing. The left-hand plot
can be used to estimate the fraction of pipe length, PL, in the system corres-
ponding to the selected value of PM.
The next step is to determine the effective length controlled by a
single flushing operation. The results of the sewer flushing experiments re-
ported in Chapters 8 through 11 indicated that flushing over 700 feet was ef-
fective in removing more than half of the deposited solids and pollutant
loadings. A predictive procedure was developed in Chapter 12 for estimating
percent mass of solids, organics and nutrients remaining in suspension at
downstream points as a result of a single upstream flush. The model projects
that about 20% of the flushed solids mass would be in suspension at a down-
stream distance of 1000 feet, and also 45% of flushed BOD and 50% of TKN mass-
es would remain in suspension at this distance. A simple rule is to assume
that a single flushing operation could control 800 feet of pipe segment with
a 30% solids removal, and 60% organics and nutrient removals. The number of
flushing sites can then be estimated and the overall cost effectiveness com-
puted. Other combinations of flushing effectiveness and controlled segment
length could be investigated. Other alternatives, such as street sweeping
and storage/treatment, could be prepared and their cost effectiveness computed.
Step C - Decision
If the "desk-top" assessments show that sewer flushing is attractive
then the next important step is to determine those segments in the collection
system where sewer flushing can be considered.
Step D - Plan & Profile Maps
Plan & profile maps of the collection system, should be assembled
and reviewed. Tabulations should be prepared of segment location and pipe
slope.
301
-------
CO
o
a.
i
CD
LU
LU
a.
a.
LU
UJ
o
tr:
LU
|
s
o
100 —
MUST USE FIGURE 82
CHAPTER 13 TO DERIVE
THIS CURVE
PL
CUMULATIVE PERCENTAGE
MASS DEPOSITED
PM 100
5PL
SLOPE
FIGURE 87
DETERMINATION OF THE CUT-OFF SLOPE FOR A
PERCENTAGE OF MASS DEPOSITED
-------
Step E - Selection of Pipe Segments
A rough-cut approach was developed in this study to identify in a
preliminary fashion, potential segments for flushing considerations. Exten-
sive statistical analyses of sewerage system pipe slopes revealed that collec-
tion system pipe slopes can be represented by an exponential probability
model. The exponential distribution can be completely defined using the mean
value. The average collection system pipe slope was computed in Step A from
an estimate of the average ground slope. Equation 51 in section 13.7.4
can be used to define the cumulative function of collection system pipe slope,
FS. A plot of a cumulative pipe slope distribution function is shown on the
right in Figure 87.
The cut-off slope for locating potential segments for sewer flush-
ing can be determined in the following manner. The percent mass, PM, to be
flushed is entered in the plot on the left in Figure 87. The corresponding
percentage of pipe length, PL, associated with PM, is then read off the curve.
The value of PL is then entered into the ordinate scale of the cumulative
pipe slope distribution function on the right side of Figure 87, and the cut-
off slope, Sp|_, is determined. The percentage of total pipe length in the
basin with pipe slope smaller than or equal to SPL is therefore known. Com-
bining the two parts in Figure 87 does not necessarily mean that the pipes
over which PM deposits all have slopes flatter than or equal to Sp[_. This
may be a reasonable working assumption and, if it is made, locations of de-
positing segments can be determined. This step can be accomplished by noting
on a sewerage system map all pipe segments with slopes equal to or smaller
than Spi_. This procedure will quantify the sewer segments associated with
the percentage PM of the total load.
Step F - Inspection
Once the segments have been identified through analysis of sewer
slope maps, they should be physically inspected to ascertain existing deposi-
tion levels and whether mechanical or water jet cleaning is necessary. Man-
holes should be inspected for soundness to determine whether remedial repairs
such as rebuilding the manhole table and/or replastering manhole walls are
necessary.
Step G - Flushing Method
The final decision is the choice of the flushing method. Procedural
guidelines useful in weighing the pro's and con's of different approaches are
discussed in section 14.5.
14.4 Sewer Flushing Costs
Realistic sewer flushing program costing information from full-
scale experience is non-existent. Preliminary "desk-top" cost estimates were
prepared by FMC (12). Results from a recently completed combined sewer man-
agement study in Fitchburg, MA are presented here (5,6). These estimates are
again only conjecture at this point in time. The sewer flushing alternative
in that study entailed flushing daily 46 combined sewer small diameter
303
-------
(10-18 inch), pipe segments. The flushing program was costed using manual tanker
operation and automated modules similar to the device successfully operated
on Shepton Street described in Chapter 11. Further testing of such devices
is necessary before recommendation for general application can be made, be-
cause no full-scale network of automated sewer flushing modules have ever
been tried in practice. Costs for this type of operation are nevertheless
presented for comparison with a manual operation using water tankers.
Initial costs included $1500 per flushing segment for site prepara-
tion work, which entailed repairing and, in some cases, installing manhole
tables, grouting the flush manhole walls, mechanical cleaning and jetting of
the flush segment and other miscellaneous contingency items. Fabrication and
installation of the devices was estimated at $7,500 each. A three-man main-
tenance crew was assumed to ensure proper operation of the configuration of
the 46 modules. Equipment component replacement costs of $300 per module per
year were also assumed. A contingency factor of $10,000 per year was included
to cover possible blockages and malfunctions. The estimated annual budget
for this operation is $80,000. The O&M cost per module per year is $1630.
The present value of O&M cost per module, assuming a 20-year discount period
at 8 percent interest, is $18,745. Total present worth cost per module is
$27,745. These costs (ENR = 2,000) are summarized in Table 80.
Alternative program costs consisting of manually flushing 46 seg-
ments, using three water-tankers and a six-man labor force, are shown in
the second half of Table 80. Total present worth costs for this alternative
are $24,880 per segment.
One of the major recommendations of this report is to perform an
additional study aimed at evaluating long-term operational performance of
a combination of different automated flushing techniques. An ancillary aim
of the envisioned study is to develop realistic operational costs associated
with sewer flushing.
14.5 Sewer Flushing Techniques - A Discussion
The following question arises: if a sewer flushing program were to
be instituted within a given community, what would be the best approach for
the given situation? Data and experiences generated during this study pro-
vide a general format for reaching answers to such questions. The following sub-
section is aimed at presenting generalized concepts, as well as known prop's
and con's, of various flushing approaches.
Basically, sewer flushing as tested in this study consists of in-
ducement of turbulent flow shock waves along a given segment which will tend
to resuspend previously deposited solids with a number of mechanisms. The
flush wave itself may be pictured in a simplistic sense as a rolling, gyrating
ball that tends to push'materials on its front side and pull materials as it
passes. The severity of the roll and bounce tends to dictate the fractions
of materials from heavy sand to light organic particles that will be resus-
pended by the wave and carried along for some distance. As the turbulence
and speed decrease, the materials resuspended or carried within the flush
304
-------
TABLE 80. ESTIMATED COSTS OF SEWER FLUSHING METHODS
NUMBER OF SEGMENTS: 46
DAILY FLUSHING PROGRAM
ALTERNATIVE 1 - AUTOMATIC FLUSHING MODULE OPERATION
CAPITAL COST:
- Site Preparation (Grout Manhole, Fix Base,
Clean Segment)
- Fabricate & Install Air-Operated Module
$1,500/Segment
7,500
$9,000/Segment
ANNUAL OPERATIONAL COST (TOTAL PROGRAM);
- 3 Men @ $15,000/Yr.
- Truck Rental, Gas, Insurance
- Equipment Component Replacement
$300/Yr/Module
- Sewer Cleaning Contingency
- Water
COST/MODULE/YR = $1,630
PRESENT WORTH/SEGMENT = $18,745
TOTAL PRESENT WORTH/COST/MODULE:
- First Cost
- O&M Cost
ALTERNATIVE 2 - MANUAL FLUSHING MODE
CAPITAL COST:
- 3 Outfitted Water Tankers @ $18,000
or ,$1100/Segment
OPERATIONAL COST (TOTAL PROGRAM):
- 6 Men @ $15,000/Yr.
