Environmental Protection Technology Series
AUTOMATIC ORGANIC
MONITORING SYSTEM FOR
STORM AND COMBINED SEWERS
National Environmental Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
EPA-670/2-75-067
June 1975
AUTOMATIC ORGANIC MONITORING SYSTEM FOR
STORM AND COMBINED SEWERS
By
Angelo Tulumello
Raytheon Company
Portsmouth, RI 02871
Contract No. 68-03-0262
Program Element No. 1BB034
Project Officer
Hugh E. Masters
Storm and Combined Sewer Section (Edison, N. J.)
Advanced Waste Treatment Research Laboratory
National Environmental Research Center
Cincinnati, Ohio 45268
NATIONAL ENVIRONMENTAL RESEARCH CENTER
OFFICE OF RESEARCH AND DEVELOPMENT
U. S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
REVIEW NOTICE
The National Environmental Research Center—Cincinnati
has reviewed this report and approved its 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 endoresement or recom-
mendation for use.
11
-------
FOREWORD
Man and his environment must be protected from the adverse effects
of pesticides, radiation, noise, and other forms of pollution, and the
unwise management of solid waste. Efforts to protect the environ-
ment require a focus that recognizes the interplay between the com-
ponents of our physical environment—air, water, and land. The
National Environmental Research Centers provide this multidisci-
plinary focus through programs engaged in:
• studies on the effects of environmental contaminants on man and the
biosphere
• a search for ways to prevent contamination and to recycle valuable
resources.
This report discusses the development and evaluation of an automatic moni-
toring system for measuring organics in storm and combined sewage. Total
organic carbon measurement was found to be the optimal method for measur-
ing storm-generated pollution.
A.W. Breidenbach, PhD
Director, National Environmental
Research Center, Cincinnati
111
-------
ABSTRACT
Early in the program to develop a stormwater total organic carbon (TOC) sys-
tem, it was established in report EPA-670/2-74-087 that continuous on-line
TOC was the best method for the measurement of stormwater pollution loading.
Hardware was assembled that would process stormwater samples containing
high suspended solids and that would obtain a continuous signal proportional to
the concentration of TOC in the sample.
Synthetic samples of municipal raw influent charged with primary sludge were
analyzed using the TOC analyzer. Data were also obtained on actual stormwater
samples collected during storm events at Boston. Further modifications were
made after these observations.
Automatic circuitry designed to provide turn on, auto-zero, auto-span, and
sample line flushing was added to the hardware, and the system was installed at
Boston Cottage Farm Storage Facility.
Automatic continuous analyses were obtained during storms on site at the
Cottage Farm Storage Facility.
This report was submitted in partial fulfillment of Program Element No. 1BB034,
Contract No. 68-03-0262 by Raytheon Company under the sponsorship of the
U. S. Environmental Protection Agency. Work was completed in March of 1975.
IV
-------
CONTENTS
Foreword iii
Abstract iv
List of Figures vi
List of Tables viii
Acknowledgements ix
Sections
I Conclusions 1
n Recommendations 3
III Introduction 4
IV Phase I—Review 7
V Experimental Work 10
VI Installation of a System in a Remote Location 33
VH Phase II -Field Test 38
VIII Control System and Assembly 40
IX Operating Experience 47
X Appendices 52
-------
FIGURES
No. Page
1 Stormwater TOC system block diagram 9
2 Raytheon TOC 11
3 Sample delivery system for simulated laboratory tests 12
4 Simulated stormwater delivery system 13
5 TOC chart for the storm of 4/23/74 18
6 Characterization of the storm of 4/23/74 19
7 TOC chart for the storm of 5/10/74 19
8 Characterization of the storm of 5/10/74 21
9 TOC chart for the storm of 5/12/74 23
10 Characterization of the storm of 5/12/74 23
11 Analog filter schematic 26
12 Integration time using volume 27
13 Effect of added mixing volume 28
14 TOC chart for the storm of 6/1/74 without added mixing volume 28
15 TOC chart for the storm of 6/1/74 with added mixing volume 29
16 Characterization of the storm of 6/1/74 29
17 TOC electrical schematic 36
18 TOC analyzer flow diagram 37
19 Cottage Farm Storage Facility stormwater TOC installation 39
20 Stormwater TOC control circuit 41
21 Manifold of the liquid flow system 43
22 Location of sample take-off pump 44
VI
-------
FIGURES (Cont)
No.
23 Installation of stormwater TOC in the Cottage Farm Storage
Facility delivery system 45
24 Sample, water, drain, and waste delivery system 46
25 Operation of the Cottage Farm Storage Facility, 10/16/74 48
26 Display of storm data of 11/19/74 for comparison with previous
day's calibration curve 49
27 Data collected during the storm of 12/2/74 50
28 The bathub curve 53
29 Reliability model for stormwater TOC analyzer 55
30 Modular manhole configuration 57
vn
-------
TABLES
No. Page
1 Hardware Used in Stormwater Simulation 12
2 Comparison Data Obtained from Simulated Storm Conditions 15
3 Data from Samples Obtained at Cottage Farm Storage Facility,
Storm of 4/23/74 20
4 Data from Samples Obtained at Cottage Farm Storage Facility,
Storm of 5/10/74 22
5 Data from Samples Obtained at Cottage Farm Storage Facility,
Storm of 5/12/74 24
6 Data from Sample Obtained at Cottage Farm Storage Facility,
Storm of 6/1/74 30
7 Maintenance Requirements 31
8 Modular Components Used in Stormwater TOC 40
9 Components Activated in the Operate Cycle 41
Vlll
-------
ACKNOWLEDGEMENTS
We wish to acknowledge the support and direction of Mr. Richard Field, Chief
of the Storm and Combined Sewer Section of the U. S. Environmental Protection
Agency and Mr. Hugh E. Masters, Project Officer.
We feel a deep sense of gratitude to Mr. Anthony Ventatuolo and his staff of the
Cranston Municipal Treatment Plant for providing us with the facilities and
assistance to obtain sewage samples and conduct the experimental work neces-
sary to establish the accuracy and repeatability of our data.
We also wish to acknowledge the cooperation of Mr. Allison C. Hayes, Director
of the Sewage Division of the Metropolitan District Commission of Boston for
allowing us to install our instrumentation at the Cottage Farm Storage Facility.
In addition, we wish to express our appreciation of the cooperation of Mr. Frank
Zinfolino, Manager and his staff of the Cottage Farm Storage Facility who pro-
vided us the necessary operating information and assistance in the installation
of hardware and the assembly of operating data.
IX
-------
SECTION I
CONCLUSIONS
1. The total system for the evaluation of storm-generated pollution
depends upon the availability of a sample delivery system and a
suitable processing and analysis system. A sample processing and
analysis system has been described. A suitable sample delivery
system is yet to be designed. At the present stage of development,
each location must have a custom sample delivery subsystem.
2. Total organic carbon (TOC) analysis is a rapid, reliable and well-
correlated method for measuring storm-generated pollution.
3. A stormwater TOC measuring system has been developed and field tested.
4. The stormwater TOC measuring system can be calibrated on potassium
acid phthalate (KHP) in accordance with the method of the Environ-
mental Protection Agency.
5. Tests at the Boston Metropolitan District Commission-Cottage Farm
Combined Sewer Overflow Storage Facility (Cottage Farm Storage Faci-
lity) indicate that the installed system assembly is capable of
accurately monitoring TOC of storm-generated pollution.
6. Unattended and automatic turn on, turn off, auto-zero, auto-span,
and the sample delivery flush systems eliminate the need for fre-
quent operator attention and permit continuous on-line readiness.
7. Manhole installation of the TOC instrumentation presents substantial
problems. The present equipment, placed in a mobile or remotely
located trailer or a shed operation, can be expected to collect re-
liable data from a suitably designed sample delivery system.
8. The cost of stormwater TOC instrumentation cannot be realistically
estimated at this time because the cost of a practical sample de-
livery subsystem is unknown. The cost of the stormwater TOC in-
strument without sample delivery subsystem is about one and a half
times the cost of the present online TOC instruments available for
other purposes.
-------
9. The reliability of the stormwater TOG instrument exclusive of the sample
delivery subsystem is adequate to the needs of stormwater control and
treatment applications.
-------
SECTION H
RECOMMENDATIONS
1. Verify the existing system with further wet-weather testing at some on-
going USEPA demonstration project where comparative data are desired.
This would include an improved sample delivery subsystem.
2. Determine control and treatment strategies utilizing the combined system
at selected sites.
3. Incorporate feed-forward control and exercise the system by utilizing the
developed control strategies.
4. Specify and deploy stormwater control and treatment systems based on the
preceding steps.
-------
SECTION in
INTRODUCTION
THE PURPOSE OF THE DEVELOPMENT
One of the major sources of pollution of our natural bodies of water is the ran-
dom occurrence of uncontrolled storm flows in both storm and combined sewer
facilities. Programs of stormwater diversion, detention, and treatment have
been suggested as a means of controlling these events.
The nature of a storm flow presents a characteristic system loading as a func-
tion of time. The first effect of the high velocity turbulent flow of water is to
scour the drainage area, cleaning the ground of organic material (leaves, grass,
asphalt), and suspending settled solids that have come to be deposited in the
sewer system. This flow entrains a rich and concentrated organic load which
cleans as it flows. Later flow may act to dilute the remaining pollution load.
To implement a timely course of action in handling this pollution source it is
necessary that one obtain a real-time measurement of the pollution level of the
storm flow. The information resulting from this measurement can be used in
the control and treatment of stormwater, in system design programs, or for
enforcement purposes.
The purpose of this contract is to develop and demonstrate an efficient automatic
monitoring device for the rapid, in situ determination of dissolved and suspended
organic loading. Several options were available in developing a suitable analyt-
ical scheme to monitor the pollution level of storm flow: biological oxygen
demand (BOD), chemical oxygen demand (COD), total oxygen demand (TOD), or
total organic carbon (TOC). After a careful evaluation, TOC was selected for
development.
-------
The Choice of TOG
Storm flows present a highly variable sample containing a wide variety of
organic materials in various forms, among these is a first flush of high concen-
tration of suspended solids. This presents a particularly severe sample from
which to obtain essentially instantaneous data. TOC was recognized as the only
practical analytical technique that could deliver the quantitative information
necessary to permit rapid, intelligent, and appropriate action that could be used
to interface with the facilities used to handle the pollution load. The TOC analy-
sis is a relatively independent analysis requiring a minimum of maintenance and
supply—this is not true of wet chemical and biological systems. Because of this,
it is possible to enclose the system in a mobile or sheltered enclosure.
