United States
Environmental
Protection Agency
Office of Water
(4601)
EPA815-C-99-002
September 1999
ANALYSIS OF GAC EFFLUENT
BLENDING DURING THE
ICR TREATMENT STUDIES
-------
This page intentionally left blank.
-------
Foreword
Between 1997 and 1999, 98 public water systems conducted treatment studies to evaluate
disinfection byproduct (DBF) precursor removal performance of granular activated carbon
(GAC) and membranes. The treatment study requirement was a part of the Information
Collection Rule (ICR) for Public Water Systems, Subpart M of the National Primary Drinking
Water Regulations, § 141.141(e).
Sixty-two public water systems evaluated GAC. The ICR required that DBF precursor removal
be evaluated in the effluent of a single contactor as a function of run time, to assess the
breakthrough of DBF precursors as the GAC was exhausted. In practice, full-scale plants can
reduce carbon usage rates by blending the effluents of multiple parallel contactors prior to
disinfection. When the treatment objective is reached in the blended effluent, the contactor with
the "oldest" GAC is taken off-line and replaced by a contactor with fresh GAC, and this
replacement occurs at regular intervals. GAC effluent blending extends the service life of each
contactor because water from contactors that exceed the treatment objective is blended with
water from contactors with fresher GAC that have effluent concentrations below the treatment
objective. The treatment objective must only be maintained in the blended contactor effluent.
A primary goal during analysis of the treatment study results by the USEPA will be to estimate
blended contactor run times to meet target regulatory treatment objectives. This information can
be used to estimate GAC treatment costs that reflect full-scale effluent blending. This study
provides the background and foundation for the analytical tools used to analyze treatment study
data, assessing the applicability and limitations of these tools.
One objective of this study was to evaluate mathematical models for representing single
contactor breakthrough data. From a data management perspective, model parameters will be
easier to manage than the entire experimental data sets comprising 8,000 to 9,000 breakthrough
curves generated by the 62 GAC treatment studies. A best-fit curve also facilitates interpolation
and extrapolation of the experimental data. Furthermore, a function that describes the single
contactor experimental data set is a prerequisite for calculating the integral breakthrough curve, a
tool for predicting blended contactor run times.
A second objective of this study was to evaluate and compare two approaches for predicting the
integral breakthrough curve. The first approach, based on application of the average value
function to the single contactor breakthrough curve, has been presented by previous researchers.
The second approach evaluated was a new computationally-simpler method developed by the
treatment study technical work group (TS-TWG).
These two objectives were applied to experimental results from bench-scale GAC runs on eight
water sources. Analytes evaluated included DBF surrogates, DBF class sums, and DBF species,
yielding an extensive experimental matrix for a thorough evaluation of model results. In
addition, GAC effluent blending was assessed experimentally to test model predictions of the
integral breakthrough curve.
-------
Disclaimer
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
-------
ANALYSIS OF GAC EFFLUENT BLENDING DURING THE ICR
TREATMENT STUDIES
Prepared for:
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF GROUND WATER AND DRINKING WATER
STANDARDS AND RISK MANAGEMENT DIVISION
Technical Support Center
26 W. Martin Luther King Drive
Cincinnati, Ohio 45268
Prepared by:
INTERNATIONAL CONSULTANTS, INC.
4134 Linden Avenue, Suite 200
Dayton, Ohio 45432
SUMMERS & HOOPER, INC.
6 Knollcrest Drive
Cincinnati, Ohio 45237
Under ICI Contract 68-C-98-051, Task Order 4
-------
This page intentionally left blank.
-VI-
-------
Acknowledgments
This document was prepared by the United States Environmental Protection Agency, Office of
Ground Water and Drinking Water, Standards and Risk Management Division, Technical
Support Center. The Task Order Project Officer was Steven Allgeier. The Contract Project
Officer was Phyllis Branson.
Technical consultants played a significant role in the research performed and the preparation of
this document. This task was conducted jointly by International Consultants, Inc. (ICI) and
Summers & Hooper, Inc. (S&H), under ICI Contract 68-C-98-051, Task Order 4. Timothy
Soward and Christopher Hill acted as Project Managers for ICI. The S&H project team was led
by Stuart Hooper, who acted as the Technical Project Leader. Mr. Hooper was responsible for
technical direction and led the data analysis effort. Vanessa Hatcher (S&H) led and conducted
the associated experimental work, and was assisted by Kristina Trenkamp and Carrie Wyrick.
Jeff Welge and Rick Song of ICI performed mathematical modeling and statistical analyses, and
provided statistical guidance when necessary. Technical review of the final report was provided
by James Westrick (USEPA-TSC), Thomas Speth (USEPA-RREL), and R. Scott Summers
(University of Colorado-Boulder).
The experimental portion of this project was performed in conjunction with treatment studies
required under the Information Collection Rule. The following public water systems were
involved: Charleston Commissioners of Public Works, City of Aurora, City of Escondido, City
of Greensboro, City of Topeka, Iowa-American Water Company, Miami-Dade Water and Sewer
Department, and Sweetwater Authority. This work could not have been completed without the
use of the data generated as part of the treatment studies funded by these public water systems.
-Vll-
-------
This page intentionally left blank.
-Vlll-
-------
Executive Summary
When granular activated carbon (GAC) contactors are utilized for disinfection byproduct (DBF)
precursor control, multiple contactors can be used more efficiently when operated in parallel
with staggered GAC replacement cycles, with the effluents of all contactors blended prior to
disinfection. By doing so, individual contactors can be operated past a point at which the
effluent exceeds a given treatment objective, because the treatment objective must only be
maintained in the blended effluent of all contactors.
The design of this study incorporated two main goals. The primary objectives were to evaluate
the ability of the logistic function to model single contactor breakthrough curve data and to
evaluate the success and limitations of predictive models used to determine the integral
breakthrough curve, a relationship between single contactor run time and blended contactor
water quality. The secondary objective of this study was to evaluate the applicability of these
models and predictive methods in the context of the Information Collection Rule (ICR) GAC
treatment study data analysis.
Full-, pilot-, and bench-scale GAC treatment studies were performed by 62 utilities in fulfillment
of ICR requirements. Regardless of scale, the ICR required that the effluent of single GAC
contactors be analyzed for DBF surrogate and formed DBF breakthrough as a function of run
time, to assess the breakthrough of DBF precursors as the GAC was exhausted. Bench-scale
GAC studies typically examined two empty-bed contact times (EBCTs) of 10 and 20 minutes
during each of four quarterly studies to account for seasonal variability in source water quality.
Pilot-scale GAC studies were typically composed of one to two sessions including 10 and 20
minute EBCT contactors. Thus, a large amount of data was generated and will be analyzed: the
62 GAC treatment studies performed will yield a total of 8,000 to 9,000 individual breakthrough
curves.
The logistic function has previously been used to model GAC breakthrough curves and is a
suitable model due to the characteristic 'S' shape of most breakthrough curves. Three
modifications to the logistic function were developed to improve its performance for modeling
single contactor data. Curve fitting involved determining which model was applicable based on
characteristics of the breakthrough curve, and applying the appropriate model to the
breakthrough curve for each parameter. These enhanced forms of the logistic function model
were able to successfully fit single contactor breakthrough curve data for all parameters,
including DBF surrogates, DBF sum class parameters, and individual trihalomethane (THM) and
haloacetic acid (HAA) species. A method was also employed to detect outlier data points and to
limit the influence of these deviant observations on the parameter estimates.
Two predictive approaches were compared for developing determining the integral breakthrough
curve, a relationship between operation time of each individual contactor and water quality in the
blended contactor effluent: the direct integration (DI) method and the surrogate correlation
approach (SCA). The DI method is a time-normalized integration of the logistic function
calculated using the average value function that yields the integral breakthrough curve. The
SCA method first utilizes the DI method to establish an integral breakthrough curve for total
organic carbon (TOC). Then data points on both the single contactor and integral breakthrough
curves at a given TOC concentration are mapped, and all other water quality parameters
-IX-
-------
associated with the single contactor effluent data set at that TOC concentration are applied to the
blended effluent curve. The SCA method is especially applicable to the ICR data analysis
requirements because it minimizes the computations necessary to estimate blended contactor run
times to treatment objectives. An assessment of the concentration of other DBFs at any given
treatment objective will be performed as part of the data analysis effort, and the SCA procedure
is also suited to this task. The SCA procedure requires that GAC breakthrough curves for all
measured parameters be represented by the logistic function model curve fit. By doing so, a
smaller amount of data are needed to represent the entire breakthrough curve experimental data
set. This procedure inherently relies on the assumption that the relationship between TOC
concentration and the other water quality parameters established in the single contactor effluent
is maintained in the blended contactor effluent. This study verified this assumption and found
that it was valid between TOC and other DBF surrogates, DBF class summation parameters, and
individual DBF species. In addition, the correlation between TOC and bromine incorporation
factors for THMs and HAAs was shown to be consistent between the single contactor effluent
and experimental blended effluent.
The results of the DI and SCA model predictions were compared to experimental data for eight
GAC runs performed on eight water sources, with varying pretreatments, influent TOC
concentrations, bromide concentrations, and simulated distribution system (SDS) chlorination
conditions to evaluate DBF formation. An analysis of the model results across all waters and
analytes showed that the prediction error for the two models was equivalent. Both models were
biased negative, indicating a tendency to underpredict the experimental data. The SCA model
had a slightly higher negative bias than did the DI model. Since the SCA method simplifies and
reduces the amount of computations necessary to estimate the integral breakthrough curve, its
use is recommended to estimate blended contactor water quality during the ICR treatment study
data analysis.
An analysis of model results for individual parameters showed that the SCA method was more
successful in predicting the integral breakthrough curve of brominated DBF species, while the DI
method was a better predictor of non-brominated DBF species breakthrough. Prediction of the
breakthrough of individual DBF species in the blended effluent is important since individual
DBFs of potential health concern will be considered during analysis of the ICR treatment study
data. For sum parameters such as total THM (TTHM) and the sum of five HAAs (HAAS), the
SCA method yielded results that were comparable to or superior to the DI method predictions.
Both predictive methods rely on the assumption that an infinite number of contactors are on-line
and operated in parallel-staggered mode. This study examined this assumption and found that
the error incurred when applying run time estimates based on the infinite contactor assumption to
run times for finite numbers of contactors is impacted by the number of contactors and the
magnitude of the treatment objective examined in relation to the asymptotic concentration
approached by the single contactor breakthrough curve. Based on the logistic function model of
the GAC effluent breakthrough profile, the infinite contactor assumption will yield estimated run
times within 10 percent of actual run times for 13 or more contactors operated in parallel-
staggered mode. For 10 contactors on-line, the infinite contactor assumption will yield run time
estimates within 12 percent of the actual run times. In all cases, run time estimates based on the
infinite contactor assumption are longer than those for a finite number of contactors, thus
providing a best case scenario for GAC performance. The applicability of the infinite contactor
assumption in this model to finite numbers of contactors is especially important for small plants.
-x-
-------
Extrapolation of the integral breakthrough curve was also examined during this study. During
the ICR treatment study data analysis, extrapolation of some TOC integral breakthrough curves
may be necessary to estimate GAC run times when target regulatory values for DBFs or their
surrogates are exceeded upon application of the SCA method. Extrapolation of runs performed
on two waters were compared to the same runs without extrapolation, yielding a 3 percent error
in the predicted blended effluent TOC concentration for a 21 percent run time extrapolation and
an 8 percent error in the predicted blended effluent TOC concentration for a 61 percent
extrapolation. The impact of extrapolation on the SCA procedure estimates of the integral
breakthrough curves for other DBF surrogates and formed DBFs was small: the mean error at
the end of the extrapolated integral breakthrough curve was 5 percent for a 21 percent
extrapolation and 9 percent for a 61 percent extrapolation. Therefore, based on the waters
examined in this study, data sets that do not exceed a given treatment objective may be
extrapolated, and the error in predicted blended contactor water quality incurred by extrapolation
up to 50 percent of the original run time should average less than 10 percent.
-XI-
-------
This page intentionally left blank.
-Xll-
-------
Table of Contents
List of Tables xvii
List of Figures xix
List of Abbreviations xxv
1 Background and Method Development 1
1.1 Optimization of GAC Operation 2
1.2 Modeling the Operation of Multiple Contactors Operated in Parallel-Staggered Mode .... 2
1.3 Single Contactor Breakthrough Curve Models 6
1.4 Direct Integration Approach 8
1.5 Surrogate Correlation Approach 9
1.6 Impact of Bromide Concentration on GAC Effluent Blending Models 11
1.7 Effluent Blending Modeling of Fewer than 10 Contactors 12
1.8 GAC Breakthrough Curve Extrapolation 12
1.9 ICR GAC Treatment Study Data Analysis Context 13
1.10 Appropriateness of Model Assumptions to Full-Scale GAC Effluent Blending 14
2 Study Objectives and Approach 15
3 Materials and Methods 17
3.1 Experimental Approach 17
3.1.1 Rapid Small-Scale Column Test 17
3.1.1.1 GAC Preparation Procedures 17
3.1.1.2 RSSCT Column Setup 19
3.1.1.3 Batch Influent Preparation 20
3.1.1.4 RSSCT Monitoring 20
3.1.2 Bench-Scale Blended Water Quality Assessment Approach 20
3.1.3 DBP Formation Assessment 21
3.1.4 Assessment of the Impact of Sampling on the Integral Breakthrough Curve 22
3.2 Data Analysis and Modeling Approach 23
3.2.1 Logistic Function Models 23
-xiii-
-------
3.2.2 Outlier Methods 25
3.2.3 Direct Integration Approach 26
3.2.4 Surrogate Correlation Approach 27
3.2.5 Comparison of Methods for Predicting the Performance of GAC Contactors
Operated in Parallel-Staggered Mode 28
3.2.6 Comparison of Single Contactor and Blended Effluent DBF Bromine Incorporation.
28
3.2.7 Breakthrough Curve Extrapolation 29
3.3 Waters Examined 29
3.3.1 Pretreatment and Water Quality 29
3.3.2 Simulated Distribution System Chlorination Conditions 31
3.4 Analytical Methods 31
3.5 Experimental QA/QC Summary 32
4 Results and Discussion 35
4.1 Overview 35
4.2 Correlation between Surrogates and DBFs in Single Contactor and Blended Contactor
Effluents 37
4.2.1 Correlation between Surrogate Concentration and DBF Formation 37
4.2.2 Correlation between Surrogates and DBF Speciation 38
4.3 Assessment of Logistic Function Fit to Single Contactor Breakthrough Curve Data 67
4.3.1 Surrogates and Class Sum Logistic Function Curve Fits 67
4.3.2 DBF Species Logistic Function Curve Fits 68
4.4 Comparison of SCA and DI Methods Used to Predict the Blended Contactor Integral
Breakthrough Curve 89
4.4.1 Surrogates and Class Sums 95
4.4.2 DBF Species 96
4.5 Analysis of Model Applicability to Finite Number of Contactors 113
4.6 Impact of Extrapolation on Integral Breakthrough Curve Prediction 121
5 Summary and Conclusions 131
6 References 135
-xiv-
-------
Appendix A: Breakthrough Curves Corrected for Impact of Sampling 137
Appendix B: SAS Code 143
Appendix C: Full- and Bench-Scale Pretreatment Schematics 147
Appendix D: Single Contactor and Blended Effluent DBF Surrogate and Formed DBF
Correlations 157
Appendix E: Logistic Function Model Curve Fits 223
Appendix F: Comparison of SCA Method to DI Approach for Integral Breakthrough Curve
Prediction 305
Appendix G: Logistic Function Model Best-Fit Parameters 387
Appendix H: Impact of Extrapolation on SCA Prediction of the Integral Breakthrough Curve393
Appendix I: Impact of Extrapolation on DI Prediction of the Integral Breakthrough Curve .... 415
-xv-
-------
This page intentionally left blank.
-xvi-
-------
List of Tables
1 Summary of RSSCT design parameters for all runs 18
2 Relationship between blended effluent sample number and RSSCT effluent sample number 21
3 Summary of pretreatment and water quality 30
4 SDS chlorination conditions 31
5 Summary of analytical methods and MRLs 32
6 Summary of laboratories conducting analyses 32
7 Summary of field duplicate precision for single contactor and blended effluent data 34
8 Frequency of logistic function model used and R2 values for all parameters and all waters 69
9 Summary of DI and SCA integral breakthrough curve prediction RSS values for Waters 1
through 4 90
10 Summary of DI and SCA integral breakthrough curve prediction RSS values for Waters 5
through 8 91
11 Summary of model prediction bias for Waters 1 through 4 92
12 Summary of model prediction bias for Waters 5 through 8 92
13 Summary of mean RSS, mean bias, normalized mean RSS, and normalized mean bias for all
waters 93
14 Summary of run times to a 0.35 treatment objective 115
15 Summary of run times to a 0.50 treatment objective 115
16 Summary of run times to a 0.65 treatment objective 116
17 Summary of run times to a 0.80 treatment objective 116
-xvii-
-------
This page intentionally left blank.
-XVlll-
-------
List of Figures
1 Schematic of blending of multiple GAC contactors operated parallel-staggered mode 3
2 Effluent water quality during quasi steady-state operation of multiple contactors in parallel-
staggered mode 3
3 Operation of multiple contactors in parallel-staggered mode to various single contactor run
times: derivation of the blended effluent integral breakthrough curve 5
4 Graphical summary of SCA procedure used for base analysis of GAC treatment studies 10
5 Logistic function model curves 24
6 Correlations based on GAC effluent TOC concentration for single contactor and blended
effluents for Water 1 40
7 Correlations based on GAC effluent TOC concentration for single contactor and blended
effluents for Water 2 41
8 Correlations based on GAC effluent TOC concentration for single contactor and blended
effluents for Water 8 42
9 THM correlations based on GAC effluent TOC concentration for single contactor and
blended effluents for Water 2 43
10 THM correlations based on GAC effluent TOC concentration for single contactor and
blended effluents for Water 3 44
11 THM correlations based on GAC effluent TOC concentration for single contactor and
blended effluents for Water 8 45
12 HAA correlations based on GAC effluent TOC concentration for single contactor and
blended effluents for Water 5 46
13 HAA correlations based on GAC effluent TOC concentration for single contactor and
blended effluents for Water 7 47
14 HAA correlations based on GAC effluent TOC concentration for single contactor and
blended effluents for Water 2 48
15 HAA correlations based on GAC effluent TOC concentration for single contactor and
blended effluents for Water 4 49
16 HAA correlations based on GAC effluent TOC concentration for single contactor and
blended effluents for Water 7 50
17 Correlation between single contactor and blended effluent TOC concentration and THM
bromine incorporation factor (n) for Water 1 51
18 Correlation between single contactor and blended effluent TOC concentration and THM
bromine incorporation factor (n) for Water 2 51
19 Correlation between single contactor and blended effluent TOC concentration and THM
bromine incorporation factor (n) for Water 3 52
20 Correlation between single contactor and blended effluent TOC concentration and THM
bromine incorporation factor (n) for Water 4 52
21 Correlation between single contactor and blended effluent TOC concentration and THM
bromine incorporation factor (n) for Water 5 53
22 Correlation between single contactor and blended effluent TOC concentration and THM
bromine incorporation factor (n) for Water 6 53
23 Correlation between single contactor and blended effluent TOC concentration and THM
bromine incorporation factor (n) for Water 7 54
24 Correlation between single contactor and blended effluent TOC concentration and THM
bromine incorporation factor (n) for Water 8 54
-xix-
-------
25 Correlation between single contactor and blended effluent TOC concentration and HAA
bromine incorporation factor (n') for Water 1 55
26 Correlation between single contactor and blended effluent TOC concentration and HAA
bromine incorporation factor (n') for Water 2 55
27 Correlation between single contactor and blended effluent TOC concentration and HAA
bromine incorporation factor (n') for Water 3 56
28 Correlation between single contactor and blended effluent TOC concentration and HAA
bromine incorporation factor (n') for Water 4 56
29 Correlation between single contactor and blended effluent TOC concentration and HAA
bromine incorporation factor (n') for Water 5 57
30 Correlation between single contactor and blended effluent TOC concentration and HAA
bromine incorporation factor (n') for Water 6 57
31 Correlation between single contactor and blended effluent TOC concentration and HAA
bromine incorporation factor (n') for Water 7 58
32 Correlation between single contactor and blended effluent TOC concentration and HAA
bromine incorporation factor (n') for Water 8 58
33 Correlation between single contactor and blended effluent UV absorbance and THM
bromine incorporation factor (n) for Water 1 59
34 Correlation between single contactor and blended effluent UV absorbance and THM
bromine incorporation factor (n) for Water 2 59
35 Correlation between single contactor and blended effluent UV absorbance and THM
bromine incorporation factor (n) for Water 3 60
36 Correlation between single contactor and blended effluent UV absorbance and THM
bromine incorporation factor (n) for Water 4 60
37 Correlation between single contactor and blended effluent UV absorbance and THM
bromine incorporation factor (n) for Water 5 61
38 Correlation between single contactor and blended effluent UV absorbance and THM
bromine incorporation factor (n) for Water 6 61
39 Correlation between single contactor and blended effluent UV absorbance and THM
bromine incorporation factor (n) for Water 7 62
40 Correlation between single contactor and blended effluent UV absorbance and THM
bromine incorporation factor (n) for Water 8 62
41 Correlation between single contactor and blended effluent UV absorbance and HAA
bromine incorporation factor (n') for Water 1 63
42 Correlation between single contactor and blended effluent UV absorbance and HAA
bromine incorporation factor (n') for Water 2 63
43 Correlation between single contactor and blended effluent UV absorbance and HAA
bromine incorporation factor (n') for Water 3 64
44 Correlation between single contactor and blended effluent UV absorbance and HAA
bromine incorporation factor (n') for Water 4 64
45 Correlation between single contactor and blended effluent UV absorbance and HAA
bromine incorporation factor (n') for Water 5 65
46 Correlation between single contactor and blended effluent UV absorbance and HAA
bromine incorporation factor (n') for Water 6 65
47 Correlation between single contactor and blended effluent UV absorbance and HAA
bromine incorporation factor (n') for Water 7 66
-xx-
-------
48 Correlation between single contactor and blended effluent UV absorbance and HAA
bromine incorporation factor (n') for Water 8 66
49 Single contactor and blended effluent TOC breakthrough curves for Water 1 71
50 Single contactor and blended effluent TOC breakthrough curves for Water 2 71
51 Single contactor and blended effluent TOC breakthrough curves for Water 3 72
52 Single contactor and blended effluent TOC breakthrough curves for Water 4 72
53 Single contactor and blended effluent TOC breakthrough curves for Water 5 73
54 Single contactor and blended effluent TOC breakthrough curves for Water 6 73
55 Single contactor and blended effluent TOC breakthrough curves for Water 7 74
56 Single contactor and blended effluent TOC breakthrough curves for Water 8 74
57 Single contactor and blended effluent UV254 breakthrough curves for Water 5 75
58 Single contactor and blended effluent UV254 breakthrough curves for Water 7 75
59 Single contactor and blended effluent SDS-TOX breakthrough curves for Water 4 76
60 Single contactor and blended effluent SDS-TOX breakthrough curves for Water 8 76
61 Single contactor and blended effluent SDS-TTHM breakthrough curves for Water 3 77
62 Single contactor and blended effluent SDS-TTHM breakthrough curves for Water 6 77
63 Single contactor and blended effluent SDS-HAA9 breakthrough curves for Water 1 78
64 Single contactor and blended effluent SDS-HAA9 breakthrough curves for Water 2 78
65 Single contactor and blended effluent SDS-CF breakthrough curves for Water 4 79
66 Single contactor and blended effluent SDS-BDCM breakthrough curves for Water 6 79
67 Single contactor and blended effluent SDS-BDCM breakthrough curves for Water 7 80
68 Single contactor and blended effluent SDS-BDCM breakthrough curves for Water 4 80
69 Single contactor and blended effluent SDS-BDCM breakthrough curves for Water 8 81
70 Single contactor and blended effluent SDS-BF breakthrough curves for Water 1 81
71 Single contactor and blended effluent SDS-BF breakthrough curves for Water 6 82
72 Single contactor and blended effluent SDS-DCAA breakthrough curves for Water 3 82
73 Single contactor and blended effluent SDS-TCAA breakthrough curves for Water 5 83
74 Single contactor and blended effluent SDS-DBAA breakthrough curves for Water 6 83
75 Single contactor and blended effluent SDS-BCAA breakthrough curves for Water 1 84
76 Single contactor and blended effluent SDS-DCBAA breakthrough curves for Water 2 84
77 Single contactor and blended effluent SDS-CDBAA breakthrough curves for Water 6 85
78 Single contactor and blended effluent SDS-TBAA breakthrough curves for Water 7 85
79 Single contactor and blended effluent SDS-BF breakthrough curves for Water 2 86
80 Single contactor and blended effluent SDS-CDB AA breakthrough curves for Water 5 86
81 Single contactor and blended effluent SDS-BF breakthrough curves for Water 3 (original
step-lag logistic function model curve fit) 87
82 Single contactor and blended effluent SDS-BF breakthrough curves for Water 3 (fit to
step-lag-peak logistic function model) 87
83 Cumulative frequency distribution plot of normalized residual sum-of-squares (RSS) for DI
and SCA model predictions 100
84 Cumulative frequency distribution plot of normalized bias for DI and SCA model
predictions 100
85 Comparison of DI and SCA methods for predicting the UV254 integral breakthrough curve
for Water 1 101
86 Comparison of DI and SCA methods for predicting the UV254 integral breakthrough curve
for Water 3 101
-xxi-
-------
87 Comparison of DI and SCA methods for predicting the SDS-TTHM integral breakthrough
curve for Water 1 102
88 Comparison of DI and SCA methods for predicting the SDS-TTHM integral breakthrough
curve for Water 7 102
89 Comparison of DI and SCA methods for predicting the SDS-HAA5 integral breakthrough
curve for Water 2 103
90 Comparison of DI and SCA methods for predicting the SDS-HAA5 integral breakthrough
curve for Water 7 103
91 Comparison of DI and SCA methods for predicting the SDS-HAA6 integral breakthrough
curve for Water 3 104
92 Comparison of DI and SCA methods for predicting the SDS-HAA6 integral breakthrough
curve for Water 5 104
93 Comparison of DI and SCA methods for predicting the SDS-HAA9 integral breakthrough
curve for Water 5 105
94 Comparison of DI and SCA methods for predicting the SDS-HAA9 integral breakthrough
curve for Water 6 105
95 Comparison of DI and SCA methods for predicting the SDS-TOX integral breakthrough
curve for Water 6 106
96 Comparison of DI and SCA methods for predicting the SDS-TOX integral breakthrough
curve for Water 7 106
97 Comparison of DI and SCA methods for predicting the SDS-CF integral breakthrough
curve for Water 2 107
98 Comparison of DI and SCA methods for predicting the SDS-CF integral breakthrough
curve for Water 6 107
99 Comparison of DI and SCA methods for predicting the SDS-BDCM integral breakthrough
curve for Water 1 108
100 Comparison of DI and SCA methods for predicting the SDS-BDCM integral breakthrough
curve for Water 8 108
101 Comparison of DI and SCA methods for predicting the SDS-DBCM integral breakthrough
curve for Water 3 109
102 Comparison of DI and SCA methods for predicting the SDS-DBCM integral breakthrough
curve for Water 7 109
103 Comparison of DI and SCA methods for predicting the SDS-BF integral breakthrough
curve for Water 1 110
104 Comparison of DI and SCA methods for predicting the SDS-BF integral breakthrough
curve for Water 5 110
105 Comparison of DI and SCA methods for predicting the SDS-DCAA integral breakthrough
curve for Water 1 Ill
106 Comparison of DI and SCA methods for predicting the SDS-DCAA integral breakthrough
curve for Water 4 Ill
107 Comparison of DI and SCA methods for predicting the SDS-DBAA integral breakthrough
curve for Water 2 112
108 Comparison of DI and SCA methods for predicting the SDS-BCAA integral breakthrough
curve for Water 8 112
109 Integral breakthrough curves for varying numbers of contactors operated in parallel-
staggered mode (B=30;D=0.1) 117
-xxii-
-------
110 Integral breakthrough curves for varying numbers of contactors operated in parallel-
staggered mode (B=30; D=0.05) 117
111 Integral breakthrough curves for varying numbers of contactors operated in parallel-
staggered mode (B=10; D=0.1) 118
112 Integral breakthrough curves for varying numbers of contactors operated in parallel-
staggered mode (B=10; D=0.05) 118
113 Integral breakthrough curves for varying numbers of contactors operated in parallel-
staggered mode (B=10; D=0.2) 119
114 Integral breakthrough curves for varying numbers of contactors operated in parallel-
staggered mode (B=30; D=0.2) 119
115 Run time as a function of number of contactors in parallel, expressed as percent of run time
for infinite n (B=30; D=0.1) 120
116 Run time as a function of number of contactors in parallel, expressed as percent of run time
for infinite n(B=30;D=0.2) 120
117 Impact of extrapolation on DI prediction of the TOC integral breakthrough for Water 5 124
118 Impact of extrapolation on DI prediction of the TOC integral breakthrough for Water 8 124
119 Impact of extrapolation on SCA prediction of the UV254 integral breakthrough for Water 5 125
120 Impact of extrapolation on SCA prediction of the SDS-TOX integral breakthrough for
Water5 125
121 Impact of extrapolation on SCA prediction of the SDS-TTHM integral breakthrough for
WaterS 126
122 Impact of extrapolation on SCA prediction of the SDS-HAA5 integral breakthrough for
WaterS 126
123 Impact of extrapolation on SCA prediction of the SDS-HAA6 integral breakthrough for
WaterS 127
124 Impact of extrapolation on SCA prediction of the SDS-HAA9 integral breakthrough for
WaterS 127
125 Impact of extrapolation on SCA prediction of the UV254 integral breakthrough for Water 8 128
126 Impact of extrapolation on SCA prediction of the SDS-TOX integral breakthrough for
WaterS 128
127 Impact of extrapolation on SCA prediction of the SDS-TTHM integral breakthrough for
WaterS 129
128 Impact of extrapolation on SCA prediction of the SDS-HAA5 integral breakthrough for
WaterS 129
129 Impact of extrapolation on SCA prediction of the SDS-HAA6 integral breakthrough for
WaterS 130
130 Impact of extrapolation on SCA prediction of the SDS-HAA9 integral breakthrough for
WaterS 130
-XXlll-
-------
This page intentionally left blank.
-xxiv-
-------
List of Abbreviations
BCAA
BDCAA
BDCM
BF
BMRL
BrTOC
C(t)
C(t)
C(tf)
C(tp)
Co
CDBAA
CF
CP
DBAA
DBCM
DBF
DCAA
DCBAA
DI
EBCT
Bromochloroacetic acid
Bromodichloroacetic acid
Bromodichloromethane
Bromoform
Below the minimum reporting level
Bromide to TOC ratio
Effluent concentration
blended effluent concentration at individual contactor run time, t
Last observed data point
Measured peak concentration
Influent concentration
Chlorodibromoacetic acid
Chloroform
Logistic function model best-fit concentration at tp
Dibromoacetic acid
Dibromochloromethane
Disinfection byproduct
Dichloroacetic acid
Dichlorobromoacetic acid
Direct integration
Empty bed contact time
-xxv-
-------
EPA
7
ft
GAC
HAA
HAAS
HAA6
HAA9
ICR
K
MBAA
MCAA
MCL
MRL
n
N
NA
nBr
Nc
n'Br
NOM
PAS
United States Environmental Protection Agency
Fraction of organic matter remaining in the combined effluent
C(t)/C0
Granular activated carbon
Haloacetic acid
Sum of five haloacetic acids: MCAA, DCAA, TCAA, MBAA, DBAA
Sum of six haloacetic acids: HAAS, BCAA
Sum of nine haloacetic acids: HAA6, DCBAA, CDBAA, TBAA
Information Collection Rule
Adsorption rate coefficient
Monobromoacetic acid
Monochloroacetic acid
Maximum contaminant level
Minimum reporting level
Freundlich isotherm parameter
Number of contactors
Not applicable
Bromine incorporation factor for THMs
Adsorption capacity coefficient
Bromine incorporation factor for HAAs
Natural organic matter
Polyaluminum sulfate
-xxvi-
-------
R2
RSS
RSSCT
RTsc
SAS
SCA
SDS
SDS-BCAA
SDS-BDCAA
SDS-BDCM
SDS-BF
SDS-CDBAA
SDS-CF
SDS-DBAA
SDS-DBCM
SDS-DBPs
SDS-DCAA
SDS-HAA5
Throughput of the individual contactor when the treatment objective is
exceeded in the single contactor
Throughput of the nth contactor at the time the treatment objective is
exceeded in the blended effluent
Coefficient of determination
Residual sum of squares
Rapid small-scale column test
Blended contactor run time
Single contactor run time
Statistical Analysis Software
Surrogate correlation approach
Simulated distribution system
Bromochloroacetic acid evaluated under SDS conditions
Bromodichloroacetic acid evaluated under SDS conditions
Bromodichloromethane evaluated under SDS conditions
Bromoform evaluated under SDS conditions
Chlorodibromoacetic acid evaluated under SDS conditions
Chloroform evaluated under SDS conditions
Dibromoacetic acid evaluated under SDS conditions
Dibromochloromethane evaluated under SDS conditions
Disinfection byproducts evaluated under SDS conditions
Dichloroacetic acid evaluated under SDS conditions
The sum of five haloacetic acids evaluated under SDS conditions
-xxvii-
-------
SDS-HAA6 The sum of six haloacetic acids evaluated under SDS conditions
SDS-HAA9 The sum of nine haloacetic acids evaluated under SDS conditions
SDS-MBAA Monobromoacetic acid evaluated under SDS conditions
SDS-MCAA Monochloroacetic acid evaluated under SDS conditions
SDS-TBAA Tribromoacetic acid evaluated under SDS conditions
SDS-TCAA Trichloroacetic acid evaluated under SDS conditions
SDS-TOX Total organic halides evaluated under SDS conditions
SDS-TTHM Total trihalomethanes evaluated under SDS conditions
SM Standard Methods
t Service time
tb Run time at which initial breakthrough above detectable levels occurs
TBAA Tribromoacetic acid
TCAA Trichloroacetic acid
THM Trihalomethane
THMs Trihalomethanes
TOC Total organic carbon
TOX Total organic halide
tp Run time at which the peak concentration occurs
TSUVA Specific ultraviolet absorbance based on TOC
TTHM Total trihalomethane
UV254 Ultraviolet absorbance at 254 nm
v Linear velocity
-XXVlll-
-------
Bed depth
-xxix-
-------
This page intentionally left blank.
-XXX-
-------
1 Background and Method Development
Sixty-two public water systems required by the Information Collection Rule (ICR) to perform
treatment studies for disinfection byproduct (DBF) precursor removal fulfilled this requirement
by evaluating granular activated carbon (GAC), while membranes were examined by 36 utilities.
GAC treatment studies included bench-, pilot-, and full-scale studies. Regardless of the scale,
the ICR required that GAC treatment studies evaluate DBF precursor removal in the effluent of a
single contactor as a function of run time, to assess the breakthrough of DBF precursors as the
GAC was exhausted. In practice, full-scale plants can optimize GAC performance and reduce
carbon usage rates by blending the effluents of multiple contactors operated in parallel prior to
disinfection. To optimize the blending process, replacement of the GAC in each contactor is
staggered at regular intervals. By doing so, the service life of each contactor is extended because
water from contactors that exceed the treatment objective is blended with water from contactors
with fresher GAC that have effluent concentrations below the treatment objective. A cost
analysis based on single contactor run times will overestimate the actual treatment costs of full-
scale operation of multiple contactors operated in parallel-staggered mode.
Previous researchers have developed mathematical relationships to characterize the water quality
in the blended effluent based on single contactor breakthrough curves. These relationships yield
a blended contactor effluent or integral "breakthrough" curve that provides a measure of the
extended run time for which each contactor can be operated to meet a treatment objective in the
blended effluent. The integral breakthrough curve does not directly represent blended effluent
water quality, but is a tool for determining the GAC replacement frequency for each individual
contactor (of multiple contactors operated in parallel-staggered mode) to maintain the blended
effluent below the treatment objective. For example, the results of a GAC study show that based
on a single contactor breakthrough curve, GAC effluent formed DBF levels are maintained
below the treatment objectives for 80 days of operation. This is the single contactor run time,
RTSc. Since full-scale implementation of GAC will involve multiple contactors operated in
parallel-staggered mode, integral breakthrough curves for each parameter are developed based on
the single contactor breakthrough curves. The integral breakthrough curves might show that the
level of DBFs formed in the GAC blended effluent are maintained below the treatment
objectives for 160 days of operation, indicating that each individual contactor can be operated for
160 days while maintaining the plant blended effluent water quality below the treatment
objective. The blended contactor run time (RTBc) is 160 days.