- Insurance, Gas, Maintenance
- Water
COST/SEGMENT/YR. = $1938
PRESENT WORTH/SEGMENT = $22,287
45,000
8,000
14,700
10,000
2,300
$80,000
9,000
18,745
$27,745
$54,000
90,000
3,000
2,000
$95,000
TOTAL PRESENT WORTH COST/SEGMENT:
- Site Preparation
- First Cost
- O&M
USE $25,000/SEGMENT (ENR = 2000)
1,500
1,100
22,290
$24,880
305
-------
wave tend to become lighter with the heavier sand and gravel particles
dropping out.
Two factors control the removals and carrying potential of induced
flush waves. These are: the flush input rate and the flush volume. Rate
may be roughly equated with wave turbulence, while volume roughly equates
with long-term carrying capacity and momentum. Contrary to classical hydrau-
lics theory, the energy dissipation of the flush wave is not rapid, maintain-
ing high degrees of turbulence for at least 300-400 feet (92-122 m). The re-
sults of this study presented in Chapter 8 tended to indicate that removals
were slightly more volume-dependent than energy or input rate dependent, but
a combination of both proved best. This statement basically pertains to man-
ual methods flushing, although if jetted modules were developed the statement
would pertain to them as well. Large volume flushing via backup and release
methods as tested show comparable removals to jetted flushes.
Selection of the optimum flushing approach or mixed approaches is
dependent on many factors. The following is a compilation of the authors'
thoughts and ideas as to the pro's and con's of various flushing approaches.
Sewer flushing programs have to be tailored to suit given situations. The
purpose of this section is to present thoughts and ideas based on considerable
field experience.
1) Sewer flushing is a mechanism to minimize, to the extent practicably
possible, deposited solids within combined sewer systems that are subject to
resuspension and potential overflow during storm events.
2) Sewer flushing is not meant to be a complete replacement for on-
going cleaning programs, in that no flushing method is truly effective with
respect to grit removal. It has been the authors' experience that once
cleaned, sewer flushing will maintain and sometimes further clean combined
sewer laterals over extended periods of time. Flushing frequencies on the
average of once every 3-4 days seem sufficient for previously highly deposit-
ing problem segments. From a pollution standpoint, it would seem that flush-
ing more frequently would be more desirable, since the return period of light-
moderate storms may be less than 3 to 4 days. Some preliminary evidence show-
ing that sanitary organic deposits are not readily suspended and transported
during moderate rain events is presented in Chapter 11. This evidence sug-
gests that the intervals between flushing could be longer and not necessarily
determined solely by rainfall considerations. More evidence is needed to
substantiate this observation.
3) Sewer flushing can and should be incorporated as part of ongoing
sewer maintenance programs to minimize additional cost, while maximizing the
maintenance program effectiveness.
4) Sewer flushing can be effectively accomplished fay manual or auto-
mated.means. Manual methods utilizing a pressurized water tank truck system
have been demonstrated to be viable and readily accomplished. On the other
hand, flushing by automated devices is in its infancy. This study represents,
to the authors' knowledge, the first real field application of an automated
mechanical flushing device other than automated siphon manholes (39).
306
-------
5) The choice of flushing by manual or automated means (assuming auto-
mated systems were perfected) poses the issue of capital vs labor intensive
programs. Present funding mechanisms heavily favor capital investments,
although the situation might change.
6) Widespread application of automated sewer flushing techniques is at
this point unrealistic. Much development needs to be done, sensing units
and mechanisms perfected and tested in an extensive program before solely
automated flushing can be considered viable.
7) Automated flushing has several potential advantages: increased
flushing frequency, possibility for centralized control and considerable
control flexibility.
8) Flushing by automated means could have several significant disadvan-
tages, including: potential for sticks and rags blocking devices infrequently
maintained, lesser degree of "hands on maintenance" and observation of the
sewer system, and complex repair problems.
9) Manual or automated flushing could readily be accomplished utilizing
sewage, natural waters or even industrial wastewaters, if suitable, in water-
scarce areas. Tankers could potentially be filled with strained sewage.
10) In areas where travel times are short, night flushing may be a good
way of load balancing the wastewater treatment plant. Flushing sewers at
night would be beneficial only if the time of travel in the collection
system were short enough to move the late evening deposits all the way to the
treatment facility before the next peak. This approach method would be
appropriate if the branching order of the collection system were low, as
occurs in most small river front towns, but not appropriate if the collection
system were broad, fan-shaped with high branching order (and large area!
depth). In general, since the tractive shear stresses are greater during
higher flow conditions, it seems more prudent to flush upstream laterals
mid-day or peak flow diurnal hours. The idea is that the suspended deposits
would pass through uninterrupted zones of high tractive shear stress
maximizing the flush travel.
11) Flushing of long, flat.segments 1000 feet (305 m) or greater could
require additional downstream or booster flushes to keep substantial deposits
from accumulating if the downstream segments do not carry enough background
sewage to provide solids carrying capacity. Intermittent upstream storage
modules may also be considered prior to locations where solids redeposition
occurs. At this time, not enough is known to define the exact location of
such booster flushes.
12) Choice of proper flushing rate/volume for a given segment is some-
what dependent on the segment characteristics, 'that is,.size, shape, slope
and nature of sediments. The' results of this study indicated that flush
volumes between 35 and 50 cf (.99' - 1.98 cubic meters) at discharge rates
in excees of 0.5 cfs (14.16 liters/sec) would be sufficient for most
combined sewer laterials.
307
-------
SECTION 15
ASSESSMENT OF TREATMENT COSTS WITH SEWER FLUSHING
15.1 Foreword
In this chapter two separate desk-top analysis dealing with the
economic implications of sewer flushing programs on wastewater management are
presented. The first analysis deals with estimation of the additional dry
weather sewage treatment operational and maintenance costs resulting from
increased pollutant loadings reaching the treatment plant during dry weather
as a result of sewer flushing. Annual cost increases are computed for three
different types of commonly encountered treatment flow sheets under seven
different levels of design flow. Types of sewage treatment plants considered
in this analysis include: primary, trickling filter and conventional
activated sludge. Design discharge plant sizes considered in this analysis
are 1, 5, 10, 25, 75 and 100 mgd. The second analysis deals with the
implication of sewer flushing in regards to wet weather storage/treatment
costs.
15.2 Dry Weather Analysis
Operational costs of wastewater treatment depend in part, on the
amount of pollutants requiring removal. Sewer flushing programs on any
significant scale would increase on the average, sewage strength beyond
levels most often used in design. For example, additional suspended solids
loadings would increase both the cost of wastewater treatment and to a great-
er degree, the cost of wastewater solids treatment. The objective of this
analysis is to roughly compute the rise in annual operational costs resulting
from differing yet reasonable sewer flushing program levels for hypothetical
communities of different sizes.
To establish these additional costs, a range of treatment plant
types will be first sized at different design flows, and the operation and
maintenance cost of their unit operations found. Cost models will then be
developed in the form of cost per unit loading versus design flow for each
unit operation. Loading parameters include the pollutant that has the
greatest impact on the cost of the operation.
If the amount of pollutant generated by the flushing is known, then
the extra treatment cost could be estimated with the cost models in this
form. The amount of pollutants flushed from the sewerage system tributary to
a particular size plant will be crudely estimated using results of the sewer
flushing experiments given in Chapter 8 and various statistical relationships
between demographic and sewershed characteristics.
308
-------
Total pipe length is estimated in the following way. First,
population is estimated from an empirical relationship between plant waste-
water flow and per capita waste flow rates (40). Next, population density is
computed from an empirical relationship between population and population
density (41). Finally, total pipe footage is computed from an empirical
relationship between population density and pipe footage (42). Thus a
particular treatment plant size is related to an estimate of total pipe
footage of the hypothetical sewerage system tributary to the treatment plant.
Experience in assessing sewer flushing potential for several
combined sewered communities in the Boston area indicates that most dry
weather deposits of any significance lie roughly within 10 to 20 %
of the total collection system pipe footage (4,6,13). In this analysis two
hypothetical sewer flushing programs are considered that would flush 100
percent of the estimated loadings assumed to deposit in the range of
collection system pipe footage cited above. It is further assumed that the
amount of flushed solids and BOD loadings reaching the treatment plant would
be 25 to 50 % of the estimated flushed loadings. This assumption
derives from the conclusions of the settleability analysis of flush waves
presented in Chapter 10. It is also assumed that flushing would be
conducted by either automated or manual means using sewage as the flushing
source. No additional hydraulic loading to the treatment plant is considered.