Instrumentation Considerations
The problem associated with the adaptation of chemical analyses to on-line
monitoring is not the ability to assemble hardware that will perform the required
operations. The main problem is to keep the hardware operating reliably in
time. Chemical systems dependent on convection to go to completion tend to
have a relatively long reaction time. The conditions that chemical systems pre-
sent are limitations to the hardware normally used to transfer solids, liquids,
and gases. The need for acids, bases, and other reactive reagents, and the
requirement of high temperature and pressure places severe strain on materials
used in sample handling. As a result of these conditions, materials and designs
have to be carefully selected and tested to provide the reliability required for
action based on continuous on-line analyses.
The operating conditions for continuous on-line monitoring instrumentation are
much more severe than those normally imposed on laboratory equipment. Labo-
ratories provide ready and experienced service when breakdown occurs. When
on-line instrumentation fails, it interrupts an established routine which places
a random strain on the available service personnel. This frequently occurs
when other pressing assignments are in process, resulting in a prolonged period
of downtime for the instrumentation. Under conditions of this kind, people lose
their confidence in the reliability of the hardware and often disarm it.
Modern material technologies have provided us with previously unrealized options
in addressing the conditions necessary for continuous on-line chemical analyses.
Quantitative peristaltic pumping provides us with a considerable improvement in
sampling suspended solids. Inconel alloy provides the high temperature furnace
environment required for a complete combustion and long service. The reliabil-
ity of modern solid-state switching, the availability of off-the-shelf tube connec-
tors and the versatility and improvement in modern plastics has made it possible
to assemble and test the hardware necessary to perform the continuous on-line
analysis of TOC in stormwater.
-------
The design of instrumentation for the analysis of stormwater and combined
sewage presents problems which result from the nature of the sample. The
sample contains solids as well as liquid in several phases (i.e., oil in water).
This requires care in sample transfer to minimize fiber hang-up and solids
coagulation sites that can occur in low velocity delivery systems. The plumbing
can provide sites for solids collection and plug formation. These sites can be
minimized by transferring sample through carefully assembled sanitary (connec-
tions without shoulders) fittings. These problems were addressed in Phase I.
The above mentioned problem areas, made it necessary to plan an on-line field
test and to provide a program of vigorous routine maintenance with documented
experience and redesign as indicated by the field experience. In this way, it
was possible to demonstrate the reliability of Raytheon's instrumentation for
installation in continuous on-line monitoring of storm flows. This was done in
Phase II.
-------
SECTION IV
PHASE I-REVIEW
During Phase I, a report was prepared entitled an "Assessment of and Develop-
ment Plans for Monitoring of Organics in Storm Flows", EPA-670/2-74-087.
It is appropriate to review some of the material contained in the report and see
how it would apply to the system employed in the laboratory portion of Phase I.
In compiling this report, several criteria were established for the judgment of
the hardware. The optimal system would be one which would operate reliably
and reproducibly in the concentration range-of-interest (5 to 1000 mg/1 TOG).
It should require a minimum of maintenance, be economical, and operate under
conditions likely to be found in a storm flow sewer or control facility. The
devices should be automated as to zero, span, startup and shutdown. In addition,
the storm flow TOG monitoring system should be capable of analyzing samples
with varying suspended solids concentrations.
These requirements are necessary to perform efficient organic carbon analyses
at remotely located storm sample sites. They have been recognized and defined
in the design goals of this contract.
Continuous operating on-line TOC analyzers have not, in general, been able to
meet these requirements. The common problems encountered in the experi-
mental effort have been substantial. The delivery of a continous representative
sample containing up to 1000 mg/1 suspended solids is a considerable accom-
plishment. The continuity of the delivery is easily interrupted because of
plugging. These problems tend to render unreliable much of the commercial
hardware designed to process and deliver representative samples containing
suspended solids. If suspended solids are removed, the samples are not'
representative and the TOC measurements are unreliable. Care must be taken
in the design of systems of continuous TOC analysis to reduce the particle size,
transfer the particles, and deliver them through sanitary fittings to an efficient
and continuous analytical system.
STATE-OF-THE-ART
The commercially available, continuous duty TOC instruments have a demonstrated
history of failure; moreover, their maintenance requirements are so substantial
that some of the instruments have been abandoned. The problems associated
-------
with the design of TOC instrumentation of this type are not only the problems
of sample delivery but also the difficulty of providing efficient continuous com-
bustion of small amounts of carbon in water. In order to accomplish this, it
has been shown that a high temperature, 950° C, is needed so that compounds
delivered to the reactor may be oxidized rapidly regardless of the nature of the
organic structure of the material. This is a problem of some magnitude in
itself. Many of the instrument suppliers have cascaded their difficulties by
incorporating flame ionization detection (FID) and catalytic reactors-. These
have their own failure modes and tend to add to the unreliability of the instru-
mentation.
A review of the literature indicated a need for a continuous TOC analyzer. The
TOC technique has several advantages over the alternative methods of organic
pollution monitoring, BOD and COD. BOD and COD are both time consuming.
BOD is non-reproducible. They are dependent upon operating conditions and
the nature of the sample.' TOC can be specific to organics; it is very repro-
ducible. Operating instrumentation can provide essentially real-time data in
time to take necessary action.
Storm events place excessive and unpredictable organic loads on sewage facili-
ties. With consideration of the storm events and the problems associated with
processing samples during these events, a stormwater TOC system was proposed
for the TOC analysis. This system is diagrammed in Figure 1. The validity
of the proposed TOC system was verified in an extensive laboratory test pro-
gram which included both simulated and actual stormwater samples.
-------
Figure 1. Stormwater TOG system, block diagram
-------
SECTION V
EXPERIMENTAL WORK
To discuss the performance of a TOC system, certain standards must be estab-
lished to demonstrate the experimental results. This section of the report,
presents the substantiating data necessary to demonstrate the device's ability
to meet the requirement for on-line measurement of organics in storm flows.
PROCEDURE
During the early stages of the laboratory test effort, simulated stormwater
samples were used to obtain the performance data. The simulation of storm-
water runoff and combined sewage overflows was accomplished by adding settled
primary sludge, containing about 25, 000 mg/1 of solids to raw municipal sewage
resulting in a net TOC of 500 mg/1 and a suspended solids of 1000 mg/1. Later
in the program, actual combined sewer overflow samples, collected from
Boston's Cottage Farm Storage Facility were used to evaluate the TOC system's
performance.
The equipment used in these simulated tests consisted of an assembly of hard-
ware designed to deliver representative samples from a sample container by
careful control of flow rates and the delivery of the sample through properly
sized sanitary fittings. The simulated storm samples were delivered continu-
ously through the Raytheon homogenizer (Refer to EPA-67 0/2-74-087, Page 41)
from which a continuous stream was taken to the Raytheon TOC analyzer (see
Figure 2). The sample delivery system is diagrammed in Figure 3. Figure 4
is a picture of the sample delivery system. The hardware used in the application
is listed in Table 1. Prior to discussing the results, it is first necessary to
define the accuracy and precision of the system and the methods of defining these
parameters.
ACCURACY AND PRECISION
The accuracy of an instrument can be established by reference to the precision
with which the instrument analyzes primary standard solutions of a known com-
position. In this case, the method of the EPA (USEPA Publication No. 16020—
07/71, Pages 221-229) is specified by contract. This method employs a Beckman
TOC analyzer, whose precision depends strongly on the sample characteristics,
10
-------
Figure 2. Raytheon TOG
11
-------
STORM
WATER
TOC
DRAIN
BAFFLE RESERVOIR
WITHSTIRRER
FLEXIBLE
IMPELLER
PUMP
Figure 3. Sample delivery system for simulated laboratory tests
Table 1. HARDWARE USED IN STORMWATER SIMULATION
Function
Hardware
Container
Motor Driven Stirrer
Sample Delivery Pump
Particle Reduction
TOC Analysis
50-gallon Fiberglas Drum
1/20 Horsepower Industrial VWR Stirrer
Barnat Variable-Speed Peristaltic Pump
Raytheon Homogenizer
Raytheon TOC
specifically on the type, concentration and preliminary treatment of the sus-
pended solids. To obtain data of any value at all with suspended solids, it is
necessary for the discrete sample to be processed through a Waring type blender
until the biggest particle can be passed through a 0. 02 inch needle in both direc-
tions without coalescing on the walls of a 20 jil syringe.
12
-------
Figure 4. Simulated stormwater delivery system
-------
The Raytheon TOC, like the Beckman TOC, is standardized, by generating a
TOC signal from a solution of known composition of a primary standard. The
material used to prepare these standards is potassium acid phthalate (KHP).
Using this material, a scale is established to operate in the concentration region-
of-interest to stormwater analysis. (0-1000 mg/1). Because the Raytheon TOC
has a continuous delivery system, the signal obtained does not depend on the
precise selection, measurement, and delivery of a discrete sample containing
solids, but delivery occurs continuously with time. If organic solids are deliv-
ered to a continuous analyzer, a high reading will appear for the period during
which the sample is being processed. The precision and accuracy of the instru-
ment is established by burning clear solutions of known composition.
When an instrument has been designed for measurement over a concentration
range, it is not generally used over a very different concentration range without
modification. The stormwater TOC was intended to operate from 0-1000 mg/1.
The error expected in any measurement with this instrument is ±2% of full scale.
Reference samples were submitted by the EPA Analytical Quality Control Lab-
oratory. It was established that the concentration of the more concentrated of
these samples was 140 mg/1 TOC; the lower concentration was 4 mg/1 TOC.
Data collected at this lower concentration are within the expected instrumentation
error and could not practically be measured with the stormwater TOC assembly.
In order for an analytical technique to be of any value to the operators, it is
necessary to have some estimate of the error associated with the application of
the technique to the problems at hand. This error can be assumed to be
associated with a normal spread of the data obtained for a particular true value
of an aliquot which can be estimated by identifying each mean value with a
standard deviation, "s". Small standard deviations result when the techniques
are capable of precise estimation of the values-of-interest. Soluble TOC is
not difficult to determine by any TOC analyzer. The EPA reports standard
deviations of the order of 8 mg/1 on soluble organics with concentration of 107
mg/1. The stormwater TOC analyzer has done somewhat better than this
throughout the range of soluble TOC. This is true with both filtered biological
materials as well as with KHP.
Most continuous TOC analyzers have difficulty in the combustion of KHP because
of the noise which results from the nonuniformity of combustion of this species.
To avoid these problems, secondary standards have been used (for example,
sucrose and ethylene glycol) as reference solutions because of their greater ease
of combustion. These secondary standards have been prepared by a previous
comparative analysis using KHP.
14
-------
In the stormwater TOC system, as in the standard method of the EPA, a solution
of KHP may be burned directly as a primary standard. This has been accom-
plished by establishing a more efficient combustion chamber. Higher tempera-
ture (950° C) and smaller sample size (2 cc/min) were found to give efficient
combustion of 425 mg/1 TOC prepared from KHP.
The combustion of suspended solids is a considerable accomplishment in itself.