The basic method utilized to determine the integral breakthrough curve has been a mathematical
or numerical integration of the single contactor breakthrough curve. A minimal amount of
experimental verification has been performed to verify this approach. Some bench-scale
experimental verification data has been presented (Chowdhury et al., 1996; Summers et al.,
1998), whereby mathematical integration was simulated experimentally by continuously
collecting the effluent from a bench-scale GAC contactor in a large reservoir, and sampling from
this reservoir over time. An extensive verification study is needed to thoroughly evaluate the
appropriateness of the integration method for predicting the integral breakthrough curve.
-1-
-------
1.1 Optimization of GAC Operation
The GAC in a contactor has to be replaced when the mass transfer zone begins to exit the column
as shown in Figure 1, and the effluent concentration exceeds the treatment objective. However,
at this point only part of the GAC bed is saturated and replacement of the GAC will result in
high carbon use rates (Snoeyink, 1990). Two common methods of lowering carbon usage rates
are to operate contactors in series or to operate multiple contactors in parallel with staggered
GAC replacement cycles.
For adsorption of micropollutants, the amount of water treated per mass GAC can be increased
by operation of two contactors in series. In this mode of operation, two contactors are operated
in series until the treatment objective is exceeded in the second contactor. At this point, the
GAC in the first contactor is replaced with virgin or reactivated GAC, and valves are switched so
that the second contactor is now operated in-line ahead of the first contactor. This cycle is
repeated to maintain effluent levels below the treatment objective. For efficient operation, the
mass transfer zone should be contained within the bed length of one contactor. This can be
achieved using reasonable bed lengths for adsorption of micropollutants, but the mass transfer
zone for TOC removal (and therefore DBF precursor removal) is usually too long. For DBF
precursor control, operation of two contactors in series does not result in significantly longer run
times over single contactor operation (Sontheimer et al., 1988).
Multiple contactors operated in parallel and staggered in terms of GAC replacement times, as
shown in Figure 1, yield blended effluent concentrations as shown in Figure 2. When the lead
contactor, which has been in operation the longest, is taken off-line and replaced with a contactor
with fresh GAC, the blended effluent concentration decreases as shown in Figure 2. The level of
this decrease is dependent on the number of contactors in operation. For two contactors, the
concentration will decrease by 50 percent, while for an infinite number of contactors the
decrease approaches zero.
Due to the blending of waters from multiple contactors producing varying levels of effluent
water quality, individual contactors can be operated past the point at which the effluent water
they are producing exceeds the treatment objective, because the treatment objective must be
maintained in the blended effluent. Therefore, evaluation of single contactor breakthrough curve
data will result in overestimates of the carbon usage rate for a full-scale system operating
multiple staggered contactors in parallel-staggered mode. For DBF precursor control, contactor
effluents are blended prior to disinfection.
1.2 Modeling the Operation of Multiple Contactors Operated in Parallel-
Staggered Mode
In modeling the operation of multiple contactors operated in parallel-staggered mode, the goal is
not to simulate the actual blended effluent water quality during normal operation (as shown in
Figure 2), but to derive the integral breakthrough curve. The integral breakthrough curve is a
tool used to determine the GAC replacement frequency for each individual contactor to maintain
the blended effluent below the treatment objective: it is a curve that relates single contactor run
time to blended contactor effluent water quality. Multiple contactor throughput to a treatment
objective can be estimated from the integral breakthrough curve by determining the operation
-2-
-------
Influent water
0000
Disinfectant
Blended effluent
^Finished
water
Figure 1 Schematic of blending of multiple GAC contactors operated in parallel-staggered
mode
Blended effluent
concentration
Blended effluent treatment objective
New contactor
placed on-line
Operation time
Figure 2 Blended effluent water quality during quasi steady-state operation of multiple
contactors in parallel-staggered mode (adapted from Summers et al., 1998)
-3-
-------
time when the curve intersects the treatment objective, as is done with single contactor
breakthrough curves.
Figure 3 shows a series of graphs that describe how the integral breakthrough curve is developed.
Graphs A through E depict eight GAC contactors on-line in parallel-staggered mode. For
simplification, identical breakthrough curves are assumed for each contactor. The interval
between GAC replacement, or the single contactor run time (RTSc), is increased from 16 to 99
days over these five graphs. At each RTSc, the single contactor effluent concentration, C(f), is
given as well as the blended contactor effluent concentration, C(Y), which is calculated by
averaging the effluent water quality of each of the eight contactors at RTsc (shown as a short
dotted line representing the intersection of effluent concentration at RTSc). The dashed line
breakthrough curve represents the contactor that replaces the first contactor when it has reached
the end of its service life. As RTSc is increased, moving from Graph A through E, the C(t) at
RTSc increases. Therefore, the GAC replacement interval affects the blended contactor effluent
water quality. Specifically, as shown graphically by the integral breakthrough curve in Graph E,
as RTSC increases, blended effluent water quality declines. Using the integral breakthrough
curve, the single contactor run time at which the treatment objective is exceeded in the blended
effluent can be determined.
Assuming a linear breakthrough curve, Westrick and Cohen (1976) modeled the impact of
parallel contactor operation on blended effluent water quality, and derived the following
equation:
<»
where TV is the number of parallel contactors, qN is the specific throughput of the Mh contactor at
the time the treatment objective is exceeded in the blended effluent, and qt is the throughput of
the individual contactor when the treatment objective is exceeded in that single contactor. For
large N, Equation 1 shows that q^ approaches twice qt: the run time of each contactor when the
blended effluent treatment objective is exceeded will approach twice that of a single contactor
when the treatment objective is exceeded in the single contactor effluent. For a finite N, such as
10 contactors, q^ is a factor of 1.8 times greater than qt.
Equation 1 establishes a relationship between the specific throughput of the Mh contactor at the
time the treatment objective was exceeded in the blended effluent and the throughput of a single
contactor to that same treatment objective, assuming a linear breakthrough curve. For large N,
Equation 1 shows that the run throughput of each contactor approaches twice that for a single
contactor:
<7~=2
-------
Identical beattiroutj curves represent ffeinbp
2 i RTsc=16days contactors operated in parallel-stagred mode
o
I
0)
u
8
o
1
HI
°_.
c
o
0)
Operation time, days
RTSc = 31 days
Operation time, davs
RTSc = 46 days
Operation time, days
C(RTSC)= 0.11 (A)
C(RTSC)= o.67
C(RTSC)= 0.31 (B)
O
c"
o
t:
o
O
RTSC = 61 days
C(RTSC) = 0.99
crRrsc; = 0.64 (D)
e s
Operation time, days
2 -, C(RTSC)= 1.00
C(RTSC)= 0.80 (E)
o
1
HI
Operation time, days
«,
C
o
RTSC = 99 days
Single contactor
breakthrough curve
Blended effluent integral
breakthrough curve
n = 8 contactors)
Operation time, days
Figure 3 Operation of multiple contactors in parallel staggered mode to various single contactor run times: derivation of
the blended effluent integral breakthrough curve
-------
Therefore, under the linear breakthrough curve assumption, when 9 contactors are operated in
parallel-staggered mode, the throughput of each contactor when the blended effluent exceeds the
treatment objective will be 90 percent of that for an infinite number of contactors operated in
parallel-staggered mode. The analysis assumes a linear breakthrough curve and shows that N90 is
independent of the magnitude of the treatment objective.
Roberts and Summers (1982) examined the impact of contactor operation in parallel-staggered
mode on the run times of individual contactors. The authors showed that the fraction of organic
matter remaining in the combined effluent, /, could be estimated from a single contactor
breakthrough curve, assuming regular GAC replacement intervals:
where TV is the number of contactors and/ is the fraction, C(t)/Co, of organic matter remaining in
the effluent of the rth contactor, determined from a breakthrough curve. A plot of the integral
breakthrough curve over operation time provides an estimate of the service time of each
contactor in a multiple contactor scenario for a treatment objective. This relationship is referred
to as the integral breakthrough curve, and the development of this curve is explained graphically
in Figure 3.
1.3 Single Contactor Breakthrough Curve Models
A model used to describe single contactor effluent experimental data is needed for several
reasons. From a data management perspective, best-fit curve parameters that adequately
describe experimental data are less memory intensive than storing the entire experimental data
set. A best-fit curve also facilitates interpolation and extrapolation necessary to estimate run
times for given treatment objectives. Use of a best-fit model curve also provides an estimate of
the scatter in the data through the coefficient of determination, and the model minimizes the
impact of this scatter on run time estimates. Finally, a function that describes the single
contactor breakthrough curve is a prerequisite for calculating the integral breakthrough curve, a
curve that relates single contactor run time to blended effluent water quality under the
assumption that an infinite number of contactors are operated in parallel-staggered mode. Run
time estimates generated by the integral breakthrough curve are more applicable to full-scale
GAC operation where multiple contactors are operated in parallel-staggered mode to increase the
service time of each individual contactor.
Many researchers have applied various forms of the logistic function to predict GAC
breakthrough curves or to fit existing breakthrough curve data. The logistic function is a
symmetric S-shaped curve with a midpoint inflection. Oulman (1980) describes the
development of the Bohart-Adams equation, published in 1920, which was used to model the
service life of activated carbon for the removal of airborne chlorine by gas masks. The Bohart-
Adams equation is:
-6-
-------
In
c(0-i
(5)
where C(f) is the effluent concentration at time t, Co is the influent concentration, K is an
adsorption rate coefficient, Nc is an adsorption capacity coefficient, x is the bed depth, v is the
linear velocity, and t is the service time. The Bohart-Adams equation is a bed depth service
model and was derived based on surface reaction theory (Clark, 1987). Equation 5 can be
rewritten in the form of the logistic function:
1
C(t) \ + e(a+bt)
where the variables a and b are defined as:
(6)
a =
-KNcx
(7)
and
(8)
If the adsorption coefficients K and Nc are known, as well as other operational parameters
(velocity and depth), the effluent concentration at time t can be predicted by Equation 6.
Because GAC breakthrough curves are generally not symmetrical, Clark (1987) proposed the use
of the generalized logistic function to model GAC breakthrough curves. This generalized model
incorporates the Freundlich isotherm parameter and is as follows:
C
n-l
n-l
Ae~
(9)
where l/n is the Freundlich isotherm parameter, r is a constant, and A, a constant, is defined as:
A =
Cn
,-rt
(10)
The derivation of this generalized logistic function is described in Clark, Symons, and Ireland
(1986). Their approach, which builds on the work of Oulman, is an effort to predict GAC
breakthrough curve profiles based on adsorption characteristics and influent concentration.
A predictive approach to single contactor GAC breakthrough curves was not examined as a part
of this study. Instead, a model was needed to curve fit experimental breakthrough data. Due to
the inherent ability of the logistic function to match the typical S-shaped breakthrough curve, and
the previous body of work that has utilized the logistic function to model GAC breakthrough
data, the logistic function was chosen as the equation used to fit GAC breakthrough data in this
study.
-7-
-------
Chowdhury et al. (1996) and Summers et al. (1998) applied the following form of the logistic
function to model experimental GAC breakthrough data:
where the values for A, B, and D are determined experimentally by a best-fit to the breakthrough
data. The parameter A represents the level to which the function approaches asymptotically.
Parameters B and D affect the shape of the curve. Equation 1 1 was found to adequately fit GAC
breakthrough curves for three water sources and the parameters total organic carbon (TOC) and
formed total trihalomethane (TTHM). With a few modifications, as described in Section 3.2.1
below, Equation 1 1 served as a basis for the single contactor breakthrough curve modeling work
contained in this study.
1.4 Direct Integration Approach
The average value function, a mathematical integration, assumes an infinite number of parallel-
staggered contactors and replaces the numerical integration required for solving Equation 4.
However, it is important to understand the impact of the infinite contactor assumption on model
results. Based on Equation 1, nine contactors in parallel will yield qN=9 within 10 percent of q^ .
Therefore, for 10 contactors or greater, the difference will be less than 10 percent, based on the
linear breakthrough curve assumption. Additionally, Westrick and Cohen (1976) found that the
carbon usage rate for individual contactors operated in parallel-staggered mode will approach
half the carbon usage rate based on meeting the treatment objective using a single contactor.
Roberts and Summers (1982) examined integral TOC breakthrough curves applied to eight case
studies, and found that GAC run times increased by a factor of two to three over those based on
the single contactor breakthrough curve, for a 50 percent TOC breakthrough objective. They
concluded that staggered multiple parallel contactor operation will lower carbon usage rates even
more than predicted by Westrick and Cohen, whose conclusions were based on a linear
breakthrough curve.
Chowdhury et al. (1996) and Summers et al. (1998) applied the average value function to show
that the integral breakthrough curve can be represented by:
C(t) = -\C(t)dt (12)
t o
where C(t) is the blended effluent concentration at individual contactor run time, t, and C(f) is
an equation that describes the single contactor effluent concentration as a function of time. This
time-averaged mathematical integration of the function that describes the breakthrough curve
yields the integral breakthrough curve, as it describes the average value of the function at any
point in time. This procedure is referred to here as the direct integration (DI) approach. While
Equation 4 provides an expression of the average blended concentration of N contactors in
parallel, Equation 12 represents a time-averaged blended effluent concentration of an infinite
number of staggered parallel contactors. Therefore, a plot of Equation 12 over operation time
represents the run time to any given treatment objective in the blended contactor effluent for
-8-
-------
each individual contactor of an infinite number of parallel-staggered contactors. This plot is not
a direct representation of blended effluent water quality, but a tool to determine GAC run times
of each contactor operated in parallel-staggered mode.
Chowdhury et al. (1996) and Summers et al. (1998) applied Equation 12 to Equation 11 to
predict the integral breakthrough curve. Application of the DI approach utilizing different forms
of the logistic function (Equation 11) is described below in Section 3.2.3.
1.5 Surrogate Correlation Approach
Theoretically, the direct integration approach is a reliable method for estimating blended effluent
water quality. However, during the analysis approach for the ICR GAC treatment study data, the
treatment study technical work group (TS-TWG) determined that it would be computationally
intensive to apply the DI approach to the large number of breakthrough curves (8,000 to 9,000)
comprising the ICR treatment study data set. Furthermore, the DI approach may be less reliable
for breakthrough curves that do not follow the typical "S" pattern, such as "peak" curves or sharp
"S" shaped curves followed by a plateau. For these reasons, another approach for estimating
blended effluent water quality from single contactor data was developed for analysis of the ICR
treatment study GAC data.
The surrogate correlation approach (SCA) was developed by the TS-TWG as an alternative to
the direct integration approach for calculating the integral breakthrough curve (USEPA, 1999).
The SCA is based on the assumption that the relationship between TOC and all other parameters
(UV254 and SDS-DBPs) in the single contactor effluent is maintained in the blended effluent of
multiple staggered contactors. In other words, the concentration and speciation of DBFs formed
after chlorination of the blended effluent of multiple contactors with a given TOC concentration
is the same as that formed after chlorination of the single contactor effluent with the same TOC
concentration. This method of estimating blended effluent water quality from single contactor
data requires that an integral breakthrough curve be determined based on the single contactor
TOC breakthrough only. Once this curve is known, along with a relationship between TOC and
all other parameters in the single contactor effluent, integral breakthrough curves are estimated
for all other parameters. This approach not only allows for the evaluation of blended contactor
run times to various breakthrough criteria, but also for the occurrence of other DBFs at that run
time. For example, it will be important to evaluate the levels of individual DBFs, such as
BDCM, at various regulatory targets under consideration. This will allow EPA to assess whether
or not regulating one DBF or DBF group can effectively control the occurrence of other DBFs or
DBF groups.
An example of the SCA procedure is described in the steps below and summarized graphically in
Figure 4.
1. Select a treatment objective (e.g., TTHM = 32 |ig/L).
2. Use the single contactor breakthrough curve for the parameter of interest and determine the
single contactor run time (RTSc) at which the treatment objective is exceeded.
-9-
-------
SDS-TTHM
C(0
RT
SDS-HAA6
sc •«-
SDS-CF
(All measured parameters)
Chosen treatment
objective "(Step I)
Run time, t
Run time, t
Run time, t
Logistic function best-fit of single
contactor experimental data
Determine single contactor run time,
RTSC, to treatment objective (Step 2)
Step 3 (repeat for all
measured parameters)
TOC integral breakthrough curve
calculated using direct integration
approach
Run time, t
Run time, t
Figure 4 Graphical summary of SCA procedure used for base analysis of GAC
treatment studies fadaoted from USEPA, 1999)
-10-
-------
3. Use the RTSc from Step 2 to determine the concentrations of all other parameters from the
single contactor breakthrough curves. (In this step, the single contactor effluent
concentrations of all parameters are "linked" through the single contactor run time, RTSc).
4. Use the RTSc from Step 2 to determine the single contactor effluent TOC concentration that
corresponds to the treatment objective.
5. From the integral TOC breakthrough curve, calculated by the DI approach, determine the
blended contactor run time (RTBc) to reach the TOC concentration calculated in step 4. This
is the only point in the analysis where it is necessary to apply the DI method to establish an
integral breakthrough curve, and only the TOC integral breakthrough curve is required. If
the integral breakthrough curve does not exceed the TOC concentration calculated in Step 4,
extrapolation of the TOC integral breakthrough curve may be required.
This analysis makes the following assumptions:
1. The logistic function model can accurately describe the breakthrough of DBF precursors and
DBF precursor surrogates,
2. The relationship between TOC and DBF precursors (concentration and speciation) observed
in the single contactor effluent is maintained in the blended effluent,
3. The TOC breakthrough curve can be extrapolated with reasonable results,
4. The TOC integral breakthrough curve accurately predicts the blended water quality of an
infinite number of multiple contactors operated in parallel-staggered mode. Furthermore, the
TOC integral breakthrough curve based on an infinite number of contactors is a reasonable
approximation to a finite numbers of contactors.
All four of these assumptions are verified as a part of this study.
1.6 Impact of Bromide Concentration on GAC Effluent Blending Models
The second assumption listed for the SCA procedure is that the relationship between TOC and
DBF precursors (concentration and speciation) observed in the single contactor effluent is
maintained in the blended effluent This is an important assumption because the SCA applies
DBF formation and speciation at a given single contactor run time and TOC concentration to the
blended contactor run time at which an equivalent TOC concentration is achieved. The relative
concentrations of TOC and bromide may influence DBF formation and speciation.
A number of researchers have discussed the impact of bromide concentration on DBF formation
and speciation. Under constant chlorination conditions, and at a constant TOC concentration, as
the bromide concentration increases the bromide to TOC ratio increases and DBF speciation
shifts to favor the formation of brominated species (Summers et al., 1993). GAC treatment does
not remove bromide, while TOC is adsorbed, resulting in higher GAC effluent bromide to TOC
ratios in the GAC effluent as compared to those in the GAC influent. Due to this increase, GAC
effluent formed DBFs may undergo shifts in speciation to higher fractions of the more
brominated DBF species. In some cases effluent formed DBF species concentrations are
measured higher than those formed in the influent. It is important to track the breakthrough
-11-
-------
behavior of specific DBF species, because some may be of potential health concern and a MCL
could be set for a specific DBF species.
The shift in DBF speciation for THMs can be measured by calculating the bromine incorporation
factor for THMs, nBr (Gould et al., 1983):
(13)
—
TTHM
where all concentrations (of species and TTHM) are expressed as molar concentrations. The
value of nBr can range from 0 (only chloroform formed) to 3 (only bromoform formed). Based
on the bromine incorporation factor for HAA6 (Shukairy et al., 1994), the bromine incorporation
factor, nrBr, for HAA9 is defined as:
1-MBAA + 1-BCAA + 1-DCBAA + 2-DBAA + 2-CDBAA + 3-TBAA
n'B= - (14)
Br HAA9
where all concentrations (of species and HAA9) are expressed as molar concentrations.
The value of n'Br can range from 0 (only MCAA, DCAA, or TCAA formed) to 3 (only TBAA
formed). Examining and comparing nBr and n'Br values between single contactor and blended
effluent will help determine whether the second SCA method assumption described in Section
1.5 is valid.
1.7 Effluent Blending Modeling of Fewer than 10 Contactors
Both the DI and SCA methods for determining integral breakthrough curves rely on the
assumption of an infinite number of contactors operated in parallel-staggered mode. For 10 or
more contactors in parallel-staggered operation, the integration presented in Equation 12
approximates blended effluent water quality within 10 percent (Roberts and Summers, 1982).
Equation 12 is utilized exclusively for the DI approach, while the SCA approach relies on
Equation 12 to establish the integral breakthrough curve for TOC as part of the analysis.
Therefore, it is important to examine the impact of the infinite number of contactors assumption
on the ability of these methods to predict blended contactor effluent water quality for a finite
number of contactors. Furthermore, for smaller plants that operate fewer than 10 contactors in
parallel-staggered mode, the approximation given by Equation 12 will be less accurate and actual
service lives will be shorter than those predicted by Equation 12.
1.8 GAC Breakthrough Curve Extrapolation
The SCA procedure is limited by the highest TOC concentration reached by the TOC integral
breakthrough curve, which is typically 40 to 70 percent of the highest single contactor TOC
concentration. Therefore, although higher single contactor TOC concentrations are associated
with formed DBF concentrations, these cannot be applied to the integral breakthrough curve
during application of the SCA procedure unless the TOC integral breakthrough curve is
extrapolated. It is therefore important to establish the impact of extrapolation of the TOC
-12-
-------
integral breakthrough curve on estimated run times to treatment objectives and blended contactor
water quality. The sensitivity of the predicted integral breakthrough curve to extrapolation was
evaluated for two waters in this study.
1.9 ICR GAC Treatment Study Data Analysis Context
The design of this study incorporated two main goals. The primary objectives were to evaluate
the ability of the logistic function to model single contactor breakthrough curve data and to
evaluate the success and limitations of predictive models used to determine the integral
breakthrough curve, a relationship between single contactor run time and blended contactor
water quality. The secondary objective of this study was to evaluate the applicability of these
models and predictive methods in the context of the ICR GAC treatment study data analysis.
A large amount of ICR treatment study data will be analyzed: the 62 GAC treatment studies
performed will generate a total of 8,000 to 9,000 individual breakthrough curves. The SCA
method is especially applicable to this data analysis procedure because it minimizes the
computations necessary to estimate blended contactor run times for treatment objectives. An
assessment of the concentration of other DBFs at any given treatment objective will be required
as part of the data analysis effort, and the SCA procedure is also suited to this task. The SCA
procedure requires that GAC breakthrough curves for all measured parameters be represented by
the logistic function model curve fit. By doing so, a smaller amount of data are needed to
represent breakthrough curve experimental data. The following steps outline the data analysis
procedure to be utilized during the ICR treatment study data analysis effort:
1. The logistic function model will be used to fit all water quality parameters in the GAC
effluent.
2. The TOC integral breakthrough curve will be determined using the DI approach. In some
cases, this curve may be extrapolated.
3. All logistic function model fit coefficients will be entered into a database to allow different
breakthrough criteria to be evaluated and queried across all studies.
4. The SCA procedure will be used to estimate run times for multiple contactors operated in
parallel-staggered mode by determining the effluent concentrations of various water quality
parameters linked to a common single contactor run time, and calculating the blended
effluent run time that corresponds to the TOC concentration. In this manner, simultaneous
treatment objectives, such as THM or HAA regulatory target concentrations, can be
evaluated to determine which parameter controls the design and operation of the process.
5. Results will be used to evaluate blended contactor run times that meet regulatory target
treatment objectives, and this information can be used to estimate costs. Simultaneous
treatment objectives will also be evaluated (e.g., HAAs and THMs, TOC and THMs, THMs
and BDCM).
-13-
-------
1.10 Appropriateness of Model Assumptions to Full-Scale GAC Effluent Blending
Both the DI and SCA methods rely on the assumption that the GAC in each contactor of an array
of contactors is replaced at regular intervals, so that the service times of all contactors are equal.
They also assume that the breakthrough curve profiles of all single contactors are identical. In a
full-scale plant, these idealized conditions will rarely occur. Variability in source water quality
may impact the run time of the contactors, depending on when they are placed in service and
DBF precursor concentrations in the GAC influent during their service life. Variability in
distribution system conditions, especially temperature, may impact the contactor service life, as
DBF levels may increase with higher temperatures.
For a plant that operates a fixed number of contactors, water demand changes during the year
may directly impact the EBCT of each contactor, or the number of contactors in operation.
Under operation of a constant number of contactors, a contactor that is placed on-line at the
beginning of the summer high water demand months may be operated under a shorter EBCT as
compared to a contactor placed on-line during the winter. Furthermore, it is less desirable to
remove contactors from service to replace GAC during high demand periods. Another approach
is to increase the number of contactors on-line as water demand increases.
-14-
-------
2 Study Objectives and Approach
This study was performed in conjunction with bench-scale GAC treatment studies that were
performed at one laboratory in fulfillment of ICR requirements for eight utilities. It was
designed to examine the following experimental objectives:
1. Assess the ability of the logistic function model to fit single contactor breakthrough data for
eight GAC runs using eight water sources and all measured parameters, including DBF
surrogates, DBF class sum parameters, and DBF species. The water sources represent a
range in TOC concentrations, DBF precursor levels, bromide concentrations, and SDS
chlorination conditions.
2. Verify through bench-scale experiments the accuracy of the direct integration (DI) method
for establishing the integral breakthrough curve, a relationship between single contactor
operation time and blended effluent water quality.
3. Examine the accuracy of the computationally-simpler surrogate correlation approach (SCA)
to predict the integral breakthrough curves of all measured parameters. Verify a basic
assumption of the SCA procedure, that the relationship between DBF formation and TOC
concentration in the single contactor effluent and in the blended effluent is maintained.
4. Examine the impact of GAC effluent blending on DBF speciation. Assess the accuracy of
models used to predict the integral breakthrough curve for individual THMs and HAAs, and
compare bromine incorporation into DBF formation in the single contactor and blended
contactor effluents.
5. Evaluate the impact of extrapolation of the integral breakthrough curve on blended effluent
water quality.
6. Verify the accuracy of the integral breakthrough curve predictive models, which assume an
infinite number of contactors, for the prediction of the service life of finite numbers of
contactors.
-15-
-------
This page intentionally left blank.
-16-
-------
3 Materials and Methods
3.1 Experimental Approach
3.1.1 Rapid Small-Scale Column Test
The performance of full-scale GAC contactors was simulated using the rapid small-scale column
test (RSSCT). Previous studies have shown that diffusion of natural organic matter (NOM) to
adsorption sites on GAC is proportional the particle size (Crittenden et al., 1989; Sontheimer et
al., 1988). Therefore, by grinding the GAC to a smaller size, rates of adsorption are increased in
proportion to the ratio of full-scale to RSSCT GAC particle sizes, or scaling factor. The scaling
factor also relates the RSSCT EBCT and superficial velocity to the full-scale contactor. If the
RSSCT utilized the full-scale contactor bed length, extremely long columns requiring very high
inlet pressures would be required. However, studies have shown that adjustments to the RSSCT
Reynold's number utilized (between 0.1 and 1.0) have a negligible impact on results. Therefore,
by decreasing the RSSCT Reynold's number, a much shorter column can be designed (with
consequently shorter superficial velocities to maintain a constant EBCT). Complete details on
the RSSCT design for precursor removal studies can be found in the literature (USEPA, 1996;
Summers et al., 1995; Summers et al., 1992; Crittenden et al., 1991).
The RSSCTs designed in this study followed the guidelines outlined in the GAC Precursor
Removal Studies section of the ICR Manual for Bench- and Pilot-Scale Treatment Studies
(USEPA, 1996). A summary of the RSSCT design used for each run is given in Table 1. For
most waters, a 20 minute full-scale EBCT contactor was simulated. However, 15 minute and 7.2
minute EBCT contactors were simulated for Waters 1 and 8, respectively. The designs were
based on the estimated or known GAC influent TOC concentration (which can directly impact
the rate of breakthrough), the full-scale water temperature at the time of sampling, the full-scale
GAC particle size, and the full-scale EBCT simulated. The full-scale bed porosity was assumed
to be 0.45 for all runs. The minimum Reynold's number used ranged from 0.48 to 0.60.
3.1.1.1 GAC Preparation Procedures
Representative batches of Filtrasorb 300 (F-300), and Filtrasorb 400 (F-400), bituminous coal-
based GAC, were obtained from the manufacturer, Calgon Carbon Corporation. The
representative batch of the reactivated GAC/virgin GAC blend used for Water 4 was obtained
directly from the utility. Using a riffle splitter, a small (30 to 50 g) representative sample of the
GAC was obtained. Using ajar mill, the GAC was ground to the needed mesh size. Care was
taken to frequently remove and sieve the GAC in the jar mill. The GAC was ground until the
entire sample passed through the upper mesh size sieve. Usually, a recovery of 25 to 30 percent
was obtained, as defined by the amount of GAC retained between the two mesh size sieves and
divided by the total amount of GAC prior to grinding.
The ground GAC was transferred to a beaker, and covered with reagent grade (adsorbed-
deionized) water. The GAC was washed by repeated additions and decantations of reagent grade
-17-
-------
Design parameter
Design value for each ater
45
oo
GAC manufacturer
GAC band name
GAC tpe
GAC mesh siE
Article diameter.d c (nm)
€kieral design parameters
Mimum avoids numbr.B Smln 1
Full-scale operatingemperature C)
HSematic viscosity vc (n 2s)
Bed porosity ec J
bhsured dr)bd density ps fim 3)
RSSCT design parameters
BTnesh siE
Article diameter.d s (nm)
Siling'actor.B
Calgn Carbn Co. Calgn Carbn Co. Calgn Carbn Co. i&oHrbj tend Calgn Carbn Co. Calgn Carbn Co. Calgn Carbn Co. Calgn Carbn Co.
F-8 F-8 F-8 F-E-8 F-8 F-8 F-8 F-8
Bituminous Bituminous Bituminous Bituminous Bituminous Bituminous Bituminous
a e a m a B a a
606086 6
3032802
H-0 E-6 B-0 1-0 E-6 E-6 1-0 E-6
Bituminous
^faulic loadingate.v G (nnr) 6
Column diameter.D s (nm) 3
Flow rate.Q s (nfnin) 8
Full-scale empt^fed contact time, EBCT c (nin) 5
Estimated full-scale run time.t c T dap) a
Estimated BTun time.t G T dap) 3
tolume water reqired.V G [. Q
bfes GAC reqired.m G & 9
B"6mpt>bd contact time, EBCT s (nin) 9
Bed lengi.l s (:m) 3
2 8 S 8 2 8 8
8 09093 9
7 9 a a 9 s a
a a a a a a z
s s a 2 e 4 $
& a M a 2 a t
a B i e s a a
8968313
3989923
8 2 S 3 8 3 2
Table 1 Summary of RSSCT design parameters for all runs
-------
water. The reagent grade water was added at a high rate and turbulence, to stir up the GAC and
release fines. The supernatant water containing GAC fines was decanted after the GAC was
allowed to settle. Towards the end of the cleaning procedure, the sample was sonicated twice for
5 to 10 seconds. The sonication step helped loosen fines that were subsequently removed by the
addition and decantation of reagent grade water.
The GAC was dried in an oven at 80 to 90°C for 6 to 12 hours. The temperature was then raised
to between 100 and 110°C and the sample was dried until it reached a constant weight. The
sample was removed and cooled inside a dessicator. Once cooled, if not immediately used, it
was stored in a glass vial sealed with a lid with Teflon-lined septum until ready for use.
The dry bed density was measured using a sample of dried and cooled GAC. Stored GAC was
dried in an oven as described above prior to the dry bed density measurement. To measure the
dry bed density, a sample of the GAC was placed inside a 10-mL glass graduated cylinder to a
level of 5 to 9 mL. The cylinder was tapped to pack the GAC. A volume was measured and
recorded. This GAC was then weighed on a balance. The volume reading of the graduated
cylinder was checked and calibrated if necessary by adding a known volume of water to it using
a 10-mL class A graduated pipette. The GAC dry bed density was calculated by dividing the
weight by the calibrated volume.
The calculated mass of GAC for each RSSCT was weighed, placed inside a clean beaker, and
covered with reagent grade water. The wetted GAC was usually allowed to sit for 12 to 24
hours, followed by placement in a vacuum for at least 1 hour to displace the air within the pores.
3.1.1.2 RSSCT Column Setup
The ground GAC used for the RSSCT was packed in glass chromatography columns. Due to the
range of GAC influent TOC concentrations, which correlates to the rate of TOC breakthrough,
columns with inner diameters ranging from 8.0 to 12.6 mm were utilized. The 8.0 and 11.0 mm
inner diameter columns were standard. Other column diameters (9.0, 10.0, and 12.6 mm) were
custom-ordered and the GAC required additional support to ensure the GAC was within the
effective length of the column. The GAC support for these special order columns consisted of a
stainless steel screen (60 or 100 mesh size), Teflon beads, glass wool, and two stainless steel
screens. The support for 8.0 and 11.0 mm inner diameter columns consisted of two stainless
steel screens placed on top of the Teflon fitting. The mesh size of the screens utilized were
based on the ground GAC mesh size. For 100x200 GAC, a 100 mesh screen and a 200 mesh
screen were used. For 140x230 GAC, a 200 mesh screen and a 325 mesh screen were used. For
all column inner diameter sizes, a small amount of glass wool was placed inside the Teflon
fitting, supported by a 60 mesh size stainless steel screen.
The GAC was added to the columns as a slurry and packed by repeatedly tapping the column
sides. The 20 minute full-scale equivalent EBCT RSSCTs were packed into two columns of the
same inner diameter placed in series. Only reagent grade water was used during the packing
process.
-19-
-------
3.1.1.3 Batch Influent Preparation
Prior to RSSCT testing, all water samples were filtered through a 1.0-jim nominal pore size glass
fiber cartridge filter. The cartridge filter was pre-rinsed with deionized water. Dilute solutions
of sulfuric acid and sodium hydroxide were used to maintain the influent pH within 0.1 pH units
of the target pH during operation of the RSSCTs.
3.1.1.4 RSSCT Monitoring
The effluent flow rates were monitored constantly to ensure that the flow rates were maintained
within 5 percent of the design flow rate. The calibration of the effluent flow rate control system
was checked at least three times daily and adjusted when a flow rate differed by more than 3
percent from the design flow rate. The system pressure was monitored daily. The effluent TOC
concentration was monitored frequently so that samples could be taken at the required 5 to 8
percent increments of the average influent TOC concentration.
3.1.2 Bench-Scale Blended Water Quality Assessment Approach
To simulate the integral breakthrough curve obtained by blending multiple full-scale contactors
operated in parallel-staggered mode, the entire effluent from a single GAC contactor was
collected in a reservoir and sampled over time. In this study, the entire effluent from the RSSCT
was collected in a clean 30 or 55 gallon drum. The only effluent water not collected in the drum
was that required for monitoring and sampling of the RSSCT effluent. This included TOC
monitoring and sampling, UV254 sampling, and SDS chlorination. Over time, the blended
effluent drum was sampled and analyzed for TOC and UV254. Samples were also taken and
chlorinated under the same target SDS conditions as those applied to the single contactor study.
During every run, TTHM and HAA9 were analyzed, and during two runs, TOX was also
analyzed, in addition to TTHM and HAA9. Ten percent of blended effluent samples taken were
sampled in duplicate and all analyses were conducted on the duplicate samples (field sample
duplicates).
The first discrete sample taken from the RSSCT effluent also constituted the first blended
sample. Subsequently, seven samples were taken from the blended effluent drum, evenly spaced
over the course of the run. The RSSCTs were operated until at least 70 percent TOC
breakthrough was reached, and the last blended effluent sample was taken at that time. Since the
ICR required that twelve samples be taken from the RSSCT effluent at even increments of TOC
breakthrough, the blended effluent sampling schedule followed the RSSCT sampling. The
second blended effluent sample was taken when the third RSSCT effluent sample was taken.
The remaining blended effluent sample points were based on the RSSCT effluent sample number
as described in Table 2. In practice, the blended effluent samples were taken from the drum just
prior to the addition of the RSSCT effluent sample listed in Table 2. This way, the use of the
RSSCT effluent sample water for SDS chlorination had a smaller impact on the blended effluent
sample. All blended samples were chlorinated under the same SDS conditions as the RSSCT
effluent samples. For two of the waters, (Waters 6 and 8) the run was extended for five
laboratory days after 70 percent breakthrough had occurred. In these instances, eight blended
-20-
-------
effluent samples were taken with the sixth at 70 percent breakthrough, the seventh 2.5 days later,
and the eighth five days later.
Blended effluent RSSCT effluent sample
sample number number
1
2
3
4
5
6
7
8
1
3
6
8
9
10
11
12
Table 2 Relationship between blended effluent sample number and RSSCT effluent
sample number
3.1.3 DBP Formation Assessment
The blended effluent samples were chlorinated under the same target conditions as the RSSCT
effluent. To minimize the volume of water sampled from the blended effluent reservoir, the
chlorine dose used was based on the single column effluent chlorine demand data, by relating
chlorine demand to TOC concentration. The sampling volume required was minimized based on
the DBFs to be analyzed. For 6 of the 8 waters, the blended effluent sample volume was 800
mL, sufficient for TOC, UV254, and SDS chlorination for the analysis of HAA9 and TTHM.