Additional pollutant loadings resulting from the two hypothetical
flushing operations are crudely estimated using 10 and 20 % of the
total estimated pipe footage and the average pollutant removals in pounds
per day per foot of pipe, scaled by the aforementioned transport efficiency
factors. It is assumed that the deposition loadings and flushing removals
characteristics of the four test segments in Dorchester would be typical of
all pipe segments where flushing is assumed in this analysis.
The estimated flushed pollutants are then used in the cost models
to determine additional operation and maintenance costs. These costs are
modified by utilization factors to account for the increased pollutant
loadings since marginal treatment costs will change.
In this analysis two sets of cost models are used. One set of
models will be used for estimating nominal operation and maintenance costs
of unit operations that are used in primary, secondary, and advanced
treatment. The other set of models will be for estimating the additional
operation and maintenance costs incurred by flushing programs for primary
and secondary plants. Flow sheets for the primary, trickling filter and
conventional activated sludge plants considered in this analysis are
shown in Figure 88.
15.2..1 Discussion of Pertinent Wastewater Treatment Plant Cost Models
Several empirical models were reviewed that present overall
operation and maintenance costs for a given type of treatment plant (43, 44,
45, 46, 47). These models were developed from regression analyses of actual
treatment plant audits. Models that estimate cost for specific unit process
309
-------
Raw
Wastewater
Influent Pumping
Preliminary
Treatment
Primary
Clarifier
Chlorinatfon Effluent
—tM->
PRIMARY PLANT
Sludge Holding
Tank
Wastewater Preliminary
Influent Pumping Treatment
Primary
Clarifier
Trickling
Filter
Secondary
Clarifier Chtori
v
Effluent
Recirculation & Sludge Return
TRICKLING FILTER PLANT
Sludge Holding
Tank
Wastewater Preliminary
Influent Pumping Treatment
Primary
Clarifier
Aeration
Tank
Secondary
Clarifier
Chtorination
Effluent
itage II StageM
Anaerobic
Digesters
Sludge
Holding
Tank
CONVENTIONAL ACTIVATE!)
SLUDGE PLANT
Sludge To
Landfill
FIGURE 88: FLOWSHEETS OF PRIMARY TRICKLING FILTER AND
CONVENTIONAL ACTIVATED SLUDGE TREATMENT PLANTS
310
-------
are more useful, since the flushed loadings have different cost implications
for the various unit operations in treatment plants.
The methodology developed by Van Note (48) costed individual
processes for primary, secondary and tertiary treatment plants. The costs
of treatment are divided into unit operation costs. Each unit operation is
depicted on a unit operation sheet which consists of the process flow diagram
the equipment and sub-unit operations making up the unit operation, the
influent and effluent wastewater quality, and the design parameters used in
the unit operation. Associated with each unit operation sheet are cost
formulae for variable and fixed operation and maintenance costs.
W. L. Patterson and R. F. Banker (46) provide a more refined
procedure in that the treatment plants are subdivided into their basic unit
operations and costs are provided for each unit operation. In their
work, operation and maintenance labor and material costs are presented in
graphical form. Furthermore costs for unit processes were plotted against
design parameters. For example, sedimentation cost was plotted against
surface area, and vacuum-filtration cost was plotted against dry solids
filtered. The costing methodology used in this study is the following. Suita-
ble design criteria were used to size the individual unit operations for a
given treatment flowsheet, so that the required independent variables were
obtained'for the cost models. The cost models in graphical form were then
used to find the annual labor manhours and the material and supply costs.
Appropriate estimates of manpower cost factors were.then used to determine
annual labor costs.
15.2.2 Development of Unit Operation - Operational Cost Models
Primary, Trickling Filter ajad Conventional Activated Sludge Plants.
Utilization of the operational cost methodology developed by Patterson and
Banker requires that pertinent design variables be specified for each unit
process. Several commonly used design criteria (50, 51, 24, 52) were used
to estimate these unit process design variables for each plant type and for
each design flow. These variables are as follows:
Operation
Raw waste pumping
Preliminary treatment
Sedimentation
Trickling filter
Aeration-diffused system
Recirculation or inter-
mediate pumping
Chlorination
Primary sludge pumping
Sludge holding tank
Variable
average wastewater flowrate pumped
(mgd)
average plant flow (mgd)
2
surface area of clarifier (ft )
p
surface area of filter (ft )
blower capacity (cfm)
pumping capacity (mgd)
chlorine used (ton/year)
pumping capacity (gpm)
sludge volume (cf)
311
-------
Sludge digestion sludge volume (cf)
Vacuum filtration-sludge dry solids filtered (ton/year)
to landfill
Various assumptions used in the preliminary treatment plant designs are
cited in Table 81.
Estimates of sewage strength as a function of design flow were
derived from several empirical relationships given by Michels et al. (40).
These results are shown in graphical form in Figure 89 and are used in this
analysis. The BOD and TSS curves were sufficiently close so that one curve
was used for both parameters. These results were used in the preliminary
design analysis" for establishing sewage strength as a function of design flow.
Design variables for each unit process were then used in Patterson's
and Banker's procedure to obtain annual payroll manhours and material and
supply costs. The earnings for nonsupervisory workers (53) in December 1976
was $190.00 per week or $4.75 per hour based on a forty-hour week. This
figure was increased to $5.00 per hour as some of the workers are supervisory.
Indirect labor cost was taken as fifteen percent of the direct labor cost.
Material and supply costs cited in Patterson's and Banker's work for
January, 1971, were adjusted using a trend factor to December 1976 level.
The trend factor used is the Wholesale Price Index (54) for industrial
commodities for December 1976 of 187.4 referenced to the January, 1971
index of 112.2.
Operational cost estimates for each unit process for all plant
types and scale, together with estimates of intra-process applied pollutant
loadings derived from the preliminary plant designs, were used to develop a
new set of operational cost curves directly relating cost to applied loadings
for different design flows. The various unit operation cost functions and
associated cost parameters developed in this analysis are summarized in
Table 82 for each of the three treatment plant types. In sum, twenty-five
operation and maintenance cost models were prepared. The sludge pumping
cost functional was later modified to reflect unit cost per pounds of sludge
settled.
Factors used to compute the applied loadings onto each unit process
from the plant influent solids loadings and BOD loadings are given in
fractional form in Table 83.
15.2.3 Operational Unit Operation - Cost Model Modifications
The operational cost models described in section 15.2.3 are in the
form of cost per applied loading versus average design flow. The additional
dry weather operation and maintenance cost resulting from sewer flushing can
be computed using these procedures. The cost models were however, modified
to reflect the fact that additional pollutant loadings would change plant
utilization and hence marginal unit-operation treatment costs.
312
-------
TABLE 81. ASSUMPTIONS USED IN PRELIMINARY DESIGN OF
PRIMARY, TRICKLING FILTER AND ACTIVATED SLUDGE PLANTS
Primary Plants
. Sedimentation overflow rate =
. Peak flow factor =
. Suspended sol ids removal by sedimentation =
. Primary sludge solids concentration =
. Sludge pumping =
. Sludge 'holding tank detention =
. Chi orination dose =
. Two stage digestion detention =
. Mean solids concentration in secondary digester
. % Volatile to digesters =
. Reduction volatile in digesters
Trickling Filter Plant
. Recirculation ratio =
. BOD removal in primary sedimentation =
. Applied organic load to filter =
. Depth of filter =
. Sludge generation (% plant influent solids) =
. Raw solids sludge concentration =
. Chorination dose =
. All other parameters same as primary plants
Conventional Activated Sludge Plant
. Applied BOD loading aeration tank =
. Mixed liquid suspended solids =
. Aeration tank detention =
. Sludge age =
. Return sludge concentration =
. Air requirement =
. Sludge generation (% plant influent solids) =
. Chlorination dose =
. All other parameters same as primary plants
700 gpd/fT
2.6
55%
5%
4 hours/day
5 days
10 mg/1
14 days/25 days
8%
70%
55%
.75
35%
45 Ib BOD/1000 cf /day
6 feet
85%
4.5%
8 mg/1
35 Ib/lOOOcf/day
2500 mg/1
5 hours
8 days
10,000 mg/1
1500 cf/ Ib BOD
90%
5 mg/1
313
-------
100
.1
.2
.3
.4
.5
.6
i i i i i i T r
Lbs/Cap/Day BOD & SS
50
30
a
0
!