In the TOC method of the EPA, it is suggested that "insoluble particulate car-
bonaceous materials" can be analyzed using the Beckman TOC. However, there
is no precision data associated with this analysis. A little experience with this
technique shows that there is considerable difficulty in obtaining very precise
data. The solids themselves are variable as are the combustion characterises
of the possible components in the stormwater system. This type of sample
results in an unpredictable variability that can be expected to be at least twice
that of the soluble TOC values.
Raytheon has demonstrated considerable success in the efficient combustion of
suspended solids as will be noted in the following section describing the storm-
water results.
STORMWATER RESULTS
Simulated Storms
A sample data sheet obtained with a synthesized storm is seen in Table 2.
establishing a simulation of storm sewer loading it becomes necessary to
In
Table 2. COMPARISON DATA OBTAINED FROM SIMULATED STORM
CONDITIONS
Characterization
Sample
Stormwater Mean
TOC Standard
Deviation
Beckman 1
TOC 2
mg/1 3
4
5
Mean
Std Deviation
Settleable
Solids 30 Mto
Suspended
Solids mg/1
Particle Size 420 |i
Distribution 210|i
mg/t llO^i
In Batch Tank
Primary
Effluent
Sample 21
130
115
120
120
120
121
5.48
1 ml/1
120
6.5
9
11
® +900 mil
Primary Sludge
Sample 22
425
425
430
425
425
426
2.24
50 ml/1
1,090
209
16
15
@ 50% Dil
of ®
Sample 23
190
185
190
185
185
187
2.74
12 ml/1
390
40
9.5
4
50% Dil
of ®
Sample 24
90
90
90
105
90
93
6.71
2 ml/1
140
12
5
5
Delivered to TOC
Primary
Effluent
Sample 21A
113.5
1.73
115
110
110
115
115
113
2.74
NR
110
1
0
0
(A) +900 ml
Primary Sludge
Sample 22A
456.8
49.7
470
450
470
460
480
466
11.4
NR
1,070
122
21
18
@ 50% Dil
of ®
Sample 23A
193.1
7.6
180
175
170
175
180
176
4.18
NR
360
53
4.5
11
50% Dil
of ®
Sample 24A
110
5
95
95
95
90
100
95
3.5
NR
200
9
1
5
Condition at the time of data collection
Peristaltic by-pass flow rate 900 cc/min
Raytheon homogenizer flow rate 800 cc/min
15
-------
rapidly add a high concentration of suspended solids to a rather modestly loaded
(200 mg/1 TOC) sanitary sewage. This increases the concentration of TOC
to about 500 mg/1 TOC. This also increases the severity of the combustion
required to produce rapid complete oxidation. The easiest solids to burn are
those that contain fairly large amounts of water structurally bound to the carbon.
Hydrated pulps make the most easily handled suspended solids for TOC analysis.
Items such as plastics, rubber, and carbon structures that have survived ultra-
violet radiation, biological degradation, rain, abrasion, and high and low tem-
peratures are among the most difficult to combust in continuous analyses. Storm
flows tend to concentrate the more resistant structures. Municipal sewage tends
to contain more hydrated pulps, but combined sewage in general is a random
mixture of all of these. As a result, there is always some level of nonuniformity
to the combustion of suspended solids.
To minimize nonuniform combustion, it was found that it is desirable to grind
the suspended solids to a size no larger that 1200 microns in largest dimension
and deliver them rapidly, uniformly, and quantitively to an efficient combustion
environment.
Our first synthesized storms were prepared by adding primary sludge to primary
effluent. This was a conveniently random sample on which to optimize our de-
livery system. Solids loading established in this way was very severe. The
data obtained showed standard deviations which were large compared to similar
data obtained using the Beckman analyzer and a Waring blender, as shown in
Table 2.
Table 2 is a presentation of data collected on four sewage samples delivered to
the stormwater TOC assembly. The purpose of this experiment was to show that
sample had been delivered to the continuous analyzer in such a way as to retain
a concentration of TOC at the continuous analyzer that was representative of the
material in the batch tank (see Figure 3). If the sample delivery system is pro-
perly assembled, the TOC values in the batch tank sample, as measured by the
Beckman TOC unit will be found to be essentially the same as the Beckman TOC
values obtained on samples from the exhaust of the Raytheon homogenizer. These in
turn will be essentially the same as the sample analyzed on the continuous ana-
lyzer. In order to simulate a storm loading condition, four compositions are
examined. The first is primary effluent which is characterized and analyzed as
seen under primary effluent. This sample is then enriched with primary sludge
to represent the sewer loading associated with the first flush effect. We can see
that the solids loading has increased with the increase in TOC and that there has
been considerable reduction in the total weight of the larger particles because
of their treatment in passing through the blender even though the suspended solids
16
-------
remain essentially the same. The next two samples show the result of water
additions to the enriched primary effluent, simulating the later stages of storm
flow.
The mean and standard deviations of the discrete TOC samples are obtained by
a simple calculation using the numbers the table. The standard deviations of
the continuous data are obtained by selecting sample data from the continuous
recording. It was found that in these data, and similar tables of data, that the
standard deviation is proportional to the concentration of suspended solids.
Actual Stormwater Samples
In some actual stormwater samples obtained during storm flows at the Cottage
Farm Storage Facility, the problem of the nonuniform combustion of suspended
solids became even more severe; in others it was not. Since there is a concern
with the general condition of stormwater, it became necessary to improve the
combustion environment so that the more severe samples might be uniformly
combusted. To do this, operating modifications were necessary to minimize
the removal of heat from the reactor, close to the point of sample delivery. The
flow rate of water at the sample delivery tube was minimized. This minimized
the loss of heat through this medium. The effect of this operation was to decrease
the nonuniformity of the combustion, and consequently minimize the noise in
the signal from samples containing high suspended solids.
Data taken onflows during the storm of April 23 seemed quite acceptable. Storm
samples collected during the storms of May 10 and 12 gave a very different pic-
ture. The difference was probably caused by the fact that on 4/23/74 much of
the carbon was present as soluble material. See Figure 5, Table 3 and
Figure 6 for the storm of April 23; Figure 7, Table 4 and Figure 8 for the
storm of May 10 and Figure 9, Table 5 and Figure 10 for the storm of May 12.
In the case of the storms of 5/10/74 and 5/12/74, the soluble organic carbon is
very low, approaching the concentration of organic carbon in deionized water
(10 mg/1). The suspended solids are quite high and contain virtually all the
organic material. As a result, the signal is noisy even in samples that have a
relatively low TOC. A study of these data indicate that stormwater is a more
severe sample than the simulated sample in which primary sludge is added to
primary effluent. This is probably because of the fact that primary sludge con-
sists largely of hydrated pulps and stormwater organic solids contain more
resistant carbon structures.
17
-------
I
it^bt^fe
-1 HR-
:»*7-
B-z3^~
E*fr
- — T-1—'
, ^V(ta
5::"*4
-*g--~*2
W?
- -* |
238 mg/l
215 mg/l
160 mg/l
-Omg/l
Figure 5. TOG chart for the storm of 4/23/74
18
-------
-f- SECKMANTOC
O RAYTHEON TOC
X SOLUBLE ORGANIC CARBON
A VOLATILE SOLIDS
• SUSPENDED SO LI OS
100
MINUTES
Figure 6. Characterization of the storm of 4/23/74
U—1 HR—J
-300 mg/l
A-230 mg/l
-140 mg/l
:-100mg/l
-80 mg/l
-Omg/l
Figure 7. TOC chart for the storm of 5/10/74
19
-------
Table 3. DATA FROM SAMPLES OBTAINED AT COTTAGE FARM STORAGE
FACILITY, STORM OF 4/23/74
Characterization
Suspended Solids
mg/l
Volatile Suspended
Solids mg/1
Soluble 1.
Organic 2,
Carbon 3.
(Beckman) 4.
5.
Mean mg/1
Total 1.
Organic 2.
Carbon 3.
TOC 4.
(Beckman) 5.
Mean mg/1
Stormwater TOC mg/1
Settleable Solids
30 Mill.
pH Stabilized
Sample Number
1
205
112
102
123
123
102
123
115
185
168
198
184
180
183
161
Trace
2.0
2
158
97
195
194
198
206
205
200
247
243
253
245
252
244
238
Trace
1.9
3
169
110
173
172
174
173
176
174
214
217
230
257
210
226
21S
Trace
2.2
4
141
92
162
158
164
144
167
159
191
195
192
194
218
198
202
Trace
2.3
5
121
86
128
128
128
110
130
125
188
163
167
175
156
170
158
Trace
1.9
6
103
76
115
121
121
110
121
119
150
140
157
165
142
151
150
Trace
2.1
7
104
71
122
121
114
111
118
117
128
157
150
145
155
147
140
Trace
2.1
8
101
70
128
125
128
114
115
122
137
140
138
131
150
" 139
142
Trace
2.0
9
103
71
120
120
125
124
118
121
140
150
153
157
153
151
142
Trace
2.0
10
82
54
18
20
15
17
17
18
131
150
135
141
125
136
149
Trace
2.3
11
84
62
117
108
120
119
119
117
132
130
120
141
130
131
135
Trace
1.7
20
-------
BECKMAN TOC
RAYTHEON TOO
SOLUBLE ORGANIC CARBON
VOLATILE SOLIDS
SUSPENDED SOLIDS
H —
MINUTES
Figure 8. Characterization of the storm of 5/10/74
21
-------
Table 4. DATA FROM SAMPLES OBTAINED AT COTTAGE FARM STORAGE
FACILITY, STORM OF 5/10/74
Suspended Solids
mg/1
Volatile Suspended
Solids mg/1
Soluble
Organic Carbon
(Beckman)
mg/1 Mean
TOC
(Beckman)
mg/1 Mean X
Std. Dev. S
Raytheon Stormwater
TOC
mg/1 Mean X
Std. Dev. S
Settleable Solids
cc/100
Sample Number
13
515
320
36
37
37
36
36
36
240
246
262
228
244
244
12.25
215
243
359
205
189
255
265
215
243
56.47
4
14
644
387
36
36
36
36
36
36
329
287
337
318
290
312
22.69
280
210
300
395
350
310
307
62.91
6
15
624
359
30
30
35
30
33
32
320
312
312
312
339
319
11.70
188
240
265
290
200
212
205
230
35.13
3.5
16
391
208
25
26
30
25
25
26
175
168
177
165
177
172
5.55
134
140
145
135
145
170
144
13.20
2.5
17
377
206
23
23
23
23
23
23
125
115
148
129
119
127
12.81
112
120
135
130
150
129
14.55
2.0
18
266
129
22
22
22
22
22
22
119
124
104
129
107
116
10.78
104
117
100
110
120
110
8.44
1
19
338
193
22
17
22
17
22
21
119
101
94
112
122
110
11.41
80
94
105
112
75
85
92
14.52
Trace
20
259
114
22
20
20
22
23
21
95
76
60
76
76
77
12.39
68
85
68
72
81
92
94
80
10.94
Trace
21
274
100
20
23
22
21
20
21
82
68
95
72
84
80
10.6
80
76
76
76
110
65
65
79
16.56
Trace
22
195
75
21
23
24
23
21
22
72
67
60
Vs
63
67
5.61
68
50
60
60
80
65
65
10.49
Trace
23
248
99
23
23
25
23
24
24
60
73
63
82
70
70
8.68
60
68
95
95
84
98
81
16.61
Trace
24
174
75
36
33
32
37
33
34
58
63
62
54
60
59.4
3.5
45
52
67
67
85
54
64
75
63
13.97
Trace
to
-------
Figure 9. TOC chart for the storm of 5/12/74
"t BECKMAN TOC
O RAYTHEON TOC
X SOLUBLE ORGANIC CARBON
A VOLATILE SOLIDS
• SUSPENDED SOLIDS
X X *
SOLUBLE OH6ANIC CARBON | |
Figure 10. Characterization of the storm of 5/12/74
23
-------
Table 5. DATA FROM SAMPLES OBTAINED AT COTTAGE FARM STORAGE
FACILITY, STORM OF 5/12/74
Characterization
Suspended Solids
mg/1
Volatile Suspended
Solids mg/1
Soluble Organic
Carbon (Beckman)
mg/1 Mean
TOC (Beckman)
mg/1 Mean X
Std. Dev. S
Raytheon
Stormwater
TOC
mg/1 Mean X
Std. Dev. S
Settleable Solids
pH *2
Sample Number
1
188
78
15
15
15
15
15
15
60
71
58
76
74
68
8.26
60
27
26
42
50
27
30
42
38
12.59
-------
To meet the severe sample loading that the stormwater sample presents to the
TOC analyzer, two modifications were explored. One of these was to improve
the combustion efficiency of the reactor; the other was to modify the signal.