When TOX was also analyzed after chlorination, a 1,300 mL sample was required.
For single contactor effluent samples, chlorine demand studies were performed on the first
effluent sample and the influent water. For each sample, three 125-mL chlorine demand-free
amber glass bottles were used. A combined phosphate/borate buffer solution was added to each
bottle (2.0 mL/L) to maintain a constant target pH during incubation. Dilute solutions of sulfuric
acid or sodium hydroxide were used to adjust the pH prior to chlorination if necessary. Three
chlorine doses were selected based on the TOC concentration of the water and the results of a 5-
minute chlorine demand study (providing a relative measure of inorganic chlorine demand).
Each bottle was filled to 80 to 90 percent of capacity. Using a pipette, a measured amount of a
standardized chlorine solution (using Standard Methods 4500-C1 B) was added to each bottle.
The bottle was then filled and capped head-space free. The bottles were placed in a constant
temperature bath in the dark for the duration of the target incubation period. A titrimetric
procedure (StandardMethods 4500-C1 F) was used to measure the free chlorine residual after the
holding time. The data generated by the chlorine demand study was used to estimate the
chlorine dose to achieve the target residual for all effluent samples by correlating TOC
-21-
-------
concentration to chlorine demand. When TTHM and HAA9 were analyzed, a clean chlorine
demand-free 500 mL amber glass bottle was used. When TOX was also analyzed, a 1000 mL
bottle was used. Caps with Teflon-lined septa were used. The order of DBF sampling was
THMs, HAAs, and TOX. DBF samples were taken in duplicate in prepared bottles with
appropriate preservatives and quenching agents based on the analytical method.
3.1.4 Assessment of the Impact of Sampling on the Integral Breakthrough Curve
Discrete sampling from the RSSCT effluent and sampling from the blended effluent reservoir
was required to obtain water quality data to generate single contactor and integral breakthrough
curves. However, the impact of both types of sampling on the integral breakthrough curve were
unknown. A model was developed to simulate the experimental procedure in an effort to assess
the impact of sampling on the integral breakthrough curve. Based on the model results, RSSCT
design and sampling procedures were optimized to minimize the impact of sampling on the
integral breakthrough curve.
The model was based on sampling the RSSCT effluent continuously in 3-L aliquots. It assumed
that these aliquots were sampled for TOC and UV254 to monitor breakthrough as needed. Based
on the TOC concentration of samples taken, effluent samples and duplicates at ICR-required
TOC increments were identified and chlorinated under SDS conditions.
All samples were added to the reservoir in the order of sampling, and at even increments over the
course of the run, samples were taken from the blended effluent reservoir for TOC and UV254
analysis and SDS chlorination. The results of these analyses comprised the integral
breakthrough curve. The sample volume required was minimized, and any unchlorinated sample
remaining after analysis was complete was reintroduced to the reservoir.
Based on these experimental procedures, a model was developed to assess the impact of
sampling on the integral breakthrough curve, prior to laboratory experiments. The model inputs
were influent TOC concentration, empty-bed contact time (EBCT), column inner diameter, full-
scale GAC mesh size, RSSCT GAC mesh size, minimum Reynold's number, and full-scale
operating temperature. Based on a correlation between influent TOC concentration and bed
volumes to 50 percent TOC breakthrough (Summers et al., 1994; Hooper et al., 1996), and
adjustments to the correlation to estimate run times to different levels of TOC breakthrough, a
TOC breakthrough curve was predicted.
From the estimated single contactor breakthrough curve, two blended effluent breakthrough
curves were generated. A theoretical integral breakthrough curve assumed no column effluent or
blended effluent sampling occurred. The experimental integral breakthrough curve assumed
typical discrete RSSCT effluent sampling, as described in the ICR Manual for Bench- and Pilot-
Scale Treatment Studies, and also included the sampling required from the blended effluent
reservoir to characterize the blended effluent water quality as described in Section 3.1.2.
Increasing the RSSCT column inner diameter (and thus increasing the ratio of volume water
passed to water required for analyses) decreased the impact of RSSCT effluent sampling on the
integral breakthrough curve. The model also demonstrated that sampling from the blended
effluent reservoir had the greatest impact on the difference between theoretical and experimental
integral breakthrough curves. Thus, by increasing the volume of water passed through the
-22-
-------
RSSCT (by increasing the column inner diameter or minimum Reynold's number) this effect
could be minimized.
After each GAC run was complete, the experimental TOC integral breakthrough curve was
compared to an integral curve from which the impact of actual single contactor and blended
effluent sampling was subtracted (corrected integral breakthrough curve). For TOC
breakthrough of each water, the two curves are compared in Appendix A. Overall, the impact of
sampling on the integral breakthrough curve was very small, yielding a mean difference in TOC
concentration at the end of the integral breakthrough curve (where the difference between the
two curves was typically maximized) of 3 percent. The difference ranged from 1 to 6 percent,
with shorter runs yielding higher percent differences. In all cases, the TOC concentration at the
end of the experimental integral curve was higher than that in the corrected integral breakthrough
curve.
3.2 Data Analysis and Modeling Approach
3.2.1 Logistic Function Models
In this study, the logistic function was used to fit GAC breakthrough data from 8 water sources
and GAC runs, and up to 20 measured GAC effluent parameters (including DBF surrogates,
formed DBF class sum parameters, and formed DBF species). In some cases, the logistic
function presented in Equation 1 1 did not satisfactorily fit the breakthrough curve data. Three
modifications were made to the logistic function model to enhance the ability of the model to fit
breakthrough curve data. The models are described in Figure 5. The first modification was to
include an additional parameter, improving the curve fit for breakthrough curves with moderate
to high immediate breakthrough levels. The step logistic function is as follows:
where C(i), B, D, and t are as defined previously in Section 1.3, Equation 11. The term A0
represents the step applied to the logistic function to match moderate to high immediate
breakthrough levels. The term A represents the asymptote to which the logistic function is
approaching. In this study, the step logistic function model was used to curve fit all single
contactor TOC breakthrough curves. The step logistic function model is also adequate for
modeling the GAC effluent chlorine demand breakthrough curve, which typically has relatively
high levels of immediate breakthrough. The addition of the step term also allows the curve to
have a negative y-intercept, which is necessary for incorporation into the step-lag logistic
function model described below.
Because many breakthrough curves (especially for SDS-DBPs) result in a relatively long initial
interval where effluent concentrations are below reportable limits prior to breakthrough, a lag
was incorporated into the function to shift the logistic function outward, allowing for a better fit
of experimental breakthrough data. The run time at which initial breakthrough above detectable
levels occurs is termed 4, and the step-lag logistic function is as follows:
0 t
-------
O
o
8
o
o
C(f) = -
a. Logistic function
O
O
o
Ctf) =
c. Step-lag logistic function
Run time, t
to
Run time,
O
c"
O
I
o
o
b. Step logistic function
O
c"
g
u-i
re
o
o
= o
(tp,cp)
d. Step-lag-peak logistic function
Run time, t
Run time, t
Figure 5 Logistic function model curves
-------
In many cases, a best-fit of the data will yield a negative value for A0 (a negative y-intercept). In
essence, this shifts the logistic function downward, so that the beginning of the curve is negative
(as shown by the dotted line on curve C in Figure 5). However, this occurs when t < tb, so by
Equation 16, the result is set to zero. For ICR treatment study data analysis, the result will be set
to 50 percent of the MRL for DBF surrogates and species, and will be set to zero for DBF sum
parameters. This modification improves the ability of the logistic function to fit breakthrough
data that are not symmetrical.
Under certain conditions, some brominated DBF species exhibit increasing and decreasing
breakthrough curves. These "peak" curves can be modeled using the logistic function to the
maximum concentration. After this point, effluent concentrations decrease, and this decrease is
modeled using a simple linear function. Prior to the point of peak concentration, C(i) is
described by the step-lag logistic function model. The step-lag-peak logistic function model is as
follows:
C(0 = 0 ttp (20)
where Cp is the logistic function model best-fit concentration at tp, the run time at which the peak
occurs, and S is the slope of the linear best-fit curve. The following algorithm was used to detect
"peak" breakthrough curves:
1. The measured peak concentration, C(tp), is at least 20 percent greater than the concentration
at the last observed data point, C(tf).
2. The run time tp is less than 80 percent of the run time of the last observed data point, tf.
3. The data point corresponding to tp is located prior to the penultimate observed data point.
In all cases, a best-fit of the logistic function model to the data was generated by least squares
minimization approach. The coefficient of determination, R2, was computed for all best-fits.
The logistic function models were fit by non-linear least-squares using PROC NLIN in version
6.12 of the SAS system (Littell et al., 1996). Additional details and SAS code are included in
Appendix B.
3.2.2 Outlier Methods
The large amount of data that will be processed during the ICR treatment studies data analysis
necessitates the use of automated curve fitting procedures. To ensure that the fitted models are
robust to extreme values, an outlier adjustment methodology that uses all experimental data
-25-
-------
points but limits the influence of deviant observations on the parameter estimates was developed
and used for this analysis.
Indiscriminate deletion of deviant observations can cause goodness of fit measures (i.e., R2) to be
unrealistically high. Rather than deleting potential outliers, suspect observations were replaced
by less extreme values by the following procedure:
1. Fit the logistic function model to the observed data and determine approximate 95 percent
prediction limits on the observations.
2. Observations that exceed the threshold Predicted + (U95 - L95)*K are adjusted to Predicted
+ (U95 - L95)*K, where U95 is the upper 95 percent confidence limit and L95 is the lower
95 percent confidence limit. The constant K determines the magnitude of the adjustment,
with larger values of K corresponding to fewer declared outliers and smaller adjustments to
those that are detected. The value for the constant K was set to 1/3 based on simulation
results that indicated a good balance between false alarms and power to identify substantial
outliers (>3 standard deviations from the best-fit prediction). Approximately 2.5 percent of
the data in the present study were identified as outliers and adjusted.
3. Refit the logistic function model using adjusted values.
3.2.3 Direct Integration Approach
The use of the integrated logistic function model for a given parameter to predict the integral
breakthrough curve for that parameter is termed the direct integration (DI) approach. The
average value function described by Equation 12 was applied to the expressions used to describe
the breakthrough curves presented in Section 3.2.1. An expression for the average value of the
step logistic function (Equation 15) is:
(21)
Application of the average value function to the step-lag logistic function model (Equations 16
and 17) yields the following equations that represent the integral breakthrough curve:
C(t) = 0 t 4 is
more complex due to the use of two functions to describe the experimental data. The application
of the average value function to the three functions used to describe the experimental data over
the range [0, t] is given by the following equation:
-26-
-------
C(t) = -\C(t)dt = -
t J t
16 P '
Jq(0^ + Jc2(0^ + jQ
(24)
where Ci(f), C^), and C3(/) represent Equations 18, 19, and 20, respectively, the three equations
used to model the experimental data. For t < tb, the integral breakthrough curve is represented
by:
t< tb
(25)
For tb tp, Equation 19 is evaluated from tb to tp, and Equation 20 is evaluated from tp to t.
Combining the two equations yields:
^
D
•Be~Dt^
+ Be
-Dtb
t>tn
(27)
The logistic function integral approach to determine the integral breakthrough curve assumes an
infinite number of contactors operated in parallel-staggered mode. This assumption was verified
for finite N = 2, 3, 4, 6, 10, and 20 contactors. For finite numbers of contactors, numerical
integration of Equations 15 through 20 was performed as described by Equation 4.
3.2.4 Surrogate Correlation Approach
The surrogate correlation approach (SCA) is a procedure that simplifies and reduces the amount
of computations necessary to estimate DBF formation in the effluent of staggered multiple
contactors (blended effluent). The SCA procedure relies on a constant relationship between
TOC concentration and DBF formation in both the single contactor effluent and the blended
effluent. To apply the SCA method to a GAC run, the single contactor TOC, UV254, SDS-DBP
class sum parameters, and SDS-DBP species breakthrough curves are fit to the appropriate
logistic function model. The SCA procedure steps are summarized in Section 1.5 and
represented schematically in Figure 4.
Typically, a linear or second order polynomial relationship exists between GAC effluent TOC
concentration and any other parameter. The SCA method does not determine or use these
correlations directly but instead assumes they exist and are constant in both the single contactor
effluent and blended effluent. This relationship is exploited by linking the concentration of all
water quality parameters at a common single contactor run time. Using Equation 21, the integral
breakthrough curve is calculated for the TOC single contactor breakthrough curve only. Then
the single contactor TOC concentration, to which the concentrations of all other water quality
-27-
-------
parameters are linked, is applied to the run time at which an equivalent TOC concentration is
reached in the TOC integral breakthrough curve.
The SCA procedure can be performed to determine blended contactor run times for a given GAC
run and blended effluent treatment objective. In this study, the procedure was verified by
comparing the SCA integral breakthrough curve against experimental data.
3.2.5 Comparison of Methods for Predicting the Performance of GAC Contactors
Operated in Parallel-Staggered Mode
During this study, the methods described in Sections 3.2.3 and 3.2.4 for predicting the integral
breakthrough curve were compared to the experimentally derived integral breakthrough curve for
eight GAC runs. Using the logistic function models, a best-fit of the experimental blended
effluent data was derived. The logistic function models were applied to the integral
breakthrough curve data because the blended effluent data curves were similar in shape to those
encountered in the single contactor effluent.
Predictions obtained by the SCA method, the DI method, and the best-fitting model were
compared using the calculated residual sums of squares (RSS) from the experimental blended
effluent data. For each parameter prediction, the prediction bias was determined by averaging
the calculated residuals for each model. An evaluation of the bias shows whether the model
tended to overpredict or underpredict the observed data. For some parameters, a limited amount
of data were measured above the minimum reporting level (MRL). In cases where fewer than
six points above the MRL were present, curve fitting was not performed.
3.2.6 Comparison of Single Contactor and Blended Effluent DBP Bromine
Incorporation
To verify the assumption that the impact of TOC on DBP bromine incorporation would be
similar in the single contactor and blended effluents, the bromine incorporation factors nBr and
n'Br were modeled as a polynomial function of TOC and UV254 concentrations for all waters
simultaneously with two regression models:
Y = Wi + (3iX + |32X2 (28)
Y = Wi + (3iX + p2X2 + yZ, (29)
where Y is nBr or n 'Br and X is TOC or UV254. The intercept W; is allowed to be different for
each water, to reflect natural differences in bromine incorporation across waters. In the second
model, different intercepts are fitted for single contactor (Z=0) and blended (Z=l) data. The
additional parameter y represents the change in average bromine incorporation associated with
blending. Since the first model is a special case of the second with y=0, the assumption that
blending will not impact average bromine incorporation would be supported if the goodness of
fit of the two models is similar. Models which allowed the linear and quadratic parameters (3i
and p2 to differ for single contactor and blending were considered in order to demonstrate that
blended and single contactor scenarios are comparable in terms of the shape of the bromine
incorporation profile as well as mean bromine incorporation (i.e., the intercepts). Although these
-28-
-------
models make efficient use of all of the data simultaneously in order to reach a general
conclusion, the presence of the water-specific random effects W; makes ordinary least-squares
inappropriate. SAS PROC MIXED was therefore used to estimate the models via general
likelihood methods (Littell, et. al., 1996).
3.2.7 Breakthrough Curve Extrapolation
For two waters, the sensitivity of blended effluent water quality to a breakthrough curve
extrapolation procedure was verified. To verify the extrapolation approach, two GAC runs were
operated for five laboratory days beyond 70 percent TOC breakthrough. Based on the scaling
factor of each run, the five days were equivalent to 49 and 69 full-scale days for Waters 5 and 8,
respectively. The appropriate logistic function model was applied to the abbreviated data set,
which reached 70 percent TOC breakthrough. Next, the appropriate logistic function model was
applied to the entire available data set for each parameter. The model fit for the abbreviated data
set was extrapolated to the total column run time, and the blended effluent water quality
predicted based on extrapolation of the abbreviated data set was compared to that using the full
experimental data set.
During the ICR treatment studies, only the TOC integral breakthrough curve will be
extrapolated. The SCA method relies on projecting water quality data associated with single
contactor effluent TOC concentration to the integral TOC breakthrough curve. Since the integral
TOC breakthrough curve typically reaches only 40 to 70 percent of the single contactor
breakthrough curve, DBF data associated with higher TOC concentrations would not be
included. By extrapolating the integral TOC breakthrough curve, the benefit of the data set is
increased, because DBF formation associated with higher GAC effluent TOC concentrations will
be included in the analysis.
3.3 Waters Examined
3.3.1 Pretreatment and Water Quality
GAC runs on eight water sources were included in this study. Pretreatment schematics for each
water source are shown in Appendix C. The influent to GAC TOC concentration for these water
sources ranged from 2.0 to 5.6 mg/L. The specific UV absorbance for TOC (TSUVA, defined as
100*UV254/TOC) ranged from 1.6 to 2.3 L/mg-m. A wide range of bromide concentrations were
measured, ranging from 28 to 300 |ig/L. The GAC influent bromide to TOC ratio (Br:TOC)
ranged from 10 to 68. Water 1, received from Miami, Florida, was a groundwater. Water 2,
received from Aurora, Illinois, was a mixture of groundwater and surface water. The other 6
waters were all surface waters. A summary of the source water, pretreatment, and treated water
quality is shown in Table 3.
-29-
-------
Water
1
2
3
4
5
6
7
8
Water
source
Miami -
Dade
County,
Florida
Aurora,
Illinois
Topeka,
Kansas
Davenport,
Iowa
Escondido,
California
Charleston,
S. Carolina
Sweetwater,
California
Greensboro,
N. Carolina
Pretreatment Treated water quality (GAC influent)
TOC UV254 TSUVA pH Alkalinity Total Bromide
(mg/L) (I/cm) (L/mg-m) (mg/L as hardness (ng/L)
CaCO3) (mg/L as
CaCO3)
Lime- 4.5 0.094 2.1 9.2 23 54 115
softening
Lime- 2.6 0.055 2.1 9.4 58 131 105
softening
Two-stage 2.4 0.048 2.0 9.0 30 133 160
softening
Conventional 3.0 0.065 2.2 7.1 127 217 29
(PAS*)
Conventional 3.1 0.051 1.7 7.4 109 225 70
(alum)
Conventional 2.6 0.060 2.3 6.3 9 29 140
(alum)
Conventional 5.6 0.109 2.0 7.6 138 221 300
(ferric)
Conventional 2.0 0.033 1.6 7.6 23 32 28
(alum)
BrTOC
(Hg/mg)
25
40
68
10
23
53
54
14
* PAS: polyaluminum sulfate
Table 3 Summary of pretreatment and water quality
-------
3.3.2 Simulated Distribution System Chlorination Conditions
Table 4 summarizes the SDS chlorination conditions for each water. The SDS conditions were
chosen to reflect site-specific distribution system conditions at the time the water was sampled.
SDS incubation times ranged from 6 to 48 hours; incubation pH ranged from 7.4 to 9.2; target
free chlorine residual ranged from 0.75 to 1.50 mg/L; incubation temperatures ranged from 18 to
26°C. All samples were buffered to maintain the target incubation pH constant during the
incubation period.
Water
1
2
3
4
5
6
7
8
Date sampled
July 24, 1998
September 22, 1998
September 1, 1998
September 23, 1998
January 7, 1999
October 12, 1998
October 22, 1998
October 6, 1998
Target SDS chlorination conditions
Incubation
time (hrs)
6
24
48
24
24
28
24
24
Incubation
temperature
(°C)
26
20
26
20
18
20
24
20
Free chlorine
residual
(mg/L)
0.75
0.80
0.80
0.75
0.80
1.50
0.75
1.00
pH
9.1
9.1
9.2
7.4
7.4
8.5
8.0
7.7
Table 4 SDS chlorination conditions
3.4 Analytical Methods
A list of all analytical methods and minimum reporting levels (MRLs) used during the study is
shown in Table 5. All analyses were conducted at Summers & Hooper, Inc. or Montgomery
Watson Laboratories as outlined in Table 6.
-31-
-------
Analyte
Alkalinity
Bromide
Calcium hardness
Chlorine dose (solution
standardization)
Chlorine residual
HAA (DCAA, TCAA, MBAA,
DBAA, BCAA, BDCAA)
HAA (MCAA, CDBAA)
HAA (TBAA)
HAA (DCAA, TCAA, MBAA,
DBAA, BCAA, BDCAA)
HAA (MCAA, CDBAA)
HAA (TBAA)
pH
Temperature
Total hardness
Total organic carbon (TOC)
Total organic halide (TOX)
THM (CF, BDCM, DBCM, BF)
UV absorbance at 254 nm (UV254)
SM: Standard Methods
NA: Not applicable
Table 5 Summary of analytical
Analyses performed
Alkalinity, chlorine dose, chlorine
residual, pH, temperature, TTHM,
TOC, TOX, UV254
HAA9
Bromide, calcium hardness, total
hardness
HAA9
Waters
All
All
All
All
All
1-4,6-8
1-4,6-8
1-4,6-8
5
5
5
All
All
All
All
All
All
All
methods
Method
SM 2320 B
EPA 300.0 A
EPA 200.7
SM 4500-C1 B
SM 4500-C1 F
EPA 552.2
EPA 552.2
EPA 552.2
SM6251B
SM6251B
SM6251B
4500-H+ B
SM2550B
SM 2340 B
SM5310C
SM 5320 B
EPA 551.1
SM5910B
and MRLs
Waters
All
1-4,6-8
All
5
Minimum reporting level (MRL)
5 mg/L as CaCO3
20^g/L
5 mg/L as CaCO3
NA
0.2 mg/L as C12
1.0 ng/L (each analyte)
2.0 ng/L (each analyte)
4.0 ng/L
1.0 (ig/L (each analyte)
2.0 ng/L (each analyte)
4.0 ng/L
NA
NA
5 mg/L as CaCO3
0.50 mg/L
25 ng/L as Cl"
1.0 ng/L (each analyte)
0.009 cm"1
Laboratory
Summers & Hooper, Inc.
Summers & Hooper, Inc.
Montgomery Watson Laboratories
Montgomery Watson Laboratories
Table 6 Summary of laboratories conducting analyses
3.5 Experimental QA/QC Summary
As a part of this study, field duplicates were performed on 10 percent of samples analyzed in the
blended effluent. The field duplicates for DBFs were generated by duplicate SDS chlorination of
split GAC effluent samples. The single contactor effluent samples were duplicated at a higher
rate, following ICR guidelines, which required that three field duplicates be collected from the
-32-
-------
effluent of each RSSCT (25 percent duplication). The results of all field duplicate analyses are
summarized in Table 7.
-33-
-------
Analyte
Count
Relative percent difference (RB)
50th
percentile
fifen
Standard
deviation
(DC
\A
BSDX
B3NA5
B3NA6
B3NA9
3
2
9
6
2
5
5
THM Species
BSDF
BSDCM
BSDBCM
2
2
HAA Species
BSGIAA
BSDCAA
BSCAA
8
0
NA
2 NA
8
BS)BAA
BSCAA
BXDBAA
BS)CBAA
BSAA
2
2
2
2
4
S
B
4
Q
8
S
01
S
%
Q
8
2
B
3
2
B: relative percent difference
NA: not applicate
Table 7 Summary of field duplicate precision for single contactor and blended effluent data
-34-
-------
4 Results and Discussion
4.1 Overview
A substantial portion of the results presented attempt to verify the underlying assumptions
behind the surrogate correlation approach (SCA) and examine the limitations of the application
of this method for determining the integral breakthrough curve, a relationship between single
contactor run time and blended contactor effluent water quality assuming contactor operation in
parallel-staggered mode. The SCA procedure requires that all GAC breakthrough data are fit to
a logistic function model. From the perspective of data analysis of the ICR GAC treatment study
data, there are several advantages to using a model to fit experimental data. Best-fit curve
parameters that adequately describe experimental data are less memory intensive than storing the
entire experimental data set. A best-fit curve also facilitates interpolation and extrapolation to
estimate run times for given treatment objectives. Use of a best-fit model curve provides an
estimate of the scatter in the data through the coefficient of determination, and the model
minimizes the impact of this scatter on run time estimates. Finally, a function that describes the
single contactor experimental data set is a prerequisite for determining the integral breakthrough
curve. Run time estimates generated by the integral breakthrough curve are more applicable to
full-scale GAC operation where multiple contactors are operated in parallel-staggered mode to
increase the service time of each individual contactor.
The SCA procedure (1) links single contactor TOC concentration to single contactor DBF
formation at a given run time (RTSc), (2) uses the direct integration (DI) approach to model the
integral TOC breakthrough curve, (3) determines the blended contactor effluent run time (RTBc)
at the TOC concentration used, and (4) applies the linked parameters to estimate DBF formation
at RTBc- This approach allows the blended contactor run time to be determined for all correlated
parameters based on the TOC integral breakthrough curve. The SCA procedure assumes that:
1. The correlation between TOC and DBF precursors (concentration and speciation) observed
in the single contactor effluent is maintained in the blended effluent. A discussion of
comparisons made to evaluate the consistency of these correlations is presented in Section
4.2.
2. The three forms of the logistic function model developed can accurately describe the range of
breakthrough of DBF precursors and DBF precursor surrogates evaluated. The results of
logistic function model fits applied to all parameters and all GAC runs are discussed in
Section 4.3.
3. The TOC integral breakthrough curve calculated by the DI approach is an accurate predictor
of blended effluent TOC concentration when multiple contactors are operated in parallel-
staggered mode. TOC integral breakthrough curves are compared to experimental data in
Section 4.3, while an evaluation of the applicability of the infinite contactor assumption
inherent by application of the DI procedure is contained in Section 4.5.
4. The integral TOC breakthrough curve can be extrapolated with reasonable results. This
assumption was verified for two GAC runs, as discussed in Section 4.6.
-35-
-------
Section 4.4 compares the integral breakthrough curves obtained by both the SCA and DI
methods against experimental data for the eight GAC runs and all measured parameters.
-36-
-------
4.2 Correlation between Surrogates and DBFs in Single Contactor and Blended
Contactor Effluents
4.2.1 Correlation between Surrogate Concentration and DBP Formation
The correlation assumption is an important underlying foundation for the SCA procedure used to
model blended effluent water quality. Verifying the assumption would demonstrate that DBP
formation is constant at a given TOC concentration in both the single contactor and blended
contactor effluent. Since bromide is not adsorbed by GAC, the bromide to TOC ratio will also
be constant, at a given TOC concentration, so DBP speciation should not change between single
contactor and blended contactor effluents.
To verify the correlation assumption, plots of paired data were generated between TOC and other
parameters, such as SDS-TTHM, SDS-HAA6, SDS-CF, SDS-DCAA, etc. Both summed DBP
classes and individual compounds were examined. For each water, two experimental data sets
were plotted for comparison: single contactor effluent and blended effluent. The entire graph set
is included in Appendix D, which includes correlations based on both TOC and UV254 as
surrogates. Overall, both parameters served well as surrogates for DBP formation, and the
correlation between single contactor and blended effluent data was approximately equivalent. A
representative sample of the results obtained are discussed in this section.
Overall, the correlations observed between TOC and formed DBFs in the blended effluent were
very similar to those observed in the single contactors for each water. Figures 6 and 7 show the
correlation of UV254, SDS-TTHM, SDS-HAA6, and SDS-TOX to TOC for both the single
contactor and blended effluent of Waters 1 and 2 (with the exception of SDS-TOX, for which
blended effluent samples were not analyzed in these waters). The correlation between TOC and
UV254 in the blended effluent was very similar to that in the single contactor effluent for these
two waters, which was typical of that observed in most cases. The greatest difference in blended
effluent and single contactor correlations for TOC and SDS-TTFDVI occurred for Water 8 (Figure
8) which showed a similar disparity between the blended effluent and single contactor
correlations for TOC and UV254. Still, at a given TOC concentration, SDS-TTHM formation in
the blended effluent was within 5 |ig/L of that in the single contactor effluent.
For the THM species, correlations between single contactor and blended effluent TOC and
formed concentrations are shown in Figures 9 and 10 for Waters 2 and 3, respectively. For all
four species, a very good agreement existed between single contactor and blended effluent TOC
correlations. This includes SDS-BF, which yielded a "peak" curve for these two runs. The peak
also occurred in the blended effluent, and the single contactor and blended effluent yielded very
similar curves when SDS-BF concentrations were plotted against TOC. The correlation of
formed THM compounds between single contactor and blended effluent was very good for all
waters. The largest difference between correlations seen in the single contactor and blended
effluent were observed for Water 8, shown in Figure 11, for which the formed THM species
concentrations were low.
Four of the predominant HAA species formed (DCAA, TCAA, DBAA, and BCAA) correlated to
TOC concentration for both single contactor and blended effluents are shown in Figures 12 and
-37-
-------
13 for Waters 5 and 7, respectively. For Water 5 both single contactor and blended effluent
formed HAAs matched very well. For Water 7, the correlations for SDS-DBAA and SDS-
BCAA also matched very well. The single contactor and blended effluent correlations between
TOC and SDS-DCAA and SDS-TCAA showed slight differences, not exceeding 3 ug/L. The
formed levels of these two compounds were low for this water.
Correlations between single contactor and blended effluent for SDS-DCBAA, SDS-CDBAA,
SDS-HAA5, and SDS-HAA9 are shown for Waters 2 and 4 in Figures 14 and 15, respectively.
Water 2 is shown as it yielded the largest difference in the correlation between single contactor
and blended effluent for these parameters, although formed DCBAA and CDBAA levels were
low. Water 4 is more representative of what was typically observed, with the correlation
between TOC and these parameters in the blended effluent matching that observed in the single
contactor very well. For SDS-TBAA, only Water 7 yielded measured levels above the MRL (4
ug/L). The correlation between TOC and formed TBAA in the blended effluent agreed well
with that observed in the single contactor effluent, as shown in Figure 16.
The use of UV254 as a surrogate for these correlations instead of TOC was investigated. In a few
cases, UV254 as a surrogate yielded better correlations between single contactor and blended
effluent formed DBFs than did TOC. However, in other instances, and approximately the same
number of cases, the correlations using TOC were superior. Because the results for the
correlations using UV254 did not improve over those using TOC, the use of TOC as the surrogate
in the SCA procedure was continued. Appendix D also summarizes the correlations observed
based on UV254.
In summary, this analysis shows that the correlation between DBF formation and TOC
concentration in the single contactor and blended contactor effluents is constant. This find is
significant because it verifies one of the underlying assumptions behind the SCA procedure, that
TOC concentration can be correlated to DBF formation in the single contactor effluent, and that
this correlation can then be applied to blended contactor effluent. Section 4.3 will address how
single contactor effluent data can be modeled as a function of run time so that this correlation
can be utilized to predict blended contactor effluent DBF formation.
4.2.2 Correlation between Surrogates and DBP Speciation
For all four models relating bromine incorporation for THMs and HAA9 (nBr and n'Br,
respectively) to TOC and UV254, the polynomial models described in Section 3.2.6 were found to
be appropriate. The model variance over all four combinations of nBr and n'Br as a function of
TOC and UV254 yielded R2 values greater than 0.92 in each case. Although the effect (y) of
blending on mean bromine incorporation was found to be statistically significant, including this
effect improved R2 values trivially (less than 1 percent for all four cases) by. There was no
evidence that different linear or quadratic terms were needed to describe the single contactor and
blended effluent data. For THMs, Figures 17 through 24 show the results of fitting second order
polynomials to the relationship between nBr and TOC, combining single and blended contactor
data for each water. For HAA9, similar results for the relationship between n'Br and TOC are
shown in Figures 25 through 32. The second order polynomial curve fits for the correlation
between nBr and UV254 and n'Br and UV254 for both single and blended contactor data are shown
in Figures 33 through 48.
-38-
-------
This analysis shows that there is no significant difference between DBF bromine incorporation
between single contactor and blended contactor effluents at equivalent TOC or UV254 values.
This conclusion supports the use of the SCA procedure, especially for individual DBF species,
because the SCA assumes that DBF formation and speciation in the single contactor effluent at a
given TOC concentration is equivalent to that in the blended contactor effluent at the same TOC
concentration.
-39-
-------
0.07 -,
0.06 •
0.05 •
"E
^ 0.04 •
- — -
a
-------
0.04 -i
0.03 •
o
°-02-
0.01 •
0.00
n Single contactor
• Blended effluent
n
0.0
0.5
1.0
TOC (mg/L)
1.5
2.0
100 -i
75-
50-
w
25'
0.0
0.5 1.0
TOC (mg/L)
1.5
2.0
30
20 •
w
Q
w
0.0
0.5
1.0
TOC (mg/L)
1.5
2.0
200 -
150 -
g
w
100 -
50-
0.0
0.5 1.0 1.5 2.0
TOC (mg/L)
Figure 7 Correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 2
-------
0.03 n
0.02 •
a
>
0.01 •
0.00
n Single contactor
• Blended effluent
0.0
0.5 1.0
TOC (mg/L)
1.5 2.0
50 n
25-
H
o.O
0.5 1.0
TOC (mg/L)
1.5 2.0
to
20 n
10-
OT
D
OT
-•-•-[•-•-
0.0 0.5
1.0 1.5
TOC (mg/L)
2.0
150 n
O 100 •
8
0.0
0.5 1.0 1.5
TOC (mg/L)
2.0
Figure 8 Correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 8
-------
25 n
20 •
0) 15 •
LL
o
10-
w
5 •
0.0
n Single contactor effluent
• Blended effluent
0.5 1.0
TOC (mg/L)
1.5
2.0
30 n
25 '
20
o 15 •
Q
m
w
Q 10 •
0.0
0.5
1.0
TOC (mg/L)
1.5
2.0
30
25 •
0 15 •
m
9
w
Q 10 •
w
5 •
0.0
0.5 1.0
TOC (mg/L)
1.5
2.0
15 -i
10-
5 -
0.0
0.5 1.0
TOC (mg/L)
1.5
2.0
Figure 9 THM correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 2
-------
20 i
15 •
1
LL 10-
Q
5 •
0 •
0
40 -1
n Single contactor effluent
35 •
• Blended effluent
n 30 .
n ^
n S25'
0 20 •
D
in
OT 15 •
,_, Q
I I £f)
10 •
•D
-• — B-rQBD — i i 1 i 1 i 1 0
fl
n
n
^°"
•
= •
•
n
0 0.5 1.0 1.5 2.0 o.O 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
30 -I
25 •
a
0 15 •
m
p
OT
D 10 •
OT
5 •
0 •
35 -,
B 30 •
D
25 •
5
LL.
D CO
• § 15'
B n w •
10 ;
BH _
o •
-• . 1 . 1 . 1 . 1 n
• „ fn
D tr
n
'
0.0
0.5 1.0 1.5
TOC (mg/L)
2.0
0.0
0.5 1.0
TOC (mg/L)
1.5 2.0
Figure 10 THM correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 3
-------
25 •
20 •
SDS-CF (pg/l
O Ol
5 •
0 •
0
s •
n Single contactor effluent n
• Blended effluent 4 .
n
=d
n ^ 3 •
o
. * OT1"
. n g
•
• D
n ° n n D n
• n
• n
n
0 0.5 1.0 1.5 2.0 o.O 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
15 -I
5 10-
O
m
D
OT
D 5 •
OT
0 •
0
2 -i
n „
• D)
— i
D LT 1 .
• n CQ
• D OT
n OT
0
• D
1
'
-1 • • !_•"• !_• — • L«-|_l rl_| !• r|_| 1_| — LJ LJ — ' 1
0.5 1.0 1.5 2.0 o.O 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
Figure 11 THM correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 8
-------
1
6 •
_i
s
J
o
Q
ch
Q
w
2 •
n .
i •
n Single contactor effluent
6 •
• Blended effluent
— i
n S
r;
o
• h- 3 •
n n w
i ^
• W 2 .
1
Q
• ' ™"° 1
-• — j~i n , . . , . . 1 1 1 1 n .
n
n
n
n
n
n
_ •
g
D
•n n
-• — .• — m~\ — rm . . . . . . . . .
0.0 0.5
1.0 1.5 2.0
TOC (mg/L)
2.5 3.0
0.0 0.5 1.0 1.5 2.0 2.5
TOC (mg/L)
3.0
ON
4 •
3 •
1 •
m
Q
i
w
Q
W
1 •
n .
H •
D°D
••• "a
f a '
•n § 4'
_ DQ
• i
• w
D Q
W
2 •
D
-• — JM — • , , , , , , , , , , n .
D
D
n n
•H"
^^
•D
rf
n
-• r« • . . . . , , , , , ,
0.0 0.5 1.0 1.5 2.0 2.5
TOC (mg/L)
3.0 0.0 0.5
1.0 1.5 2.0
TOC (mg/L)
2.5 3.0
Figure 12 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 5
-------
• 0
OT
5
E
n
n
n n
n n
• "
D
i1 "°
• . •, — . — . — . — . — . — . — . — . — . — . — . — . — . — . — . — .