10
«/>
0
111
3
&
Ul
5
Lbs/Cap/Day
BOD & SS
B MS/L
° BOD & SS
C Gal/Cap/Day
i i
100 200
Mg/L BOD, SS, AND FLOW/CAPITA/DAY
3QO
FIGURE 89. SEWAGE STRENGTH VERSUS DESIGN FLOW
314
-------
TABLE 82. PARAMETERS FOR OPERATION AND MAINTENANCE COST MODELS
PRIMARY, TRICKLING FILTER AND CONVENTIONAL ACTIVATED SLUDGE PLANTS
.OPERATION
Wastewater Pumping
Primary Sludge Pumping
Recirculation Pumping
Sedimentation
Sludge Digestion
Vacuum Filtration
Sludge Holding Tank
Trickling Filter
Activated Sludge Aeration
Aeration
Chlorination
Yardwork
Laboratory
Administration and
General
COST PARAMETER* ALL PLANTS
Cents/1000 gallons X
Cents/1000 gallons
Cents/1000 gallons
Cents/1 b sludge settled
Cents/1 b sludge applied to
digestion
Cents/1 b sludge filtered
Cents/1 b sludge applied to
tank
Cents/1 b BOD applied to
filter
Cents/Ib BOD applied to
aeration
Cents/1000 gallons X
Cents/1000 gallons X
Cents/1000 gallons
Cents/1000 gallons X
TRICKLING
PRIMARY FILTER
X X
X
X X
X X
X X
X X
X
x **
ACTIVATED
SLUDGE
X
X
X
X
X
X
X
X
CO
_J
en
*Cost model in form of cost parameter versus average design flow (mgd).
icic
Laboratory cost function for trickling filter same as for Primary Plant.
-------
TABLE 82. PARAMETERS FOR OPERATION AND MAINTENANCE COST MODELS
PRIMARY, TRICKLING FILTER AND CONVENTIONAL ACTIVATED SLUDGE PLANTS
.OPERATION
Wastewater Pumping
Primary Sludge Pumping
Recirculation Pumping
Sedimentation
Sludge Digestion
Vacuum Filtration
Sludge Holding Tank
Trickling Filter
Activated Sludge Aeration
Aeration
Chlorination
Yardwork
Laboratory
Administration and
General
COST PARAMETER* ALL PLANTS
Cents/1000 gallons X
Cents/1000 gallons
Cents/1000 gallons
Cents/1 b sludge settled
Cents/1 b sludge applied to
digestion
Cents/1 b sludge filtered
Cents/1 b sludge applied to
tank
Cents/1 b BOD applied to
filter
Cents/Ib BOD applied to
aeration
Cents/1000 gallons X
Cents/1000 gallons X
Cents/1000 gallons
Cents/1000 gallons X
TRICKLING
PRIMARY FILTER
X X
X
X X
X X
X X
X X
X
x **
ACTIVATED
SLUDGE
X
X
X
X
X
X
X
X
CO
_J
en
*Cost model in form of cost parameter versus average design flow (mgd).
icic
Laboratory cost function for trickling filter same as for Primary Plant.
-------
TABLE 83. FACTORS USED TO COMPUTE APPLIED LOADINGS PER UNIT PROCESS
Influent Susp
Solids Load
inasd Influent BOD Loadings
Plant Type Plant Type
ABC ABC
Preliminary Treatment
Sedimentation
Trickling Filter Bed
Activated Sludge Aeration
Sludge Pumping
Sludge Holding Tank
Anaerobic Digesters
Vacuum Filters
A - Primary Treatment Plant
B - Trickling Filter Plant
C - Conventional Activated
1.
.55
.55
.55
.55
.34
Sludge
1.
.85
.85
.85
.49
Plant
1. - - -
.9 - - -
- .65 -
- .65 -
.90 - - -
.90 - - -
.90 - - -
.55 - - -
The results of an analysis (55) depicting the variation
operation and maintenance cost per unit treated for discharge levels other
than design flows for 2.5 and 10.0 mgd primary, trickling filter and
activated sludge plants were used to prepare Figure 90. The curve shows how
the utilization factor changes with plant utilization. The utilization
factor is the cost per unit treated at a particular utilization divided by
the cost per unit treated at 100% utilization. It was assumed that this
formulation would be applicable for approximating marginal cost changes under
conditions of only increased pollutant loadings since the sewer flushing
programs would not increase hydraulic loadings. Utilization factors
derived from this curve were later used in estimating the additional
flushing costs to modify the operation and maintenance cost model estimates
which were prepared for 100% utilization.
15.2.4 Estimation of Additional Treatment Costs
Total collection system footage tributary to each of the seven
different sized treatment plants were estimated in the following way.
First, Figure 89 was used to relate design flow to per 'capita waste rate
for an estimated population. Population density for communities served by
combined sewers was computed using the following equation 41 :
304
Population Density (Capita/Acre) = 0.3 (Population)'
Footage of installed sewers per capita (42) was computed from population
density using the following equation:
Feet of Installed Sewer/Capita = 54 (Capita/Acre)"0'65
Additional treatment plant pollutant loadings resulting from the
sewer flushing programs were computed in the following manner. First,
an estimate of the daily flushed load per foot of sewer was derived by
averaging the pollutant removal results of the four test segments in the
316
-------
150
140
130
120
110
100
90
80
70
60
50
O
N
^
I—
3
.4
UTILIZATION FACTOR
.6
.8
1.2
1.4
1.6
1.8
FIGURE 90. TREATMENT PLANT UTILIZATION VERSUS UTILIZATION FACTOR
3T7
-------
first phase.. The average normalized pollutant mass removals in pounds per
day for each street were divided by test segment footages. These mean results
per street were then averaged to yield the factors used in this analysis.
The overall average flushed TSS, VSS and BOD loadings used in this analysis
are 0.0205, 0.0143 and .0084 Ibs/foot of sewer flushed/day. These factors
were then adjusted by a factor of 22% for suspended solids and 45% for BOD to
reflect the findings of the settleability experiments.* These factors were
then multiplied by 10 and 20% of the total _estimated collection system foot-
age tributary to each of the seven different size treatment plants.
Additional operation and maintenance costs were computed for each unit
operation within each of the three treatment plants for the seven design
flow conditions using the estimated dry weather loadings resulting from the
hypothetical sewer flushing programs. The unit process operation and
maintenance cost model estimates described in section 15.2.3 were modified
to reflect increased plant loading utilization using the utilization versus
utilization factor curve given in Figure 90. Nominal plant utilization
for pre-flushing conditions was assumed at 100%.
Summary cost results for the three treatment plants types are
given in Table 84. The estimated additional annual operation and
maintenance cost resulting from the increased pollutant loadings from sewer
flushing programs for the seven design flow conditions are presented along
with the relative increase in the total estimated annual yearly plant
operation and .maintenance cost. The total annual yearly cost included
wastewater pumping, chlorination yardwork, administration and laboratory
operational costs besides the operational costs for those particular unit
processes affected by the increased solids and organic sewer flushing
loadings. The results show that the two hypothetical sewer flushing
programs would have neglible impact on the operational cost of dry weather
sewage treatment. The cost increases range from 3 to 6% for the 1 mgd
plant and decrease to around 1% for larger sized plants.
The distribution of the additional operational costs for the
various unit operations relative to the total additional increase revealed
that the vacuum filtration units for the primary and trickling filter plants
exhibited the highest overall increase, averaging about 42%. This is not
surprising since this process has a high chemical and power cost. The
largest relative cost increase in the conventional activated sludge plant
occurred for the aeration tank which ranged from 25 to 35 % depending on
the plant size. The vacuum filtration operational cost increase for the
activated sludge plant about equalled the aeration tank cost increase.