Both techniques were investigated. Between the two alternatives, an improve-
ment in combustion characteristics was preferred.
In order to improve the combustion characterisitcs of the reactor, two design
changes were initiated; the sample delivery tube cooling jacket was insulated
from the reactor body. A cooling jacket temperature control was installed in
the instrument.
Another method of producing a signal representative of the stormwater TOC
content is by integrating the signal either electronically, or physically in time.
It would appear that this approach would be valid if it were implemented since
Raytheon is displaying means and the means of the continuous analyzer tend to
approach the means of the discrete Beckman samples, even though the variance
might be quite large. This is particularly clear in the storm of 5/10/74 (see
Figure 7 and 8, and Table 2). In early tests, an analog filter was designed as
seen in Figure 11. It worked well but was more expensive than the gas mixing
volume added to the outlet of the condenser. The gas mixing volume provides an
integration time of four minutes. The effect of this added gas mixing on the
integration time can be seen in Figure 12. This 4-minute integration time in
conjunction with intensive combustion techniques produced the improved data in
Figure 13. These data were collected on the storm samples from Boston on
June 1 and are presented in Figure 14, 15 and 16.
Signals were developed with and without the addition of a gas mixing volume.
The noise was diminished dramatically. The standard deviation of the data
collected in the absence of the mixing volume had been about 12%. With the
added volume, this was reduced to about 3%. This can be seen by an examination
of the data in Table 6. The data obtained without the additional gas mixing
volume are presented in Figure 14; they are separated for the purpose of defin-
ing the discontinuity between samples. Figure 15 presents the same samples
run with the added gas mixing volume. These are presented as a continuity to
better indicate how such a signal would appear in the continuum of an actual
storm event. The complete characterization is seen in Figure 16. These curves
show the close proximity of the Beckman means to the stormwater TOC means,
both with and without added mixing volume.
In summary, the problem of suspended solids has been solved by effective homo-
genization, careful transport and delivery, and efficient combustion and pro-
cessing techniques that lead to a uniform signal that is a quantitative measure
of the stormwater TOC. Unlike other commercial devices, the developed TOC
25
-------
NJ
RV2.
\ VAEG
pAAA-»—vV\A-» 1( t-WV^-»-vV\A-f r
L 110 Is I s 1 ci
HI PASS
OUTPUT
INPUT \
t
FRE.Q EAMCi P^S^
*VDA OUTPUT
FLL
ACTWE FILTER
iU^R-BROWW MODE.U VJf^?-
FREQ
LO PA.-5.?>
OUTPUT
t
\MPUT a COMMON (+> I -1
14
5
-V
5V -\
5V
V
-4—< O-SV OUTPUT
1 H,0 >5.\MEG
> Kio S i;«i
>A'3.9K S ±5*
Uj
1— 1
J4
1 1 — 1
V
< O - 5 WJ OUTPVJT
15V POV(ER SUPPLY
MOOEL W1IADIS-.100
^
MOTES'.
UML.fa^ CTTHETOMVit
I CAPAC\TOR
Z. ELECTRICAL
RESISTOR RKTm&=, KRE \/\OWKTT tllk
1 PART1NL HE.FEREV4CE Ofi\aNKT\OVft
ARE SHOWN. FOR COMPLETE
NUMBER AV4D '
Figure 11. Analog filter schematic
-------
1000 I—
423mg/1 POTASSIUM ACID PTHALATE
-X
900 —
800 —
700 —
600 —
5
<
ui
DC
111
500 —
400 —
300 —
200
100
X MIXING
\ VOLUME
\ OF 375cc ADDED
X WITH MIXING VOLUME
• WITHOUT MIXING VOLUME
5 6
MINUTES
Figure 12. Integration Time Using Mixing Volume
27
-------
NOISY SAMPLE
NO Ml XING VOLUME
375 cc Ml XING
VOLUME ADDED
1000 cc MIXING
VOLUME ADDED
1000 cc MIXING VOLUMES-
PLUS MIXING BEADS
Omg/l
Figure 13. Effect of added mixing volume
|-425 mg/l
Figure 14. TOO chart for the storm of 6/1/74 without added mixing volume
28
-------
U-1 HR—J
j ^ :i i j-425 mg/l
-212 mg/l
-106 mg/l
Omg/l
Figure 15. TOC chart for the storm of 6/1/74 with added mixing volume
BECKMANTOC
RAYTHEON TOC
SOLUBLE ORGANIC CARBON
VOLATILE SOLIDS
SUSPENDED SOLIDS
30 40
MINUTES
Figure 16. Characterization of the storm of 6/1/74
29
-------
Table 6. DATA FROM SAMPLE OBTAINED AT COTTAGE FARM STORAGE
FACILITY, STORM OF 6/1/74
Characterization
Suspended Solids mg/1
Volatile Suspended
Solids mg/1
Soluble
Organic Carbon
(Beckman)
mg/1 Mean X
Std. Dev. S
TOC
(Beckman)
mg/1 Mean x"
Std. Dev. S
Raytheon Stormwater
TOC
mg/1 Mean X
Std. Dev. S
Stormwater TOC
Gas Mixing Volume
mg/1 Mean x"
Std. Dev. S
Sample Number
13
260
90
25
25
20
21
23
23
2.28
176
188
212
216
207
200
17.09
163
170
172
183
168
193
168
193
168
153
172
171
11.38
187
181
176
176
187
181
5.50
14
263
101
21
19
18
17
17
18
1.67
211
162
149
128
119
154
36.19
163
140
131
133
144
140
133
131
131
128
128
131
128
135
9.71
165
160
156
150
150
156
6.50
15
224
77
14
14
15
15
14
14
0.55
115
94
70
90
82
90
16.62
126
100
122
117
140
104
91
79
83
87
87
103
20.24
121
116
116
116
111
116
3.54
16
254
108
15
14
15
15
14
15
0.55
53
73
84
63
75_
70
11.91
91
75
75
67
67
87
70
58
8T_
75
11.10
98
93
89
89
89_
92
3.97
17
123
8
13
13
12
14
16
14
1.52
86
50
73
75
60
69
13.99
59
67
55
55
55
59
55
59
51
51
57
4.70 •'
85
80
85
80
85
83
2.74
18
106
27
28
28
28
25
28_
27
1.34
82
71
62
71
80
73
8.04
75
79
• 71
67
67
63
108
71
75
83
76
12.76
85
85
80
85
80
83
2.74
19
145
71
34
32
35
34
34
34
1.10
107
75
73
73
77
81
14.63
75
75
83
83
95
75
71
11
79
8.12
89
98
93
89
85
91
4.92
20
102
66
39
38
41
38
38
39
1.30
84
84
75
75
64
76
8.26
87
91
75
79
79
75
87
75
71
80
6.86
72
80
80
80
85
79
4.67
30
-------
system can be standardized on KHP. The technique of signal smoothing by gas
mixing volume has been demonstrated on actual stormwater samples. The gas
mixing volume was employed in all later testing.
MAINTENANCE
The successful operation of an on-line TOG depends upon the maintenance pro-
gram, and an adequate supply of spare parts necessary to replace worn or con-
sumed components. The remote location of most stormwater monitoring sites
means that the prescribed maintenance needs must be minimized. With this in
mind, the stormwater TOC system was designed with a maintenance program
that will make it possible to obtain long and continuous operation of the on-line
TOC. This maintenance program can be minimized by providing automatic
features which will permit remote operation. Among these features, auto-zero
and auto-span have been provided for the Cottage Farm Storage Facility. The
maintenance tasks and frequency are shown in Table 7.
Table 7. MAINTENANCE REQUIREMENTS
MONTHLY
1) Replace sample delivery veins
2) Check reactor pressure
3) Chart paper and ink supply
4) Replace span gas
5) Clean reactor well
6) Replace homogenizer gland
7) Lubricate peristaltic pump motors
TRI-MONTHLY
1) Clean Lira and replace filter
2) Renew acid supply
SPARE PARTS
1) Replacement reactor
2) Replacement sample delivery motor
3) Replacement Lira filter
4) Span gas
5) Asco solenoid
6) Replacement gland for homogenizer
31
-------
In laboratory studies, Raytheon established that under present operating con-
ditions, it is possible to substitute CO2 free air for oxygen. The preparation
of this material can be automated and would eliminate the need for periodic
oxygen replacement. Since oxygen is inexpensive and readily available in suit-
able purity, it was used for the field demonstration program. Our maintenance
experience shows that plugging by refractory solids (salts and lime) has not
occurred in six months of operation. Although reactor flushing is not needed,
the field demonstration model contains flushing equipment which is also used for
rapid reactor cooldown.
A program providing biweekly maintenance and exercise of the instrumentation
on-line at the Cottage Farm Storage Facility was instituted.
The assembly of instrumentation has operated at theCranston Municipal treat-
ment plant for a continuous period of 5000 hours. If a storm lasts an average
of 2 1/2 hours, 5000 hours can be equated with 200 storm-event equivalents.
Clearly the life of the instrument in the stormwater installation will be deter-
mined by conditions other than instrument use.