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
TOC (mg/L)
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
TOC (mg/L)
Figure 13 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 7
-------
oo
3 -
„ .
0) 2-
<:
m
0
Q
W ,.
Q 1 '
W
n Single contactor effluent
• Blended effluent n
n
• • D —
™ n _i
0)2-
• n D <:
• n m '
D D
O
D £/)
D 1 '
OT
^
•
n n
_
Oi — — u — u u .,.,., u -, m — m |_j H _j |_|-«i LB rl_l— LJ U LJ-i 1 1
0.0 0.5 1.0 1.5 2.0 o.O 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
20 -, 30 -,
15 -
5"
01
in
<, 10-
I
>
D
OT
5 -
0 -
n
25
n
^20
n n o)
• -5
• n en
< 15
• ~r
,
n co
§10
•
5
^H
••
n
n
• D
n
i
n
1
n
• n
• D
• D
D
u -t ™ • . - . - . - . u i m , 1 , 1 , 1 , 1
0.0 0.5 1.0 1.5 2.0 o.O 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
Figure 14 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 2
-------
6 -I
"B) 4 •
^.
DQ
O
Q
ih 0
Q z '
W
0 -
0
n Single contactor effluent D
• Blended effluent
n n 5"
0)
§
n °m w
n Q
• n" • w
• •HI M •!—!•• 1— 1 1— • 1— 1 1— 1 1— 1 1— 1 1— 1
0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 2.5
TOC (mg/L) TOC (mg/L)
40 -i
30 -
LO
<• 20-
X
ch
Q
w
10 •
0 •
0
50
n
40
D 2 30
D en
X 20
n Q
• D w
• D 10
• " n
. D
n
DD
D
• D
D °
m B n"
--a
0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 2.5
TOC (mg/L) TOC (mg/L)
Figure 15 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 4
-------
12 -i
10 •
^
0) 8 •
^
^ c
CO D •
O
9
w
Q 4 •
w
2 •
0 •
n Single contactor effluent n
• Blended effluent
IT
d "ra
^
n <
D
• n V
n M
• 0
n OT
•n
_D
t] "
-B-~ — •, •! , . . . . . . . . . . . . . .
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
TOC (mg/L)
8 -i
6 •
"en
?4-
CQ
1
1
OT
D
OT
2 •
0 •
7 -,
6 •
5 •
4 •
3 •
2 •
1 .
n n
n
n
D n
u
m
a
a
a
I • ' — VI MJ !• — ' • 1 ' 1 ' 1 ' 1 ' 1 ' 1 ' 1
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
TOC (mg/L)
80 i
D
D
• •
n
1
° 1
*t
OT
D
OT
D
• . •. — n — .• . • . — . — , — . — , — . — , — . — , — . — , — . — ,
70
60
50
40
30
20
10
n
a
a
n
D n
fm
•n
rj
D •
•
f
m I
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
TOC (mg/L)
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
TOC (mg/L)
Figure 16 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 7
-------
2.5
C/5
^
I
2.0 -
1.5 -
o
Q_
O
O
CD
1.0 -
•| 0.5 -
o
m
0.0
Water 1
TOC: 4.5 mg/L
Bromide: 115 ug/L
D Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.948)
TOC concentration (mg/L)
Figure 17 Correlation between single contactor and blended effluent TOC
concentration and THM bromine incorporation factor (n) for Water 1
3.0
2.5 -
2.0 -
o
CD
1.5 -
o
Q_
O
o
1.0 -
CD
g
E
2
oo
0.5 -
0.0
Water 2
TOC: 2.6 mg/L
Bromide: 105 ug/L
0.0
n Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.772)
0.5 1.0
TOC concentration (mg/L)
1.5
2.0
Figure 18 Correlation between single contactor and blended effluent TOC
concentration and THM bromine incorporation factor (n) for Water 2
-51-
-------
CO
^
I
3.0
2.5 -
2.0 -
o
CD
1.5 -
o
Q_
O
O
1.0 -
CD
c
E
o
m
0.5 -
0.0
Water 3
TOC: 2.4 mg/L
Bromide: 160 pg/L
0.0
D Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.978)
0.5 1.0
TOC concentration (mg/L)
1.5
2.0
Figure 19 Correlation between single contactor and blended effluent TOC
concentration and THM bromine incorporation factor (n) for Water 3
1.00
to
^
I
0.75 -
o
CD
o 0.50 -
"CD
o
Q_
o
o
o> 0.
g
E
£
oo
.25 -
0.00
Water 4
TOC: 3 mg/L
Bromide: 29 ug/L
0.0
0.5
D Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.874)
1.0 1.5
TOC concentration (mg/L)
2.0
2.5
Figure 20 Correlation between single contactor and blended effluent TOC
concentration and THM bromine incorporation factor (n) for Water 4
-52-
-------
2.0
I- 1.5 -
(_
£
t_
o
£
o 1.0 -
"H—'
E
o
e-
8
c
CD
C
E
2
GO
0.5 -
0.0
Water 5
TOC: 3.1 mg/L
Bromide: 70 ug/L
0.0
0.5
n Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.406)
1.0 1.5 2.0
TOC concentration (mg/L)
2.5
3.0
Figure 21 Correlation between single contactor and blended effluent TOC
concentration and THM bromine incorporation factor (n) for Water 5
2.5
to
1 2.0 H
£
i_
o
1.5 -
o 1.0-1
8
CD
00
o.o
Water 6
TOC: 2.6 mg/L
Bromide: 140 ug/L
0.0
0.5
n Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.781)
1.0 1.5
TOC concentration (mg/L)
2.0
2.5
Figure 22 Correlation between single contactor and blended effluent TOC
concentration and THM bromine incorporation factor (n) for Water 6
-53-
-------
2.5
2.0 -
£
t_
o
1.5 -
E
o 1.0 H
8
O>
| 0.5 H
2
GO
0.0
TOC: 5.6 mg/L
Bromide: 300 ug/L
D Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.981)
1
2 3
TOC concentration (mg/L)
Figure 23 Correlation between single contactor and blended effluent TOC
concentration and THM bromine incorporation factor (n) for Water 7
1.00
0.75 -
£
i_
o
o 0.50 -
'-4—I
O
e-
8
'^ 0.25 -
I
2
00
0.00
Water 8
TOC: 2 mg/L
Bromide: 28 ug/L
D Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.896)
0.0
0.5 1.0
TOC concentration (mg/L)
1.5
2.0
Figure 24 Correlation between single contactor and blended effluent TOC
concentration and THM bromine incorporation factor (n) for Water 8
-54-
-------
2.5
o>
< 2.0 -I
I
•§ 1.5 -
o
"co
_
o
o
o>
•| 0.5 -\
£
00
0.0
Water 1
TOC: 4.5 mg/L
Bromide: 115 ug/L
D Single contactor effluent
• Blended effluent
— Best fit (RA2 = 0.751)
0123
TOC concentration (mg/L)
Figure 25 Correlation between single contactor and blended effluent TOC
concentration and HAA9 bromine incorporation factor (n1) for Water 1
3.0
< 2-5 •
X
! 2-0 H
o
•s
§ 1.5 -
o
Q.
O
O
1.0 -
o>
g
E
o
m
0.5 -
0.0
Water 2
TOC: 2.6 mg/L
Bromide: 105 ug/L
0.0
D Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.423)
0.5 1.0
TOC concentration (mg/L)
1.5
2.0
Figure 26 Correlation between single contactor and blended effluent TOC
concentration and HAA9 bromine incorporation factor (n1) for Water 2
-55-
-------
3.0
2.5 -
t: 2.0 -
o
•s
o
"co
t_
o
Q.
O
O
g
o>
c
E
£
m
1.5 -
1.0 -
0.5 -
0.0
Water 3
TOC: 2.4 mg/L
Bromide: 160 ug/L
0.0
D Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.900)
0.5 1.0
TOC concentration (mg/L)
1.5
2.0
Figure 27 Correlation between single contactor and blended effluent TOC
concentration and HAA9 bromine incorporation factor (n1) for Water 3
1.25
o>
•§ 0.75 -
0.50 H
_
o
o
O)
•| 0.25 H
o
m
0.00
Water 4
D Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.504)
TOC: 3 mg/L
Bromide: 29 ug/L
0.0
0.5
1.0 1.5
TOC concentration (mg/L)
2.0
2.5
Figure 28 Correlation between single contactor and blended effluent TOC
concentration and HAA9 bromine incorporation factor (n1) for Water 4
-56-
-------
1.5
o>
<
<
o
"G
g
"co
t_
o
Q.
8
g
o>
g
E
O
(_
m
o.o
Water 5
TOC: 3.1 mg/L
Bromide: 70 ug/L
n Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.283)
0.0
0.5
1.0 1.5 2.0
TOC concentration (mg/L)
2.5
3.0
Figure 29 Correlation between single contactor and blended effluent TOC concentratioi
bromine incorporation factor (n1) for Water 5
2.5
2.0 -
a
o
1.5 -
o 1.0 -
e-
o
o
0.5 -
E
2
m
0.0
Water 6
Co
TOC: 2.6 mg/L
Bromide: 140 ug/L
n Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.926)
0.0
0.5
1.0 1.5
TOC concentration (mg/L)
2.0
2.5
Figure 30 Correlation between single contactor and blended effluent TOC concentratioi
bromine incorporation factor (n1) for Water 6
-57-
-------
2.5
o>
3 2.0
I
o
•Q 1.5
o
'-4—'
CD
o 1.0
t_
o
o
o>
I 0.5
m
0.0
Water 7
TOC: 5.6 mg/L
Bromide: 300 ug/L
D Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.422)
1
2 3
TOC concentration (mg/L)
Figure 31 Correlation between single contactor and blended effluent TOC concentratioi
bromine incorporation factor (n1) for Water 7
0.5
o>
<
<
o
Q_
O
O
o>
0.4 -
0.3 -
0.2 -
.S 0.1 -
E
£
00
0.0
0.0
Water 8
TOC: 2 mg/L
Bromide: 28 ug/L
D Single contactor effluent
• Blended effluent
— Best fit (RA2 = 0.891)
0.5 1.0
TOC concentration (mg/L)
1.5
2.0
Figure 32 Correlation between single contactor and blended effluent TOC concentratioi
bromine incorporation factor (n1) for Water 8
-58-
-------
2.5
CO
^
I
2.0 -
1.5 -
o
Q_
O
O
CD
1.0 -
•| 0.5 -
o
m
0.0
UV-254: 0.094 1/cm
Bromide: 115 ug/L
D Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.962)
0.00 0.01 0.02 0.03 0.04 0.05
UV absorbance at 254 nm, UV-254 (1/cm)
0.06
0.07
Figure 33 Correlation between single contactor and blended effluent UV
absorbance and THM bromine incorporation factor (n) for Water 1
3.0
2.5 -
I
t_
o
2.0 -
o
CD
5 1.5 -
CD
o
Q_
1.0 -
f 0.6H
00
0.0
Water 2
UV-254: 0.055 1/cm
Bromide: 105 ug/L
n Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.781)
0.00
0.01 0.02 0.03
UV absorbance at 254 nm, UV-254 (1/cm)
0.04
Figure 34 Correlation between single contactor and blended effluent UV
absorbance and THM bromine incorporation factor (n) for Water 2
-59-
-------
CO
^
I
3.0
2.5 -
2.0 -
o
CD
1.5 -
o
Q_
O
O
1.0 -
CD
c
E
o
m
0.5 -
0.0
Water 3
UV-254: 0.048 1/cm
Bromide: 160 pg/L
D Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.974)
0.00
0.01 0.02
UV absorbance at 254 nm, UV-254 (1/cm)
0.03
Figure 35 Correlation between single contactor and blended effluent UV
absorbance and THM bromine incorporation factor (n) for Water 3
1.00
to
^
I
0.75 -
o
CD
o
"CD
o
o.
o
o
c
CD
I
m
0.50 -
0.25 -
0.00
Water 4
UV-254: 0.065 1/cm
Bromide: 29 ug/L
n Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.900)
0.00
0.01 0.02 0.03
UV absorbance at 254 nm, UV-254 (1/cm)
0.04
Figure 36 Correlation between single contactor and blended effluent UV
absorbance and THM bromine incorporation factor (n) for Water 4
-60-
-------
2.0
I- 1.5 -
(_
£
t_
o
£
o 1.0 -
"H—'
E
o
e-
8
g
'CD 0.5 -
E
m
0.0
Water 5
UV-254: 0.051 1/cm
Bromide: 70 ug/L
0.00
D Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.410)
0.01 0.02 0.03
UV absorbance at 254 nm, UV-254 (1/cm)
0.04
Figure 37 Correlation between single contactor and blended effluent UV
absorbance and THM bromine incorporation factor (n) for Water 5
2.5
to
^
I
1_
£
i_
o
2.0 -
1.5 -
o
e-
8
c
CD
1.0 -
00
o.o
Water 6
UV-254: 0.06 1/cm
Bromide: 140 ug/L
0.00
n Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.783)
0.01 0.02 0.03
UV absorbance at 254 nm, UV-254 (1/cm)
0.04
Figure 38 Correlation between single contactor and blended effluent UV
absorbance and THM bromine incorporation factor (n) for Water 6
-61-
-------
2.5
2.0 -
£
t_
o
1.5 -
E
o 1.0-1
8
c
o>
•| 0.5 H
2
m
o.o
UV-254: 0.109 1/cm
Bromide: 300 ug/L
D Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.982)
0.00 0.01 0.02 0.03 0.04 0.05
UV absorbance at 254 nm, UV-254 (1/cm)
0.06
0.07
Figure 39 Correlation between single contactor and blended effluent UV
absorbance and THM bromine incorporation factor (n) for Water 7
1.00
0.75 -
£
i_
o
5 0.50 -
o
e-
8
O)
g
E
£
00
0.25 -
0.00
Water 8
UV-254: 0.033 1/cm
Bromide: 28 ug/L
n Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.880)
0.00
0.01 0.02
UV absorbance at 254 nm, UV-254 (1/cm)
0.03
Figure 40 Correlation between single contactor and blended effluent UV
absorbance and THM bromine incorporation factor (n) for Water 8
-62-
-------
2.5
o>
< 2.0 -I
I
•§ 1.5 -
o
"co
o 1.0 H
l_
o
o
g
O>
•| 0.5 -|
£
m
o.o
Water 1
UV-254: 0.094 1/cm
Bromide: 115 ug/L
D Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.746)
0.00 0.01 0.02 0.03 0.04 0.05
UV absorbance at 254 nm, UV-254 (1/cm)
0.06
0.07
Figure 41 Correlation between single contactor and blended effluent UV
absorbance and HAA9 bromine incorporation factor (n1) for Water 1
2.5
o>
X
1_
£
&_
o
"G
o
Q.
O
O
o>
2.0 -
1.5 -
1.0 -
I °-5 ^
o
m
0.0
Water 2
UV-254: 0.055 1/cm
Bromide: 105 ug/L
D Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.435)
0.00
0.01 0.02 0.03
UV absorbance at 254 nm, UV-254 (1/cm)
0.04
Figure 42 Correlation between single contactor and blended effluent UV
absorbance and HAAS bromine incorporation factor (n') for Water 2
-63-
-------
3.0
2.5 -
2.0
o
•s
1.5 -
o
Q.
8 1-0
O>
I 0.5
00
0.0
Water 3
UV-254: 0.048 1/cm
Bromide: 160 pg/L
0.00
D Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.906)
0.01 0.02
UV absorbance at 254 nm, UV-254 (1/cm)
0.03
Figure 43 Correlation between single contactor and blended effluent UV
absorbance and HAA9 bromine incorporation factor (n1) for Water 3
1.25
o>
I
1_
£
t_
o
•s
o
Q.
O
O
o>
1.00 -
0.75 -
0.50 -
~ 0.25 H
o
m
0.00
Water 4
c0l
UV-254: 0.065 1/cm
Bromide: 29 ug/L
D Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.484)
0.00
0.01 0.02 0.03
UV absorbance at 254 nm, UV-254 (1/cm)
0.04
Figure 44 Correlation between single contactor and blended effluent UV
absorbance and HAAS bromine incorporation factor (n') for Water 4
-64-
-------
1.5
o>
<
<
t 1.0-
o
•s
o
'-4—'
CD
t_
O
Q.
O
O
g
CD
g
E
2
m
0.5 -
0.0
Water 5
D
UV-254: 0.051 1/cm
Bromide: 70 |jg/L
n Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.269)
0.00
0.01 0.02 0.03
UV absorbance at 254 nm, UV-254 (1/cm)
0.04
Figure 45 Correlation between single contactor and blended effluent UV
absorbance and HAA9 bromine incorporation factor (n1) for Water 5
2.5
en
2.0 -
1.5 -
o
Q.
O
o
CD
1.0 -
.E 0.5 H
o
m
0.0
Water 6
-e-
UV-254: 0.06 1/cm
Bromide: 140 ug/L
n Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.935)
0.00
0.01 0.02 0.03
UV absorbance at 254 nm, UV-254 (1/cm)
0.04
Figure 46 Correlation between single contactor and blended effluent UV
absorbance and HAA9 bromine incorporation factor (n1) for Water 6
-65-
-------
2.5
o>
2.0 -I
_
•§ 1.5
|1.0H
8
O>
i 0.5 -
GO
0.0
Water 7
UV-254: 0.109 1/cm
Bromide: 300 pg/L
n Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.410)
0.00 0.01 0.02 0.03 0.04 0.05
UV absorbance at 254 nm, UV-254 (1/cm)
0.06
0.07
Figure 47 Correlation between single contactor and blended effluent UV
absorbance and HAA9 bromine incorporation factor (n1) for Water 7
0.5
c
o>~
I
i_
.o
0.4 -
o
"G 0.3 H
o 0.2-I
t_
o
o
o>
I °-1 H
2
CO
0.0
UV-254: 0.033 1/cm
Bromide: 28 ug/L
0.00
n Single contactor effluent
• Blended effluent
Best fit (RA2 = 0.918)
0.01 0.02
UV absorbance at 254 nm, UV-254 (1/cm)
0.03
Figure 48 Correlation between single contactor and blended effluent UV
absorbance and HAA9 bromine incorporation factor (n1) for Water 8
-66-
-------
4.3 Assessment of Logistic Function Fit to Single Contactor Breakthrough
Curve Data
A model used to describe single contactor effluent experimental data is needed for several
reasons. From a data management perspective, best-fit curve parameters that adequately
describe experimental data are less memory intensive than storing the entire experimental data
set. A best-fit curve also facilitates interpolation and extrapolation to estimate run times for
given treatment objectives. Use of a best-fit model curve provides an estimate of the scatter in
the data through the coefficient of determination, and the model minimizes the impact of this
scatter on run time estimates. Finally, a function that describes the single contactor experimental
data set is a prerequisite for determining the integral breakthrough curve, a curve that relates
single contactor run time to blended contactor water quality under the assumption that contactors
are operated in parallel-staggered mode. Run time estimates generated by the integral
breakthrough curve are more applicable to full-scale GAC operation.
The appropriate logistic curve function model (step, step-lag, or step-lag-peak, as described in
Section 3.2.1) was fit to the single contactor effluent data for eight GAC runs comprised of up to
20 parameters each. The GAC runs were performed using conventionally-treated water from
eight waters sources, including surface and ground waters. The parameters examined included
precursors or surrogates (TOC, UV254, and SDS-TOX), DBF class sums (TTHM, HAAS, HAA6,
and HAA9), and DBF species (CF, BDCM, DBCM, BF, MCAA, DCAA, TCAA, MBAA,
DBAA, BCAA, DCBAA, CDBAA, and TBAA). For some DBF species, samples taken during
the initial portion of the breakthrough curve were measured below the MRL. Curve fits were not
performed if fewer than six effluent data points were reported above the MRL.
Due to the large amount of graphs generated, all plots are summarized in Appendix E, and a
selection of plots are included together with this analysis as examples. The plots contain the
single contactor GAC effluent parameter concentration plotted against scaled operation time. A
line representing the logistic function model best-fit is included. In addition, the experimental
blended effluent data points are included in each plot, along with a dotted line representing the
DI method prediction of the integral breakthrough curve, to demonstrate the benefit obtained by
blending the effluents of multiple contactors operated in parallel staggered mode. An analysis of
the integral breakthrough curve predictive models is deferred until Section 4.4.
4.3.1 Surrogates and Class Sum Logistic Function Curve Fits
The step logistic function model was used to fit single contactor effluent TOC breakthrough
curves for all waters, as shown in Figures 49 through 56. GAC run times ranged from 76 to 287
days, and the GAC influent TOC concentration ranged from 2.0 to 5.6 mg/L. For these waters,
the step logistic function model provided excellent curve fit approximations: the R2 values for
the curve fits ranged from 0.966 to 0.992.
Using the step-lag logistic function model, excellent curve fits were also obtained for single
contactor effluent UV254 breakthrough curves: the R2 values for the curve fits ranged from 0.982
to 0.998. The measured GAC influent UV254 ranged from 0.033 to 0.109 I/cm for the waters
-67-
-------
examined. The results for Waters 5 and 7 are shown in Figures 57 and 58. The results for the
remaining waters can be found in Appendix E.
SDS-TOX breakthrough curves were also modeled using the step-lag logistic function with
excellent results: the R2 values for the curve fits ranged from 0.990 to 0.999. The GAC influent
SDS-TOX ranged from 156 to 486 |ig/L as Cl". Figures 59 and 60 show examples of the data
obtained for Waters 4 and 8.
Both SDS-TTHM and the SDS-HAA sums (HAAS, HAA6, and HAA9) were modeled using the
step-lag logistic function model. Again, single contactor effluent data were well-represented by
the model used, as shown for SDS-TTHM in Figures 61 and 62 for Waters 3 and 6. The R2
values for SDS-TTHM curve fits ranged from 0.977 to 0.995. Overall, the step-lag logistic
function model was also successful when used to fit all three SDS-HAA species sums, with R2
values ranging from 0.952 to 0.994. Figures 63 and 64 show the step-lag logistic function model
curve fits applied to single contactor effluent SDS-HAA9 data for Waters 1 and 2.
Table 8 summarizes the R2 values measured for all curve fits, including DBF surrogates, DBF
sum class parameters, and DBF species. For all waters and all parameters the mean R2 was
0.973 ± 0.046, indicating that all breakthrough curve data were successfully fit using the logistic
function models. For DBF surrogates and DBF sum class parameters only, the mean R2 value
was 0.982 ±0.012.
4.3.2 DBP Species Logistic Function Curve Fits
THM Species. For all waters, the application of the step-lag logistic function resulted in good
curve fits for SDS-CF, as shown by R2 values ranging between 0.922 and 0.998. Typically, the
SDS-CF breakthrough curve shape was similar to that of SDS-TTHM, as shown in the example
given for Water 4 in Figure 65. For most SDS-BDCM and SDS-DBCM breakthrough curves,
the step-lag logistic function model was used and yielded good results: R2 values ranged from
0.946 to 0.998. An example of the SDS-BDCM breakthrough curve is shown in Figure 66 for
Water 6 and an example of the SDS-DBCM breakthrough curve is shown in Figure 67 for Water
7. The step-lag logistic function model was able to adequately fit the sharp 'S' shape (steep
breakthrough followed by flat plateau) observed for the SDS-BDCM experimental data in Figure
66 for Water 6 (R2 = 0.979). Water 2 also exhibited this behavior, as did Waters 1 and 3 to a less
pronounced extent, but in all cases the data were successfully fit by the model. For Waters 4 and
8, a peak breakthrough curve for SDS-BDCM was detected by the curve fit algorithm. These
curves were fit using the step-lag-peak logistic function model, as shown in Figures 68 and 69.
The curve fit procedure successfully fit the single contactor peak curves; the R2 values were
greater than 0.94 for both waters. For 3 out of 6 waters with SDS-BF levels measured above the
MRL, a peak curve was detected. The step-lag-peak logistic function model also successfully fit
these peak curves, as shown in Figures 70 and 71 for Waters 1 and 6. The R2 values for SDS-BF
curve fits ranged from 0.910 to 0.984. Summary plots of the model fits for all THM species and
all waters are included in Appendix E.
HAA Species. Of the non-brominated HAA species, no curve fits were applied to SDS-MCAA
because only for a few samples was MCAA measured above the MRL (2.0 |ig/L) in the GAC
effluent. The experimental breakthrough curves for SDS-DCAA and SDS-TCAA were well
-68-
-------
Analyte
TOC
UV-254
SDS-TOX
SDS-TTHM
SDS-HAA5
SDS-HAA6
SDS-HAA9
SDS-CF
SDS-BDCM
SDS-DBCM
SDS-BF
SDS-MCAA
SDS-DCAA
SDS-TCAA
SDS-MBAA
SDS-DBAA
SDS-BCAA
SDS-CDBAA
SDS-DCBAA
SDS-TBAA
Total
Number
Step
8
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
9
of each type
Step-lag
0
8
8
8
7
8
7
7
6
8
2
0
8
5
0
6
8
2
7
0
105
of curve fit used for all
Step-lag-peak
0
0
0
0
1
0
1
0
2
0
4
0
0
0
0
0
0
1
0
1
10
waters
No fit*
0
0
0
0
0
0
0
0
0
0
2
8
0
3
8
2
0
5
1
7
36
Curve fit
Mean
0.980
0.993
0.995
0.985
0.975
0.974
0.976
0.991
0.980
0.992
0.964
NA
0.975
0.973
NA
0.957
0.970
0.946
0.907
0.976
R2 value
SD
0.009
0.005
0.003
0.007
0.012
0.010
0.013
0.004
0.016
0.005
0.027
NA
0.023
0.027
NA
0.050
0.012
0.032
0.167
NA
*Curve fits were not performed on breakthrough curves with fewer than 6 points measured above the MRL
SD: Standard deviation
NA: Not applicable
Table 8 Frequency of logistic function model used and R2 values for all parameters and all waters
-------
modeled by the step-lag logistic function. Examples of the curve fits are shown in Figure 72
(SDS-DCAA, Water 3) and Figure 73 (SDS-TCAA, Water 5). For all waters, R2 values ranged
from 0.922 to 0.999. SDS-MBAA was not measured above the MRL (1.0 jig/L) in the GAC
effluent during any of the studies. Due to low or BMRL GAC effluent SDS-DBAA levels for
Waters 4 and 8 curve fitting procedures were not applied. For the remaining waters, the SDS-
DBAA breakthrough curve typically showed a sharp 'S' shape (steep breakthrough followed by
flat plateau) and was successfully fit by the step-lag logistic function model (R2 values ranging
from 0.859 to 0.992). An example of SDS-DBAA breakthrough and model fit is shown in
Figure 74 for Water 6. The breakthrough curves for SDS-BCAA were also well-fit by the step-
lag logistic function model (R2 values ranging from 0.952 to 0.987), such as that for Water 1,
shown in Figure 75. Examples of curve fits for SDS-DCBAA, SDS-CDBAA, and SDS-TBAA
are shown in Figures 76, 77, and 78, respectively. For these three compounds, R2 values ranged
from 0.913 to 0.987 except for the curve fit for SDS-DCBAA for Water 3, which yielded an R2
value of 0.531. For this run, GAC effluent SDS-DCBAA concentrations were low, measured
between 1.0 and 2.0 |ig/L.
A peak curve was detected in 10 cases out of 126 total curve fits. Every water examined yielded
at least one peak curve. In general, the step-lag-peak logistic function model was able to
adequately fit the peak curve data. Additional examples of peak curve fits are shown in Figures
79 and 80. Figure 81 (SDS-BF for Water 3) shows an example of single contactor data that
qualitatively shows a peak, but the peak curve algorithm did not detect it as such. The logistic
function under predicted experimental data at the peak by up to 4 |ig/L, and over predicted it at
the end of the run by up to 3 |ig/L. The R2 value for this curve fit was 0.856. When refit to the
step-lag-peak logistic function model, Figure 82, the R2 value improved to 0.984.
In general, the logistic function models successfully characterized the variety of breakthrough
curves observed in this study. Three logistic function models were utilized: step, step-lag, and
step-lag-peak. These models were able to match not only typical S-shaped breakthrough curves
but also peak curves observed for some parameters. Table 8 also summarizes the frequency of
use for each of the three types of logistic function models used. The step-lag logistic function
was the most commonly used, utilized in 66 percent of the curve fits performed. For the DBF
surrogates and class sums, the step-lag logistic function model was utilized for 82 percent of
curve fits. The step function was almost exclusively utilized for TOC breakthrough curves. For
23 percent of all data sets, no curve fit was performed because fewer than six data points were
measured above the MRL. Most of the parameters for which no curve fit was performed were
HAA species. For all data sets that were modeled, the mean R2 was 0.974, indicating that
overall, the models were able to successfully fit the GAC breakthrough profiles.
-70-
-------
3 -
o
'•a 2
CD
O
O
O
1 -
TOO
D Single contactor effluent
Logistic function best fit (RA2 = 0.966)
O Blended effluent
Dl prediction (RSS = 0.152)
O
EBCT = 20 min.
c0 = 4.54 mg/L
20 40
Scaled operation time (days)
60
80
Figure 49 Single contactor and blended effluent TOC breakthrough curves
for Water 1
2.5
2.0 -
o
'-4—'
CD
o
O
1.5 -
1.0 H
0.5 -
0.0
TOC
D Single contactor effluent
Logistic function best fit (RA2 = 0.972)
O Blended effluent
Dl prediction (RSS = 0.057)
50
EBCT = 20 min.
c0 = 2.6 mg/L
100 150
Scaled operation time (days)
200
250
Figure 50 Single contactor and blended effluent TOC breakthrough curves
for Water 2
-71-
-------
2.0
1.5 -
o
'•a
O>
o
c
o
O
0.5 -
0.0
TOO
D Single contactor effluent
Logistic function best fit (RA2 = 0.986)
O Blended effluent
Dl prediction (RSS = 0.044)
D D
EBCT = 20 min.
c0 = 2.35 mg/L
50
100 150 200
Scaled operation time (days)
250
300
Figure 51 Single contactor and blended effluent TOC breakthrough curves
for Water 3
2.5
2.0 -
O)
1.5 H
1.0 H
o
O
0.5 -
0.0
TOC
D Single contactor effluent
Logistic function best fit (RA2 = 0.984)
O Blended effluent
Dl prediction (RSS = 0.05)
EBCT = 20 min.
c0 = 2.98 mg/L
50 100 150
Scaled operation time (days)
200
Figure 52 Single contactor and blended effluent TOC breakthrough curves
for Water 4
-72-
-------
3.0
2.5 -
2.0 -
o
'•a
O>
o
o 1.0
O
0.5 -
0.0
TOO
D Single contactor effluent
Logistic function best fit (RA2 = 0.984)
O Blended effluent
Dl prediction (RSS = 0.035)
.---o"
.---O
EBCT = 20 min.
c0 = 3.08 mg/L
50 100 150 200 250
Scaled operation time (days)
300
350
Figure 53 Single contactor and blended effluent TOC breakthrough curves
for Water 5
2.5
2.0 -
O)
1.5 H
1.0 H
o
O
0.5 -
0.0
TOC
D Single contactor effluent
Logistic function best fit (RA2 = 0.992)
O Blended effluent
Dl prediction (RSS = 0.015)
50
100 150 200
Scaled operation time (days)
EBCT = 20 min.
c0 = 2.64 mg/L
250
300
Figure 54 Single contactor and blended effluent TOC breakthrough
curves for Water 6
-73-
-------
4 -
3 -
.g
"co
o> 9
o *-
c
o
O
1 -
TOO
D Single contactor effluent
Logistic function best fit (RA2 = 0.979)
O Blended effluent
Dl prediction (RSS = 0.065)
EBCT = 20 min.
c0 = 5.58 mg/L
25
50 75 100
Scaled operation time (days)
125
150
Figure 55 Single contactor and blended effluent TOC breakthrough
curves for Water 7
2.0
1.5 -
'•a 1-0 H
o>
o
c
o
O
0.5 -
0.0
TOC
D Single contactor effluent
Logistic function best fit (RA2 = 0.974)
O Blended effluent
Dl prediction (RSS = 0.033)
EBCT = 7.2 min.
C0 = 2.02 mg/L
50 100
Scaled operation time (days)
150
200
Figure 56 Single contactor and blended effluent TOC breakthrough
curves for Water 8
-74-
-------
0.040
0.035 -
0.030 -
0.025 -
CD
UV254
D Single contactor effluent
Logistic function best fit (RA2 = 0.992)
O Blended effluent
Dl prediction (RSS = 0.002)
EBCT = 20 min.
c0 = 0.051 1/cm
50
100 150 200 250
Scaled operation time (days)
300
350
Figure 57 Single contactor and blended effluent UV254 breakthrough
curves for Water 5
0.07
0.06 -
0.05 -
.o
^ 0.04 -
o>
o
c
CD
•9 0.03 -
o
0.02 H
0.01 -
0.00
UV254
D Single contactor effluent
Logistic function best fit (RA2 = 0.994)
O Blended effluent
Dl prediction (RSS = 0.0015)
O. -
25
50 75 100
Scaled operation time (days)
EBCT = 20 min.
c0 = 0.109 1/cm
125
150
Figure 58 Single contactor and blended effluent UV254 breakthrough
curves for Water 7
-75-
-------
175
150 -
-T 125 -
O
100 -
o
'-4—'
TO
75 -
SDS-TOX
D Single contactor effluent
Logistic function best fit (RA2 = 0.999)
O Blended effluent
Dl prediction (RSS = NA)
EBCT = 20 min.
C0 = 288 ug/L Cl-
50 100
Scaled operation time (days)
150
200
Figure 59 Single contactor and blended effluent SDS-TOX breakthrough
curves for Water 4
125
O
o
'-4— '
TO
0>
O
c
o
O
100 -
75 -
50 -
25 -
SDS-TOX
D Single contactor effluent
Logistic function best fit (RA2 = 0.99)
O Blended effluent
Dl prediction (RSS = NA)
EBCT = 7.2 min.
c0= 156 ug/LCI-
50 100
Scaled operation time (days)
150
200
Figure 60 Single contactor and blended effluent SDS-TOX breakthrough
curves for Water 8
-76-
-------
125
100 -
75 -
s
-I—'
§ 50
c
o
O
25 -
SDS-TTHM
D Single contactor effluent
Logistic function best fit (RSS 0.23A2 = 0.986)
O Blended effluent
Dl prediction (RSS = 2.87)
EBCT = 20 min.
c0 = 154 ug/L
50
100 150 200
Scaled operation time (days)
250
300
Figure 61 Single contactor and blended effluent SDS-TTHM breakthrough
curves for Water 3
100
75 -
o
'« 50 H
o>
o
c
o
o
25 -
0 ^
SDS-TTHM
D Single contactor effluent
Logistic function best fit (RA2 = 0.98)
O Blended effluent
Dl prediction (RSS = 6.43)
EBCT = 20 min.
c0= 128 ug/L
50
100 150 200
Scaled operation time (days)
250
300
Figure 62 Single contactor and blended effluent SDS-TTHM breakthrough
curves for Water 6
-77-
-------
30
25 -
20 -
o
|5 15 H
-I—•
§ "
I 10^
5 -
SDS-HAA9
D Single contactor effluent
Logistic function best fit (RA2 = 0.982)
O Blended effluent
Dl prediction (RSS = 0.78)
0 40
0
20 40
Scaled operation time (days)
EBCT = 20 min.
c0 = 29 ug/L
60
80
Figure 63 Single contactor and blended effluent SDS-HAA9 breakthrough
curves for Water 1
30
25 -
20 -
SDS-HAA9
EBCT = 20 min
c0 = 37 M9/L
D Single contactor effluent
Logistic function best fit (RA2 = 0.988)
O Blended effluent
Dl prediction (RSS = 3.64)
50
100 150
Scaled operation time (days)
200
250
Figure 64 Single contactor and blended effluent SDS-HAA9 breakthrough
curves for Water 2
-78-
-------
30
25 -
§20-
3.
c
o
TS 15 H
SDS-CF
D Single contactor effluent
Logistic function best fit (RA2 = 0.998)
O Blended effluent
Dl prediction (RSS = 0.41)
EBCT = 20 min.
c0 = 50.7 ug/L
50 100
Scaled operation time (days)
150
200
Figure 65 Single contactor and blended effluent SDS-CF breakthrough
curves for Water 4
35
30 -
25 -
20 -
o
'-4—'
cc
o>
o
o
O
15 -
10 -
5 -
0 -I
SDS-BDCM
EBCT = 20 min
C0 = 27.4 ug/L
50
_n
O
Single contactor effluent
Logistic function best fit (RA2 = 0.979)
O Blended effluent
Dl prediction (RSS = 3.63)
100 150 200
Scaled operation time (days)
250
300
Figure 66 Single contactor and blended effluent SDS-BDCM breakthrough
curves for Water 6
-79-
-------
50
40 -
30 -
o
'-4—'
cc
20-1
o
O
10 -
0 -I
SDS-DBCM
D Single contactor effluent
Logistic function best fit (RA2 = 0.996)
O Blended effluent
Dl prediction (RSS = 0.99)
25
' O
o-'
50 75 100
Scaled operation time (days)
' o
EBCT = 20 min.
c0 = 66.2 ug/L
125
150
Figure 67 Single contactor and blended effluent SDS-DBCM breakthrough
curves for Water 7
3 -
o
T5 2 H
o>
o
c
o
O
SDS-BDCM
EBCT = 20 min
c0 = 2.2 ug/L
D Single contactor effluent
Logistic function best fit (RA2 = 0.984)
O Blended effluent
Dl prediction (RSS = 0.58)
100
Scaled operation time (days)
150
200
Figure 68 Single contactor and blended effluent SDS-BDCM breakthrough
curves for Water 4
-80-
-------
4 -
3 -
g
'-4—'
CD
CD
O
c
O
O
2 -
1 -
0 -O
SDS-BDCM
EBCT = 7.2min.