*
These estimates were crudely taken as the percent TSS and BOD remaining
in suspension 800 feet downstream from a flushing manhole. See Figure 70,
Chapter 12.
318
-------
TABLE 84. SUMMARY OF ADDITIONAL TREATMENT OPERATION AND MAINTENANCE COSTS
FOR PRIMARY, TRICKLING FILTER AND ACTIVATED SLUDGE PLANTS
DESIGN FLOW
(mgd)
A. Flushing Program
(10% Segments)
1
5
10
25
50
75
100
B. Flushing Program
(20% Segments)
1
5
10
25
50
75
100
PRIMARY TREATMENT PLANT
A B
1820
2060
2375
3090
3950
4700
5090
3350
3990
4380
6060
7750
9300
10070
3.4
1.4
1.0
0.7
0.5
0.5
0.4
6.2
2.8
1.9
1.4
1.0
1.0
0.8
TRICKLING FILTER PLANT
A B
2410
2970
3420
4480
5810
6400
-
4460
5760
6640
8800
11390
12660
-
3.3
1.5
1.1
0.8
0.6
0.5
-
6.1
3.0
2.2
1.5
1.1
1.0
-
A
3020
4070
4925
6260
8300
9100
10600
5620
7910
9550
12260
16315
18470
21020
B
3.0
1.7
1.3
0.9
0.7
0.6
0.5
5.7
3.3
2.5
1.7
1.3
1.1
1.0
to
to
A - Additional Annual Operation & Maintenance Treatment Costs Resulting From Flushing Programs,--(I/year)
B - Percentage of Increase Relative to Normal Total Operation and Maintenance Cost
-------
15.3 Wet Heather Considerations
The results of the sewer flushing R&D program conducted over the
last two years clearly showed that flushing combined sewer laterals definitely
removed and transported on a remarkedly consistent basis pollution related
contaminants from both single pipe segments and from serial segments. The
program did not, however, document the long-term mass pollutant removal
reduction from a reasonably s'ized collection system. It can be inferred,
however, that if pollutant removals over, say, a 1000 feet stretch of lateral
were documented (which they were), and if these laterals discharged into
trunk sewers with adequate shear stress transport capacity, then in a hypo-
thetical sense, removal of dry weather pollution deposition loadings on a
larger scale could in fact be projected with" reasonable certainty. Intense
sedimentation can still occur in -flat trunk sewers.
If, however, sewer flushing would only partially transport pollu-
tants down through a system, that is, stopping short of the sewage treatment
facilities, there is still a strong benefit -occurring from even this limited
effectiveness. Except for extreme storm events characterized by intense
frontal waves, "first flushes" would have a higher likelihood of being trans-
ported either to the treatment headworks before overflows occur and/or to
storage/treatment capture points. Storage designed to capture "first flushes"
would invariably be more effectively utilized and, accordingly, the costs
would be reduced.
320
-------
SECTION 16
USER DETAILS
16.1 Foreword
This chapter contains examples of two procedures briefly described
earlier in this report, represented with more explanatory detail* The
first procedure is the multi-segment dry weather sewage deposition model de-
scribed in section 12.4. An example of the computations is presented using
the Walnut Street collection subsystem to illustrate the computations. Details
of this example are provided in section 16.2. The second procedure described
in this chapter is the application of the empirical model described in section
12.2 for predicting the amount of pollutants remaining in suspension at
downstream points resulting from an upstream flush. This procedure is de-
scribed in section 16.3. This procedure is extremely useful in the
estimation of flushed pollutant loads along a small subsystem. Its use on
a broad spatial scale is not recommended.
16.2 Application of Multi-Segment Deposition Model
A description of the deposition model fundamentals and a short de-
scription of a hypothetical example illustrating some of the aspects of the
methodology used in the model were presented in section 12.4. For this sec-
tion, a more detailed example is presented with equations and computation of
numerical- values. In presenting this example the output of one of the runs
made for the Walnut Street test segment is described and some of the numerical
values shown in the output table are verified by manual computations using
the appropriate equations, as described in section 12.4. It should be noted
that the manually computed numbers will not always match exactly those shown
in the table, due to differences in the number of significant digits used
in the manual computations and those used by the IBM 370/168 which computed
the table.
Figure 91 presents a schematic of the Walnut collection system.
Table 85 is a sample output of the deposition model, computed for a per capita
waste flow rate of 100 gpcd. The columns labeled A through R in Table 85 are
sufficiently clear so that a description of each column is not necessary.
The schematic and some of the .labeled columns shown in the table will be
used as needed in illustrating the series of computational steps described in
section 12.4.3. The steps described below follow closely the steps presented
in section 12.4.3.
*
Pertinent equations are repeated here for convenience.
321
-------
o
12"
(5)
12'
0 •
7)
O
12"
(48)
(4)
15"
(0)
0
15"
(14)
O
LEGEND
(7) LINK NUMBER
(5) POPULATION CONTRIBUTING TO LINK
12" PIPE SIZE (CIRCULAR)
FIGURE 91 SCHEMATIC OF COLLECTION SYSTEM - WALNUT STREET
322
-------
CO
I\D
CO
LOCATION t
TABLE 85. EXAMPLE OF APPLICATION OF DEPOSITION MODEL
PER CAPITA»100.
SEWER FLUSHING RESEARCH - DEPOSITION ANALYSIS t WALNUT STREET (COMBINED SEWER)
• DENOTES TRUNK LINE
A
B
C
D E
F
G
H
HYDRAULIC CHARACTERISTICS
COttP.
NO.
* i
* 2
3
* 4
* 9
SITE
WALNUT
WALNUT
SQUARE
WALNUT
WALNUT
SLOPE
0.0028
0.0027
0.0192
0.0048
0.0013
WIDTH HEIGHT
(IN) (IN)
12.00 0.0
12.00 0.0
12.00 0.0
15.00 0.0
15.00 0.0
LENGTH
(FT)
82.
259.
184.
141.
136.
TYPE
CIRC
CIRC
CIRC
CIRC
CIRC
POP.
SERVED
5.
53.
4.
57.
71.
I
FLOW
QFULL
(CFS)
1.64
1.61
4.29
3.90
2.05
J
ANALYSIS
QAVE
(CFSI
0.001
0.008
0.001
0.009
0.011
K
QHAX
(CFS)
0.001
0.011
0.001
0.012
0.015
L
-— -
CEPTH
(INCH!
0.3
0.8
0.2
0.7
1.0
M N 0 P Q
DEPOSITION ANALYSIS
SHEAR PERCENT LOAD/DAY SUM LOAD OVERALL
(PSFI DEPOSIT ^f PER DAY FRAC.DEP
0.003 40.00 1.00 1.00 0.400
0.007 20.39 4.89 5.89 0.222
0.010 12.99 0.26 0.26 0.130
0.011 12.32 0.0 6.15 0.216
0.004 35.63 6.60 12.76 0.359
* lbs./day
R
— ...
DEC
OEP
HIGH
HIGH
NOD
NOO
HIGH
-------
1.
2.
The collection system was segmented into 5 links, as shown in Figure
91, each link corresponding to one actual pipe.segment between two
consecutive manholes;
A list of downstream links that convey wastes from each indicated
link is as follows:
Link
1
2
3
4
5
Downstream Conveying
Links
2, 4, 5
4, 5
4, 5
5
-
3. The cumulative population at the downstream end of each link is
presented in column A of Table 85;
4. Using the cumulative populations of column H and the per capita
waste flow rate of 100 gpcd, the average daily dry weather flows at
the downstream end of each link, were computed, expressed in cfs, and
are presented in column J of Table 85. The numbers shown in column
J were rounded off for printing purposes only. Considering, for
example, link no. 5, the daily average flow in cfs can be estimated
as:
71 persons*x 100 gpcd x 0.1337 ft3/gallon * 86400 sec = 0.011 cfs
day
5. The maximum daily dry weather flows, QMflY at each link were compu-
ted using equation:
Q
MAX
!MAX
= aPP
-b
(53)
The computed values are presented in column 11 of Table 85. The
values of the coefficients a and b above were fixed at:
a = 1.157
b = -0.08
and PP is the cumulative population at the downstream end of the
link, in thousands. The value of b was established in a previous
study conducted in the Dorchester Bay area (4). The value of a
was adjusted so that the peak coefficient at the downstream end of
link 5 equals the average value of 1.43, determined from data at
that point, and presented in Table 65, Chapter 12.