The limitation on the usefulness of the stored and empowered instrument is the
vein life of the Tygon tubing in the peristaltic pumps. This material undergoes
cold flow on standing and experiences wear in use. In the standby condition, use
of the pumps minimizes the effect of cold flow. This can be programmed into
the control circuit if necessary. The standby condition does require the con-
tinuous operation and control of heaters for the reactor as well as the infrared
detector. Heaters have a limited life and would be expected to limit the life of
the hardware. Another factor that might be expected to present a problem for
the instrument itself is corrosion of various components in the system. This
would be dependent on how the system is used, protected, ventilated, and main-
tained.
To get a sense of reliability of the stormwater TOC installation, a reliability
estimate was made as shown in Appendix 1. In this estimate, it is determined
that the mean-time-between-failures is 3,557 hours. Factors that are more
likely to limit the collection of useful data during storms are likely to be: inade-
quate maintenance, failure of the sample delivery system to the instrumentation,
or neglect of the hardware after accidental shutdown, as will occur during power
failure or temporary loss of water.
32
-------
SECTION VI
INSTALLATION OF A SYSTEM IN A REMOTE LOCATION
The ability of a system of instrumentation to operate in a remote location
with a high degree of reliability depends upon the probabilities of failure
of each essential component in the system train. In the case of a new
and advanced technology requiring a cascading of subsystems, these failure
probabilities can combine to be quite large and very often result in a total unreliability
(50% probability of failure) of the system. It might be valuable to digress at
this point to provide an example of this type of occurrence. Consider that
there is an assembly of orifices transporting a liquid (typically stormwater) in
which there is a 5% probability of plugging each orifice; now cascade ten such
orifices in a sample transfer system and the chances are 40% that when the
system is called upon to respond it will be in an unready state. This would be
a nuisance in the laboratory but a serious problem for a remote monitoring
system. This is what one might expect in taking representative sample of
stormwater and delivering it to the reactor of an operating TOC.
A stormwater monitoring system installation must be guided by an intention to
simplify the liquid transfer process, using minimum number of components.
Special care is necessary at tube connectors so hangups on shoulders will not
cause plugging. Complexity that would be expected from stream selection
systems, flushing, and switching would have the effect of increasing the prob-
ability of failure. To obtain a dependable measurement, it is necessary to
keep the assembly as simple as possible, consistentwith the needs of the storm-
water monitoring application.
SAMPLE DELIVERY SYSTEM
With this problem in mind, it is essential that an operating TOC of high reli-
ability be furnished with a sample from a highly reliable sample delivery system.
The operation at the Cottage Farm Storage Facility is intended to take advantage
of the sample delivery system consisting of a submerged pump located below the
level of flow in the entering channel, that has been installed on-site since con-
struction of the facility. The sample can be delivered to the stormwater TOC
analyzer1 s homogenizer at high velocity. A stream can be separated for con-
tinuous homogenization and delivery of the TOC.
33
-------
Since the design of the Cottage Farm Storage Facility, new technology has been
brought to bear on the design of stormwater sampling systems. Much of this has
been as a result of the work of Dr. Phillip Shelley and his team at Hydrospace-
Challenger. The Hydrospace-Challenger design has provided solutions for many
of the problems of stormwater sample delivery and redundancy has been provided
in the system. Air and water purging, possible with the Hydrospace-Challenger
delivery system, are valuable features that might interface very well with the
stormwater TOC system. At this time, work is still in process which will result
in the optimization and testing of the Hydrospace-Challenger's current design
configuration. When the current testing programs are complete, it may become
appropriate to bring the sample delivery technology and stormwater TOC tech-
nology together to produce a highly reliable stormwater monitoring system.
The limitation on the usefulness of the stored and unpowered instrument is the
vein life of the tygon tubing in the peristaltic pumps. This material undergoes
cold flow on standing and experiences wear in use. In the standby condition, use
of the pumps minimizes the effect of cold flow. Under continuous standby use,
veins must be changed at approximately one-month intervals. Other practical
life limitations like heater life and corrosion are considerably longer term
issues.
To get a sense of reliability of the stormwater TOC installation, a reliability
estimate was made as shown in Appendix 1. In this estimate, it is determined
that the mean-time-between-failures is 3,557 hours. Factors that are more
likely to limit the collection of useful data during storms are inadequate main-
tenance, failure of the sample delivery system to the instrumentation, or neglect
of the hardware after accidental shutdown, as will occur during power failure or
temporary loss of water.
MANHOLE OPERATION
The prospect of installing the hardware in a manhole or in a water flooding
situation was kept in mind during the design and laboratory testing of the storm-
water TOC monitoring system. Considerable packaging would be required to do
this (see Appendix 2). Nothing in the developed TOC monitoring system rules
out the possibility of lowering the unit into a manhole.
There are, however, other manhole operational configurations that would require
very little modification; for example, it would not be difficult to drop a sample
delivery system into a manhole in proximity to an instrument shack containing
the TOC analyser and necessary water, power, and oxygen. From this man-
hole, a sample could be delivered which would be processed periodically or
continuously for TOC.
34
-------
Alternatively, a trailer might be outfitted for the measurement of TOG. The
trailer could be delivered to a manhole. Power and water could be provided
from a local supply or from a gasoline generator and recirculating water supply.
This would require a portable sample delivery system which could be lowered
into the manhole. It should not be difficult to be on-line with very short notice.
Either the shed or trailer will provide a useful system at the inconvenience of
above ground installation. Both alternatives will provide useful data.
DOCUMENTATION
The documentation that is necessary to describe stormwater TOC consists of
five essential parts:
• Circuit diagram (Figure 17)
• Flow diagram (Figure 18)
• Stormwater TOC control circuit (Figure 20)
• Installation diagram (Figure 22)
• Flow manifold (Figure 23)
• Design specification of stormwater TOC (Appendix 3).
35
-------
OJ
Figure 17. TOC Electrical Schematic
-------
SOLENOID
•VALVES
/VALVES v
I/R ANALYZER
RECORDER
ROTOMETER
CONTROL
VALVE
EXTERNAL
02SUPPLY
SAMPLE FROM
HOMOGENIZER
FILTER j"j
FLOW
CONTROL
"o
CHECK
VALVE
CONTROL
VALVE
SPAN
GAS
PRESSURE
SENSOR
DRAIN
FLUSH
PUMP
GAS
TRAP
r
(2 CC/MIN)
CONDENSER
REACTOR
SPARGE
SPARGE ROTOMETER
GAS ~
3-WAY
VALVE
CALIBRATED
SAMPLE
SAMPLE
PUMP
(6 CC/MIN)
ACID
PUMP
(0.05 CC/MIN)
Figure 18. TOC analyzer flow diagram
37
-------
SECTION VII
PHASE II-FIELD TEST
FIELD TEST INSTALLATION OF A STORMWATER TOC
On July 15, 1974, after modeling and testing a stormwater TOC system in Phase
1, the assembly was found to be satisfactory and the program was continued into
Phase II.
Phase II objectives were:
1. Installation of device(s) at a suitable combined sewer site, perhaps in
conjunction with an existing stormwater project, and/or
2. Installation at a sewage treatment plant influent sewer or junction over-
flow point where comparative analytical data could be be obtained during
storm flows
3. On-site demonstration and evaluation to establish comparisons, adapt-
ability, reliability, maintenance, and operational requirements, and
accuracy under actual field conditions
4. A final report which would summarize all aspects of the program.
An automated TOC was assembled to obtain data at the Cottage Farm Storage
Facility in Boston (see Figure 19). This installation was selected because the
site was well maintained and provided for sample collection which could be of
value in confirming and interpreting the data obtained by the instrumentation.
As stated in the design goals, it was of prime importance that the final design of
the stormwater TOC hardware must be fully automated so that data could be
collected with a minimum of operator attention. The Cottage Farm installation
was selected to test this automated operating mode as well as the TOC design
and operation.
38
-------
Figure 19. Cottage Farm Storage Facility stormwater TOC installation
39
-------
SECTION VIII
CONTROL SYSTEM AND ASSEMBLY
AUTOMATION AND INSTALLATION OF THE STORMWATER TOC
The stormwater TOC consists of a number of assemblies selected, modified,
and designed to give the automated performance necessary for the stormwater
TOC analysis (see Figure 20). The components used in this assembly are
following are shown in Table 8.
Table 8. MODULAR COMPONENTS USED IN STORMWATER TOC
Component
Stormwater Modifications
Raytheon TOC
Raytheon Homogenizer
MSA Auto-zero/Auto-span
Turn on and Operate
Flush and turn off
Control Unit
Modified for intermittent sample handling
Installed as delivered
Purchased and Installed as delivered
Designed for stormwater application
Designed for stormwater application
Designed for stormwater application
In order to deliver power and provide the necessary sequencing of events, a
control unit was designed which was capable of handling the turn on transient
power required by the assembled hardware. All power (117 Vac, single phase,
20A) was delivered to the control box. Power was delivered through 35A rated
mercury power relays to the control system operating three operational modes;
1. standby mode
2. operate mode
3. automatic flush and shutdown mode.
In the standby mode, power is delivered and controlled to the reactor heater
at all times. The infrared detector is powered at all times. All other compon-
ents are in the off condition until a "system on" signal is soised.
40
-------
CHASSIS
GND
NOTE' All reliyi BIB ihown in ttw u up owe red pontu
Figure 20. Stormwater TOG control circuit
In operating this instrumentation, it is desirable to obtain a sample as early as
possible in the storm event. In order to do this at the Cottage Farm Storage
Facility, the signal used to start up the Cottage Farm Sample Pump was also
used to start the Stormwater TOC instrumentation. In other operations other
signals might be used. The signal needs to be early enough to activate the auto-
zero and auto-span functions, and yet process the first flush loading of the storm
event. This turn on signal activates all components listed in Table 9 simulta-
neously. The first display is the auto-zero/auto-span data which operates from
the 5-minute time delay relay circuit.
Table 9. COMPONENTS ACTIVATED IN THE OPERATE CYCLE
Component
Operation
Auto-zero/Auto-span
Time delay relay (t + 5 min)
Chart Recorder
Homogenizer
Delivers 0 for automatic zero signal
£
Delivers . 9% COg for automatic span signal
Provides activation signal for auto-zero
Powers recorder for readout
Begins the grinding action on the Stormwater
sample
41
-------
Table 9. COMPONENTS ACTIVATED IN THE OPERATE CYCLE (Cont)
Component
Operation
Sample Transfer Pump
Acid Pump
Sample Delivery Pump
Drain pump
Normally closed condenser
solenoid valve
Normally open sample
cooling solenoid valve
Normally closed oxygen and
sparge gas delivery valve
Delivers sample to the sparge system
Liberates inorganic CO2 dissolved in the
sample and removed by sparge gas
Delivers sample to the reactor
Removes water as it condenses
Increases flow of water to condenser
Decreases flow of water to sample cooling
jacket
Provides oxygen for sparging and combustion
In addition to the operations described in Table 9, the 20-second time delay
relay is deactivated when the start-up signal is delivered. This action completes
the circuit to the flush cycle which is now ready to receive the "turn off" signal
and proceed with the 1-hour flush cycle when the storm is over.