C0 = 2.6 |jg/L
O
D
O
O
D
O
D Single contactor effluent
Logistic function best fit (RA2 = 0.946)
O Blended effluent
Dl prediction (RSS = 1.2)
50 100
Scaled operation time (days)
150
200
Figure 69 Single contactor and blended effluent SDS-BDCM breakthrough
curves for Water 8
D Single contactor effluent
Logistic function best fit (RA2 = 0.971)
O Blended effluent
Dl prediction (RSS = 2.76)
EBCT = 20 min.
C0 = 3.7
20 40
Scaled operation time (days)
60
80
Figure 70 Single contactor and blended effluent SDS-BF breakthrough
curves for Water 1
-81-
-------
20
15 -
g
'•£ 10 H
O>
o
c
o
O
5 -
0 -D
0
SDS-BF
EBCT = 20 min.
c0 = 3.3 |jg/L
50
D Single contactor effluent
Logistic function best fit (RA2 = 0.978)
O Blended effluent
Dl prediction (RSS = 3.47)
100 150 200
Scaled operation time (days)
250
300
Figure 71 Single contactor and blended effluent SDS-BF breakthrough
curves for Water 6
6 -
g
'•SS 4H
o>
o
c
o
O
2 -
SDS-DCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.995)
O Blended effluent
Dl prediction (RSS = 0.4)
O—O—r-
50
100 150 200
Scaled operation time (days)
EBCT = 20 min.
c0= 15.7ug/L
250
300
Figure 72 Single contactor and blended effluent SDS-DCAA breakthrough
curves for Water 3
-82-
-------
6 -
SDS-TCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.989)
O Blended effluent
Dl prediction (RSS = 0.28)
EBCT = 20 min.
C0 = 12.7 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure 73 Single contactor and blended effluent SDS-TCAA breakthrough
curves for Water 5
10
6 -
o
O
2 -
SDS-DBAA
EBCT = 20 min.
c0 = 5.7 ug/L
0 -t>
0
O
D Single contactor effluent
Logistic function best fit (RA2 = 0.952)
O Blended effluent
Dl prediction (RSS = 1.31)
50
100 150 200
Scaled operation time (days)
250
300
Figure 74 Single contactor and blended effluent SDS-DBAA breakthrough
curves for Water 6
-83-
-------
6 -
O>
o
c
o
O
SDS-BCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.981)
O Blended effluent
Dl prediction (RSS = 0.38)
20
40
Scaled operation time (days)
EBCT = 20 min.
c0= 7 |jg/L
60
80
Figure 75 Single contactor and blended effluent SDS-BCAA breakthrough
curves for Water 1
3.0
2.5 -
O)
o
73
o>
o
2.0 -
o 1.0 H
0.5 -
0.0 -I
SDS-DCBAA
EBCT = 20 min.
c0 = 3 |jg/L
50
D
-O
O
D Single contactor effluent
Logistic function best fit (RA2 = 0.952)
O Blended effluent
Dl prediction (RSS = 0.48)
100 150
Scaled operation time (days)
200
250
Figure 76 Single contactor and blended effluent SDS-DCBAA breakthrough
curves for Water 2
-84-
-------
3.5
3.0 -
2.5 -
c 2.0 H
o
g 1-5 H
o
c
o
0 1.0 H
0.5 -
0.0 -I
SDS-CDBAA
EBCT = 20 min.
c0 = 3 |jg/L
50
D D
. •" D Single contactor effluent
Logistic function best fit (RA2 = 0.913)
O Blended effluent
Dl prediction (RSS = 0.72)
100 150 200
Scaled operation time (days)
250
Figure 77 Single contactor and blended effluent SDS-CDBAA breakthrough
curves for Water 6
300
6 -
SDS-TBAA
EBCT = 20 min
C0 = BMRL
25
D Single contactor effluent
Logistic function best fit (RA2 = 0.976)
O Blended effluent
Dl prediction (RSS = 1.92)
50 75 100
Scaled operation time (days)
125
150
Figure 78 Single contactor and blended effluent SDS-TBAA breakthrough
curves for Water 7
-85-
-------
14
12 -
10 -
g
'-4—'
CD
CD
O
c
O
O
6 -
4 -
2 -
SDS-BF
O
EBCT = 20 min.
C0 = 3.7 |jg/L
D Single contactor effluent
Logistic function best fit (RA2 = 0.91)
O Blended effluent
Dl prediction (RSS = 3.73)
50
100 150
Scaled operation time (days)
200
250
Figure 79 Single contactor and blended effluent SDS-BF breakthrough
curves for Water 2
5 -
4 -
o
•^ ^
CD °
t_
CD
O
I2
1 -
SDS-CDBAA
EBCT = 20 min.
c0 = 3.7 |jg/L
50
D Single contactor effluent
Logistic function best fit (RA2 = 0.95)
O Blended effluent
Dl prediction (RSS = 0.88)
100 150 200 250
Scaled operation time (days)
300
350
Figure 80 Single contactor and blended effluent SDS-CDBAA breakthrough
curves for Water 5
-86-
-------
40
30 -
g
'•£ 20 H
O>
o
c
o
O
10 -
SDS-BF
EBCT = 20 min.
C0 = 11.8 |jg/L n
D
o
D
n n
o
D Single contactor effluent
Logistic function best fit (RA2 = 0.856)
O Blended effluent
Dl prediction (RSS = 6.3)
50
100 150 200
Scaled operation time (days)
250
300
Figure 81 Single contactor and blended effluent SDS-BF breakthrough
curves for Water 3 (original step-lag logistic function model curve fit)
40
30 -
o
'•£ 20 H
o>
o
c
o
O
10 -
SDS-BF
EBCT = 20 min.
c0= 11.8 ug/L
D Single contactor effluent
Logistic function best fit (RA2 = 0.984)
O Blended effluent
Dl prediction (RSS = 5.67)
50
100 150 200
Scaled operation time (days)
250
300
Figure 82 Single contactor and blended effluent SDS-BF breakthrough
curves for Water 3 (fit to step-lag-peak logistic function model)
-87-
-------
This page intentionally left blank.
-88-
-------
4.4 Comparison of SCA and Dl Methods Used to Predict the Blended Contactor
Integral Breakthrough Curve
Two methods were used to predict the integral breakthrough curve, which is used to estimate
blended contactor run times: direct integration (DI) and surrogate correlation approach (SCA).
The DI procedure was explained in Sections 1.4 and 3.2.3, while the steps followed by the SCA
procedure were described in Sections 1.5 and 3.2.4. The DI method has been presented in the
literature and is the traditional method for predicting the integral breakthrough curve, the curve
that relates blended contactor effluent water quality to single contactor run time. However,
previous verification of the DI method was limited to DBF surrogates and class sums.
Verification of the DI method here is expanded to include more GAC runs and water sources,
and a larger experimental matrix, including application to DBF species. The SCA method
developed in this study is verified against experimental data and compared to results obtained for
the DI procedure. The SCA method is computationally simpler than the DI procedure, an
important consideration when selecting a predictive method for application to the ICR GAC
treatment study data set, which may require 8,000 to 9,000 curve fits.
The results of these predictive models were compared to the experimentally-obtained blended
effluent data set. A best-fit curve was used to describe the observed data. The best-fit was
derived using the same logistic function models used for the single contactor data, as the shapes
observed for many of the blended effluent curves were similar to those encountered during single
contactor breakthrough curve analysis. The models were evaluated by comparing the predictions
to the experimental data (not the best-fit curve) and calculating the residual sum of squares
(RSS) and model bias. The model bias is defined as the mean of the residuals, calculated
between the model prediction and experimental results. A summary of the calculated model RSS
values is shown in Tables 9 and 10, while a summary of the calculated bias values is shown in
Tables 11 and 12. Note that the RSS and bias values have units equivalent to the units of the
parameter from which they are calculated; therefore, the magnitudes of these values for different
parameters may vary widely and are not directly comparable. A summary of average RSS and
bias values is given in Table 13. Based on direct comparisons between the RSS values for each
predictive approach, and across all parameters (110 comparisons) the two methods were equally
successful in predicting the observed data: for 52 percent of the predictions, the SCA method
RSS value was lower than that for the DI method.
To examine the performance of each model for predicting the integral breakthrough curve across
all water sources and water quality parameters, cumulative frequency distribution plots were
developed for the RSS and bias data. To provide a consistent basis of comparison of RSS and
bias data across different parameters, the data was normalized. This was accomplished by
dividing the RSS and bias values measured for each parameter and water by the average
concentration of that parameter in the blended effluent during each run. This procedure did
result in some extremely high normalized values, and overall, the average normalized values
were relatively high. However, this is due to the relatively low average effluent concentrations
of many parameters, by which the RSS and bias values were normalized. The normalized RSS
and bias values are tools that allow comparison of the two predictive models across all available
data in this study, and are not intended to provide an indication of model performance outside of
-89-
-------
VO
o
§ra meter
TOO (ngL)
UV-254 flfcm)
TOX &L Cl ')
TTHM gL)
HAAS gL)
HAA6 gL)
HAA9 gL)
CFgL)
BDCM gL)
DBCM gL)
BFgL)
MCAA gL)
DCAA gL)
TCAA gL)
MBAA gL)
DBAA gL)
BCAA gL)
DCBAAgL)
CDBAAgL)
TBAA gL)
Dl
0.152
0.0011
NA
3.93
0.38
0.63
0.78
0.39
1.52
0.40
2.76
NA
0.35
NA
NA
0.46
0.38
NA
NA
NA
MfeM
SCA
NA
0.0017
NA
2.55
1.58
1.26
1.04
0.65
1.03
0.73
1.31
NA
1.46
NA
NA
0.39
0.29
NA
NA
NA
Residual sum
Hter2
Lower
NA
Dl
NA
SCA
Dl
Dl
Dl
Dl
SCA
Dl
SCA
NA
Dl
NA
NA
SCA
SCA
NA
NA
NA
Dl
0.057
0.0012
NA
8.49
1.69
2.50
3.64
0.57
4.35
2.18
3.73
NA
0.49
0.27
NA
1.10
0.82
0.48
NA
NA
Dl 76CA: 5
SCA
NA
0.0015
NA
4.38
1.48
2.25
3.43
1.59
1.98
1.72
2.18
NA
0.68
0.34
NA
0.60
0.78
0.35
NA
NA
Dl
Lower
NA
Dl
NA
SCA
SCA
SCA
SCA
Dl
SCA
SCA
SCA
NA
Dl
Dl
NA
SCA
SCA
SCA
NA
NA
4SCA: 10
of sqares RSS)
HterS
Dl
0.044
0.0010
NA
2.87
2.48
2.29
4.36
0.85
1.05
1.04
5.67
NA
0.40
NA
NA
1.02
0.63
0.68
NA
NA
SCA
NA
0.0010
NA
2.64
1.13
2.30
2.20
0.40
2.57
0.68
3.14
NA
0.37
NA
NA
1.56
0.63
0.72
NA
NA
Dl
Lower
NA
Dl
NA
SCA
SCA
Dl
SCA
SCA
Dl
SCA
SCA
NA
SCA
NA
NA
Dl
Dl
Dl
NA
NA
6SCA: 7
Hter4
Dl
0.050
0.0015
NA
1.39
1.24
1.34
1.70
0.41
0.58
0.59
NA
NA
0.32
0.96
NA
NA
0.31
0.49
NA
NA
SCA
NA
0.0023
NA
2.03
0.87
1.07
1.12
1.33
0.55
0.55
NA
NA
0.47
0.64
NA
NA
0.22
0.50
NA
NA
Dl
Lower
NA
Dl
NA
Dl
SCA
SCA
SCA
Dl
SCA
SCA
NA
NA
Dl
SCA
NA
NA
SCA
Dl
NA
NA
5SCA: 7
NA: not applicable
Table SSummary of Dland SCA integral breakthrough curve prediction RSS values for Hfers through 4
-------
grameter
TOO (ngL)
UV-254(lfcm)
TOXgLCI ')
TTHM gL)
HAAS gL)
HAA6 gL)
HAA9 gL)
CFgL)
BDCM gL)
DBCM gL)
BFgL)
MCAA gL)
DCAA gL)
TCAA gL)
MBAAgL)
DBAA gL)
BCAA gL)
DCBAAgL)
CD BAA gL)
TBAA gL)
Dl
0.035
0.0020
NA
1.89
0.56
0.82
1.19
0.31
0.85
0.54
0.55
NA
0.49
0.28
NA
0.47
0.40
0.40
0.88
NA
HterS
SCA
NA
Lower
NA
0.0022 Dl
NA
1.30
0.52
0.55
1.13
0.65
0.34
0.72
0.26
NA
0.59
0.39
NA
0.20
0.21
0.34
1.10
NA
NA
SCA
SCA
SCA
SCA
Dl
SCA
Dl
SCA
NA
Dl
Dl
NA
SCA
SCA
SCA
Dl
NA
Dl 6SCA: 9
Dl
0.015
0.0017
2.39
6.43
1.02
1.30
1.88
1.37
3.63
1.02
3.47
NA
0.47
NA
NA
1.31
0.43
0.34
0.72
NA
Hter6
SCA
NA
0.0026
6.26
5.91
1.54
1.92
2.41
0.74
2.31
1.84
1.86
NA
0.31
NA
NA
1.11
0.44
0.34
0.67
NA
Dl
Residual sum
Lower
NA
Dl
Dl
SCA
Dl
Dl
Dl
SCA
SCA
Dl
SCA
NA
SCA
NA
NA
SCA
Dl
Dl
SCA
NA
8SCA: 7
Dl
0.065
0.0015
6.92
15.22
2.02
2.09
4.52
0.85
5.57
0.99
11.53
NA
0.59
0.40
NA
2.21
0.35
1.13
0.25
1.92
of sqares RSS)
tffer7
SCA
NA
0.0021
9.30
9.72
1.52
2.25
4.70
0.46
3.38
1.10
5.39
NA
0.63
1.34
NA
0.97
1.10
0.85
1.02
1.84
Dl
Lower
NA
Dl
Dl
SCA
SCA
Dl
Dl
SCA
SCA
Dl
SCA
NA
Dl
Dl
NA
SCA
Dl
SCA
Dl
SCA
9SCA: 8
Dl
0.033
0.0017
NA
2.59
1.35
1.83
2.20
0.58
1.20
1.15
NA
NA
0.68
0.73
NA
NA
0.59
0.46
NA
NA
HterS
SCA
NA
0.0017
NA
3.87
1.84
1.86
1.85
1.62
0.81
1.19
NA
NA
0.75
1.00
NA
NA
0.31
0.29
NA
NA
Dl
All waters lower RSS
Lower
NA
Dl
NA
Dl
Dl
Dl
SCA
Dl
SCA
Dl
NA
NA
Dl
Dl
NA
NA
SCA
SCA
NA
NA
8SCA: 4
Dl
0
8
2
2
3
5
3
5
1
5
0
0
6
4
0
1
3
3
2
0
53
SCA
0
0
0
6
5
3
5
3
7
3
6
0
2
1
0
5
5
4
1
1
57
NA: not applicable
Table OSummary of Dland SCA integral breakthrough curve prediction RSS values forttfers 3hrough 8
-------
grameter
TOC (ngL)
UV-254flcm)
TOX gL Cl ")
TTHM gL)
HAAS gL)
HAA6 gL)
HAA9 gL)
CFgL)
BDCM gL)
DBCM gL)
BFgL)
MCAA gL)
DCAA gL)
TCAA gL)
MBAAgL)
DBAA gL)
BCAA gL)
DCBAAgL)
CDBAAgL)
TBAA gL)
Model Bias
fetter 1
Dl
8.074
-0.0007
NA
-2.54
-0.15
-0.16
SCA
NA
-0.0012
NA
-1.65
8.29
8.15
-0.18 -0.03
8.18
-1.15
8.07
-2.17 -0
NA
-0.39
-0.74
-0.48
.33
NA
8.11 8.28
NA
NA
-0.21
8.04
NA
NA
NA
NA
NA
8.22
-0.06
NA
NA
NA
fetter
Dl
9.027
-0.0004
NA
-7.69
-1.45
-2.16
-3.07
-0.15
-3.83
-1.63
-2.91
NA
-0.38
-0.04
NA
-0.90
-0.72
-0.35
NA
NA
2
SCA
NA
-0.0005
NA
-4.09
-1.20
-1.85
-2.87
-0.92
-1.61
-1.15
-1.50
NA
-0.57
9.00
NA
-0.35
-0.66
-0.24
NA
NA
fetter 3
Dl
9.038
-0.0008
NA
-2.55
1-.41
1-.48
2.82
9.57
-0.35
9.59
-5.23
NA
9.29
NA
NA
-0.03
9.36
9.58
NA
NA
SCA
NA
-0.0007
NA
9.89
9.22
9.65
1-.23
-0.22
1-.53
-0.08
-1.17
NA
-0.22
NA
NA
9.52
9.17
9.62
NA
NA
fetter 4
Dl
9.037
SCA
NA
-0.0009 -0.0016
NA
-0.81
9.56
9.44
NA
-1.40
-0.57
-0.65
9.64 -0.36
-0.05
-0.47
-0.36
NA
NA
9.04
9.36
NA
NA
-0.13
9.34
NA
NA
-0.89
-0.09
-0.38
NA
NA
-0.33
-0.44
NA
NA
-0.09
9.38
NA
NA
NA: not applicable
Table 1 Summary of model
ifameter
TOC (ngL)
UV-254flcm)
TOX gL Cl ")
TTHM gL)
HAAS gL)
HAA6 gL)
HAA9 gL)
CFgL)
BDCM gL)
DBCM gL)
BFgL)
MCAA gL)
DCAA gL)
TCAA gL)
MBAAgL)
DBAA gL)
BCAA gL)
DCBAAgL)
CDBAAgL)
TBAA gL)
prediction
felterS
Dl
8.023
-0.0013
NA
-1.29
-0.22
-0.31
-0.86 -0
-0.04
-0.62
-0.25
-0.45 -0
NA
-0.35
8.00
NA
-0.27
-0.10
-0.18
-0.28
NA
SCA
NA
-0.0015
NA
-0.89
-0.03
8.03
.44
-0.44
9.04
-0.40
.12
NA
-0.44
-0.25
NA
9.10
9.06
-0.21
-0.02
NA
bias for ttfers through
fetter
Dl
-0.001
-0.0012
-1.53
-4.83
-0.55
-0.48
-0.10
9.55
-2.30
-0.21
-2.73
NA
9.34
NA
NA
-0.92
9.07
9.18
9.34
NA
4
Model Bias
6 fetter 7
SCA
NA
-0.0019
-4.89
-4.11
-1.14
-1.44
-1.50
-0.33
-1.47
-1.27
-0.76
NA
-0.04
NA
NA
-0.81
-0.25
-0.10
9.32
NA
Dl
9.044
-0.0007
-6.24
-11.39
-1.35
-1.20
-0.54
9.59
-3.28
9.71
-9.15
NA
9.48
9.03
NA
-1.65
9.17
9.50
9.05
9.08
SCA
NA
-0.0008
-5.48
-7.42
-0.23
9.33
1-.91
-0.30
-1.98
-0.59
-4.18
NA
9.53
-0.81
NA
-0.07
9.58
9.05
9.62
1-.02
fetter 8
Dl
-0.019
SCA
NA
-0.0001 -0.0008
NA
-2.01
-0.20
-0.43
-0.43
-0.25
-0.85
-0.90
NA
NA
-0.13
-0.09
NA
NA
-0.25
9.00
NA
NA
NA
-3.15
-0.78
-0.87
-0.79
-1.29
-0.58
-1.03
NA
NA
-0.35
-0.39
NA
NA
-0.10
9.14
NA
NA
NA: not applicable
Table 1 Summary of model prediction bias for ttfers through 8
-92-
-------
§ra meter
TOC (ngL)
UV-254 (l£m)
TOX &L Cl ')
TTHM &L)
HAAS &L)
HAA6 &L)
HAA9 &L)
CF&L)
BDCM &L)
DBCM &L)
BFfiL)
MCAA &L)
DCAA &L)
TCAA &L)
MBAA&L)
DBAA &L)
BCAA &L)
DC BAA &L)
CDBAA &L)
TBAA &L)
Mean RSS
Dl
0.06
0.0015
4.7
5.4
1.3
1.6
2.5
0.7
2.3
1.0
4.6
NA
0.5
0.5
NA
1.1
0.5
0.6
0.6
1.9
SCA
NA
0.0019
7.8
4.1
1.3
1.7
2.2
0.9
1.6
1.1
2.4
NA
0.7
0.7
NA
0.8
0.5
0.5
0.9
1.8
Mean normaliEd RSS Mean bias
%
Dl
7.2
18
8.5
19
26
23
29
36
29
19
38
NA
40
39
NA
28
28
119
98
87
SCA
NA
23
16
18
28
25
26
36
20
21
20
NA
44
61
NA
21
24
115
111
84
Dl
8.028
-0.0008
-3.88
-4.14
-0.24
-0.35
-0.22
8.17
-1.61
-0.25
-3.77
NA
8.05
8.05
NA
-0.66
-0.07
8.15
8.04
8.08
SCA
NA
-0.0011
-5.19
-2.73
-0.43
-0.46
-0.36
-0.60
-0.61
-0.67
-1.34
NA
-0.14
-0.38
NA
-0.07
-0.04
8.09
8.31
4.02
Mean normaliEd bias
%
Dl
3.7
-8.2
-7.4
-15
-4.4
-4.7
-2.3
6.5
-16
-4.6
-30
NA
3.2
4.1
NA
-17
-3.5
15
2.8
3.8
SCA
NA
-12
-9.8
-10
-7.9
-6.1
-3.9
-22
-6.1
-13
-11
NA
-8.9
-29
NA
-1.6
-2.2
8.4
24
46
Count
8
8
2
8
8
8
8
8
8
8
6
0
8
5
0
6
8
7
3
1
NA: not applicable
Table 3Summary of mean RSS,mean bias.normalied mean RSS,and normlied mean bias for all
waters
-93-
-------
the context of the actual magnitude of the RSS and bias values. For each parameter, the average
normalized RSS and bias across all waters are summarized in Table 13.
The cumulative frequency distribution plot of normalized RSS values is shown in Figure 83.
The distribution of normalized RSS values for both predictive approaches is similar, indicating
that there was little difference in the relative success of the two methods to predict experimental
data over the entire data set. The 25th to 75th percentile range of normalized RSS values for DI
predictions was 16 to 39 percent, while that for SCA predictions was similar, 16 to 38 percent.
The 10th to 90th percentile range of normalized RSS values for both methods was also similar, at
11 to 57 percent for the SCA method and 12 to 57 percent for the DI method.
Figure 84 shows the cumulative frequency distribution of normalized bias values. Across all
waters and water quality parameters, both predictive methods tended to underestimate the
experimental data. The median of the distribution for the DI method was -6 percent, while that
for the SCA method was -10 percent. The 25th to 75th percentile range of the distribution was -
16 to +3 percent for the DI method. The SCA method more often underpredicted the
experimental data as indicated by a 25th to 75th percentile range of the distribution of -20 to -1
percent. The 10th to 90th percentile range of the distribution for the DI method was -25 to +30
percent, while that for the SCA method was -28 to +10 percent.
Across all waters and analytes, the distribution of prediction error was shown to be similar based
on the cumulative frequency distribution of RSS. However, an evaluation of the results for each
parameter is still needed and is addressed below in Sections 4.4.1 and 4.4.2. By comparing the
error, as measured by RSS and bias, for each parameter across all waters, the relative
performance of each model for each individual analyte can be assessed. A summary of all model
predictions for all parameters and all waters is included in Appendix F. For this discussion,
examples of DI and SCA model results were selected from the eight runs.
4.4.1 Surrogates and Class Sums
As shown in Figures 49 through 56 (Section 4.3), the DI method was able to successfully predict
the integral breakthrough curve for TOC for the eight runs examined. The average DI model
RSS for all eight runs was 0.055 mg/L, which compares to a value of 0.025 mg/L for the average
of all eight best-fit curves applied to the data. The DI method prediction average RSS was
slightly more than twice that for a best curve fit, but both values were very low. Expressed as a
fraction of the average TOC concentration of the experimental blended effluent data for each
water, the normalized RSS for the DI prediction was low, 7 percent. The average bias for the DI
method prediction was low and positive, +0.028 mg/L. The mean normalized bias (bias
expressed as a fraction of the average blended effluent experimental data) was also low, 4
percent.
The success of the DI model for predicting the TOC integral breakthrough curve is an important
verification prior to the application of the SCA method, as this method relies on the TOC integral
breakthrough curve as a basis for predicting the integral breakthrough curves of all other
parameters. Although an excellent predictor of the TOC integral breakthrough curve, the DI
method was not as successful in predicting the integral breakthrough curve for DBFs, especially
brominated DBF species, as shown by the examples given in Section 4.3 (Figures 61 through
-94-
-------
82). The results of integral breakthrough curves predicted by the computationally-simpler SCA
procedure will be compared to those predicted by the DI method.
Figures 85 and 86 compare the DI and SCA UV254 integral breakthrough predictions against
experimental data for Waters 1 and 3. During both these runs, the RSS of the predictions were
very low (all less than 0.002 I/cm). For Water 1, the DI method was able to match experimental
results early in the run very well, while the SCA method underpredicted the observed data. Later
in the run, both methods slightly underpredicted the experimental integral breakthrough curve.
The DI method yielded a closer prediction based on RSS values for Water 1. The bias for both
predictions was negative (DI: -0.0007 I/cm; SCA: -0.0012 I/cm), indicating quantitatively that
the models slightly underpredicted the experimental data. For Water 3, both methods
underpredicted blended effluent UV254 early in the run. Beyond the midpoint of the run, both
methods matched the experimental results more closely. The RSS values for each predictive
approach were equivalent, and the bias values were both slightly negative (DI: -0.0008 I/cm;
SCA: -0.0007 I/cm). Overall, the UV254 DI prediction results yielded lower RSS values for
seven of the eight runs.
The integral breakthrough curve predictions for SDS-TTHM are given in Figures 87 and 88 for
Waters 1 and 7. Based on the calculated RSS values, the SCA method yielded better predictions
of the observed data than the DI method (Water 1 DI RSS: 3.9 |ig/L; Water 1 SCA RSS: 2.6
|ig/L; Water 7 DI RSS: 15 jig/L; Water 7 SCA RSS: 9.7 |ig/L), as was the case with six of the
eight runs. Early in the run, at the point of initial breakthrough, the SCA method underpredicted
experimental results, while the DI method matched the experimental data more closely.
However, later in the run, at higher breakthrough levels, the SCA method was able to better
predict the experimental data. This pattern was repeated during most runs. The more
pronounced deviation early in the run is preferable over later in the run, because the target
treatment objective for SDS-TTHM is usually exceeded later in the run, where the SCA
prediction is closest to the experimental data. The prediction bias for Water 1 for both methods
was negative (DI: -2.5 |J.g/L; SCA: -1.6 |ig/L). Both predictive methods also yielded a negative
bias for Water 7 (DI: -11 |ig/L; SCA: -7.4 |ig/L), indicating that the experimental results were
underpredicted by both predictive models. The average bias for all runs was negative for both
methods.
Neither predictive method was consistently more successful for the prediction of the integral
breakthrough curve for SDS-HAA5, SDS-HAA6 and SDS-HAA9. For SDS-HAA5 and SDS-
HAA9, the DI method was slightly more successful, with closer matches to experimental data
based on RSS values in five out of eight cases for each parameter. For SDS-HAA6, the SCA
method was a better predictor in five out of eight cases. The average bias for both model
predictions for SDS-HAA5, SDS-HAA6, and SDS-HAA9 for all runs was negative.
Figures 89 and 90 show the model prediction results for SDS-HAA5 for Waters 2 and 7,
respectively. For these two waters, the SCA method yielded a better prediction of the observed
blended effluent data by comparison of calculated RSS values (Water 2 DI RSS: 1.7 |ig/L; Water
2 SCA RSS: 1.5 |ig/L; Water 7 DI RSS: 2.0 |ig/L; Water 7 SCA RSS: 1.5 |ig/L). For Water 2,
both methods underpredicted experimental results throughout the run, while for Water 7, the DI
method underpredicted the observed data throughout the run. The SCA method initially
underpredicted experimental data, but between the midpoint and end of the run was a close
match to the data. The calculated bias for both waters and both methods was negative (Water 2
-95-
-------
DI bias: -1.4 ug/L; Water 2 SCA bias: -1.2 ug/L; Water 7 DI bias: -1.3 ug/L; Water 7 SCA
bias: -0.2 ug/L).
For Waters 3 and 5, Figures 91 and 92, respectively, compare the model prediction results for
SDS-HAA6. For Water 3, both models overpredicted blended effluent concentrations, especially
towards the end of the run (DI bias: +1.5 ug/L; SCA bias: +0.7 ug/L), and the RSS values were
similar (DI RSS: 2.3 ug/L; SCA RSS: 2.3 ug/L). For Water 5, both models matched the
observed results well, with the SCA method resulting in the closest match based on RSS values
(DI RSS: 0.8; SCA RSS: 0.6). The bias calculated for the DI prediction was negative (-0.3
ug/L), while that for the SCA prediction was slightly positive (+0.03 ug/L).
For Waters 5 and 6, the model predictions for blended contactor effluent SDS-HAA9 are shown
in Figures 93 and 94, respectively. The bias for both model predictions for both waters was
negative (Water 5 DI bias: -0.9 ug/L; Water 5 SCA bias: -0.4 ug/L; Water 6 DI bias: -0.1 ug/L;
Water 6 SCA bias: -1.5 ug/L). Although the SCA underpredicted blended effluent
concentrations early in the run, it yielded a very close match to the experimental data towards the
end of the run. For Water 5, the RSS value for the SCA method prediction (1.1 ug/L) was
slightly lower than that for the DI method prediction (1.2 ug/L). For Water 6, the DI method
yielded a better prediction of the integral breakthrough curve than did the SCA method, based on
RSS values (DIRSS: 1.9 ug/L; SCA RSS: 2.4 ug/L).
For two waters (Waters 6 and 7), SDS-TOX was also analyzed in the blended effluent. For these
two cases, the DI method was a better predictor of the integral breakthrough curve than the SCA
method. For Water 6 (Figure 95), the SCA method underpredicted experimental data throughout
most of the run, while for Water 7 (Figure 96), the SCA method underpredicted the experimental
results during the initial portion of the run. Later in the run the SCA method resulted in a very
good prediction of the integral breakthrough curve. For both waters, the RSS values for the DI
method were lower than those for the SCA method (Water 6 DI RSS: 2.4 ug/L Cl"; Water 6 SCA
RSS: 6.3 ug/L Cl"; Water 7 DI RSS: 6.9 ug/L Cl"; Water 7 SCA RSS: 9.3 ug/L Cl"). The bias
values for both runs and both model predictions were negative (Water 6 DI bias: -1.5 ug/L Cl";
Water 6 SCA bias: -4.9 ug/L Cl"; Water 7 DI bias: -6.2 ug/L Cl"; Water 7 SCA bias: -5.5 ug/L
Cl").
4.4.2 DBP Species
With a few exceptions, the DI method was able to better predict non-brominated DBP species,
while the SCA method was a better predictor of the brominated species. This is due in part to the
marginal ability of the DI method to predict blended effluent concentrations that result when the
single contactor curve shows a peak, or a steep breakthrough followed by a flat plateau. The
influence of the bromide to TOC ratio on the single contactor breakthrough curve is not captured
by the DI method, while the SCA method, which inherently correlates DBP formation to TOC
concentration, was better able to predict the integral breakthrough curve for brominated DBP
species. The relationship between TOC concentration and DBP formation and speciation in both
the single and blended contactor effluents is discussed in Section 4.2.
For SDS-CF, the DI method yielded lower RSS values for experimental predictions in five out of
eight cases. Figures 97 and 98 show examples of model predictions for Waters 2 and 6,
-96-
-------
respectively. For these two examples, the SCA method underpredicted the integral breakthrough
curve at the point of initial breakthrough and towards the end of the run. The SCA prediction
bias was negative for both runs (Water 2: -0.9 |ig/L; Water 6: -0.3 |ig/L). The DI method
yielded a better match to the observed for Water 2 (DI RSS: 0.6 |ig/L; SCA RSS: 1.6 |ig/L), but
the SCA method was a better predictor of the SDS-CF integral breakthrough curve for Water 6
(DI RSS: 1.4 |ig/L; SCA RSS: 0.7 |ig/L). The DI model bias was negative for Water 2 (-0.1
|ig/L) and positive for Water 6 (+0.5 |ig/L). Overall, the average bias for all runs was slightly
positive for the DI prediction, while that for the SCA prediction was slightly negative.
Figures 99 and 100 show the DI and SCA method predictions of SDS-BDCM experimental
results for Waters 1 and 8, respectively. Overall, the SCA method was a more successful model
for predicting observed data, with RSS values lower than those for the DI method for six of the
eight runs. In most cases, and as shown in the two examples, the DI method underpredicted the
experimental blended effluent concentration throughout the entire run. This was also reflected
by the negative bias values for the prediction results. The SDS-BDCM single contactor
breakthrough curves typically exhibited sharp breakthrough curves followed by a flat plateau or
peak curves. For the two examples given, the SCA method prediction provided a closer match to
the experimental data (Water 1 DI RSS: 1.5 |ig/L; Water 1 SCA RSS: 1.0 |ig/L; Water 8 DI
RSS: 1.2 |ig/L; Water 8 SCA RSS: 0.8 |ig/L). The SCA prediction bias was negative (Water 1:
-0.7 |ig/L; Water 8: -0.6 |ig/L), as was that for the DI predictions (Water 1: -1.1 |ig/L; Water 8:
-0.8 |ig/L). For all runs, the average model bias for both models was negative, although the SCA
prediction bias was lower in magnitude as compared to that for the DI method.
For SDS-DBCM, the DI method yielded a better fit to the observed data in five of eight cases.
Although a brominated DBF species, the breakthrough of SDS-DBCM was typically shaped
similar to that of SDS-CF: a gradual increase in concentration over time, without the "sharp
increase and plateau" or "peak" curve shapes observed for other brominated species. Figures 101
and 102 show examples of the results obtained for Waters 3 and 7, respectively. For Water 3
(Figure 101), both methods slightly underpredicted effluent concentrations at the beginning of
the run. The DI method overpredicted effluent concentrations towards the end of the run, while
the SCA method closely matched experimental data after the first third of the run. The SCA
prediction bias was positive (+0.6 |ig/L), while that for the DI prediction was slightly negative (-
0.1 |ig/L). The RSS value for the SCA prediction (0.7 jig/L) was lower than that for the DI
prediction (1.0 |ig/L). For Water 7, shown in Figure 102, the SCA method underpredicted
blended effluent SDS-DBCM levels at the point of initial breakthrough and towards the end of
the run, while the DI method either overpredicted or matched effluent levels. In this case, the DI
prediction RSS value (1.0 |ig/L) was slightly lower than those for the SCA prediction (1.1 |ig/L).
For these two examples, a positive model bias was measured for the DI prediction (+0.7 |ig/L),
while a negative model bias occurred for the SCA prediction (-0.6 |ig/L). For all waters, the
mean prediction bias for both methods was negative. The magnitude of the mean bias was lower
for the DI procedure.
The SCA method resulted in closer matches to the observed data than did the DI method, based
on calculated RSS values, for all SDS-BF integral breakthrough curve predictions (six of the
eight runs yielded data above the MRL). Several single contactor effluent SDS-BF data formed
peak curves, and some of the integral breakthrough curves were also peak curves. The DI
method underpredicted blended effluent SDS-BF concentrations. The SCA method was able to
better match the peak-shaped integral breakthrough curves that were sometimes observed.
-97-
-------
Examples of the model prediction results are given in Figures 103 and 104 for Waters 1 and 5,
respectively. The DI method underpredicted the experimental results throughout the entire run
for both waters. Although the SCA method initially underpredicted blended effluent
concentrations, at higher concentrations the accuracy of the predicted results increased,
especially relative to the DI method results. The measured bias values for the SCA prediction
(Water 1: -0.3 |ig/L; Water 5: -0.1 |ig/L) were lower in magnitude than those for the DI method
(Water 1: -2.2 ng/L; Water 5: -0.5 ng/L), and bias values were negative for both models. The
average model bias for the six runs analyzed (DI: -3.8 ^ g/L; SCA: -1.3 |ig/L) was negative for
both methods, and the magnitude of the DI method bias was greater than that for the SCA
method by a factor of 2.8.