324
-------
Again, for link no. 5, the peak daily flow computed by equation 53 is:
QMAX = a * PR-b * QAV or
QMAX = 1>157 * (°-071)"' * 0-011 = 0.0157 cfs
6. The shear stress is computed by the equation:
T = prs (54)
where: p = specific weight of water
r = hydraulic radius (r=flow area/wetted parameter)
s = energy slope
Equation 54 requires knowledge of the hydraulic radius corresponding
to the maximum daily dry weather flow. An iterative procedure is
used to compute the uniform flow depth corresponding to the maximum
daily flow rate, given the pipe shape, size and slope. The flow
depth is then used in computing the hydraulic radius, r.
The iterative procedure used in the model determines flow depths
corresponding to the estimated maximum daily discharges for each
link, which are presented in column L of Table 85. Using link no. 5
once more as a numerical example, the depth of 1 inch of flow in a
15-inch diameter pipe implies a hydraulic radius of
r = 0.0538 ft.
Therefore, the shear stress can be estimated by equation 54 as:
Jl
ft
T = 62.4 lb/ft3 * 0.0538 ft * 0.001335 ft = 0.00448 lb/ft2
7. The deposition rates are computed as a function of the shear stress
by equations 55 and 56, which are:
Z = 40 (0-504)~]'2 for T > °-004 Psf (55)
Z = 40 for T < 0.004 psf (56)
The suspended solids deposition rates estimated by the equations
above are presented in column N of Table 85. In verifying the
deposition rate for link no. 5, equation 55 is used, yielding:
325
-------
7 =
L
0.004
8. The daily loads of suspended solids generated in each link are com-
puted using the contributary population of each link and an esti-
mated solids waste rate which is defined in section 12.4.2 as
0.5 Ib/cap/day. These values are not printed in Table 85, and
therefore are presented below.
Link
No.
1
2
3
4
5
Population
(Capita)
5
48
4
0
14
*
Solids Generated
Daily (Ib/day)
2.5
24.0
2.0
0.0
7.0
at 0.5 Ib/cap/day
9. Starting at link no. 1 and using the estimated deposition rate of
40.00% shown in column N of Table 85, the deposition load in that
link is:
0.4 * 2.5 Ib/day = 1.0 Ib/day
10. Searching the list of segments downstream from link no. 1, that is,
links 2, 4 and 5 (refer to step no. 2 above), it can be observed
from column N in Table 85 that none has a higher deposition rate
than link 1. This implies that a fraction (40%) of the load gener-
ated in link no. 1 will deposit in that link, but none of the parti-
cles that reach the end of link 1 will settle in downstream segments
since they all have shear stresses higher than that of link 1.
11. The search now moves to link 2. Searching the list of links down-
stream from link no. 2, namely links 4 and 5, it can be observed that
link 5 has a higher deposition rate than that of link 2, respectively
35.63% against 20.39%. Therefore, 20.39% of the load generated in
link 2 will deposit in that link, whereas 35.65%-20.39% = 14.24% of
that same load will deposit in link 5. This logic can be explained
by noting that each particle size is associated with a shear stress
value, by equation: 9/o
T = O.OZp^6 (57)
where T = fluid shear stress, psf
p = particle diameter, mm
326
-------
12.
Use of equation 57 implies that all particles having sizes larger
than or equal to p will settle at the corresponding shear stress T,
and smaller particles will be carried downstream in suspension. In
this manner, the downstream segment with the lowest shear stress
(conversely with the highest deposition rate) defines the smallest
particle size generated in an upstream link which will deposit in
the entire system. In this case, link 5, with a shear stress of
0.004 psf, defines the smallest particle size, p, of the load gen-
erated in link 2, which will deposit in the entire system. Given
an assumed distribution for the particle sizes in the sewage,,repre-
sented indirectly by equations 55 and 56, it was estimated that the
particles that deposit under the shear stress of 0.00448 psf at link
5 account for 35.63% of the total mass generated in link 2. A por-
tion of the coarser particles generated in link 2 will settle in
that link, which has a shear stress of 0.007 psf. These coarser
particle sizes that deposit in link 2, which account for 20.39% of
the total load generated in that link, will not be available for
deposition in link no. 5. Therefore link 5 can only deposit the dif-
ference of 35.63%-20.39% = 15.24%, of the total mass generated in
link 2.
Following a similar approach for all downstream links, the results
can be summarized as follows:
FRACTION OF LOADS GENERATED IN THE LINKS NUMBERED
HORIZONTALLY THAT WILL DEPOSIT IN THE LINKS NUMBERED VERTICALLY
DEPO-
SITION
LINKS
1
2
3
4
5
Links Originating Loads
1
2
3
4
5
Loads generated by Link (Ib/day)
2.5
0.4
0.0
0.0
0.0
0.0
24.0
0.2039
0.0
0.0
0.1524
2.0
0.1299
0.0
0.2264
0.0
0.1232
0.2331
7.0
0.3563
Total
load
per
link*
(Ib/day!
1.0
4.89
0.26
0.0
6.6
* Computed as the horizontal summation of the products of the fractions
shown in the table times the daily deposition loads appearing at the top of
the table. For link 5, for example, the total load is:
0.1524*24.0 + 0.2264*2.0 + 0.2331*0.0 + 0.3563*7.0 = 6.6Ib/day
327
-------
16:3 Downstream Transport of Solids in Suspension
This section is intended to provide an example utilizing the metho-
dology described in section 12.2. For simplicity, the verification case pre-
sented in section 12.2.6 for the second phase flushes at Port Norfolk Street
is repeated in greater detail.
Assume that the estimated mean deposits from the second phase
flushing program along the three successive pipe segments in Port Norfolk
Street are known to be 4.22 kg, 6.09 kg and 7.40 kg (a schematic of the three
pipe segments is presented in Figure 92). This procedure provides a rough
estimate of the loads flushed out of the last pipe segment if an average
flush (in volume and in injection rate) is injected into the most upstream
manhole.
Using the values of coefficients a and b of equation 24, which were
described in section 12.2.2, equation 24 can be written as
V = 0.5 e0-789(l+e-°-000634L) (58)
where: V = average velocity of the flush wave over the
distance L, fps;
L = downstream distance from the flush injection
point, ft; and
e = 2.71828.
The estimation of the masses from the first pipe segment that re-
main in suspension at the end of the first, second and third segments, that
is, at downstream distances of 162, 409 and 659 ft, respectively, is accom-
plished by the following steps:
1. Solve equation 58 for values of L equal to 162 ft, 409 ft and
659 ft, thus determining three values of average wave velocity, V;
2. Use some iterative procedure associated with Manning's equation, or
any other dynamic equation, to determine the three water depths,
h, associated with the wave velocities above;
3. With these three sets of values of V, h and L, solve equation 59 :
vsett ' T1 ' <59>
determining three values of settling velocities which will define
the particles that remain in suspension after the first, second
and third downstream manholes;
328
-------
PHASE 2 PROGRAM
co
ro
10
a
ID
INJECTION
PRIOR SCRAPING
(PHASE 2)
t
SAMPLING_
~AND GAGING
I62.51
246.91
250.0'
FIGURE 92 SCHEMATIC OF SECOND PHASE FIELD OPERATIONS
-------
4. Enter the values of the settling velocities determined above in the
"Conservative Turbulent Settling Curve" shown in Figure 69 of Chap-
ter 12, or use equation 60 :
P = 424. xlO5 V - 960. -xlO3 V + 123..xl02V + 10.8 (60)
g tt - . - ..s6tt
where P is percentage of mass remaining in suspension.