The 1-hour flush cycle acts to deliver water to the stormwater TOC sample
delivery system at the input of the homogenizer (see Figure 21). The water is
processed through the sample system so as to dissolve and entrain any residue
that is held up during the processing of storm sample. This action prepares
the instrument for the next storm.
The flush cycle is activated by a clock timer that delivers power to the operating
units and signals a water delivery system that flushes the sample line and
operates the unit for a period of an hour before final shutdown. The stormwater
TOC system remains in this standby condition until the system command signal
is again activated.
Sample Transfer and Water Flow Systems
One of the substantial problems associated with the accurate measurement of
organic pollution is the problem of transferring a representative sample from
an inhomogeneous flow stream. Not only do the high velocity and turbulent flow
characteristics in storm or combined sewers present problems of solid segre-
gation and non-representative sample delivery, but the entrained solids (i.e.,
ladders, rocks, tires, etc.) tend to tear out the sample takeoffs and associated
insertions in the sewer system.
42
-------
WATER
SAMPLE FLUSH
SOLENOID
N -
SAMPLE
IN
ACID IN
SPARGE GAS IN
n,
"2
WATER IN
SAMPLE
RAYTHEON
HOMOGENIZER
EXCESS
SAMPLE
J DELIVERY u-
, COOLING r
WATER
NEEDLE VALVE
FLOW CONTROLLER
TO
DRAIN
Figure 21. Manifold of the liquid flow system
There have been a number of assemblies designed to address these problems.
They have been studied in Dr. Shelley's previously mentioned work. All of the
commercially available sample delivery systems have been shown to have some
problems associated with their operation. This was also the case at the Cottage
Farm Storage Facility. Nevertheless, a sample delivery system was necessary,
and there are several reasons for applying it to the measurement. The installed
hardware had a basic simplicity that recommended it in the absence of available
and demonstrably more reliable hardware. This simplicity of design is such as
to permit repair on-site by personnel after the storm.
The sample delivery system at theCottage Farm Facility consists of a Kenco
Model 93 pump rated to deliver 600 gallons per hour to a height of 60 feet. This
pump is located in the flume to the wet well behind a cement support in the direct
path of the flow to the wet well (see Figure 22) and beyond the bar screen of the
Cottage Farm system. Even in this protected area, the pump takes severe
abuse and has had a considerable repair history.
In Phase I simulation of storm flow, the sample delivery system in Figure 3,
was assembled and found to give satisfactory quantitatively representative
sample delivery. In proceeding with Phase II, it was intended that the simulated
43
-------
Figure 22. Location of sample take-off pump
44
-------
system be followed as clasely as possible. Since the Cottage Farm Facility
sample delivery system is of primary use in providing sample for the Protech
sample collection unit, representative samples from this delivery system were
used to compare the stormwater TOC data with the characterized discrete
sample from the Protech unit (see Figure 23).
Sample, Water, and Waste Manifold
To transfer and control stormwater, cooling water, and waste in the automated
stormwater TOC installation, a manifold and flow control system was devised
which is uniquely applicable to standby TOC instrumentation. During standby,
it is desirable that a minimum of water flow be provided for the condenser and
a larger flow of water be provided to cool the sample delivery tube, minimizing
the risk of corrosion and erosion at higher temperatures. The flow manifold
describing the liquid flow and drain system is diagrammed in Figure 23. A
photograph showing the drain and delivery assembly is shown in Figure 24. The
diagram in Figure 21 shows the flow pattern for all liquids in the stormwater
TOC with the solenoids in their normal, unactivated condition during the standby
operation. In the operate cycle, the deactivated solenoids in the cooling water
flow are actuated. The normally closed valve in the sample line is actuated to
deliver cleaning water during the flush cycle.
Figure 23. Installation of stormwater TOC in the Cottage Farm Storage Facility
delivery system
45
-------
Figure 24. Sample, water, drain, and waste delivery system
46
-------
SECTION IX
OPERATING EXPERIENCE
INSTALLATION, MODIFICATION, AND MAINTENANCE
After design of the assembly described in the previous section, the instrumentation
was delivered to the Cottage Farm Storage Facility. The final installation was
completed by October 1, 1974, after which it was found that the reliability of the
Cottage Farm Facility sample delivery system was a limitation on our ability
to collect data reliably. This was not surprising in view of the state-of-the-art
of reliable sample delivery.
The stormwater TOC experienced one hardware failure in which the sparge
pump developed bearing failure. As a result of this event, sample was not
delivered during the storm of October 31. Because of this occurrence, the pump
was removed from the system and sparging accomplished from the oxygen supply.
This modification represented a considerable improvement in expected reliability
without serious sacrifice in cost for expendable oxygen.
In order to maintain the instrumentation the program of maintenance in Table 7
(Section 4) was instituted, eliminating the need for daily zero and span adjust-
ment. In addition, exercise of the system was provided once every two weeks.
During this period, the system was calibrated. Gas supplies were checked, and
the sample and signal delivery system were tested to establish proper operation
of the equipment. It was not found necessary to replace either the sample delivery
veins or the homogenizer glands during the intermittent operation at the Cottage
Farm Storage Facility, which spanned a ten-week period.
DATA COLLECTION
The stormwater TOC instrumentation was on-line at Cottage Farm from October
1, 1974 to December 10, 1974. The first operation of the Stormwater TOC
assembly took place on October 16, 1974. The data collected during that event
are reported in Figure 25. On this occasion, the instrument was turned on
manually some time late in the storm so that the operation of the hardware
could be viewed in the presence of the skilled technician responsible for its
47
-------
- 860 mg/l
;-400 mg/l
200 mg/l
80 mg/l
10PM 9PM 8PM 7PM 6PM 5PM 4PM 3PM
OC EC
E E
:H-
1*
i
•*tV-jbTr±-
4-fr-^Vr-T-
= t : ..iqz.
^ I - , — 1
:=i
i
'±^-
--f^
rfc
WATER SHUT OFF
FLUSH 3:45 AM
2AM
1 AM
4a
-
- 860 mg/l
El- 400 mg/l
mg/l
- 80 mg/l
12 M 11PM
Figure 25. Operation of the Cottage Farm Storage Facility, 10/16/74
48
-------
installation and maintenance. The severity and duration of the storm flow of
October 16 was unusual. During this event, a considerable amount of rock and
debris was entrained in the sample which resulted in damage to the Cottage
Farm Storage Facility sample delivery system requiring replacement of the
sample delivery pump.
The damage to the Cottage Farm Storage Facility delivery system required
repair and replacement of several system components among them was the
sample pump. These changes required turn on and turn off operations which
activated our system in the absence of a storm repeatedly and without sample
being present. This resulted in a number of unexpected and unreported start-
ups at the time. During this activity, the bearing seized on the sparge pump
and the span gas was depleted. Rules for reporting operation was associated
with the occurrence of storm events. The problem was not recognized until a
demonstration was attempted on 10/30/74 and could not be repaired in time for
the storm of 10/31/74.
The next storm occurred for a half-hour during the evening of November 19, 1974.
During this storm, the Raytheon instrumentation came on-line as expected. Auto-
mated features operated as designed. Data were collected as shown in Figure
26. In this figure a standardization curve is shown in which standard solutions
STORM OF 11/19/74
— 800 mg,
— 400mg/l
— 200mg/l
STANDARDIZATION ON KHP
11/18/74
Figure 26. Display of storm data of 11/19/74 for comparison with previous day's
calibration curve
49
-------
of potassium acid phthalate, KHP, were passed through the TOC unit. Here we
see the uniform combustion characteristics that are obtained with this stand-
ardization. This procedure scales the unit at three points, 201 mg/1 TOC,
402 mg/1 TOC, and 804 mg/1 TOC.
On December 2, 1974, a very severe storm hit the Cottage Farm Storage Facility
(see Figure 27). The system turned on and operated in its automatic mode for
a period of about 15 minutes. At that point in time, sample delivery to the
stormwater TOC failed. The signal that is used to start up the pumping station
is also used to start the TOC analyzer. The frequent number of zero and span
signals indicate that the pump had several starts and stops during the storm of
December 2. The operation of the instrumentation was described to the con-
tractor by the Cottage Farm Storage Facility plant manager; at that time, a storm
was occurring and the instrumentation was reading zero. The contractor's
technician arrived during the storm and reported that a pump failure had occurred.
The stormwater TOC unit was standardized to verify that it was operating and that
it could have processed a storm if sample had been delivered. When the operator
of the Cottage Farm Storage Facility realized that the pump was not functioning,
he shut off pump power, which in turn, shut off the stormwater TOC system.
This was the last operation of the stormwater TOC system during storms at
Boston's Cottage Farm Storage Facility.
800 mg/l
Figure 27. Data collected during the storm of 12/2/74
50
-------
The data collected on the unit during the test period are encouraging. The
reliability of the measurement in this particular test was largely limited by the
ability of the sample transferring system to deliver a sample continuously
during the severe conditions that were associated with storm flows.
Several successful experiences in data retrieved suggest that this stormwater
TOC could be used as a permanent feature in a stormwater center such as that
at the Cottage Farm Storage Facility.
51
-------
SECTION X
APPENDICES
52
-------
APPENDIX 1
RELIABILITY
The failure rate of instrumentation when plotted against time tends to follow
the typical bathtub curve in which failure, at first, may occur with some high
probability determined in large part by assembly errors or component failures
associated with manufacturing quality control. Problems of this type are
corrected by repair and replacement as necessary. When these initial problems
are corrected, the instrumentation enters a period in which the reliability can
be assessed by referencing the failure rate of the individual components. In
the course of continued operation, wear, sooner or later, becomes a problem.
Failure rate increases because of wear and decay of various kinds. This com-
pletes the bathtub curve (Figure 28).
The initial failure modes are generally covered under manufacturers' guarantees
and respond to improved quality control and engineering. The final failures
caused by wear and corrosion cannot be assessed until more experience is
obtained with the hardware operating as an assembly.
NUMBER
OF
FAILURES
FAILURE
DUETO
DEFECTIVE
COMPONENTS
FAILURE
DUETO
WEAR
PERIOD OF CONSTANT RELIABILITY
TIME
Figure 28. The bathub curve
53
-------
The contractor has addressed the reliability during the interval period of con-
stant failure rate. Data for this evaluation was obtained placing the operating
components of the stormwater TOC in sequence and summing the failure rates
of these individual components so as to obtain an estimate of the mean-time-
between-failures (see Figure 29).