The DI method continued to perform well for non-brominated HAA species, while the SCA
method was superior when predicting the integral breakthrough curve for brominated species.
Exceptions to this trend occurred for SDS-CDBAA and SDS-TBAA, for which the DI method
more often yielded better predictions. However, only four total cases were available for
evaluation of both species, due to concentrations not exceeding the MRLs during many runs.
Furthermore, analysis was complicated because formed levels were typically only slightly higher
than the MRL. Values below the MRL were assigned a value of zero, which was not the case for
the prediction methods. For the one case available to compare SDS-TBAA blended effluent
concentrations, the SCA method more closely matched effluent levels (three points) that were
measured above the MRL (4.0 |ig/L). However, the overall calculated RSS for the SCA method
was slightly higher than that for the DI method. For SDS-DCBAA, comparisons were possible
during only three runs, and the SCA method was a more accurate predictor of the integral
breakthrough curve in all three cases. No comparisons were possible for SDS-MCAA or SDS-
MBAA.
For SDS-DCAA, the DI method resulted in lower RSS values in six of the eight runs. Examples
of the results are shown in Figures 105 and 106 for Waters 1 and 4, respectively. While the DI
method well-predicted the data throughout the entire run for Water 1 (RSS: 0.4 |ig/L), the SCA
method initially underpredicted and then overpredicted the observed data (RSS: 1.5 |ig/L). For
Water 4, the DI method again was a successful predictor of the SDS-DCAA integral
breakthrough curve (RSS: 0.3 |ig/L), while the SCA method consistently underpredicted
experimental data (RSS: 0.5 |ig/L). For Water 1, both predictions yielded a positive bias (DI:
+0.1 ng/L; SCA: +0.3 |ig/L), while for Water 4, the bias for the SCA prediction was negative (-
0.3 |ig/L), and that for the DI prediction was low and positive (+0.04 |ig/L). Similar results were
obtained for SDS-TCAA predictions based on the two methods: for four of five possible
comparisons, the RSS value for the DI method was lower than that for the SCA method,
indicating a better prediction. The average bias for the DI method prediction (+0.1 |ig/L) was
low and positive, while that for the SCA method was negative (-0.4 |ig/L).
For SDS-DBAA and SDS-BCAA, whose single contactor breakthrough curve typically exhibit
sharp breakthrough curves followed by a flat plateau or peak curves, the SCA method
outperformed the DI method for integral breakthrough curve prediction. RSS values for the SCA
prediction were lower in four out six cases for SDS-DBAA, and in six out of eight cases for
SDS-BCAA. The average model bias was negative for both parameters. For SDS-DBAA, the
average bias magnitude was larger for DI predictions (DI: -0.7 |ig/L; SCA: -0.1 |ig/L), while for
SDS-BCAA, the average bias for both methods was low (DI: -0.07 |ig/L; SCA: -0.04 |ig/L).
Figure 107 shows the SDS-DBAA results obtained for Water 2 (DI RSS: 1.1 |ig/L; SCA RSS:
-98-
-------
0.6 |ig/L), while Figure 108 shows the SDS-BCAA results obtained for Water 8 (DI RSS: 0.6
|ig/L; SCA RSS: 0.3 |ig/L). In both these examples, the SCA prediction was more accurate than
the DI prediction based on RSS values. The model bias for these two examples was negative for
both models, and lower in magnitude for SCA method predictions as compared to DI method
predictions .
This analysis showed that the DI prediction was successful in predicting the integral
breakthrough curve for TOC, which is an important step for use of the SCA method. The
cumulative frequency distribution comparisons of the SCA and DI model results showed that the
two methods were equivalent in their ability to predict the integral breakthrough curve, based on
a comparison across all waters and water quality parameters. Over the entire data set, both
methods were biased negative in their predictions of the experimental data. Both the DI and
SCA predictions agreed well with experimental data for surrogates and class sums. For
chlorinated DBF species, the DI method was a better predictor of the integral breakthrough curve
than the SCA method. However, SCA method predictions outperformed those determined by the
DI procedure for brominated DBF species integral breakthrough curves.
The main advantage of the SCA method over the DI procedure is that predictions using the SCA
minimize the number of calculations necessary to predict blended contactor water quality as a
function of single contactor run time. Based on the results of this study, use of the SCA
procedure for predicting the integral breakthrough curve is recommended during ICR GAC
treatment study data analysis.
-99-
-------
CD
100
80 -
o 60
CD
.Q
O
-I—'
CD
CD
CL
40 -
To
» Dl prediction
o SCA prediction
10 100
Normalized residual sum-of-squares, RSS (%)
1000
Figure 83 Cumulateive frequency distribution plot of normalized residual sum-of-squares
(RSS) for Dl and SCA model predictions
100
CD
80 -
o 60
CD
.Q
_
O
-i—'
CD
CD
CL
40 -
20 -
-100
-50
0 50
Normalized Bias (%)
» Dl prediction
o SCA prediction
100
150
Figure 84 Cumulative frequency distribution plot of normalized bias for Dl and SCA model
predictions
-100-
-------
0.035
0.030 -
0.025 -
0.000
UV254
EBCT = 15 min.
c0 = 0.094 1/cm
• Observed data
Best fit
Dl prediction (RSS = 0.0011)
SCA prediction (RSS = 0.0017)
0 25 50
Scaled operation time (days)
Figure 85 Comparison of Dl and SCA methods for predicting the UV254
integral breakthrough curve for Water 1
75
0.020
0.015 -
o
o>
o
c
CD
.Q
O
5
.0
0.010 -
0.005 -
0.000
UV254
EBCT = 20 min.
co= 0.048 1/cm
• Observed data
Best fit
Dl prediction (RSS = 0.001)
SCA prediction (RSS = 0.001)
50
100 150 200
Scaled operation time (days)
250
300
Figure 86 Comparison of Dl and SCA methods for predicting the UV254
integral breakthrough curve for Water 3
-101-
-------
• Observed data
Best fit
Dl prediction (RSS = 3.93)
SCA prediction (RSS = 2.55)
0 25 50
Scaled operation time (days)
Figure 87 Comparison of Dl and SCA methods for predicting the SDS-TTHM
integral breakthrough curve for Water 1
75
125
100 -
a 75 H
o
'-4—'
S
| 50
o
O
25 -
SDS-TTHM
EBCT = 20 min.
co = 200
• Observed data
Best fit
Dl prediction (RSS = 15.22)
SCA prediction (RSS = 9.72)
25
50 75 100
Scaled operation time (days)
125
150
Figure 88 Comparison of Dl and SCA methods for predicting the SDS-
TTHM integral breakthrough curve for Water 7
-102-
-------
15
10
O)
o
"H—•
2
-i—'
I
O c J
O 5 H
SDS-HAA5
EBCT = 20 min
c° = 24 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.69)
SCA prediction (RSS = 1.48)
50
100 150
Scaled operation time (days)
200
250
Figure 89 Comparison of Dl and SCA methods for predicting the SDS-
HAA5 integral breakthrough curve for Water 2
25
20 -
15 -
CD
O
c
o
O
5 -
SDS-HAA5
EBCT = 20 min.
CD = 65 pg/L
• Observed data
Best fit
Dl prediction (RSS = 2.02)
SCA prediction (RSS = 1.52)
25
50 75 100
Scaled operation time (days)
125
150
Figure 90 Comparison of Dl and SCA methods for predicting the SDS-
HAA5 integral breakthrough curve for Water 7
-103-
-------
20
15 -
o
'•a 10
O>
o
o
O
5 -
SDS-HAA6
EBCT = 20 min
co = 43 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 2.29)
SCA prediction (RSS = 2.3)
50
100 150 200
Scaled operation time (days)
250
300
Figure 91 Comparison of Dl and SCA methods for predicting the SDS-
HAA6 integral breakthrough curve for Water 3
16
14 -
12 -
1 1°-
o
'•a 8
O>
c 6
o
O
4 -
2 -
SDS-HAA6
EBCT = 20 min.
c0 = 34 M9/L-
• Observed data
Best fit
Dl prediction (RSS = 0.82)
SCA prediction (RSS = 0.55)
50
100 150 200 250
Scaled operation time (days)
300
350
Figure 92 Comparison of Dl and SCA methods for predicting the SDS-
HAA6 integral breakthrough curve for Water 5
-104-
-------
25
20 -
15 -\
§ 10-
o
O
5 -
SDS-HAA9
EBCT = 20 min
C0 = 48 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.19)
SCA prediction (RSS = 1.13)
50
100 150 200 250
Scaled operation time (days)
300
350
Figure 93 Comparison of Dl and SCA methods for predicting the SDS-
HAA9 integral breakthrough curve for Water 5
25
20 -
a 15 H
c
o
'-4—'
S
§ 10
c
o
O
5 -
SDS-HAA9
EBCT = 20 min.
Co = 61 |jg/L
Observed data
Best fit
Dl prediction (RSS = 1.88)
SCA prediction (RSS = 2.41)
50
100 150 200
Scaled operation time (days)
250
300
Figure 94 Comparison of Dl and SCA methods for predicting the SDS-
HAA9 integral breakthrough curve for Water 6
-105-
-------
o
o
'-4—'
CD
CD
O
C
O
O
100
80 -
60 H
40 -
20 -
SDS-TOX
EBCT = 20 min.
co = 305 |jg/L Cl-
• Observed data
Best fit
Dl prediction (RSS = 2.39)
SCA prediction (RSS = 6.26)
50
100 150 200
Scaled operation time (days)
250
300
Figure 95 Comparison of Dl and SCA methods for predicting the SDS-TOX
integral breakthrough curve for Water 6
200
150 -
O
.2 100 -
"CD
CD
o
o
0
50 H
SDS-TOX
EBCT = 20 min.
co = 486 |jg/L Cl
• Observed data
Best fit
Dl prediction (RSS = 6.92)
SCA prediction (RSS = 9.3)
25
50 75 100
Scaled operation time (days)
125
150
Figure 96 Comparison of Dl and SCA methods for predicting the SDS-
TOX integral breakthrough curve for Water 7
-106-
-------
15
O)
o
'-4—'
TO
o>
o
c
o
O
10 -
SDS-CF
EBCT = 20 min.
c0= 41.9 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.57)
SCA prediction (RSS = 1.59)
50
100 150
Scaled operation time (days)
200
250
Figure 97 Comparison of Dl and SCA methods for predicting the SDS-CF
integral breakthrough curve for Water 2
6 -
O
'•£ 4
o>
o
c
o
O
SDS-CF
EBCT = 20 min
c0 = 55.3 pg/L
50
• Observed data
Best fit
Dl prediction (RSS = 1.37)
SCA prediction (RSS = 0.74)
100 150 200
Scaled operation time (days)
250
300
Figure 98 Comparison of Dl and SCA methods for predicting the SDS-CF
integral breakthrough curve for Water 6
-107-
-------
20
16 -
O)
a 12 H
c
o
'-4—'
S
§ 8
c
o
O
4 -
SDS-BDCM
EBCT = 15 min.
c0 = 19.3|jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.52)
SCA prediction (RSS = 1.03)
25 50
Scaled operation time (days)
75
Figure 99 Comparison of Dl and SCA methods for predicting the SDS-
BDCM integral breakthrough curve for Water 1
4 -
3 -
8 2
o
O
1 -
SDS-BDCM
EBCT = 7.2 min.
c0 = 2.6 |jg/L
0 -!•-
0
• Observed data
Best fit
Dl prediction (RSS = 1.2)
SCA prediction (RSS = 0.81)
50 100
Scaled operation time (days)
150
200
Figure 100 Comparison of Dl and SCA methods for predicting the SDS-
BDCM integral breakthrough curve for Water 8
-108-
-------
o
I
I
o
O
16
14 -
12 -
10 -
8 -
6 -
4 -
2 -
SDS-DBCM
EBCT = 20 min
co = 44.5 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.04)
SCA prediction (RSS = 0.68)
50
100 150 200
Scaled operation time (days)
250
300
Figure 101 Comparison of Dl and SCA methods for predicting the SDS-
DBCM integral breakthrough curve for Water 3
25
20 -
15 -
§ 10
c
o
O
5 -
SDS-DBCM
EBCT = 20 min.
c0 = 66.2 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.99)
SCA prediction (RSS = 1.1]
25
50 75 100
Scaled operation time (days)
125
150
Figure 102 Comparison of Dl and SCA methods for predicting the SDS-
DBCM integral breakthrough curve for Water 7
-109-
-------
16
14 -
12 -
10 -
o
'-4—'
CD
CD
O
c
o
O
SDS-BF
EBCT = 15 min
c0= 3.7|jg/L
• Observed data
Best fit
Dl prediction (RSS = 2.76)
SCA prediction (RSS = 1.31)
25 50
Scaled operation time (days)
75
Figure 103 Comparison of Dl and SCA methods for predicting the SDS-BF
integral breakthrough curve for Water 1
3.0
2.5 -
2.0 -
o
'•SS 1-5 H
o 1.0 H
0.5 -
0.0
SDS-BF
EBCT = 20 min.
c0= 1-2M9/L
• Observed data
Best fit
Dl prediction (RSS = 0.55)
SCA prediction (RSS = 0.26)
50
100 150 200 250
Scaled operation time (days)
300
350
Figure 104 Comparison of Dl and SCA methods for predicting the SDS-BF
integral breakthrough curve for Water 5
-110-
-------
10
6 -
o
"CD
O> 4 .
O ^ I
c
o
O
2 -
SDS-DCAA
EBCT = 15 min.
c0 - 12.5 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.35)
SCA prediction (RSS = 1.46)
0 25 50
Scaled operation time (days)
Figure 105 Comparison of Dl and SCA methods for predicting the SDS-
DCAA integral breakthrough curve for Water 1
75
6 -
5 -
4 -
o
'-4—'
CD
SDS-DCAA
EBCT = 20 min.
co = 20.3 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.32)
SCA prediction (RSS = 0.47)
50
100
Scaled operation time (days)
150
200
Figure 106 Comparison of Dl and SCA methods for predicting the SDS-
DCAA integral breakthrough curve for Water 4
-111-
-------
6 -
o
14H
I
o
O
SDS-DBAA
EBCT = 20 min
c0 = 5 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.1)
SCA prediction (RSS = 0.6)
100 150
Scaled operation time (days)
200
250
Figure 107 Comparison of Dl and SCA methods for predicting the SDS-
DBAA integral breakthrough curve for Water 2
3.5
3.0 -
2.0 H
O
0 1.0 H
0.5 -
0.0
SDS-BCAA
EBCT = 7.2min.
CD = 3 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.59)
SCA prediction (RSS = 0.31)
50 100
Scaled operation time (days)
150
200
Figure 108 Comparison of Dl and SCA methods for predicting the SDS-
BCAA integral breakthrough curve for Water 8
-112-
-------
4.5 Analysis of Model Applicability to Finite Number of Contactors
Both the DI and SCA models for predicting the integral breakthrough curve, a relationship
between single contactor run time and blended contactor effluent water quality, rely on the
assumption that an infinite number of contactors are operated on-line in parallel-staggered mode.
The application of the DI procedure to any parameter relies on the infinite contactor assumption;
the DI prediction of the TOC integral breakthrough curve is a preliminary step utilized by the
SCA method for predicting the integral breakthrough curves of all other water quality
parameters. The application of these methods to situations involving finite numbers of
contactors may be limited by this assumption. The following analysis was developed to
determine the acceptability of the error incurred by the infinite contactor assumption.
Using the step logistic function model, a series of integral breakthrough curves were developed
and plotted for the following numbers of contactors in parallel, N: 2, 3, 4, 6, 10, 20, and infinite.
For the case of an infinite number of contactors Equation 12 (Section 1.4) was used, while in all
other cases the numerical integration shown in Equation 4 (Section 1.2) was used. The single
contactor breakthrough curve from which the integral curves were developed was also plotted.
The results of this modeling are shown in Figures 109 through 114 for varying values of B and
D, logistic function parameters that affect the shape of the curve. Values used for B ranged from
10 to 30, while values for D ranged from 0.05 to 0.20. These value ranges for B and D were
chosen to reflect the range typically seen for fits of GAC breakthrough curve data. Set values
were used instead of experimental data to examine the impact of B and D while A and A0 were
held constant. For the curve fits performed in this study, the 25th and 75th percentiles for the
values of B were 5 and 25, respectively, while the 25th and 75th percentiles for the values of D
were 0.02 and 0.07, respectively. The best-fit parameters for all curve fits performed in this
study are included in Appendix G. The parameter values used for A0 and A were held constant
for all six series of breakthrough curves, at values of 0 and 1, respectively. (The logistic function
parameter A represents the asymptote to which the logistic function is approaching, while A0
represents a step value given to the function.) All the data were plotted over 100 full-scale
operation days.
The graphs in Figures 109 through 114 show that as the number of contactors operated in
parallel-staggered mode increases, the integral curve approaches the model results for an infinite
number of contactors. The largest incremental benefit afforded by operating contactors in
parallel-staggered mode over single contactor operation occurs when two contactors are
operated. The benefit realized by adding an additional contactor decreases as the number of
contactors on-line increases.
To quantitatively compare the results, the run time to a range of treatment objectives expressed
as a fraction of the parameter A, the value to which the breakthrough curve approaches
asymptotically, was estimated based on the curves developed for N contactors. These are not
percent breakthrough values; they represent instead a fraction of the concentration that the single
contactor curve approaches asymptotically (asymptotic concentration). Values of 0.35, 0.50, and
0.65 were utilized as treatment objectives. Note that each value corresponds to a lower percent
breakthrough value than that calculated based on the GAC influent concentration. An extreme
case of a 0.80 treatment objective was also examined. This analysis is summarized in Tables 14
through 17.
-113-
-------
The goal of this analysis was to establish what the smallest value for TV was for which the integral
curve (derived based on the infinite number of contactors assumption) would yield a maximum
10 percent error in the estimated blended contactor run times, assuming this level of error is
acceptable. The parameter N90 will represent this breakpoint and is defined as the minimum N at
which point the throughput of each contactor operated in parallel-staggered mode is 90 percent
of that for an infinite number of contactors on-line. This is equivalent to the point at which the
run time estimated using an assumption of an infinite number of contactors on-line is within 10
percent of the actual run time of each of TV contactors.
Figures 115 and 116 establish a relationship between integral breakthrough curve run time
estimates based on an infinite number of contactors and that based on a finite number of
contactors. As shown in Figures 115 and 116, the minimum number of contactors operated in
parallel-staggered mode necessary so that the integral breakthrough curve for finite A/90 is within
10 percent of the integral breakthrough curve for infinite contactors varies with treatment
objective. For B = 30 and D = 0.10 and treatment objectives ranging between 0.35 and 0.65, the
A/9o ranged from 7 to 13 contactors. As the value used for the treatment objective was increased
(a less stringent treatment objective relative to GAC effluent concentration), A/90 increased.
Similar results were obtained for B = 30 and D = 0.20. As the treatment objective was varied
between 0.35 and 0.65, the Ngo ranged from 7 to 13 contactors. The Ngo to meet the given
treatment objectives did not vary as B and D were varied: for treatment objectives between 0.35
and 0.65, the A/90 varied from 7 to 13 contactors for both pairs of B and D values modeled.
Therefore, the error between run times estimated based on an infinite number of contactors and
that estimated for N contactors is independent of the shape of the curve, although it is dependent
on the extent of breakthrough achieved. For the extreme asymptotic concentration fraction value
of 0.80, A/9o was 22 contactors.
The assumption of an infinite number of contactors operated in parallel-staggered mode
simplifies the analysis of the integral breakthrough curve, as does the assumption of a linear
breakthrough curve. However, the data presented here indicate that for breakthrough profiles
that follow the logistic function form, N90 varies between 7 and 13 contactors, as the treatment
objective is varied between 35 and 65 percent of the asymptotic concentration. Thus, when
using the infinite contactor assumption to model the integral breakthrough curve for less than 14
contactors in parallel, the magnitude of the treatment objective in relation to the single contactor
breakthrough curve is important in minimizing the error between qN, the specific throughput of a
finite number of contactors, A/, and q^, the specific throughput of each contactor assuming an
infinite number of contactors on-line.
Based on the use of the logistic function model and for the range of treatment objectives
evaluated in this analysis, the integral breakthrough curve developed from the infinite contactor
assumption will yield estimated run times within 10 percent of actual run times for 13 or more
contactors operated in parallel-staggered mode. For 10 contactors on-line, the infinite contactor
assumption will yield run time estimates within 12 percent of the actual run times. In all cases,
run time estimates based on the infinite contactor assumption are longer than those for a finite
number of contactors, and thus yields a best case estimate of GAC performance. The
relationship developed between integral breakthrough curve run time estimates based on an
infinite number of contactors and that based on a finite number of contactors can be applied to
-114-
-------
ICR treatment study data to obtain GAC run time estimates that are more applicable to GAC
applications with a small number of parallel contactors.
Number of
contactors,
N
1
2
3
4
6
10
20
Run time to
5 = 30;
£> = 0.10
56
71
78
83
88
92
96
Table 14 Summary of run
Number of
contactors,
N
1
2
3
4
6
10
20
Run time to
5 = 30;
D = 0.10
50
67
75
80
86
91
95
treatment
5 = 30;
£> = 0.05
56
71
78
83
88
92
96
objective (0.35), expressed as percentage of run
infinite number of contactors
5=10;
£> = 0.10
54
70
78
82
87
92
96
times to a 0.35 treatment
treatment
5 = 30;
£> = 0.05
NA
NA
NA
NA
NA
NA
NA
5=10;
£> = 0.05
54
70
78
82
87
92
96
objective
5=10;
D = 0.2
54
70
78
82
87
92
96
objective (0.50), expressed as percentage of run
infinite number of contactors
5=10;
D = 0.10
50
67
75
80
86
91
95
5=10;
£> = 0.05
50
67
75
80
86
91
95
5=10;
D = Q.2
50
67
75
80
86
91
95
time for
5 = 30;
D = 0.2
56
71
78
83
88
92
96
time for
5 = 30;
D = 0.2
50
67
75
80
86
91
95
NA: not applicable, treatment objective not exceeded
Table 15 Summary of run times to a 0.50 treatment objective
-115-
-------
Number of Run time to
contactors,
N
1
2
3
4
6
10
20
NA: not
Table 16
5 = 30;
£> = 0.10
41
59
69
74
81
88
94
treatment
5 = 30;
£> = 0.05
NA
NA
NA
NA
NA
NA
NA
objective (0.65), expressed as percentage of run
infinite number of contactors
5=10;
£> = 0.10
43
60
70
75
82
88
94
5=10;
£> = 0.05
NA
NA
NA
NA
NA
NA
NA
5=10;
D = 0.2
43
60
70
75
82
88
94
time for
5 = 30;
£> = 0.2
41
59
69
74
81
88
94
applicable, treatment objective not exceeded
Summary of run
Number of Run time to
contactors,
N
1
2
3
4
6
10
20
5 = 30;
£> = 0.10
NA
NA
NA
NA
NA
NA
NA
times to a 0.65 treatment
treatment
5 = 30;
£> = 0.05
NA
NA
NA
NA
NA
NA
NA
objective
objective (0.80), expressed as percentage of run
infinite number of contactors
5=10;
£> = 0.10
NA
NA
NA
NA
NA
NA
NA
5=10;
£> = 0.05
NA
NA
NA
NA
NA
NA
NA
5=10;
D = 0.2
31
48
58
64
73
82
90
time for
5 = 30;
£> = 0.2
28
45
56
63
71
81
89
NA: not applicable, treatment objective not exceeded
Table 17 Summary of run times to a 0.80 treatment objective
-116-
-------
1.2
1.0 -
o
ro 0.8 -
CD
o
_
CD
-I—'
CD
0.4 -
CD
CL
0.2 -
0.0
A0: 0
A: 1
B: 30
D: 0.1
25
50 75
Run time (days)
100
125
Figure 109 Integral breakthrough curves for varying numbers of contactors
operated in parallel-staggered mode (B = 30; D = 0.1)
1.0
0.8 -
CD
O
c
o
o
t_
CD
-I—'
CD
2.
CD
CL
0.6 -
0.4 -
0.2 -
0.0
25
50 75
Run time (days)
100
125
Figure 110 Integral breakthrough curves for varying numbers of contactors
operated in parallel-staggered mode (B = 30; D = 0.05)
-117-
-------
1.2
1.0 -
o
ro 0.8 -
CD
o
§
E
& 0.4 -
CD
CL
0.2 -
0.0
A0: 0
A: 1
B: 10
D: 0.1
25
50 75
Run time (days)
Number of
contactors
-2
-4
10
-20 - - - Inf.
100
125
Figure 111 Integral breakthrough curves for varying numbers of contactors
operated in parallel-staggered mode (B = 10; D = 0.1)
1.0
0.8 -
g °-6J
o
o
o
CD
"CD 0.4
E
CD
CD
CL
0.2 -
0.0
A0: 0
A: 1
B: 10
D: 0.05
25
50 75
Run time (days)
-20 - - - 'Inf.
100
125
Figure 112 Integral breakthrough curves for varying numbers of contactors
operated in parallel-staggered mode (B = 10; D = 0.05)
-118-
-------
1.2
1.0 -
o
'CD 0.8 -
CD
o
8 0.6 -
"CD
2
CD
CL
0.4 -
0.2 -
0.0
25
50 75
Run time (days)
Number of
contactors
•1 -
>3 -
6 --
20 -
2
'—4
•-- 10
- 'Inf.
100
125
Figure 113 Integral breakthrough curves for varying numbers of contactors
operated in parallel-staggered mode (B = 10; D = 0.2)
1.2
1.0 -
o
|5 0.8 -
CD
O
8 0.6 H
CD
CD 0.4 -
CD
CL
0.2 -
0.0
25
50 75
Run time (days)
Number of
contactors
•1 -
•3 -
-6 --
•20 -
2
•—4
•-- 10
- 'Inf.
100
125
Figure 114 Integral breakthrough curves for varying numbers of contactors
operated in parallel-staggered mode (B = 30; D = 0.2)
-119-
-------
100
c
O)
^ 90 --
80 -
o
"CD
t_
en
CD
£ 70 -
CD
O
t_
CD
Q.
CD"
60 -
50 -
40
90 percent
10 15
Number of contactors, n
Treatment
objective
0.35
0.50
0.65
20
25
Figure 115 Run time as a function of number of contactors in parallel,
expressed as percent of run time for infinite n (B = 30; D = 0.1)
100
g
i_
O
90 --
80 -
O
"CD
t_
O)
B 70
c
60 -
o
I
CD
Q.
CD
E 50
a:
40
90 percent
10 15
Number of contactors, n
20
25
Figure 116 Run time as a function of number of contactors in parallel,
expressed as percent of run time for infinite n (B = 30; D = 0.2)
-120-
-------
4.6 Impact of Extrapolation on Integral Breakthrough Curve Prediction
Extrapolation of the integral TOC breakthrough curve may be necessary during ICR treatment
study data analysis because the SCA procedure is limited by the highest TOC concentration
estimated from the direct integration of the single contactor TOC breakthrough curve. The
highest blended effluent TOC concentration is typically 40 to 70 percent of the highest single
contactor TOC concentration. Although higher single contactor TOC concentrations are
associated with formed DBF concentrations, these cannot be applied to the integral breakthrough
curve during application of the SCA method unless the TOC integral breakthrough curve is
extrapolated. The impact of breakthrough curve extrapolation was analyzed for two GAC runs
to determine whether the error caused by extrapolation was acceptable for ICR data analysis.
For two water sources (Waters 5 and 8), the RSSCTs were operated longer than required by the
ICR. For Water 5, the run was extended 49 full-scale days after 70 percent TOC breakthrough
was reached, equivalent to a 21 percent extension of the required run time. For Water 8, the
extension beyond the required run time was 69 days, equivalent to a 61 percent increase. During
the extended run time, two additional single contactor GAC effluent samples were taken. Two
additional blended effluent samples were also taken. The first 11 GAC effluent data points that
comprised a normal ICR treatment study run reaching 70 percent TOC breakthrough were
modeled separately from the entire 13-point GAC effluent data set. Based on the logistic
function model best-fit of the first 11 data points, the water quality at the end of the run (21 or 61
percent extrapolation) was predicted by extending the model fit.
Using the DI method, the TOC integral breakthrough curve was calculated based on the
truncated logistic function model fit and then extrapolated to predict water quality at the end of
the entire run. The TOC integral breakthrough curve was also calculated based on the logistic
function best-fit of the entire single contactor effluent data set. For the other water quality
parameters, the SCA prediction of blended contactor effluent water quality based on the
extrapolated data set and was compared to that based on the entire data set. The impact of
extrapolation on the TOC integral breakthrough curve is of special interest, due to the application
of this extrapolation as a part of the SCA procedure during data analysis of the ICR treatment
studies.
The impact of extrapolation on integral breakthrough curve predictions is shown in Figures 117
through 130. In these figures, the single contactor and integral breakthrough curves are plotted
against scaled operation time. The data points used for the extrapolated single contactor data set
are plotted with filled symbols, to differentiate these from the open symbols used for the first 11
data points. For the single contactor data, the solid best-fit line represents the logistic function
best-fit using the entire data set, while the dashed line represents the best-fit of the first 11 points
and extrapolation to the end of the run. Similarly, a dashed line also represents the extrapolated
integral breakthrough curve. The R2 value is given for both best-fits of the single contactor data.
Figures 117 and 118 show the impact of extrapolation on the TOC single contactor logistic
function curve fit and the TOC integral breakthrough curve predicted by the DI method for
Waters 5 and 8, respectively. At the end of both runs, the extrapolated logistic function curve fit
underpredicted single contactor effluent data. Based on the extrapolated logistic function, the
TOC concentration at the end of the run was 8 percent lower than that based on the logistic
-121-
-------
function best-fit of all available single contactor data for Water 5. For Water 8, the extrapolated
predicted TOC concentration at the end of the run was 12 percent lower.
This underprediction of the single contactor effluent data by the extrapolated logistic function
best-fit had a smaller impact on the TOC integral breakthrough curve prediction. The error
associated with extrapolation of the TOC integral breakthrough curve at the end of the run was 3
percent (0.05 mg/L) for Water 5. For Water 8, the error was slightly larger, 8 percent (0.08
mg/L).
Application of the SCA procedure to the extrapolated single contactor breakthrough curves for
all parameters yielded an average error in the predicted blended contactor concentration at the
end of the run of 5 ± 3 percent for Water 5. Figures 119 through 124 show the impact of
extrapolation on SCA prediction of the blended contactor effluent for UV254, SDS-TOX, SDS-
TTHM, SDS-HAA5, SDS-HAA6, and SDS-HAA9. For these DBF surrogates and class sums,
the average error in the SCA prediction due to extrapolation was also 5 percent, (the run time
was extrapolated by 21 percent) and in all cases, the extrapolated prediction was lower than the
prediction using the entire data set. Comparisons of all water quality parameters, including DBF
species, are given in Appendix H.
For Water 8, the average error in the predicted blended contactor concentration at the end of the
run based on the SCA prediction of the extrapolated integral breakthrough curve was 9 ± 5
percent. This higher mean observed error for Water 8 as compared to Water 5 can be attributed
in part to the larger percent extrapolation, 61 percent, performed on Water 8 as compared to that
for Water 5, 21 percent. For UV254, SDS-TOX, SDS-TTHM, SDS-HAA5, SDS-HAA6, and
SDS-HAA9 the average error was 7 percent. Figures 125 through 130 show the impact of
extrapolation on the SCA prediction of the blended contactor concentration of these DBF
surrogates and class sums. The blended contactor effluent concentration at the end of the run
after extrapolation was lower than that based on the entire data set, except for SDS-HAA5.
An analysis of Water 8 was also made using an extrapolation of 29 percent, which is similar to
the extrapolation applied to Water 5. For this shorter level of extrapolation, the mean error in the
integral breakthrough curve at the end of the run for all parameters was 8 percent, only slightly
lower than that observed for an extrapolation of 61 percent. For UV254, SDS-TOX, SDS-TTFDVI,
SDS-HAA5, SDS-HAA6, and SDS-HAA9 the average error was 6 percent. Therefore, for
Water 8, extending the extrapolation from 29 to 61 percent of the run time did not proportionally
increase the mean observed error in the predicted integral breakthrough curve at the end of the
run.
Under the SCA procedure, the error in the extrapolated integral breakthrough curve for any
parameter (other than TOC) is dependent on the error in the extrapolated DI prediction of the
TOC integral breakthrough curve as well as on the difference between the parameter single
contactor curve fits using the entire data set and the truncated data set. This difference typically
increases over the course of the run, and can be especially large in the extrapolated portion of the
curve. However, the error at the end of the single contactor curve may not impact the prediction
of the integral breakthrough curve by the SCA method because the maximum TOC concentration
at the end of the extrapolated integral breakthrough curve is still less than the maximum TOC
concentration at the end of the single contactor curve. This upper bound on TOC limits the
-122-
-------
concentrations of all other parameters to the initial and middle portions of the single contactor
curves when using the SCA procedure to predict integral breakthrough curves.
Conversely, the DI method prediction of the integral breakthrough curve relies on the entire
single contactor breakthrough curve for each parameter. Therefore, the error at the end of the
run due to extrapolation may have a larger impact on integral breakthrough curve predictions by
the DI method than predictions by the SCA procedure. Appendix I summarizes the integral
breakthrough curve predictions after extrapolation based on the DI method for all parameters.
The average difference at the end of the run between DI integral breakthrough curve predictions
with and without extrapolation was 3 percent for Water 5, which is slightly lower than the
average obtained using the SCA method. For Water 8, the average error was slightly higher, 10
percent.
Based on the two runs examined, extrapolation up to 61 percent beyond available experimental
data yielded a mean 9 percent error in the predicted water quality based on the extrapolated
integral breakthrough curve as compared to that based on the complete data set. For a shorter
extrapolation, 21 percent, the mean error was 5 percent. An understanding and acceptance of the
error due to breakthrough curve extrapolation is important because extrapolation may be used in
many cases during the ICR treatment study data analysis to gain additional information from
GAC breakthrough data sets. In particular, data sets that do not exceed a given treatment
objective may be extrapolated by up to 50 percent in an attempt to determine a GAC run time for
the treatment objective. Based on the two waters examined in this study, the error incurred by
extrapolation up to 50 percent of the original run time should average less than 10 percent.
-123-
-------
3.0
2.5 -
2.0 -
o
•^= -I c
CO '•*•
O>
o
o 1.0 H
O
0.5 -
0.0
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.984)
- - Extrapolated logistic function best fit (RA2 = 0.979)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
TOO
EBCT = 20 min.
c0 = 3.08 mg/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure 117 Impact of extrapolation on Dl prediction of the TOC integral
breakthrough curve for Water 5
2.0
1.5 -
O)
'• 1.0 H
o>
o
c
o
O
0.5 -
0.0
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.974)
- - Extrapolated logistic function best fit (RA2 = 0.961)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
TOC
EBCT = 7.2 min.
c0 = 2.02 mg/L
50 100
Scaled operation time (days)
150
200
Figure 118 Impact of extrapolation on Dl prediction of the TOC integral
breakthrough curve for Water 8
-124-
-------
0.035
0.030 -
0.025 -
o
o>
o
0.020 -
0.000 -i
0
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.992)
- - Extrapolated logistic function best fit (RA2 = 0.984)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
UV254
EBCT = 20 min.
c0 = 0.051 1/cm
50 100 150 200 250
Scaled operation time (days)
300
350
Figure 119 Impact of extrapolation on SCA prediction of the UV254 integral
breakthrough curve for Water 5
150
O
125 -
100 -
o 75 -
"CD
CD
£
o
o
50 -
25 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.995)
- - - Extrapolated logistic function best fit (RA2 = 0.99)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-TOX
EBCT = 20 min.
C0 = 205 |jg/L Cl-
50 100 150 200 250
Scaled operation time (days)
300
350
Figure 120 Impact of extrapolation on SCA prediction of the SDS-TOX
integral breakthrough curve for Water 5
-125-
-------
50
40 -
30 -
o
'-4—'
CD
20 -
o
O
10 -
o 4
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.979)
- - Extrapolated logistic function best fit (RA2 = 0.975)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-TTHM
EBCT = 20 min.
C0 = 58 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure 121 Impact of extrapolation on SCA prediction of the SDS-TTHM
integral breakthrough curve for Water 5
20
EBCT = 20 mm.