Then determine the percentages of material suspended from the first
pipe segment that will remain in suspension at the end of the first,
second and third segments. These values were computed as being
respectively 67.5%, 34.2% and 24.2% and are presented on the first
line of item number 2 in Table 86.
TABLE 86. COMPARISON OF MEASURED VERSUS ESTIMATED AVERAGE
TSS MASS REMOVALS - PORT NORFOLK ST. (SECOND PHASE)
1.
2.
3.
4.
Estimated mass
% remaining in
Segmerit
Segment
Segment
Estimated mass
Measured mass
deposited
suspension
1
2
3
(kg)
from:
flushed out (kg)
flushed out
(kg)
4.22
67.5
-
-
2.85
2.85
6.
34.
46.
-
4.
4.
09
2
8
29
03
7.
24.
27.
41.
5.
7.
40
2
9
2
77*
39
*0.242 x 4.22 + 0.279 x 6.09 + 0.412 x 7.40 = 5.77.
The percent removals from the successive pipe segments are computed
in a similar manner except for the fact that the settling velocity values
given by equation 59 corresponding to the distance L from the flush
injection, should be corrected for the length down to the new pipe segment
being considered. In this case equation 31 , Chapter. 12 applies and is
given by:
where
L = 0 ft. for materials removed from the first segment;
L = 162 ft. for materials removed from the second segment; and
L = 409 ft. for materials removed from the third segment.
It should be noted that discretization by pipe segment as described above is
330
-------
crude. A finer discretization, that is using smaller pipe sections would
be more desirable.
The percentages of the loads from the second pipe segment still in
suspension at the end of the second and third pipe segments were estimated
as 46.8% and 27.9%. The percentage of the loads from the third pipe segment
still in suspension at the end of the third segment was estimated to be 41.2%.
All those percentages are shown under item 2 of Table 86. The masses of
solids transported out of the first, second and third pipe segments can now
be estimated as:
First segment (162 ft.): 0.675 x 4.22 = 2.85 kg
Second segment (409 ft.): 0.342 x 4.22 + 0.468 x 6.09 = 4.29 kg
Third segment (659 ft.): 0.242 x 4.22 + 0.279 x 6.09 + 0.412 x 7.40 =
5.77 kg
These values compare reasonably well with the average measured values shown
as item 4 in Table 86. The first value, 2.85 kg was modified by adjusting
the mass composited in the first pipe segment, by 12.5% as was described in
Section 12.2.6.
The procedure is summarized as follows: (1) computation of the
average velocity V; (2) solution of Manning's equation for h; (3) solution
of equation 59 for V$ett' and (4) solution of equation 60 for P. A more
expedited solution can Be obtained by using the curves of Figure 69, or, in
the case of TSS, directly solving equation 30 , section 12.2.2, which is:
POO = - 1.33 (InL)3 + 33.86 (InL)2 - 288.57 (InL) + 834.32 (62)
To obtain the percentages of solids from the first pipe segment that are
carried beyond the end of the first, second and third segments, equation
62 is solved for L equal to 162 ft., 409 ft. and 659 ft., respectively,
obtaining values of 67.5%, 34.2% and 24.2%, as before. To obtain the
percentage of solids from the second pipe segment leaving in suspension the
end of the second and third pipe segments, equation 62 is solved for L
equal to 247 ft. and 497 ft, yielding 49.84% and 29.64%, respectively.
Finally, equation 62 is solved for L = 250 ft to obtain the percentage of
mass from the third pipe segment remaining in suspension at the end of the
third pipe segment. This estimate is 49.40%. With these new values, the
masses of solids leaving the end of the first, second and third pipe segments
are evaluated as:
First segment (162 ft): 0.675 x 4.22 = 2.85 kg
Second segment (409 ft): 0.342 x 4.22 + 0.498 x 6.09 = 4.48 kg
Third segment (659 ft): 0.242 x 4.22 + 0.296 x 6.09 + 0.494 x 7.40 = 6.50 kg
It was pointed out in Section 12.2.6 that the use of the curves in
Figure 69, or equation 62 for TSS provides the same results for the first
segment as the 4 step procedure described before, nonetheless ft overestimates
the percentages of materials removed from the second and third pipe segments.
Such application is equivalent to assuming that the upstream end of these
downstream pipe segments is at the point of the flush injection.
331
-------
REFERENCES
1. Turner, B.G., et. al. 1976 Survey Needs for Control of Pollution from
Combined Sewer Overflows and Stormwater Discharge. EPA-430/9-76-012,
U.S. EPA, February, 1977.
2. Energy & Environmental Analysis, Inc. Assessment of the National Pollu-
tant Discharge Elimination System of Public Law 92-500. National Study
Commission on Water Quality, May, 1975.
3. Field, R. and E.J. Struzeski, Jr. Management and Control of Combined
Sewer Overflows. Jour. Water Poll. Control Fed., 44: 1393, 1972.
4. Process Research, Inc. A Study of Pollution Control Alternatives for
Dorchester Bay. Metropolitan District Commission, Boston, Ma., December,
1976.
5. Environmental Design & Planning, Inc. Section 208: Areawide Wastewater
Combined Sewer Management Plan for City of Fitchburg. Montachusett
Regional Planning Commission, Fitchburg, Ma., December, 1977.
6. Pisano, W.C. Case Study: Best Management Practice (BMP) Solution for
a Combined Sewer Program. Urban Stormwater Management Workshop Proceed-
ings. EPA-600/9-78-017, U.S. EPA, August, 1978.
7. Gameson, A.L.H. and R. Davidson. Stormwater Investigations at North-
hampton. Annual Conference of the Institute of Sewage Purification,
Llandudno, England, June, 1962.
8. Rhett, John T. Program Guidance Memorandum - 61. Grants for Treatment and
Control of Combined Sewer Overflows and Stormwater Discharges. Water
Program Operations. U.S. EPA, December 16, 1975.
9. Inaba, K. Extent of Pollution by Stormwater Overflows and Measures for
Its Control. 7th International Water Pollution Research Conference,
San Francisco, 1972.
10. Ogden, H.N. Transactions, ASCE, Vol. 40, 1898.
11. FMC Corp., Central Engineering Laboratory. Feasibility of a Periodic
Flushing System for Combined Sewer Cleansing. 11020DN08/67, U.S. EPA,
August, 1967.
12. FMC Corp. A Flushing System for Combined Sewer Cleansing. 11020DNO-
03/72, U.S. EPA, March, 1972.
332
-------
13. Pisano, W.C. Cost-Effective Approach for Combined and Storm Sewer
Cleanup. Proceedings of Urban Stormwater Management Seminars. WPD-03-
76-04, U.S. EPA, January, 1976.
14. Lager, J.A. and W.G. Smith. Urban Stormwater Management and Technology
- An Assessment. EPA-670/2-74-040, U.S. EPA, December, 1974.
15. Operating and Service Manual, Model III Fluorometer. G.K. Turner
Associates, Palo Alto, California.
16. Kodak Filters for Scientific and Technical Uses. Kodak Publication
No. B-3, 7-70, U.S.A.
17. Zanoni, A.E. and M.W. Blomquist. Column Settling Tests for Flocculant
Suspensions. Journal Environmental Engineering Div., ASCE 101: EE3,
June, 1975.
18. Standard Methods for the Examination of Water and Wastewater, 14th
Edition. APHA-AWWA-WPCF, 1975.
19. Wilcox, E.R. A Comparative Test of the Flow of Water in 8-Inch Concrete
and Vitrified Clay Sewer Pipe. University of Washington, Exp. Sta.
Series Bull. 27, 1924.
20. Yarnell, D.L. and S.M. Woodward. The Flow of Water in Drain Tile.
Dept. of Agriculture Bull. No. 854, U.S. Govt. Printing Office,
Washington, D.C., 1920.
21. Johnson, C.F. Determination of Kutter's n for Sewers Partly Filled.
Transactions, ASCE Journal, Vofl. 109, 1944.
22. Camp, T.R. Discussion - Determination of Kutter's n for Sewers Partly
Filled, by C.F. Johnson. Transactions, ASCE Journal, Vol. 109, 1944.