Failure rates were derived from the following sources:
1. Electrical Components (Switches, Relays, Diodes, etc. )— Failure rates
have been taken from MIL-HDBK-217B and the RADC Reliability Note-
book, Volume II, assuming commercial grade components and ground-
fixed environment
2. Mechanical Components (Pumps, Valves, etc.)— Failure rates have been
taken from AVCO Reliability Engineering Data Series. Failure rates
used are the mean generic failure rates. Wearout rates are not included
3. Recorder and Analyzer Assemblies— Failure rates for the recorder and
analyzer are estimates based upon assemblies of similar complexity.
The mean-time-between-failures of the assembled stormwater TOC instrumen-
tation was calculated to be 3,558 hours.
The model and individual failure rates of the components are described in
Figure 29. The sum of the individual failure rates result in an estimate of 281
failures/million hours = \.
-r- =3, 558 hours = mean-time-between-failures
The reliability, R, is a probability funtion varying between zero and one in
which
where "p" is the probability of failure, and " X" is the failure rate.
These data indicate a mean-time-between-failures of approximately five months
for the final stormwater TOC assembly.
54
-------
RESERVOIR
& MOTOR
STIRRER
.58 f/106 HRS
13.5 f/106 MRS 13.5 f/106 MRS 13.5f/10BHRS
.5f/106HRS 13.5 f/106 HRS 3.5 f/106 MRS 2.3 f/106 MRS 4.9 f/106 MRS
13.5 f/106 HRS 1.7 f/106 HRS .3 f/106 HRS
11.0 f/106 HRS 11.0 f/106 HRS
75 f/106 HRS
50 f/106 HRS
25.5 f/106 HRS
S\=281.08f/106HRS MTBF = 3557 HRS
NOTE: f/106 HRS = FAILURES PER MILLION HOURS
MTBF =MEAN-TIME-BETWEEN-FAILURES
Figure 29. Reliability model for stormwater TOC analyzer
55
-------
APPENDIX 2
MANHOLE MOUNTED CONFIGURATION
As has been mentioned in the body of the report, the installation of the TOC
instrumentation in a manhole would require considerable engineering design.
This design might well be approached by selecting hardware and assembling it
in a way that will accomplish the operations detailed in Figure 1 by means of a
modular design (Figure 30). The purpose of this assembly mode is to separate
and isolate the operational components into discrete functional units so that
failure in one of the components will not transmit its destructive effect beyond
its own module. For this purpose, three modules have been chosen:
1. a pumping and sample preparation module
2. a reactor and condenser module
3. an electronic module.
This whole assembly is sized so as to be encased in a corrosion resistant steel
envelope that can be sealed at the top by a compressed gasket that will protect the
expensive equipment from corrosion and provide a suitable port for utilities to
be delivered and heat to be removed.
If a modular design was not used, a failure in one component could do extensive
damage to other components. Consider two fairly probable failures; worn pump
veins and a corroded or faulty reactor. Both of these components might fail
because of wear or corrosion. In a non-modular assembly, a broken pump would
cause flooding of the total assembly and confine the flood so as to eventually get
liquid and/or vapor to the expensive electronics and the reactor. The general
effect would be to destroy the equipment totally. If the reactor should fail, the
confined space of a module would allow this event to be sensed early in the storm
and shut the total sample delivery and operation down.
Routine maintenance will still be necessary on this instrumentation. Because of
the added complexity and exposure of the manhole-operated equipment, the mean-
time-between-failures would be less than the 3, 558 hours in our present assembly.
56
-------
8" ACCESS COVER
ELECTRONICS
SIGNAL PROCESSOR
INFRARED DET
TOCCONTROL
POWER CONTROL
GAS FILTERS
SAMPLE
OUT
COOLING
COIL
DRAIN
Figure 30. Modular manhole configuration
57
-------
In our present assembly, the probability of failure in one month is 17%; the
probability of failure in two months is 27%. This is large enough so that service
should be provided at least once every two months, and would require removing
the hardware from the envelope and repairing it, or replacing it with an inspected
and tested assembly. This presents the problems of introduction and connection
of utilities. The reliability model, which was used to describe the above-ground
instrumentation, assumes that routine maintenance is provided the hardware.
Raytheon has provided this maintenance once a week, and more recently once
every two weeks without serious problems, but it is doubtful that this could be
extended longer than once a month. It can be concluded that a design goal to
improve MTBF of the manhole equipment by an order of magnitude over the
above ground configuration should be established.
Some redesign will be necessary in the manhole configuration. The components
must be sized so as to be able to be inserted into a manhole with a diameter of
3 feet. This requires that all components greater than that size have to be
oriented so as to fit the confined space. This will require redesign of the
motor bearings in order that they can be loaded vertically instead of horizontally
as is the case with the present blender configuration. In an effort to confine the
heat to the inside of the reactor, the reactor should be redesigned with heat being
provided internally to a ceramic shell. This would have the effect of maximizing
the thermal gradient at the outside of the reactor allowing for a minimum heating
of air with external heaters. Alternatively, inductive heating might be adapted
to this configuration. It would be much more expensive, but would provide the
necessary energy only when needed. In most cases, this would not exceed three
or four hours, but it would have to be sized so as to operate for the possibility
of a long and extended storm. In any case, provision would have to be made for
removal of the excess heat by cooling. This need for cooling would be less during
standby and would have to be increased when heat is transferred during operation.
A closed Freon system could be used for all cooling needs, including the condenser.
This would simplify the design and eliminate the need for water in the system.
Flushing of the sample delivery line would be provided externally. This has
already been designed into sample delivery systems, such as that of Hydro-
space-Challenger. These redesigns are difficult to say the least, and in the
case of the reactor, the new design would test the state-of-the-art in this appli-
cation. They are required because of the necessary total enclosure of the appli-
cation. A trailer or permanent structure above ground would not require this
redesign because it could be vented to the ambient atmosphere and drained with
the aid of gravity. These conditions minimize the possibility of the destruction
of one component by the neglected failure of another.
58
-------
Freezing would not be a problem in the environment of a heated instrument in a
manhole; overheating is a problem. The three possibilities for removing heat
from the system are: by transfer of heat to the ambient sewage, the ambient
atmosphere at the surface of the manhole or externally delivered water. Sewage
would foul the heat exchanger surfaces. In certain localities, air might be pro-
vided through the manhole cover itself, but this might fail during freezing periods
in the winter. Water is, by far, the most reliable heat exchanger for the purpose
and might be provided externally to the instrumentation.
59
-------
APPENDIX 3
DESIGN SPECIFICATION ON STORMWATER TOC
I. GENERAL
A. This specification applies to the design of a total organic carbon moni-
toring system to operate intermittently and on signal to measure the
TOC of a sample of stormwater.
B. Applications
1. Natural streams and rivers
2. Storm and combined wastewater treatment plants and delivery
systems.
C. The stormwater TOC system consists of essentially three interconnected
components. One of these is a Raytheon Homogenizer used in sample
grinding and delivery. The second is the singly enclosed and modified
stormwater TOC unit. The third component is the MSA auto-zero, auto-
span unit. The single enclosure will have front access sufficient to
accomplish all calibration and routine maintenance. All required ser-
vices, power, and signal connections will be made at the rear of the
unit.
D. The standard output signal will be an analog voltage, 0—5 volts. A
local meter display on the front panel will read directly in milligrams
of carbon per liter of sample.
E. A local strip chart recorder will be provided to deliver continuous and
permanent record of data as it is collected in time. A chart speed of
1-1/2 inches per hour will be used,
II. MEASUREMENT TECHNIQUE
A. Basic Approach
The stormwater TOC analyzer measures organic carbon content of a
sample stream by mechanizing the following steps:
60
-------
1. Automatically turning on all hardware from the standby mode on
command.
2. Automatically zeroing and spanning the system.
3. Grinding the delivered stormwater sample containing particles no
larger than 1/4 inch.
4. Removing inorganic carbon from the sample.
5. Volatilizing of the sample.
6. Completely combusting the volatilized organics.
7. Removing water vapor from the gas stream.
•
8. Measuring CO concentration in the gas stream as an index of the
organic carbon concentration in the original sample.
B. Automated features will be implemented by using the control system in
Figures 20 and 23.
C. Grinding will be accomplished using a continuous motor-driven stone
grinder.
D. Inorganic Carbon Removal
The sample stream is drawn from the sample line by the intake pump.
Sulfuric acid is added to the stream from an internal supply container
by the acid pump, converting the inorganic carbonates to CO~ gas.
Oxygen is then deliverd into the sample stream. This vigorous flow
serves the dual purpose of transporting the liquid sample rapidly up to
the sample pump, and stripping the CO2 from the liquid phase. CO ,
O2 and excess sample are led off to drain at this point. This technique
for CO2 removal is known as sparging.
E. Sample Volatization and Combustion
The sample pump, delivers a metered flow of sample from the sparging
stream for delivery to the input tube of the combustion reactor. A
precisely metered and controlled flow of oxygen is delivered to this
point.
The multifinned, fixed-bed, noncatalytic reactor and its related heater
system is required to deliver large amounts of energy on a rapid and
continuous basis, and with relatively uniform distribution. On entering
61
-------
the reactor, the sample vaporizes rapidly, and the volatilized organics
are mixed with the oxygen so that complete combustion takes place.
the end products of this process are water vapor, CO2, and a large
amount of excess oxygen, which exit the reactor as a gas stream under
pressure.
F. Water Vapor Removal
The gas stream is then passed through a water-cooled condenser which
removes all of the water vapor by cooling the stream below 100° C. The
condensate is collected in a trap below the condenser which is provided
with an overflow to drain. This trap also provides a pressure relief
for the gas stream when the path to the gas analyzer is blocked during
calibration.
G. Carbon Dioxide Concentration Measurement
The gas stream from the condenser passes through a filter which removes
dirt particles entrained in the stream and is followed by an integrating
volume to smooth out random fluctuations in COo concentration. The
stream is then delivered to the gas analyzer.
The gas analyzer is a nondispersive infrared analyzer (NDIR) which
measures the concentration of CO2 in the gas stream by measuring the
absorption of infrared energy over a given path length in the gas stream
at a wavelength selected to correspond with a principal abosrption band
for CO .
£l
By proper choice of oxygen flow rate and NDIR Analyzer sensitivity, the
NDIR output signal can be calibrated in terms of the amount of organic
carbon in the original sample.
I. Range Flexibility
Full-scale range of the stormwater TOG can be changed within limits by
a change of NDIR analyzer sensitivity, oxygen flow rate, or sample flow
rate, or a combination of the three.
III. DESIGN OBJECTIVES
A. Calibration
Calibration of the stormwater TOG is accomplished by using the auto-
zero/auto-span accessory available from MSA. In this.case, available
oxygen is used as the zero gas and span gas must be provided and
installed as specified by the manufacturer.