C0 = 28 ug/L
15 -
o
'•CD 10
CD
O
c
o
O
5 -
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = 0.965)
Extrapolated logistic function best fit (RA2 = 0.959)
Blended effluent
-SCA prediction
• SCA prediction - extrapolated
SDS-HAA5
100
150 200
Scaled operation time (days)
250
300
350
Figure 122 Impact of extrapolation on SCA prediction of the SDS-HAA5
integral breakthrough curve for Water 5
-126-
-------
25
20 -
o
'-4—'
CD
o
O
15 -
10 -
5 -
0 -i
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.976)
- - Extrapolated logistic function best fit (RA2 = 0.971)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-HAA6
EBCT = 20 min.
c0 = 34 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure 123 Impact of extrapolation on SCA prediction of the SDS-HAA6
integral breakthrough curve for Water 5
40
35 -
30 -
I 25-
c
o
'•CD 20 \
CD
*
o
O
15 -
10 -
5 -
0 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.99)
- - Extrapolated logistic function best fit (RA2 = 0.988)
O Blended effluent
SCA prediction
SCA prediction - extrapolated __^j»*» " n
SDS-HAA9
EBCT = 20 min.
C0 = 48 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure 124 Impact of extrapolation on SCA prediction of the SDS-HAA9
integral breakthrough curve for Water 5
-127-
-------
0.025
0.020 -
o
^ 0.015
CD
O
CD
o 0.010
0.005 -
0.000
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = 0.994)
Extrapolated logistic function best fit (RA2 =
Blended effluent
-SCA prediction
• SCA prediction - extrapolated
UV254
EBCT = 7.2min.
C0 = 0.033 1/cm
50 100
Scaled operation time (days)
150
200
Figure 125 Impact of extrapolation on SCA prediction of the UV254 integral
breakthrough curve for Water 8
125
O
o
'-4—'
CD
CD
O
C
O
O
100 -
75 H
50 -
25 -
o -b
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = 0.99)
Extrapolated logistic function best fit (RA2 = 0.98)_
Blended effluent
-SCA prediction
SCA prediction - extrapolated
SDS-TOX
EBCT = 7.2 min.
c0= 156 ug/LCI-
50 100
Scaled operation time (days)
150
200
Figure 126 Impact of extrapolation on SCA prediction of the SDS-TOX
integral breakthrough curve for Water 8
-128-
-------
40
30 -
Single contactor effluent
Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.98)
Extrapolated logistic function best fit (RA2 = 0.966)
Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-TTHM
EBCT = 7.2min.
c0 = 42 ug/L
50 100
Scaled operation time (days)
150
200
Figure 127 Impact of extrapolation on SCA prediction of the SDS-TTHM
integral breakthrough curve for Water 8
16
14 -
12 -
10 -
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = 0.955)
Extrapolated logistic function best fit (RA2 = 0.911)
Blended effluent
-SCA prediction
• SCA prediction - extrapolated
SDS-HAA5
EBCT = 7.2min.
c0 = 23 ug/L
50
100
Scaled operation time (days)
150
200
Figure 128 Impact of extrapolation on SCA prediction of the SDS-HAA5
integral breakthrough curve for Water 8
-129-
-------
20
15 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.956)
- - - Extrapolated logistic function best fit (RA2 = 0.923)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
05
o
'• 10
o>
o
c
o
O
5 H
SDS-HAA6
50 100
Scaled operation time (days)
150
200
Figure 129 Impact of extrapolation on SCA prediction of the SDS-HAA6
integral breakthrough curve for Water 8
20
15 -
o
'•S3 10
o>
o
c
o
O
5 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.952)
- - Extrapolated logistic function best fit (RA2 = 0.924)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
50
SDS-HAA9
EBCT = 7.2min.
c0 = 30 ug/L
100
Scaled operation time (days)
150
200
Figure 130 Impact of extrapolation on SCA prediction of the SDS-HAA9
integral breakthrough curve for Water 8
-130-
-------
5 Summary and Conclusions
The design of this study incorporated two main goals. The primary objectives were to evaluate
the ability of the logistic function to model single contactor breakthrough curve data and to
evaluate the success and limitations of predictive models used to determine the integral
breakthrough curve, a relationship between single contactor run time and blended contactor
water quality. The secondary objective of this study was to evaluate the applicability of these
models and predictive methods in the context of the ICR GAC treatment study data analysis. A
large amount of data will be analyzed: the 62 GAC treatment studies performed will generate
8,000 to 9,000 individual breakthrough curves. Experimental verification was performed on data
from eight bench-scale GAC runs with varying water sources, pretreatments, DBF precursor
concentrations, bromide concentrations, and SDS chlorination conditions. The GAC runs
utilized the rapid small-scale column test (RSSCT) and were performed according to ICR
requirements.
A primary requirement for the ICR GAC treatment study data analysis procedure is to model the
single contactor effluent breakthrough data, for all parameters analyzed, including DBF
surrogates, DBF class sums, and DBF species. A model used to describe single contactor
effluent experimental data is needed for several reasons. From a data management perspective,
best-fit curve parameters that adequately describe experimental data are less memory intensive
than storing the entire experimental data set. A best-fit curve also facilitates interpolation and
extrapolation to estimate run times for given treatment objectives. Use of a best-fit model curve
also provides an estimate of the scatter in the data through the coefficient of determination, and
the model minimizes the impact of this scatter on run time estimates. Finally, a function that
describes the single contactor experimental data set is a prerequisite for determining the integral
breakthrough curve, a curve that relates single contactor run time to blended effluent water
quality under the assumption that contactors are operated in parallel-staggered mode. Run time
estimates generated by the integral breakthrough curve are more applicable to full-scale GAC
operation.
Previous researchers have used various forms of the logistic function to predict and model GAC
breakthrough curve data, and the logistic function was found to be an adequate model in this
study, where it was applied to data from eight GAC runs (160 potential curve fits, but only 126
were performed because concentrations in the GAC effluent were below the MRL for some
parameters). However, poor curve fits occurred for some parameters, especially very sharp "S"
shaped breakthrough curves, and "peak" curves (breakthrough curves for brominated species that
increased and then decreased in concentration over the course of the run due to changes in the
bromide to TOC ratio). To improve the performance of the logistic function for modeling single
contactor data, three enhancements were made, yielding the step, step-lag, and step-lag-peak
logistic function models. The step logistic function model is applicable for GAC breakthrough
curves with relatively high levels of immediate breakthrough. The step-lag logistic function
model incorporates an initial phase where the model output is set to zero, to better fit DBF
breakthrough curves with relatively long initial intervals of effluent concentrations reported as
BMRL, prior to breakthrough. The step-lag-peak logistic function model incorporates a linear
decay term and is applied to "peak" breakthrough curves, which sometimes occur for brominated
DBF species.
-131-
-------
Curve fitting involved determining which model was applicable, and applying it to the
breakthrough curve for each parameter. These enhanced forms of the logistic function model
were able to successfully fit single contactor breakthrough curve data for all parameters. A
method was also developed to detect outlier data points that limited the influence of deviant
observations on the parameter estimates. In general, the application of the three enhanced forms
of the logistic function curve to single contactor effluent breakthrough data was successful, with
a mean R2 of 0.974.
Two predictive approaches for determining the integral breakthrough curve were examined: the
direct integration (DI) and the surrogate correlation approach (SCA). The results of these
procedures were compared to experimental results. Experimental data were obtained by
collecting the entire effluent from each of eight bench-scale GAC experiments in separate
reservoirs and sampling from these reservoirs over time. This experimental procedure simulates
the integral breakthrough curve for an infinite number of contactors operated in parallel-
staggered mode.
The DI procedure applied the average value function to the logistic function model fit of the
experimental single contactor data, and is based on the assumption of an infinite number of
contactors operated in parallel-staggered mode with regular GAC replacement frequencies. For
DBF surrogates (TOC, UV254, and SDS-TOX), the DI procedure yielded excellent results in
comparison to experimental data. For class sums, such as SDS-TTHM and SDS-HAA5,
predictions were usually adequate. However, for individual DBF compounds, especially
brominated species, the DI approach did not always result in accurate predictions. The
inaccuracy of the DI approach for the prediction of some DBF species is problematic since
individual DBFs of potential health concern will be considered during analysis of the treatment
study data.
The SCA procedure was developed to reduce the computational requirements of estimating
blended contactor run times for any given regulatory treatment objective evaluated during the
ICR GAC treatment study data analysis. The SCA procedure relies on the DI method to
establish an integral breakthrough curve for TOC only. The DI method yielded excellent
predictions of the TOC integral breakthrough curve for the waters examined in this study, with a
mean error quantified by the residual sum of squares (RSS) of 0.055 mg/L, and a mean bias of
+0.028 mg/L. Once the TOC integral breakthrough curve is obtained, data points on both the
single contactor and integral breakthrough curves at a constant TOC concentration are mapped,
and all other water quality parameters associated with the single contactor effluent data set at that
TOC concentration are applied to the blended contactor integral breakthrough curve. The SCA
procedure requires that single contactor effluent data be adequately modeled using the logistic
function models. For eight GAC runs and eight water sources, this study showed that the
enhanced logistic function models successfully simulated a wide variety of GAC breakthrough
curve profiles. The SCA procedure inherently relies on the assumption that the relationship
between TOC and the other organic precursors, SDS DBF class sums, and SDS DBF species
established in the single contactor effluent is maintained in the blended contactor effluent. This
study found that this assumption is valid for the eight waters examined. In addition, the
correlation between TOC and bromine incorporation factors for THMs and HAAs was shown to
be consistent between the single contactor effluent and experimental blended effluent.
-132-
-------
Application of the SCA procedure to the experimental GAC breakthrough curves for eight water
sources showed that the overall accuracy of the model was equivalent to the DI method for
predicting the integral breakthrough curve. This analysis was performed by calculating the RSS
and bias between model predictions and experimental data. The cumulative frequency
distribution of the normalized RSS showed that across all waters and all analytes, the prediction
error for the two models was equivalent. Both models were biased negative, indicating a
tendency to underpredict the experimental data. The SCA model had a slightly higher negative
bias than did the DI model. Based on these results, application of the computationally-simpler
SCA procedure to ICR treatment study data is recommended.
A comparative analysis of the two predictive models was also performed for each individual
analyte. The SCA method was more accurate than the DI method when applied to SDS-TTHM,
and was equally accurate when applied to SDS-HAA5, SDS-HAA6, and SDS-HAA9. For the
two waters that examined SDS-TOX, the DI method was a better predictor of experimental
results. For the predominant THM and HAA species (for which comparisons could be made for
six or more of the eight runs), the SCA method outperformed the DI method when applied to
brominated DBF species, with the exception of SDS-DBCM. The DI method generated more
accurate predictions of the non-brominated DBF species (SDS-CF, SDS-DCAA, and SDS-
TCAA). The impact of changing bromide to TOC ratio on the shape of the breakthrough curve
for brominated DBF species, yielding peak curves or very sharp curves followed by a plateau,
typically resulted in underpredictions by the DI method of the observed data. The SCA method
was able to better predict the integral breakthrough curves for the brominated species because it
relied on the relationship between TOC and DBF formation in the single contactor effluent,
which inherently accounts for changing bromide to TOC ratios.
An important limitation of the DI and SCA procedures is that they rely on the assumption of an
infinite number of contactors operated in parallel-staggered mode. Previous work has
maintained that for 10 or more contactors in operation, actual blended effluent run times are
within 10 percent of those estimated based on the infinite contactor assumption. This study
found that the error incurred when applying run time estimates based on the infinite contactor
assumption to run times for finite numbers of contactors is impacted by the number of contactors
and the magnitude of the treatment objective examined in relation to the asymptotic
concentration approached by the single contactor breakthrough curve. Based on the logistic
function model, the infinite contactor assumption will yield estimated run times within 10
percent of actual run times for 13 or more contactors operated in parallel-staggered mode. For
10 contactors on-line, the infinite contactor assumption will yield run time estimates within 12
percent of the actual run times. In all cases, run time estimates based on the infinite contactor
assumption are longer than those for a finite number of contactors, thus providing a best case
scenario for GAC performance.
The applicability of the infinite contactor assumption in this model to finite numbers of
contactors is especially important for small plants operating fewer than 10 contactors on-line.
Also evident was that the largest incremental benefit afforded by operating contactors in parallel-
staggered mode occurs when two contactors are operated as compared to a single contactor. The
benefit realized by adding an additional contactor decreases as the number of contactors on-line
increases. The relationship developed between integral breakthrough curve run time estimates
based on an infinite number of contactors and that based on a finite number of contactors can be
-133-
-------
applied to ICR treatment study data to estimate the performance of large and small GAC
systems.
During the total run time for any given single contactor TOC breakthrough curve, the blended
contactor integral breakthrough curve determined by the DI approach will typically reach
concentrations ranging from 40 to 70 percent of those measured at the end of the single contactor
GAC run. Therefore, when using the SCA procedure, the latter portion of the single contactor
breakthrough curve (consisting of higher surrogate and formed DBF levels) will not usually be
applied to the blended effluent data set. This may be an issue during ICR treatment study data
analysis if a significant number of integral breakthrough curves estimated by the SCA procedure
do not reach regulatory treatment objectives. For these runs, extrapolation of the TOC integral
breakthrough curve would increase the usefulness of the entire data set.
For two waters, experimental evaluation of the sensitivity predicted water quality to
extrapolation of the integral breakthrough curve prediction was performed. For one water,
extrapolation by 21 percent yielded a 3 percent error in the predicted integral TOC concentration
at the end of the run and a mean 5 percent error at the end of the run for all analytes predicted by
the SCA method. For a second water, which was extrapolated by 61 percent, the error in
predicted TOC concentration was 8 percent at the end of the run, and the mean error at the end of
the extrapolated run was 9 percent for all analytes.
An understanding and acceptance of the error due to breakthrough curve extrapolation is
important because extrapolation may be used in many cases during the ICR treatment study data
analysis to gain additional information from GAC breakthrough data sets. In particular, data sets
that do not exceed a given treatment objective may be extrapolated by a reasonable extent (up to
50 percent) in an attempt to determine a GAC run time for the treatment objective. Based on the
two waters examined in this study, the error incurred by extrapolation up to 50 percent of the
original run time should average less than 10 percent. This error may be acceptable for ICR
treatment study data analysis, given the benefit afforded by the extrapolation.
The models used to simulate GAC operation of multiple contactors in parallel-staggered mode
rely on the assumption that the GAC is replaced at regular intervals, so that the service times of
all contactors are equal. They also assume that the breakthrough curve profiles of all single
contactors are identical. In a full-scale plant, this idealized situation will rarely occur.
Variability in source water quality will impact the run time of the contactors, depending on when
they are placed in service and the concentration of DBF precursors in the GAC influent during
their service life. Variability in distribution system conditions, especially temperature, will
impact the contactor service life, as formed DBF levels increase with higher temperatures.
Furthermore, for a plant that operates a fixed number of contactors, water demand changes
during the year may directly impact the EBCT of each contactor. A contactor that is placed on-
line at the beginning of the summer high demand months may be operated under a shorter EBCT
as compared to a contactor placed on-line during the winter. Furthermore, it is less desirable to
remove contactors from service to replace GAC during high demand periods. Another approach
is to increase the number of contactors on-line as water demand increases. These full-scale
issues should be considered when interpreting the results of the ICR treatment study data
analysis.
-134-
-------
6 References
Clark, R.M. 1987. Modeling TOC Removal by GAC: The General Logistic Function. Jour.
AWWA. (79:1:33).
Chowdhury, Z.K., G. Solarik, D.M. Owen, S.M. Hooper, and R.S. Summers. 1996. NOM
Removal by GAC Adsorption: Implications of Blending. In Proc. of the AWWA Annual
Conference, Toronto, Ontario, Canada.
Clark, R.M., J.M. Symons, and J.C. Ireland. Evaluating Field Scale GAC Systems for Drinking
Water. Jour. Environ. Engr. (112:4:744).
Crittenden, J.C., P.S. Reddy, D.W Hand and H. Arora. 1989. Prediction of GAC Performance
Using Rapid Small-Scale Column Tests. AWWARF/AWWA.
Crittenden, J.C., P.S. Reddy, H. Arora, J. Trynoski, D.W. Hand, D.L. Perram, and R.S.
Summers. 1991. Predicting GAC Performance with Rapid Small-Scale Column Tests.
Jour. AWWA. (83:1:77).
Gould, J.P., L.E. Fitchorn, and E. Urheim. 1983. Formation of Brominated Trihalomethanes:
Extent and Kinetics. Water Chlorination: Environmental Impact and Health Effects.
Vol. 4 (R.L. Jolley et al., eds.), Ann Arbor Sci. Publ., Ann Arbor, Michigan.
Hooper, S.M., R.S. Summers, G. Solarik, and S. Hong. 1996. GAC Performance for DBF
Control: Effect of Influent Concentration, Seasonal Variation, and Pretreatment. In
Proc. of the AWWA Annual Conference, Toronto, Ontario, Canada.
Littell, R.C., G.A. Milliken, W.W. Stroup, and R.D. Wolfmger. 1996. SAS System for Mixed
Models. Cary, NC (SAS Institute Inc.)
Oulman, C.S. 1980. The Logistic Curve as a Model for Carbon Bed Design. Jour. AWWA
(72:1:50).
Roberts, P.V. and R.S. Summers. Performance of Granular Activated Carbon for Total Organic
Carbon Removal. Jour. AWWA. (74:2:113).
Shukairy, H.M., Miltner, R.J., and Summers, R.S. 1994. Bromide's Effect on DBF Formation,
Speciation and Control: Part 1, Ozonation. Jour. AWWA. (86:6:72).
Snoeyink, V.L. 1990. Adsorption of Organic Compounds. Water Quality and Treatment: A
Handbook for Community Supplies. 4th ed. (F.W. Pontius, ed.), AWWA and McGraw-
Hill, Inc.
Sontheimer, H., J.C. Crittenden, and R.S. Summers. 1988. Activated Carbon for Water
Treatment. 2nd ed., DVGW-Forschungsstelle, Universitat Karlsruhe, Karlsruhe,
Germany.
Standard Methods for the Examination of Water and Wastewater. 1998. APHA, AWWA, and
WEF. Washington D.C. (20th ed.)
-135-
-------
Summers, R.S., D.M. Owen, Z.K. Chowdhury, S.M. Hooper, G. Solarik, and K. Gray. 1998.
Removal of DBF Precursors by GAC Adsorption. AWWARF and AWWA, Denver,
Colorado.
Summers, R.S., S.M. Hooper, G. Solarik, D.M. Owen, and S. Hong. 1995. Bench-Scale
Evaluation of GAC for NOM Control. Jour. AWWA. (87:8:69).
Summers, R.S., S. Hong, S.M. Hooper, and G. Solarik. 1994. Adsorption of Natural Organic
Matter and Disinfection By-Product Precursors. In Proc. of the AWWA Annual
Conference, New York, NY.
Summers, R.S., M.A. Benz, H.M. Shukairy, and L. Cummings. 1993. Effect of Separation
Processes on the Formation of Brominated THMs. Jour. AWWA. (85:1:88).
Summers, R.S., L. Cummings, J. DeMarco, DJ. Hartman, D.H. Metz, E. Howe, B. MacLeod,
and M. Simpson. 1992. Standardized Protocol for the Evaluation of GAC. AWWARF
and AWWA, Denver, Colorado.
USEPA. 1999. GAC Base Analysis Approach for the ICR Treatment Study Database.
USEPA. 1996. ICR Manual for Bench-and Pilot-Scale Treatment Studies. EPA 814-B-96-003.
Technical Support Division, Office of Ground Water and Drinking Water, Cincinnati,
Ohio.
Westrick, JJ. and J.M. Cohen. 1976. Comparative Effects of Chemical Pretreatment on Carbon
Adsorption.
-136-
-------
Appendix A: Breakthrough Curves Corrected for Impact of Sampling
-137-
-------
4.0
3.5 -
3.0 -
2.5 -
O)
c
o
en
•£ 2.0 H
o>
o
§ 1-5 H
O
O
1.0 -
0.5 -
0.0
EBCT = 15 min
C0 = 4.54 mg/L
0
Single contactor effluent
Blended effluent - experimental
Blended effluent - adjusted for sampling
25 50
Scaled operation time (days)
75
100
Figure A-1 Comparison of blended effluent TOC adjusted for experimental
sampling to blended effluent experimental results for Water 1
2.0
1.5 -
O)
c
o
1°
~ 1.0
§
o
o
O
O
0.5 -
0.0
EBCT = 20 min
c0 = 2.6 mg/L
Single contactor effluent
Blended effluent - experimental
Blended effluent - adjusted for sampling
50
100 150
Scaled operation time (days)
200
250
Figure A-2 Comparison of blended effluent TOC adjusted for experimental
sampling to blended effluent experimental results for Water 2
-138-
-------
2.0
1.5 -
co
£ 1.0
CD
O
c
O
O
O
O
0.5 -
0.0
EBCT = 20 min.
C0 = 2.35 mg/L
Single contactor effluent
Blended effluent - experimental
Blended effluent - adjusted for sampling
50
100 150 200
Scaled operation time (days)
250
300
Figure A-3 Comparison of blended effluent TOC adjusted for experimental
sampling to blended effluent experimental results for Water 3
2.5
2.0 -
1.5 -
CD
O
O
O
1.0 -
0.5 -
0.0
EBCT = 20 min
c0 = 2.98 mg/L
Single contactor effluent
Blended effluent - experimental
Blended effluent - adjusted for sampling
50 100
Scaled operation time (days)
150
200
Figure A-4 Comparison of blended effluent TOC adjusted for experimental
sampling to blended effluent experimental results for Water 4
-139-
-------
3.0
2.5 -
2'0
~ 1.5 -
o>
o
c
o
o
O
O
0.5 -
0.0
EBCT = 20 min.
C0 = 3.08 mg/L
Single contactor effluent
Blended effluent - experimental
Blended effluent - adjusted for sampling
50
100 150 200 250
Scaled operation time (days)
300
350
Figure A-5 Comparison of blended effluent TOC adjusted for experimental
sampling to blended effluent experimental results for Water 5
2.5
2.0 -
§ 1.5 H
CD
O
§ 1.0 H
o
O
O
0.5 -
0.0
EBCT = 20 min.
c0 = 2.64 mg/L
Single contactor effluent
Blended effluent - experimental
Blended effluent - adjusted for sampling
50
100 150 200
Scaled operation time (days)
250
300
Figure A-6 Comparison of blended effluent TOC adjusted for experimental
sampling to blended effluent experimental results for Water 6
-140-
-------
4.5
4.0 -
3.5 -
O)
E 3.0 -
o>
o
o
o
O
O
1.5 -
1.0 -
0.5 -
0.0
EBCT = 20 min
c0 = 5.58 mg/L
0
Single contactor effluent
Blended effluent - experimental
Blended effluent - adjusted for sampling
25
50 75 100
Scaled operation time (days)
125
150
Figure A-7 Comparison of blended effluent TOC adjusted for experimental
sampling to blended effluent experimental results for Water?
2.0
1.5 -
O)
E
1.0 -
0>
o
o
o
O
O
I- 0.5 -
0.0
EBCT = 7.2 min
c0 = 2.02 mg/L
Single contactor effluent
Blended effluent - experimental
Blended effluent - adjusted for sampling
50 100
Scaled operation time (days)
150
200
Figure A-8 Comparison of blended effluent TOC adjusted for experimental
sampling to blended effluent experimental results for Water 8
-141-
-------
This page intentionally left blank.
-142-
-------
Appendix B: SAS Code
SAS code to fit the step-lag-peak logistic model and perform outlier
adjustment.
This program fits data with the following functions:
1. Cs(t) = 0 for t <= tO
2. Cs(t) = AO+A/(l+B*exp(-D*t)) for tO < t <= Tmax
3. Cs(t) = Cmax+S*(t-Tmax) for Tmax < t
where tO = (1/D)*LN(-AO*B/(AO+A)), Cmax = Cs(Tmax),
Tmax is the runtime at which effluent reaches its maximum.
Bounds: -Cmax/2 <= AO <= Cmax/2, 0 < A <= 1.5*Cmax, 0
-------
der.A = (l+BEXP+A*BEXP*D*dA)/BEXP2;
der.B =-A*BEXP*(l/B-D*dB)/BEXP2;
der.D = A*BEXP*(tO+D*dD)/BEXP2;
end;
else if time < &xTmax then do;
BEXP = B*exp(-D*time);
model &Analyte = AO+A/(1+BEXP);
der.AO= 1;
der.A = I/ (1 + BEXP) ;
der.B =-A*BEXP/(B*(1+BEXP)**2);
der.D = A*BEXP*time/(1 + BEXP)**2 ;
end;
else do;
BEXPmax = B*exp(-D*&xTmax);
model &Analyte = AO+A/(1+BEXPmax)+S*(time-&xTmax);
der.AO= 1;
der.A = I/(1+BEXPmax);
der.B =-A*BEXPmax/(B*(1+BEXPmax)**2);
der.D = A*BEXPmax*time/(1+BEXPmax)**2;
der.S = time-ScxTmax;
end;
output out=outp p=pred u95=u95 195=195;
*proc print;
data outp; set outp;
winsor=&Analyte;
delta=(U95-195)/3 ;
low=0;
high=0;
if &Analyte ne . and &Analytepred+delta then do;
outlier=&Analyte; winsor=pred+delta; high=l; end;
keep time &Analyte winsor outlier low high;
proc means data=outp noprint;
var low high;
output out=outout sum=outlow outhigh;
data outout; set outout;
keep outlow outhigh;
*proc print;
PROC NLIN DATA=outp outest=outest NOPRINT;
parms A0=0 A=&xCmax B=10 to 50 by 10 D=0.05 to 0.15 by 0.02 S=0;
bounds -&xCmax05 <= AO <= &xCmax05, 0< A <= &xCmax!5, 0
-------
der.B =-A*BEXP*(l/B-D*dB)/BEXP2;
der.D = A*BEXP*(tO+D*dD)/BEXP2;
end;
else if time < &xTmax then do;
BEXP = B*exp(-D*time);
model &Analyte = AO+A/(1+BEXP);
der.AO= 1;
der.A = I/ (1 + BEXP) ;
der.B =-A*BEXP/(B*(1 + BEXP)**2 );
der.D = A*BEXP*time/(1+BEXP)**2 ,•
end;
else do;
BEXPmax = B*exp(-D*&xTmax);
model &Analyte = AO+A/(1+BEXPmax)+S*(time-&xTmax)
der.AO= 1;
der.A = I/(1+BEXPmax);
der.B =-A*BEXPmax/(B*(1 + BEXPmax)**2) ;
der.D = A*BEXPmax*time/(1+BEXPmax)**2;
der.S = time-ScxTmax;
end;
output out=outq p=pred r=redis u95m=u95m 195m=195m;
*proc print data=outest;
proc reg data=outq outest=r2 adjrsq noprint;
model winsor = pred / noint;
data r2; set r2;
keep _rsq_ _adjrsq_;
data outq; set outq;
keep time pred redis 195m u95m;
data &Analyte; merge outp outq; by time;
if &Analyte=. then missing=pred;
if outlier=. then winsor=.;
if 195m<0 then 195m=0;
RUN;
data std; set outest;
if _type_='COVB1;
retain std_aO std_a std_b std_d std_s;
if _name_='AO' then std_aO=sqrt(aO);
if _name_='A' then std_a=sqrt(a);
if _name_='B' then std_b=sqrt(b);
if _name_='D' then std_d=sqrt(d);
if _name_='S' then std_s=sqrt(s);
if _name_='S';
keep std_aO std_a std_b std_d std_s;
data estimate; set outest;
if _type_='ITER' or _type_='FINAL';
proc sort data=estimate; by _type_;
data estimate; set estimate; by _type_;
if last._type_;
proc means data=estimate noprint;
var _iter_ _sse_ aO a b d s;
output out=outest mean=iter sse aO a b d s;
data &FitInfo; merge outest std r2 outout;
-145-
-------
analyte="&Analyte";
if _freq_=l then converge='NO ';
else converge='YES';
if std_aO=0 or std_a=0 or std_b=0 or std_d=0 or std_s=0 then
singular='YES';
else singular='NO ';
if AO+A/(1+B) >= 0 then tO=0; else t0=(1/D)*log(-AO*B/(AO+A));
keep analyte aO a b d s to std_aO std_a std_b std_d std_s
_rsq_ converge singular outlow outhigh;
run;
%mend;
/*
%FitF012(RawData=Raw, Baslnfo=Bas!nfo, Analyte=BDCM, Fitlnfo=Fit!nfo)
proc print data=Fit!nfo;
run;
*/
-146-
-------
Appendix C: Full- and Bench-Scale Pretreatment Schematics
-147-
-------
Full-Scale
• ^^ ^^ ^^ ^^ ^^
Bench-Scale
Influent
Biscayne Aquifer
Full-Scale Softening
Lime
Activated alumina
Bench-Scale
Cartridge Filtration
Bench-Scale Mixing
Sulfuric acid
Bench-Scale GAC
Adsorption
Figure C-l Full- and bench-scale pretreatment schematic for Water 1
-148-
-------
Full-Scale
^^ ^^ ^^ ^^ ^^ •
Bench-Scale
Influent
Influent
Claricone
(softening/clarification)
Fox River
PAC
Potassium permanganate
• Alum
Cationic polymer
Deep/shallow wells
Bench-Scale
Cartridge Filtration
Bench-Scale Mixing
Bench-Scale GAC
Adsorption
Sulfuric acid or
sodium hydroxide
Figure C-2 Full- and bench-scale pretreatment schematic for Water 2
-149-
-------
Full-Scale
• ^H ^H ^H ^H ^H
Bench-Scale
Influent
Full-Scale Rapid Mix
i
r
Full-Scale Primary
Presedimentation
Full-Scale Secondary
Presedimentation
Full-Scale Rapid Mix
Full-Scale Primary
Sedimentation
Full-Scale Secondary
Sedimentation
Kansas River
Organic polymer
Organic polymer
Alum
CaO
SOC
Bench-Scale
Cartridge Filtration
Bench-Scale Mixing
Bench-Scale GAC
Adsorption
Sulfuric acid
Figure C-3 Full- and bench-scale pretreatment schematic for Water 3
-150-
-------
Full-Scale
^^ ^^ ^^ ^^ ^^ •
Bench-Scale
Influent
Full-Scale Rapid Mix
Full-Scale Solids
Contact Clarification
-I-
Mississippi River
Potassium permanganate
Cationic polymer
Polyaluminum sulfate
Anionic Emulsion Polymer
Bench-Scale
Cartridge Filtration
Bench-Scale GAC
Adsorption
Figure C-4 Full- and bench-scale pretreatment schematic for Water 4
-151-
-------
Influent
Full-Scale Rapid Mix
Lake Dixon, Lake
Wohlford and/or
SDCWA
Alum
Organic polymer
Potassium permanganate
Full-Scale
Flocculation
Full-Scale
Full-Scale
Sedimentation
Bench-Scale
Bench-Scale
Cartridge Filtration
Bench-Scale GAC
Adsorption
Figure C-5 Full- and bench-scale pretreatment schematic for Water 5
-152-
-------
Full-Scale
Influent
Full-Scale Rapid Mix
Full-Scale
Flocculation
Full-Scale
Sedimentation
Edisto River or Bushy
Park Reservoir
Alum
Organic polymer
Bench-Scale
Bench-Scale
Cartridge Filtration
Bench-Scale GAC
Adsorption
Figure C-6 Full- and bench-scale pretreatment schematic for Water 6
-153-
-------
Full-Scale
Influent
Intake Tower
<-
Flash Mix
Flocculation
Sedimentation
Filtration
Sweetwater
Reservoir
Chlorine (shut off when
sampling filter effluent)
• Potassium permanganate
• Ferric chloride
Organic polymer
. Ammonia
PAC
Bench-Scale
Bench-Scale
Cartridge Filtration
Bench-Scale GAC
Adsorption
Figure C-7 Full- and bench-scale pretreatment schematic for Water 7
-154-
-------
Influent
Full-Scale Rapid Mix
i
r
Lake Daniel Reservoir
and Lake Brandt
Reservoir
- Potassium permanganate
Alum
Full-Scale
Flocculation
Full-Scale
• ^^ ^^ ^^ ^^ ^^
Bench-Scale
Full-Scale
Sedimentation
Bench-Scale
Cartridge Filtration
Bench-Scale GAC
Adsorption
Figure C-8 Full- and bench-scale pretreatment schematic for Water 8
-155-
-------
This page intentionally left blank.
-156-
-------
Appendix D: Single Contactor and Blended Effluent DBF Surrogate
and Formed DBF Correlations
-157-
-------
0.07-1
0.06-
0.05 -
f
^ 0.04 -
—
-*•
CN 0.03 -
0.02 -
0.01 -
n nn -
D Single contactor
• Blended effluent n
D
D
D
n
B °
• D
D
j"D
• 'D
75 -i
=d 50-
D)
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
TOC (mg/L)
4.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
TOC (mg/L)
3.5 4.0
oo
30 -
20-
I
CO
Q
CO
150 n
O 100 -
X
O
CO
Q
CO
0-K3-,
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
TOC (mg/L)
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
TOC (mg/L)
Figure D-1 Correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 1
-------
0.04 n
0.03-
8
0.02 -
0.01 -
0.00
100 i
D Single contactor
• Blended effluent
75-
d
D)
50-
0.0 0.5
1.0
TOC (mg/L)
1.5 2.0
o.O
0.5 1.0
TOC (mg/L)
1.5 2.0
30
D)
CD
X
CO
20-
200-1
150-
O
o
CO
50-
-D r-
0.0
0.5 1.0 1.5
TOC (mg/L)
2.0 o.O
0.5 1.0
TOC (mg/L)
1.5 2.0
Figure D-2 Correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 2
-------
0.03 n
8
. 0.02 -
0.01 -
0.00
D Single contactor
• Blended effluent
0.0
0.5 1.0 1.5
TOC (mg/L)
75-
CO
25-
B
2.0 o.O
0.5
1.0
TOC (mg/L)
1.5 2.0
Oi
O
30 -
20-
w
810
0.0 0.5
1.0 1.5
TOC (mg/L)
150-1
0 100-
X
O
Q 50-
S
2.0 0.0
0.5 1.0 1.5
TOC (mg/L)
2.0
Figure D-3 Correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 3
-------
0.04 -
0.03-
1"
_o
^ 0.02 -
8
0.01 -
0 00 -
su -
D Single contactor
• Blended effluent g
-
D 2-
"5)
D 5
D x 25 -
• jz
BDO ° |
• D
• D
D
D
D
D
•
D
D
1 °
M
•D
^ ^
U.UU ' I ' I ' I ' I ' I u -I H-V1 1 1 1 1 1 1 1 1 1
0.0 0.5 1.0 1.5 2.0 2.5 o.O 0.5 1.0 1.5 2.0 2.5
TOC (mg/L) TOC (mg/L)
50 -i
40-
.~~.
^
S 30-
i •
X
W 20-
Q
W
10-
n -
200
n
_ 150
T~
O
^j
D D a
X 10°
O
n W
D Q
D D 50
. -5
0 D
• • ^
D
D
D
D
D
D
D
D
D
D °
n
0.0 0.5 1.0 1.5 2.0 2.5 o.O 0.5 1.0 1.5 2.0 2.5
TOC (mg/L) TOC (mg/L)
Figure D-4 Correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 4
-------
0.04 n
0.03-
8
0.02 -
0.01 -
0.00
D Single contactor
• Blended effluent
0.0 0.5
1.0 1.5 2.0
TOC (mg/L)
50 i
8
2.5 3.0 o.O 0.5
1.0 1.5 2.0
TOC (mg/L)
2.5 3.0
Oi
30 -
20 -
I
w
w
m -
nu
0.0 0.5 1.0 1.5 2.0 2.5
TOC (mg/L)
3.0
150 n
O 100 -
% 50 H
W
n D
0.0 0.5
1.0 1.5 2.0
TOC (mg/L)
2.5 3.0
Figure D-5 Correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 5
-------
ON
0.04 -
0.03-
_o
^ 0.02 -
$
0.01 -
o.c
10 -
1UU •
D Single contactor n
• Blended effluent
75-
D ^.
D j|
D 1 50-
. o t:
CO
D Q
• °
• D
_ D
••'
Kj
D
D
D
1 D D
n
D
%-"*
0.0 0.5 1.0 1.5 2.0 2.5 o.O 0.5 1.0 1.5 2.0 2.5
TOC (mg/L) TOC (mg/L)
40 -|
30-
o o
-------
0.07-1
0.06-
0.05 -
'E
« 0.04-
-*•
tN 0.03 -
0.02 -
0.01 -
0.00 -
D Single contactor
• Blended effluent "-^
D
D
D
•D
• D
B *
i
;r
^
H
w
Q
w
175 -i
150 -
125 -
100-
75-
50-
25 -
n -
D
D D
m a
" *D °
D
M
D
D
•
1
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
TOC (mg/L)
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
TOC (mg/L)
ON
W
50 i
40-
30-
20 -
Q
W
10-
300 -i
250-
O 200-
150 -
W
W
100-
50-
• D
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
TOC (mg/L)
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
TOC (mg/L)
Figure D-7 Correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 7
-------
ON
0.03 -
—. 0.02 -
_o
si
3 0.01 -
o.c
0 -
SU -
D Single contactor
• Blended effluent
D
D 5~
D j|
^
• D X 25 -
i
"D w
• D
• D
.*B °
• . n
D
D°
B
• D
• D
D
" D
m n
0.0 0.5 1.0 1.5 2.0 o.O 0.5 1.0 1.5 2.