23. Hooke, R. and R.A. Jeeves. Direct Search Solution of Numerical and
Statistical Problems. J. Assoc. Comp. Mach., 8:212-229, 1961.
24. Metcalf and Eddy, Inc. Wastewater Engineering Collection, Treatment
and Disposal. McGraw-Hill, 1972.
25. Ackers, P. and A.J.M. Harrison. Attenuation of Flood Waves in Part-
Full Pipes. In: Proceedings of the Institution of Civil Engineers,
London. Vol. 28, July 1964.
26. Mitchell, J.S. Comparison of Mathematical Versus Experimental Flood-
Wave .Attenuation in Part-Full Pipes. Master Thesis, Colorado State
Univeristy, Fort Collins, April, 1967.
27. Martin, C.S. and F.G. DeFazio. Open-Channel Surge Simulation by
Digital Computer. ASCE Journal of the Hydraulics Division, HY6,
November, 1969.
333
-------
28. Sonnen, M.B., Water Resources Engineers. Abatement of Deposition and
Scour in Sewers. EPA-600/2-77-212, U.S. EPA, November, 1977.
29. Metcalf & Eddy, Inc. University of Florida, Water Resources Engineers,
Inc. Storm Water Management Model, Volume I. 11024DOC07/71, U.S. EPA,
July, 1971.
30. Yao, K.M. Sewer Line Design on Critical Shear Stress. Journal Environ-
mental Engineering Div., ASCE, 100: E22, April, 1974.
31. Camp, T.R. Design of Sewers to Facilitate Flow. Sewage Works Journal,
18:3, January, 1946.
32. Graf, W.H. Hydraulics of Sediment Transport. McGraw-Hill, 1971.
33. Hughmark, G.A. Aqueous Transport of Settling Slurries. Industrial and
Engineering Chemistry, Vol. 53, May, 1961.
34. Raths, C.H. and R.F. McCauley. Deposition in a Sanitary Sewer.
W&SW No. 1962, pp. 238-243.
35. FMC Corp. Relationship of Sewage Characteristics to Carrying Velocity
for Pressure Sewer. American Society of Civil Engineering, No. R-2598,
August, 1969.
36. Pisano, W.C. Pollution Control Alternatives for Dorchester Bay.
Unpublished, September, 1976.
37. Pisano, W.C. and C.S. Queiroz. Procedures for Estimating Dry Weather -
Pollutant Deposition in Sewerage Systems. EPA-600/2-77-120, U.S. EPA,
July, 1977.
38. Environmental Design & Planning, Inc. Infiltration/Inflow Study for
West Roxbury-Brookline-Newton-Dedham Area. Metropolitan District
Commission.
39. Steel, E.W. Water Supply and Sewerage. McGraw-Hill, 1960.
40. Michels, R.L., Pelmoter, A.L. and R.C. Palange. Operation and Mainte-
nance of Municipal Waste Treatment Plants. J.W.P.C.F., 41:335, March,
1969.
41. Combined Sewer Regulation and Management: A Manual of Practice.
1122DMU08/70, U.S. EPA, July, 1970.
42. Smith, R. and R.G. Eilers. Cost to the Consumer of Collecting and
Treating Wastewater in the United States. 17090—07/70, U.S. EPA,
July, 1970..
43. Michels, Robert L. Cost and Manpower for Municipal Wastewater Treatment
Plant Operation and Maintenance, 1965-1968. J.W.P.C.F., 42:1883,
November, 1970.
334
-------
44. Rowan, P.P., Jenkins, K.L. and D.H. Howell. Estimating Sewer Treatment
Plants' Operating and Maintenance Cost. J.W.P.C.F., 33:111, February,
1961.
45. Logan, J.A., Hatfield, W.D., Russel, G.S. and W.R. Lynn. Analysis of
the Economics of Wastewater Treatment. J.W.P.C.F., 34:860, September,
1962.
46. Smith, Robert. Cost of Conventional and Advanced Treatment of Waste-
water. J.W.P.C.F., 40:1546, September, 1968.
47. Shah, K.L. and 6.W. Reid. Techniques for Estimating Construction Cost
of Wastewater Treatment Plants. J.W.P.C.F., 42:776, 1970.
48. Van Note, R.H., et. al. A Guide to Selection of Cost Effective Waste-
water Treatment Systems, EPA 430/9-75-002, U.S. EPA:S Office of Water
Programs Operations, July, 1975.
49. Patterson, W.L. and R.F. Banker. Estimating Costs and Manpower Require-
ments for Conventional Wastewater Treatment Facilities. 17090DAN10/71,
U.S. EPA, October, 1971.
50. Guides for Design of Wastewater treatment Works. Technical Advisory
Board of New England Interstate Water Pollution Control Commission.
51. Recommended Standards for Sewer Works. Upper Mississippi River Board
of State Engineers, 1972.
52. Clark, J.W., Viessman, Jr., W. and M.J. Hammer. Water Resources and
Pollution Control. International Testbook Company, 1971.
53. Chart Book on Prices, Wages and Productivity. U.S. Dept. of Labor,
Bureau of Labor Statistics, January, 1977.
54. Wholesale Prices and Price Index. U.S. Dept. of Labor, Bureau of Labor
Statistics, May, 1977.
55. Cost Effectiveness and Clean Water, Annual Report to the Congress.
U.S. EPA, Water Quality Office, January, 1971.
335
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-79-133
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
DRY-WEATHER DEPOSITION AND FLUSHING FOR COMBINED
SEWER OVERFLOW POLLUTION CONTROL
5. REPORT DATE
August 1979 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
.AUTHOR(S)Wnl1am c> P1san0i Gerald L. Aronson, Celso S.
Queiroz, Frederic C. Blanc, James C. O'Shaughnessy
8. PERFORMING ORGANIZATION REPORT NO.
260 Huntington Avenue
Boston, Mass. 02115
ADDRESSSubcontractor:
Environmental Design &
Planning, Inc.
257 Vassar Street
Cambridge. Mass. 02139
10. PROGRAM ELEMENT NO.
1BC822 SOS 1 Task 43
11. CONTRACT/GRANT NO.
R804578
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory—Cin.,OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final 7-76 - 12-78
14. SPONSORING AGENCY CODE
EPA/600/14
15. SUPPLEMENTARY NOTEsproj-ect officers: Richard Field and Richard P. Traver (201) 321-
6674 (8-340-6674). Supplement to EPA-600/2-77-120, "Procedures for Estimating Dry
Weather Pollutant Deposition in Sewerage Systems"
16.ABSTRACT jj-jis report' summarizes the results of a two year study aimed at addressing
the feasibility, cost-effectiveness and ease of application of upstream solids control
as an integral part of overall combined sewer management. The project was functionally
divided into four major phases. In the first phase four test segments on different
streets in the Boston sewerage system were field flushed over an extended period to
quantify the effectiveness of flushing deposition accumulation from a single pipe seg-
ment, and to estimate deposition characteristics within collection system laterals.
The second phase was concerned with the problem of flushing a long flat stretch of
combined sewer laterals. Flushes were injected into the uppermost manhole and pollu-
tant levels in the flush wave passing three downstream manholes were monitored, provi-
ding insights into flushing effectiveness over three segments of pipe. Settleability
tests were also performed for the purpose of estimating how far beyond the flushing
monitoring manholes would the materials be carried. In the third phase, an automatic
sewer flushing module was designed, fabricated, installed and operated on a single seg-
ment for an extended period. The purpose of this work was to begin to develop opera-
tional experience, using automated flushing equipment. In the fourth phase, various
predictive deposition loading and flushing criteria were generated from the large data
base developed during the field programs. These formalisms allow for scanning of
large-scale sewer systems to identify problem pipes with respect to deposition.
17.
KEY WORDS AND DOCUMENT ANALYSIS
a.
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
Combined sewers, Flushing, Water pollution,
Water quality, Economics, Mathematical
models, Methodology, Maintenance, Regres-
sion analysis
Field measurements, Pol-
lution abatement, Sewer
flushing
13B
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. OF PAGES
362
RELEASE TO PUBLIC
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (Rev. 4-77)
336
* U S. GOVERNMENT PRINTING OFFICE: 1979 -657-060/5459
------- |