62
-------
B. Alarms
If an operative failure should occur so as to result in a no-sample-
delivery condition, this should be indicated by a temperature sensor at
the reactor exit and be indicated by a low temperature signal. If a
failure should occur in the sample delivery cooling jacket so as to
deliver a large amount of cooling water to the reactor, this shall be
detected by a high temperature signal. A thermocouple sensor con-
nected to a meter relay shall be used to sound an alarm and turn on a
light, which will indicate the problem with the instrument during
operation.
C. Protective Devices
A high-pressure sensor rated to operate at 5 psi shall be installed in
the oxygen delivery system so as to remove power from the pump if the
pressure inside the reactor exceeds the preset value.
In the event of electrical failure of any of the operating components, a
circuit breaker will operate to remove all power from the operating
components.
D. Reactor Cleaning and Repacking
The combustion reactor assembly will be designed to permit
removal, disassembly, cleaning, repacking, reassembly, and reinstal-
lation in the unit within a period of 2 hours after the reactor temperature
has cooled to a safe handling level, by a trained service technician.
Special tools required, if any, will be supplied with the unit and will be
mounted for storage within the cabinet.
IV. PERFORMANCE SPECIFICATIONS
A. TOG Range (Standard)
The stormwater TOG shall operate in the range 0—1000 mg/liter.
B. Repeatability
±2% of full scale
I. Defined as twice the standard deviation of a statistical sample of 10
or more readings taken over a twenty-minute period.
63
-------
C. Linearity
The stormwater TOG shall be linear to ± 1% of full scale.
D. Accuracy
±2% of full scale.
E. Zero Drift
±1% of full scale in 24 hours.
F. Span Drift
±1% of full scale in 24 hours.
G. Noise
Less than 1% of full scale.
H. Transport Delay
Five minutes from introduction of sample at intake fitting until output
signal rises to 10% of final value.
I. Rise Time
Five minutes for output signal to rise from 10% of final value to 90% of
final value.
J. Response Time
Ten minutes from introduction of sample at intake fitting until output
signal reaches 90% of final value.
K. Warmup Time on Installation
Two hours.
V. ELECTRICAL SPECIFICATIONS
A. Input Power
117 Vac ±10%, 60 Hz + Hz, 1 4>, 25 A.
B. Signal Output
0—5 Vdc, into 1,000 ohms minimum.
64
-------
VI. MECHANICAL SPECIFICATIONS
A. Overall Size and Weight (Approximate)
Height: 70 inches (178 cm) + 24 inches (61 cm) for auto-zero
Width: 29 inches ( 74 cm) + 12 inches (30. 5 cm) for homogenizer
Depth: 26 inches ( 66 cm) + 9 inches (22.9 cm) for controller
Weight: 575 Ib.
B. Mounting
The TOC unit is provided with an integral caster base for floor mounting
and convenient handling during installation and service. Clearances
required when installed: The auto-zero will be bracket-mounted.
Front: 25 inches Min (64 cm)
Rear: 10 inches Min (26 cm)
Sides: 12 inches Min (31 cm) for homogenizer and manifold
C. Accessibility
All routine service and calibration may be performed from the front of
the unit. Internal components in the lower and central chambers of the
unit are reached through two hinged access doors provided with safety
locks. The NDIR analyzer is mounted behind a fixed upper panel, and
withdraws from the front for service. The strip chart recorder is also
mounted on this upper panel, and also withdraws from the front for
service.
D. Cabinet Classification
The cabinet is designed for general purpose indoor installation, equiva-
lent to NEMA type 1. It is not drip proof, dust tight, or weatherproof.
When installations requiring these properties are made, the unit must
be housed or enclosed in a manner sufficient to buffer the unit from the
severe service conditions.
E. Finish
Cabinet: Epoxy paint, with primer
Panels: Epoxy paint, with primer
All internal aluminum: Clear anodized.
F. Plumbing Connections
Plumbing connections for external services and drain are located on a
panel at the left rear bottom of the unit. Swagelock bulkhead fittings
(1/4 inch) are employed. An exterior manifold is provided made of PVC
plumbing hardware to provide drainage for sample, water and drainage.
65
-------
G. Electrical Connections
1. ac Power: An 8-foot type "SO" power cord will be supplied with the
unit, providing three #12 conductors and a standard three prong
grounded male plug. Strain relief will be provided at the point
where the cord emerges from the cabinet. Spade lug terminations
will be used at a barrier strip provided within the cabinet.
2. Signal and Control Wiring: An internal barrier strip will be pro-
vided for spade lug termination of all customer-installed signal
and control cable.
Vn. SAMPLE REQUIREMENTS
A. Suspended Solids
Up to 1000 mg/1, no particles larger than 1/4 inch in any dimension,
may be passed through the homogenizer.
B. Flow Rate
6 cc/min to the TOC unit
Note
This flow rate represents the sample actually drawn
into the TOC analyzer. Lines delivering sample to
the system should provide sufficient flow to maintain
at least 2 feet per second (60 cm per second) velocity,
with excess sample being bypassed to drain. Flow
through the homogenizer must be maintained at 4
gallons/hour (15 liter/hours)minimum.
C. Pressure
2
10 to 30 psig (700 to 2,100 gm/cm ) using the homogenizer.
D. Temperature
32° F to 140± F (nonfreezing) (° to 60° C).
VIII. ADDITIONAL SERVICES REQUIRED
A. Oxygen
2
1. Pressure: Regulated to 16 to 24 psig (1,125 to 1,700 gm/cm )
2. Flow: Approx. 4 CFH to maintain sparge & combustion gas
66
-------
B. Span Gas (If Required)
CO2/N2 mixture of appropriate concentrations, in internally-mounted
cylinder.
5 to 15 psig (350 to 1050 gm/cm )
Approx. 300 cc/min when calibrating
10 to 50 psig (700 to 3, 500 gm/cm )
3 to 6 gpm (11.4 to 23 liters/min)
1. Pressure:
2. Flow:
C. Cooling Water
1. Pressure:
2. Flow:
3. Temperature: 90° F Max. (32° C Max)
D. Sulfuric Acid
10% H2 SC>4 by weight. Internal container provided.
Consumption: 0. 05 cc/min.
E. Drain (gravity) 6 gpm
A persistaltic pump shall be provided to remove water from the sump
at a flow rate of 5cc/min.
F. Calibration Solution
Provision for external delivery of calibration solutions shall be provided.
DC. CONTROLS AND INDICATORS
A. Front Panel
1. Unit OFF:
2. Unit Start-up:
3. Unit Operate:
4. NDIR ON/OFF:
5. NDIR ON:
B. Internal Panel
1. Lamp Test:
2. Reactor Flush:
3. Sparge Gas Flow:
4. Carrier Gas Reg:
5. Carrier Gas Flow:
6. Reactor Temp:
Indie ator/ Pu shbutton
Indicator/Pushbutton
Indicator/Pushbutton
Toggle
Indicator
Pushbutton
Pushbutton
Indicator/Controller
100 Turn Controller
Rotameter
Meter, 0 - 2,000°
67
-------
7. Reactor Temp Adj: Potentiometer (Knob)
8. High Temp Limit: Potentiometer (Knob)
9. Main Power On: Indicator
10. Heater Breaker Trip: Indicator
C. Other Internal
1. Main Power On/Off: Circuit Breaker
2. Heater On/Off: Circuit Breaker
3. Sample Tube Cooling: Needle Valve
4. Calibrate Valve: Manual Valve
X. ENVIRONMENTAL
A. Ambient Temperature
32 ° F to 104° F (nonfreezing): (0° - 40° C)
B. Humidity
0 - 95% RH noncondensing
C. Vibration and Shock
Capable of withstanding normal shock and vibration in shipping.
XI. SAFETY PROVISIONS
A. Circuit breakers on Main ac and Heater ac.
B. Primary ac Fuses on power supply
C. High Voltage notices and shields on all terminal over 30 Vdc and 30 Vac.
D. Separate ground lug on rear of cabinet.
E. ac leakage current is less than 0. 5 mA when measured at 115 Vac, 60 Hz.
F. High temperature notices on all hot surfaces over 55° C.
G. Dangerous chemical notice on inside of door.
H. Flush procedure instruction card on inside of lower swing-out door.
I. Automatic alarm and action provisions, as follows:
68
-------
Condition
Possible Cause
Action
High Reactor Temp
High Reactor Pressure
Loss of Reactor
Cooling Water
High Internal Instru-
ment Temp
1. Faulty thermocouple
2. Solid state relay
failure
3. Faulty temperature
controller
1. Reactor plug
2. Input tube plug
1. Main water shut off
2. Hose leak
3. Faulty hose clamps
1. Reactor malfunction
2. Loss of cooling fans
Complete instrument
shut-down
Switch to Start-Up Mode
(reactor heat maintained)
Complete shutdown
Complete shutdown
69
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before comp1-—T-
1. REPORT NO.
EPA-670/2-75-067
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Automatic Organic Monitoring
System for Storm and Combined Sewers
5. REPORT DATE
June 1975 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Angelo Tulumello
8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORG \NIZATION NAME AND ADDRESS
Raytheon Company
Portsmouth, R.I. 02871
10. PROGRAM ELEMENT NO.
1BB034ROAP 21-ASY; Task 038
11. CONTRACT/JSJtflflsW NO.
68-03-0262
12. SPONSORING AGENCY NAME AND ADDRESS
National Environmental Research Center
Office of Research and Development
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
Supplement to EPA-670/2-74-087, "Assessment and Development Plan for Monitoring of
Qrganics in Storm Flow," NTIS PB-238 810/AS _______
16. ABSTRACT
Early in the program to develop a stormwater TOC (total organic carbon) system, it was
established in report EPA-670/2-74-087 that continuous on-line TOC was the best method for
the measurement of stormwater pollution loading. Hardware was assembled that would
process stormwater samples containing high suspended solids and that would obtain a con-
tinuous signal proportional to the concentration of TOC in the sample.
Synthetic samples of municipal raw influent charged with primary sludge were analyzed using
the TOC analyzer. Data were also obtained on actual stormwater samples collected during
storm events at Boston. Further modifications were made after these observations.
Automatic circuitry designed to provide turn on, auto-zero, auto-span and sample line flushing
was added to the hardware, and the system was installed at Boston Cottage Farm Storage
Facility.
Automatic continuous analyses were obtained during storms on site at the Cottage Farm
Storage Facility.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Combustion, Instruments,
Recording instruments,
Carbon dioxide,
Homogenizing
Infrared, Total organic carbon, On-line
alarms, Continuous monitoring, .Sample
taking, Sample transport, Stormwater
TOC, Organic pollution, Combined sewage,
Storm-related wastewaters, Stormwater
environment
13B
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. OF PAGES
80
20. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
70
U. S. GOVERNMENT PRINTING OFFICE: 1975-657-59't/5ll09 Region No. 5-11
------- |