TOC (mg/L) TOC (mg/L)
20 -i
"5)
^ 10 -
X
w
Q
w
0 -
0
150
D
D _
B o 100
• D ^
° D g
H
D W
_• Q 50
cr w
D
J
D
D
D
D
D
D
D
D
D
D
n D
0 0.5 1.0 1.5 2.0 o.O 0.5 1.0 1.5 2.
TOC (mg/L)
TOC (mg/L)
Figure D-8 Correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 8
-------
15
10
O
Q
D Single contactor effluent
• Blended effluent
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
TOC (mg/L)
25 •
20-
o 15
Q
m
w
% 10
D D
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
TOC (mg/L)
a\
o\
25 -i
20-
a isH
o
m
§10
Q
W
5 -
j'1
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
TOC (mg/L)
20-i
15-
u. ioH
5-
• e
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
TOC (mg/L)
Figure D-9 THM correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 1
-------
ON
25
20
"S) 15
LJ_
O
g 10
w
5
0
c
D Single contactor effluent D
• Blended effluent 25 •
D
D -2°"
-3
• D 0 15 -
D m
• w
B J 5 -
• BD
•
• D Q
• D D D
D
• D
D
•D
•D
^
).0 0.5 1.0 1.5 2.0 0.0 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
30 -|
25-
0 15-
m
Q
w
Q 10 -
w
5-
0 -
0.(
''I
D
D
• D
• D 5~ 10~
D)
• D ^
D g
• Q
m D « 5-
D
•D
• . . n
•
D ° -D
D D
D
D
'
'•'•'•' 0 -i 1 1 1 1 1 1 1 1
3 0-5 1.0 1.5 2.0 o.O 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
Figure D-10 THM correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 2
-------
1
Oi
OO
i
20
15
1
LL10
CO
Q
CO
5
0
C
-, 40-i
D Single contactor effluent
35-
• Blended effluent
D ^n -
D _ JU
"3> oc .
D 5
0 20 -
Q
m
CO 15-
D CO
10 :
• — • — BI-QMZI — i 1 1 1 1 1 1 o
D
D
D.
•
B •
D
.
).0 0.5 1.0 1.5 2.0 o.O 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
30 -|
25-
S" 20-
D)
^_
0 15-
m
Q
CO
s 1°-
5-
0 -
35-i
B 30 -
D
25 -
LL
D m
• CO 15-
• Q
_ D CO '
•D 10~
-• 1 1 1 1 1 1 1 1 n -
i 5 • • BD D
D
• n o
D f
D
•
0.0
0.5 1.0 1.5
TOC (mg/L)
2.0
0.0
0.5 1.0 1.5
TOC (mg/L)
2.0
Figure D-11 THM correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 3
-------
1
Oi
VO
30
25
2-20
LL 15
W
Q
W 10
5
0
C
4 -
D Single contactor effluent n
• Blended effluent
- ° ^3"
D J? '
D 02-
• s
D W •
D W
• D 1 -
• D
•-» 1 1 1 1 1 1 1 1 1 n
D D
ODD D n
D
i i
• D
D
).0 0.5 1.0 1.5 2.0 2.5 Q.O 0.5 1.0 1.5 2.0 2.5
TOC (mg/L) TOC (mg/L)
15-,
S" 10-
D)
^_
^
O
m
Q
w
S 5"
0 -
-
D
D ^f
a ~s>
D -3
• LL-
D 9
W
H D W
•D"
3 0.5 1.0 1.5 2.0 2.5 Q.O 0.5 1.0 1.5 2.0 2.5
TOC (mg/L) TOC (mg/L)
Figure D-12 THM correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 4
-------
15
O
Q
W 5
D Single contactor effluent
• Blended effluent
o 10 -
Q
m
w -I
5-
D
D D
D D
0.0 0.5
1.0 1.5 2.0
TOC (mg/L)
2.5 3.0 o.O 0.5
1.0 1.5 2.0
TOC (mg/L)
2.5 3.0
20 -
15 -
O 10
m
Q
w
Q
W
5
m
1 -
0.0 0.5
1.0 1.5 2.0
TOC (mg/L)
2-5 3.0 o.O 0.5
1.0 1.5 2.0
TOC (mg/L)
2.5 3.0
Figure D-13 THM correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 5
-------
25
20
"S) 15
LJ_
O
g 10
w
5
n
-, 35-i
D Single contactor effluent n
30 -
• Blended effluent
_ 25 -
D 1?
E 2°"
_ Q
D CO 15-
W
Q
D W 10-
B
m D 5"
. m — rm — n-^n ........ n
D
D
• oo
D
DP
D
•fl
*"
0.0 0.5 1.0 1.5 2.0 2.5 o.O 0.5 1.0 1.5 2.0 2.5
TOC (mg/L) TOC (mg/L)
35 -|
30 -
^ 25-
S
^ 20-
O
m
Q 15-
w
Q
w 10-
5-
0 -
20-i
D
n
15 -
D 5~
D)
-~i
D fe 10"
BW
Q
W
D
5 -
• D'
D
Q •
D D
•
n" D
D
n
m* Q
B
^
0.0 0.5 1.0 1.5 2.0 2.5 o.O 0.5 1.0 1.5 2.0 2.5
TOC (mg/L) TOC (mg/L)
Figure D-14 THM correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 6
-------
to
20
15
„ — ,
— i
D)
^
u- 10
O
w
Q
W
5
o
c
50 -i
45 -
40-
3- 35-
3 30-
0 25-
CQ
§ 20~
Q
W 15 -
10 -
5 -
0 -
-, 75 n
D Single contactor effluent 70 -
n 65 -
• Blended effluent u 6Q
55 •
=d 50-
D)
3 45-
D o
Q 35-
D $ 3° '
S 25'
D 20 •
BD 15
B ^D 10-
B 5 •
).0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 0
TOC (mg/L)
50-i
45 -
D
40 -
35 -
D ,-j.
n -&30-
LJ- 25 -
D 9
_ g 20-
n W
*D 15-
•
D 10 -
• i •. i — i 1 1 1 1 1 1 1 1 1 1 1 1 1 , n -
D
D D
D
I
•-i ^
B
Q
• D
B
D
HB
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
TOC (mg/L)
g
^ n D
D
D
•
D
D
*
s
•
0.0 0.5 1.0
1.5 2.0 2.5 3.0
TOC (mg/L)
3.5 4.0 4.5
0.0 0.5 1.0
1.5 2.0 2.5 3.0
TOC (mg/L)
3.5 4.0 4.5
Figure D-15 THM correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 7
-------
25
20
"S) 15
LJ_
O
g 10
w
5
0
c
D Single contactor effluent D
• Blended effluent 4 .
D _
5
D -1 3 -
D °
. S 3"
• D W
.. o°
•-•-n 1 1 1 1 1 1 1 n
•
• D
D D D D ° D
D
• D
D
).0 0.5 1.0 1.5 2.0 o.O 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
15-,
5~ 10 -
D)
SDS-DBCIV
01
0 -
2 -|
S
• §1
. D u. 1 -
• D m
• D W
D g
• D
.8°
"
D 0.5 1.0 1.5 2.0 o.O 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
Figure D-16 THM correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 8
-------
10
8 -
6 -
o
Q
Q
W
D Single contactor effluent
• Blended effluent
O
w
Q
1 -
-rO-i-
0.0
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
TOC (mg/L)
0.0 0.5 1.0
1.5 2.0 2.5 3.0 3.5 4.0
TOC (mg/L)
8 -
6 -
"5)
f 4-
Q
W
Q
W
2 -
•
0
8 -
D 6 '
D D D =?
D " -
mo m ^ 4-
• o
m
D W .
m D Q
2 -
D
•
• .• i . 1 . 1 . 1 . 1 . 1 . 1 . , n-
D
D
D
D
•
• °
D
• D
H D
D
-•-r^wi rm— , . . . . . . . . . . , .
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
TOC (mg/L)
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
TOC (mg/L)
Figure D-17 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 1
-------
10 -
8 •
3~
"5)
-3 6 -
^
o
Q
Q
W
2 •
0 •
D Single contactor effluent
• Blended effluent D
D ? 2 -
D 5
• D | .
7^
• D 'r
• D Q
w n "
• D
B ^
— • n n ...... n
D
D
Q
B
D
B
D
B
0.0 0.5 1.0 1.5 2.0 o.O 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
8-,
•
•
"5)
^_
1 — '
*4-
m
Q
w
Q
w
2 -
10
D 8
. D n
D D D
— I
_ "I C
H -^
5
• D S
_ D 9 ,
• W 4
Q
D W
D 2
D
D
• D
n
B D
• D
B D
B D
•D °
0.0 0.5 1.0 1.5 2.0 o.O 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
Figure D-18 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 2
-------
ON
8 -
6-
"5)
f«-
Q
W
Q
W
2 -
0 -
0
D Single contactor effluent
• Blended effluent
D
D
D ? 2-
D)
n.
D 0
w
Q -I .
BD W
• • nan M • n i n
D
D
D
D
• ILJBLJ • • IU 1 1 1 1 1 0 | m B-rl_m_l •— • •!_!• — LJ— 1 1 1 1 1
0 0.5 1.0 1.5 2.0 o.O 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
16 -i
14 -
12 -
[IT
]» 10 -
Q
if) 6 -
W
4 -
2 -
0 -
0
10
D
8
D
D -p
D 56
••D i
• w 4
• in w
D
2
H ^ 0
D
D o
D
DB
• •
B H D
0 0.5 1.0 1.5 2.0 0.0 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
Figure D-19 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 3
-------
16 -i
14 •
12 -
"™ 10 •
<^
O
9
W 6 •
Q
W
4 -
2 -
0 •
25
D Single contactor effluent
• Blended effluent D 20
D 5 15
D ^
o
D w 10
• D Q
W
• D
• D 5
D
0
— • •.• ......... n
D
D
D
D o'°
_ " D ™
^ ^ ^n
0.0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 2.5
TOC (mg/L) TOC (mg/L)
5-
"5)
n_
m
Q
w
Q
W
0 -
4 -
? '
"5)
5 3-
D <
D 0 '
S 2-
Q
W
D D
1 -
— • •.• — •. • ru — • . • . n . n . n . n
D
D D
• H
B n
0
^ ^ ^n
0.0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 2.5
TOC (mg/L) TOC (mg/L)
Figure D-20 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 4
-------
oo
8 -
_
6-
3~
S
1 '
£4-
0
Q
W
Q
W
2 -
0 •
o
4 -i
'
3 -
;3"
"o>
3.
1"
Q
W
Q
W
1 -
•
0 -
D Single contactor effluent
• Blended effluent
DD „ 5 -
D S
-^ 4 .
3 .
• °
DO w
• Q
• W p .
g D
D
• • ™" 1 -
0 0.5 1.0 1.5 2.0 2.5 3.0 Q
TOC (mg/L)
8-1
D i—iD
"^~T
J" 1 '
•D 0 4"
en
S •
w
2 .
D
-• — r«— • 1 , 1 , 1 , 1 , 1 , n
D
D
D
D
D
• D
B
B
_
m
a
•D D
H H H~l ^^~
0 0.5 1.0 1.5 2.0 2.5 3.0
TOC (mg/L)
DD
D
D
_ D D
• B
™
•'-'
J
D
• .• — m, ..........
0.0 0.5 1.0 1.5 2.0 2.5
TOC (mg/L)
3.0 0.0 0.5 1.0 1.5 2.0
TOC (mg/L)
2.5 3.0
Figure D-21 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 5
-------
VO
10
8
"5)
-3 6
o
Q
W 4
Q
W
2
0
C
1 6-1
D Single contactor effluent
• Blended effluent D 5 •
? 4-
D)
Jsi
D O
W
Q 2 -
D W
o B D
1 -
D
D
D
" D
• • •_••(• OH D 1 1 1 1 1 1 1 0 -|-B-B_BBIB — LJ-B— LB LJ-1— LJ LJ 1 1 1 1 1
).0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 2.
TOC (mg/L) TOC (mg/L)
10 -i
8 -
3-
"5)
5 6-
m
§ 4-
Q
W
2 -
0 -
10-|
D
• D 8
° D D D ° 0-
1 6
• • m
w 4
D W
• D
• n «
D
D
D
• n
D D
.o"D
Ul 1 ' 1 ' 1 ' 1 ' 1 0 -TB-B-BBLJ 1 1 1 1 1 1 1 1 1
D 0.5 1.0 1.5 2.0 2.5 o.O 0.5 1.0 1.5 2.0 2.
TOC (mg/L) TOC (mg/L)
Figure D-22 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 6
-------
10
8
"5)
-3 6
o
9 4
Q
W
2
o
C
1
oo
0 20 -|
Ij~
| 10-
Q
W
Q
W
5 -
0 -
1 8-1
D Single contactor effluent n
7-
• Blended effluent
6-
D 2~ '
S5'
o o o D ° I4:
: ° .- I3;
• 2-
B1 % 1 -
• i •. , , , , , , , , , , , , , , , , n
).0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 0.
TOC (mg/L)
20
D
D
1
• |
B° < 10
m
D w
• 8
* 5
B
«-• 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 n
D
D D
D
. .
D
D
D 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
TOC (mg/L)
D
D
D D
D D
• "
Q °
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
TOC (mg/L)
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
TOC (mg/L)
Figure D-23 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 7
-------
oo
8 -
6-
3~
^
—
3<-
Q
W
Q
W
2 -
•
0 •
o
D Single contactor effluent
n
• Blended effluent D
5 -
• D ?
D "ra
• — 4-
5 $ .
n o
H 3-
n C/5
• Q
D W 2-
n
^j 1 •
•— •— TB-Bl , , , , , , . Q
0 0.5 1.0 1.5 2.0 o
D
D
• n
D a
D
D
D
CP
D
^ ^ P^K ^ P^K
0 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
2 -|
^^
i
§1
m
Q
w
Q
W
0 -
3]
=d 2 -
D)
n.
m
c/b
Q -i .
w
— •—•_!••-• — TB — m—rm-n HI — IB ,n n— n n . r,
0.0 0.5 1.0 1.5 2.0 o
D D
D °
D
D
D
B
B
D
0 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
Figure D-24 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 8
-------
2 -,
m
o
Q
CO
Q
CO
1 1
D Single contactor effluent
• Blended effluent
in
Q
o
co
Q
co
oo
K> 20-1
2-15"
"5)
n.
l£>
<: 10 •
X
CO
Q
CO
5-
0 -
25-i
D 20-
D
D 3.
D "5)
D D 3 15-
I;
• D Q
CO
• '• i • 1 ' 1 ' 1 ' 1 ' 1 ' 1 ' . 0-
D
D
D D
0°'
• D
• D
D
• .• i . 1 . 1 . 1 . 1 . 1 . 1 . 1
0.0 0.5 1.0 1.5 2.0 2.5
TOC (mg/L)
3.0 3.5 4.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
TOC (mg/L)
Figure D-25 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 1
-------
3n
2-
m
o
Q
co .
Q 1
CO
D Single contactor effluent
• Blended effluent
0.0
0.5
1.0
TOC (mg/L)
1.5
2.0
2-
in
Q
o
co
Q 1
co
D D
-•D WO DH
-rO-O—D D-r-
05 10
TOC (mg/L)
15
2.0
OO
20 -
15 -
10 -
CO
Q
CO
5 -
0.0
D D
0.5 1.0
TOC (mg/L)
1.5
2.0
30 n
25-
20 -
15-
5-
m D
0.0
0.5 1.0
TOC (mg/L)
1.5
2.0
Figure D-26 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 2
-------
oo
3 -
-
I'"
m
o
Q
CO .
Q 1 -
CO
0 -
0
D Single contactor effluent
• Blended effluent
D D J2"
D ^
D D o
D D D • co 1 _
CO
• • r-\m M • • • n i n
D
D
mii-im • • • • u i i i i i 0 1 • — m-n-imu — •— • — m-im — um-i < ILJ < 1
0 0.5 1.0 1.5 2.0 o.O 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
25 -i
20 -
l£>
I 10_
Q
CO
5-
0 -
0
35 -|
30
D
D
?25
5 20
Q)
^f
m "^ 15
D1 0
—
5
H ^ 0
D
D
B
D
,
D
• D IB
D
0 0.5 1.0 1.5 2.0 o.o 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
Figure D-27 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 3
-------
6n
4 -
m
o
Q
82
CO
D Single contactor effluent
• Blended effluent
D D
0.0 0.5
1.0 1.5
TOC (mg/L)
2.0 2.5
in
Q
Q.O 0.5
1.0 1.5
TOC (mg/L)
2.0
2.5
oo
40 -
30 -
D)
n.
l£>
20-
CO
Q
CO
10 -
0.0
0.5
1.0 1.5
TOC (mg/L)
2.0 2.5
CO
50 n
40 -
30 -
20 -
Q.O 0.5
1.0 1.5
TOC (mg/L)
2.0 2.5
Figure D-28 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 4
-------
1
oo
Oi
10
8
3~
•3 6
m
o
9 4
CO
Q
CO
2
n
b -
D Single contactor effluent
• Blended effluent D DD 5"
D _ •
a 4"
• D D ^3-
• O
_•• o
D «o
Q 2 ~
D co
• D
D 1 -
D
• • •
D
D
D DQ
D
• i
• r%
U _ ._ _ . i . i . i . i . i u -|— « _ |_j LB LJ-, , 1 , 1 , 1 , 1
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
TOC (mg/L) TOC (mg/L)
20 -|
15 -
2~
"5)
^
LO
<; 10-
X
CO
Q
CO
5-
0 -
40
35
DD 30
D 2~
n "§> 25
D 3
• ° D | 20
D X
• • co 15
• co
D BD ° 10
• 5
n
-• rQ 1 1 1 1 1 1 1 1 1 1 1 n
DD
D
p.
D
B"
*
D
"D
m
-m fl P
0.0 0.5 1.0 1.5 2.0 2.5
TOC (mg/L)
3.0
0.0 0.5
1.0 1.5 2.0
TOC (mg/L)
2.5 3.0
Figure D-29 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 5
-------
oo
6 -
-
S 4"
^r
m
o
Q
82-
CO
0 -
0
4 -
D Single contactor effluent D
• Blended effluent
D _ 3-
D ~^
a g
D CO
Q
D D co
1 -
D D
D
D D
D
^ •rn^^f^ r-i ^ r^m
0 0.5 1.0 1.5 2.0 2.5 Q.O 0.5 1.0 1.5 2.0 2.5
TOC (mg/L) TOC (mg/L)
25 -i
20 -
2~
l£>
5 1°-
Q
CO
5-
0 -
0
45
40
D
35
?30
D)
D cf 25
" D x 20
D D D W
CO 15
• 10
• n L ^
D
D
D
• D
D D °
•
_ •••"
Ul ' • ,.,.,., 0 -,-_ u, , , , , , , , , ,
0 0.5 1.0 1.5 2.0 2.5 Q.O 0.5 1.0 1.5 2.0 2.5
TOC (mg/L) TOC (mg/L)
Figure D-30 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 6
-------
12 -
10 •
D) 8-
CD 6 -
O
9
CO .
Q 4 •
CO
2 •
0 •
0
i
OO
00 8-1
"
6 -
2~
"5)
n.
^T
m 4 "
I—
CO
Q
CO
2 -
.
0 --•
D Single contactor effluent Q
• Blended effluent
n ?
I
Q
3
• Q
D co
•D
ft m°
m i •. — m — i — i — i — i — i 1 — i — i — i — i — i — i — i , — i
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
JOC (mg/L)
D
D
• •
D
_ D n 3"
D "D 2
D
<
X
CO
Q
CO
D
-i-»i — n — .• i • i — i — i — i — i — i — i — i — i — i — i — i — ,
•
6 -
•
5 -
4 -
3 -
2 -
•
-1 _
i
0.
80
70
60
50
40
30
20
10
n
D D
D
D
D D
D
D '
D
3 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
JOC (mg/L)
D
D
n D
D D
JD g
g
D
D •
•
f
_ •
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
TOC (mg/L)
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
TOC (mg/L)
Figure D-31 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 7
-------
3 -
-
S 2"
m
o
Q
CO .
Q 1 -
CO
0 -
0
i -
D Single contactor effluent
• Blended effluent D
n i
* • rj ~5)
D D <
m
1
CO
• • PB • FB i n
B LB B LB 1 1 1 1 1 1 0 H B-B-LB-BI LB — B ILB-LJ BJ IB TLJ LJ— LJ LJ— " 1
0 0.5 1.0 1.5 2.0 o.O 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
i
oo
^P 151
? 10-
D)
i •
CO
CO 5 "
0 -
0
20
D
D
15
D ^)
B D 3
qj
r°
CO
D Q
B CO
f 5
D
• • FB • L "
D
D
" . D °
D
D
D
D
f
• LB • i i i i i i i o -i — m-m-Lm-m 1 1 1 1 1 1 1
0 0.5 1.0 1.5 2.0 o.o 0.5 1.0 1.5 2.0
TOC (mg/L) TOC (mg/L)
Figure D-32 HAA correlations based on GAC effluent TOC concentration for single contactor and blended effluents for Water 8
-------
75 -,
=d 50
D)
W
g
D Single contactor
• Blended effluent
0.00
0.02 0.04 0.06
UV-254(1/cm)
0.08
VO
o
30 -,
D)
CD
Q
W
20 •
0.00
0.02 0.04 0.06
UV-254(1/cm)
0.08
150-1
0 100-
g
50-
W
oo
0.00
0.02 0.04 0.06
UV-254(1/cm)
0.08
Figure D-33 Correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 1
-------
100
75 -
X 50 •
W
Q
W
25 -
D Single contactor
• Blended effluent
0.00
0.01 0.02 0.03
UV-254(1/cm)
0.04
30 n
D)
CD
X
W
20 •
0.00
0.01 0.02 0.03
UV-254(1/cm)
0.04
X
o
w
100-
50-
0.00
0.01 0.02 0.03
UV-254(1/cm)
0.04
Figure D-34 Correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 2
-------
125
100 -
75
CO
Q
CO
50 -
25 -
D Single contactor
• Blended effluent
0.00
B
0.01 0.02
UV-254(1/cm)
0.03
VO
3U •
5~ 20 •
D)
CD
X
CO
CO 10-
•
n .
IbU -
D
D
a
D _
0 100 -
I
D)
D • O
B B 'r
n Q 50-
• D co
D * °
. ...... n .
B
D
D
D
D
D
D
D
D
D
D
0.00
0.01 0.02
UV-254(1/cm)
0.03
0.00
0.01 0.02
UV-254(1/cm)
0.03
Figure D-35 Correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 3
-------
50 -I
X 25-
w
Q
W
D Single contactor
• Blended effluent
0.00
0.01
0.02
UV-254(1/cm)
0.03
0.04
vo
40 •
330-I
3 1
Q
w
10 -
0.00
D C
rj •
0.01 0.02 0.03
UV-254 (1/cm)
0.04
200 n
150-
O
X
o
w
Q
100 -
50 -
0-K3-
0.00
0.01 0.02 0.03
UV-254 (1/cm)
0.04
Figure D-36 Correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 4
-------
50 -,
X 25 -
W
Q
W
D Single contactor
• Blended effluent
0.00 0.01
0.02 0.03
UV-254(1/cm)
0.04
VO
30
20 -
X
w
Q
W
-in .
1U
™
DD
0.00
0.01 0.02 0.03
UV-254(1/cm)
0.04
150-i
O 100 -
15>
Q 50-
0.00 0.01 0.02 0.03
UV-254(1/cm)
0.04
Figure D-37 Correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 5
-------
100 n
75 -
X 50-
W
Q
W
25-
D Single contactor
• Blended effluent
0.00
0.01 0.02 0.03
UV-254(1/cm)
0.04
VO
30-
5~
O>
CD
< 20-
X
w
Q
W
10-
n -
iiSU •
200-
D _
D °
=d 150-
D)
D 5
X
B D 1— 1 00 •
D D D £
W
• 50-
D
D
D
D
D
D
D B
D
i "
i"
0.00 0.01
0.02 0.03
UV-254(1/cm)
0.04
0.00
0.01 0.02 0.03 0.04
UV-254(1/cm)
Figure D-38 Correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 6
-------
W
Q
175
150 -
125
100 -
75
50
25
D Single contactor
• Blended effluent
0.00 0.01 0.02 0.03 0.04 0.05
UV-254(1/cm)
0.06 0.07
vo
bU •
40 •
•3-
D)
3. 30 •
CD
| '
co 20 •
Q
CO
10 •
.
n .
3UU -
D
D 25°-
D
D 0 200 -
1 . | •
^T 150 -
• s*\
D °
I —
D CO .,,-„-.
Q 100-
• CO
B 50-
i
• n .
D
D
D
D
D
m
ff
m
f
a
, •
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07
UV-254(1/cm)
0.00 0.01 0.02 0.03 0.04 0.05 0.06
UV-254(1/cm)
0.07
Figure D-39 Correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 7
-------
50 -I
X 25 -
W
Q
W
D Single contactor
• Blended effluent
0.00
0.01 0.02
UV-254(1/cm)
0.03
VO
7-1 20-1
3~
"5)
CD
5 10 •
X
w
Q
W
n <
150-i
D
D „
_ 0 100-
• _l
^0 ^-
D D 0
D. ^50-
D " W
B
D
D "
D
D
D
D
D
D
D
D
D
D
D
4
0.00
0.01 0.02
UV-254(1/cm)
0.03
0.00
0.01 0.02
UV-254(1/cm)
0.03
Figure D-40 Correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 8
-------
15 -
5-10-
3
LJ_
o
Q
W 5-
0 i
0.
D Single contactor effluent
• Blended effluent 25 •
D
|2°-
5
0 15 -
Q
0 %
Q 10 -
• D" 5-
• D D
D D
D
D D
D
D
D
D
Bi
•
DO 0.02 0.04 0.06 0.08 o.OO 0.02 0.04 0.06 O.C
UV-254 (1/cm) UV-254 (1/cm)
VO
0° 25-,
20-
;Zr
~3>
3 15-
0
m
§10-
Q
w
5-
0 1
20-i
D
D 15-
2~
"3)
D 5
D " Q •
"
• 5 •
D
D
MB 1 1 1 1 n i
a
'"U
D D
HD " D D
' D
D
D
•
"
i . . . .
0.00
0.02 0.04 0.06
UV-254 (1/cm)
0.08
o.OO
0.02 0.04 0.06
UV-254 (1/cm)
0.08
Figure D-41 THM correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 1
-------
25 -
20 -
"S) 15 -
LJ_
O
g lO-
c/3
5-
0 -
0.
D Single contactor effluent D
• Blended effluent 25 -
D
I20:
• D 0 15 -
D m
• w
•— n — i 1 1 1 1 1 1 1 n
• "D ° D D
D
• D
D
ft
^
DO 0.01 0.02 0.03 0.04 0.00 0.01 0.02 0.03 O.C
UV-254 (1/cm) UV-254 (1/cm)
VO
^P 301
25-
0 15-
m
Q
w
Q 10 -
w
5-
0 -
15-
D
D
• D
• D 5~ 10~
D)
~1
f .
D
• Q
m 0 w 5-
D
-• 1 1 1 1 1 1 1 1 n
•
D °- D
D D
D
D
'
0.00
0.01 0.02 0.03
UV-254 (1/cm)
0.04
o.OO
0.01 0.02 0.03
UV-254 (1/cm)
0.04
Figure D-42 THM correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 2
-------
20
15
1
LL10
w
Q
W
5
0
0
-, 40-i
D Single contactor effluent
35-
• Blended effluent
° a-30:
D 1 25 '
0 20 -
Q
m
W 15 -
D w
10 :
n" " ° 5-
D •"
• • — n — m\ 1 1 1 , 1 Q
D
D
B
D
•
00 0.01 0.02 0.03 o.OO 0.01 0.02 0.03
UV-254(1/cm) UV-254(1/cm)
to
O
0 30-.
25-
3 .
0 15-
m
Q
w
Q 10 -
w
5-
0 -
35-i
H 30 -
D
25 -
3~
n o nri
-3 20 "
LL
D OQ
• (/) 15 -
• Q
. D W '
10;
D B 5 -
g]
• 1 , 1 , 1 , n
Da" . • ° "°
• • n
% Q
D
•
0.00
0.01 0.02
UV-254(1/cm)
0.03
0.00
0.01 0.02
UV-254(1/cm)
0.03
Figure D-43 THM correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 3
-------
30
25 -
20
- 15 H
w
Q
W 10
5 -
0.00
D Single contactor effluent
• Blended effluent
0.01
0.02
UV-254 (1/cm)
0.03
0.04
4 -i
02-
s
8
W
1 -
o.OO
D D
0.01
0.02
UV-254 (1/cm)
0.03
0.04
to
15-,
0
m
Q
w
Q 5-
w
0.00
0.01
0.02
UV-254 (1/cm)
0.03
0.04
1 -
W
o.OO
-CBrOBQ rOB D D-. D—i—
0.01 0.02 0.03
UV-254 (1/cm)
0.04
Figure D-44 THM correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 4
-------
15
O
W
Q
W 5
D Single contactor effluent
• Blended effluent
DD
mS
o •
0.00
0.01 0.02 0.03
UV-254(1/cm)
0.04
O 10 -
Q
m
w
Q
5 -
D D
o.OO
D
D D
0.01 0.02 0.03
UV-254(1/cm)
0.04
to
O
20 -
15 -
O 10
m
Q
w
Q
w
5-
0.00
0.01 0.02 0.03
UV-254(1/cm)
0.04
3-1
LL
m
w
1 -
0.00
0.01 0.02 0.03
UV-254(1/cm)
0.04
Figure D-45 THM correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 5
-------
to
25
20
3~
"S) 15
LJ_
o
g 10
w
5
n
-, 35-i
D Single contactor effluent n
30 -
• Blended effluent
_ 25 -
D S
-20-
O
Q
D op 15 -
th
Q
D W 10-
D
D • 5;
' • "• * "D
.•n-an-ji ....... r.
D
D
° ' D
D
D •
D
••
§"
^
0.00 0.01 0.02 0.03 0.04 0.00 0.01 0.02 0.03 0.04
UV-254 (1/cm) UV-254 (1/cm)
35 -|
30 -
2-25-
;§ 20-
0
m
Q 15-
w
Q
W 1Q .
5-
0 -
20-,
D
D 15-
D 5
D)
D u- 10 -
Bl
(/")
D Q
W
D
5 -
S '
•n ........ n
D D •
D D
•
° '
D
• D
^
0.00 0.01 0.02 0.03 0.04 o.OO 0.01 0.02 0.03 0.04
UV-254 (1/cm) UV-254 (1/cm)
Figure D-46 THM correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 6
-------
20
15
3~
LL10
w
Q
W
5
o
0
to
o
fk 501
45 -
40-
3- 35-
a 30-
0 25-
cn
Q
W 15 -
10 -
5 -
0 -
0.0
-, 75 n
D Single contactor effluent 70 -
n 65 -
• Blended effluent u 60 ;
55 -
i>50'-.
a 45-
i-i ^ 40 -
D o
Q 35-
D C/5
D 20 -
_ rf 10:
• a 5:
00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 Q.
UV-254(1/cm)
50 -|
45 -
D
40-
D 3-35:
n o) 30-
3_
LJ- 25 -
D 9
W 20 -
m W 15;
D 10-
._ 1 * * 5
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 o.
UV-254(1/cm)
D
D D
D
• •
cm
a
DO 0.01 0.02 0.03 0.04 0.05 0.06 0.07
UV-254(1/cm)
[f D
• D D
D
D
D
•
D
•
1
DO 0.01 0.02 0.03 0.04 0.05 0.06 0.07
UV-254(1/cm)
Figure D-47 THM correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 7
-------
25
20 -
D) 15 -
10
w
D Single contactor effluent
• Blended effluent
0.00
0.01 0.02
UV-254 (1/cm)
0
D D
0.03 o.OO
0.01 0.02
UV-254 (1/cm)
0.03
to
15-,
0
m
Q
w
Q 5-
w
0.00
0.01 0.02
UV-254 (1/cm)
2 -i
u. 1 -
o
8
0.03 o.OO
-•D D—iD—D-
0.01 0.02
UV-254 (1/cm)
0.03
Figure D-48 THM correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 8
-------
10 -
8 -
"5)
-3 6 -
o
Q
Q
W
2 •
0 <
0.
D Single contactor effluent
• Blended effluepj D
D
D „
D ^
• 3
i1"
w
Q
. S
" D
D
D
D
30 0.02 0.04 0.06 0.08 o.OO 0.02 0.04 0.06 0.08
UV-254 (1 /cm) UV-254 (1 /cm)
to
O
ON 8 -i
6 -
"5)
I-
Q
W
Q
W
2 -
0 1
0.
8-1
D _6-
D D ° D ?
D " —
a • <: 4-
• o
m
D w
2 -
D
DO
D
D
D
D
•
D
M
BD
0.02 0.04 0.06 0.08 o.OO 0.02 0.04 0.06 0.08
UV-254 (1 /cm) UV-254 (1 /cm)
Figure D-49 HAA correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 1
-------
10 -
8 -
3~
"5)
-3 6 -
o
Q
Q
W
2 •
n .
D Single contactor effluent
• Blended effluent D
D ^ 2-
D ^
5 .
D £
w 1 "
• D
B ^
• n . n . n
D
D
D
• D"
"
0.00 0.01 0.02 0.03 0.04 o.OO 0.01 0.02 0.03 0.04
UV-254 (1 /cm) UV-254 (1 /cm)
to
O
-------
8 -
6-
"5)
O
Q
CO
Q
CO
2 -
0 -
0.
D Single contactor effluent
• Blended effluent
D
D
D 5~2-
D)
D 0
CO
Q 1 .
BD CO '
• — • — •}— D-Bfl — D 1 1 1 1 o
00 0.01 0.02 0.03 QC
D
D
D
D
0 0.01 0.02 0.03
UV-254 (1 /cm) UV-254 (1 /cm)
to
O
00 16-,
14 -
12 -
"5> 10 .
<
m
Q
CO 6-
CO
4 -
2 -
0 -
10
D
8
D
D -j-
"5)
Q 3; 6
. D <:
• °
• D H W
2
•
•• 1 1 1 1 1 , n
D
° Q
D
D
D H
0.00
0.01 0.02
UV-254 (1/cm)
0.03
0.00
0.01 0.02
UV-254 (1/cm)
0.03
Figure D-51 HAA correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 3
-------
16 -I
14 •
12 -
"5> 10 •
^
^ 8
O
9
W 6 •
Q
W
4 -
2 •
0 •
25
D Single contactor effluent
• Blended effluent D 20
D "B)
=1- 15
D ^
O
I
D W 1°
D" Q
(f)
nBD
if 5
•• ........ n
D
rj
D
D
n D"
. ^
ff D
0.00 0.01 0.02 0.03 0.04 o.OO 0.01 0.02 0.03 0.04
UV-254 (1/cm) UV-254 (1/cm)
i
to
O
vp 2-
2-
"5)
^:
5i-
m
Q
w
Q
w
•
0 -
5"
4 -
3- '
"5)
3; 3 -
D D <;
o
m
W 2-
Q
W
D '-'
1 -
D
D D
D D
B
D D
[f "
B
*
0.00 0.01 0.02 0.03 0.04 o.OO 0.01 0.02 0.03 0.04
UV-254 (1/cm) UV-254 (1/cm)
Figure D-52 HAA correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 4
-------
1
to
o
8 -
_
6-
3~
S '
I-
Q
W
Q
W
2 -
0 •
D S\ng\e contactor effluent
• Blended effluent
DD „ 5 -
D J
D ^ •
_ O
• 1- 3 -
D D W
• <~i
J 9 -
•p
_ D
•— n — n. r,
a
D
D
D
D
0
g
B
_
D
BD
0.00 0.01 0.02 0.03 0.04 Q.OO 0.01 0.02 0.03 0.04
UV-254 (1 /cm) UV-254 (1 /cm)
4 -i
3 -
"5)
n.
?2-
m
Q
w
Q
w
-1 .
1
•
0 -
8-1
D 1 D ° ^ 6
_ • • -^
D " JL '
%
m o
• s .
D Q
W
2 -
D
D
D
g D
_
!-•
D •
8
a
0.00 0.01 0.02 0.03 0.04 Q.OO 0.01 0.02 0.03 0.04
UV-254 (1 /cm) UV-254 (1 /cm)
Figure D-53 HAA correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 5
-------
10
8
"5)
-3 6
^
O
Q
W 4
Q
W
2
0
0
b -
D Single contactor effluent
• Blended effluent D 5 -
? 4 -
D)
D n_
<
D 0
W '
Q 2 -
D W
0° '
1 -
D
D
D
" D
00 0.01 0.02 0.03 0.04 o.OO 0.01 0.02 0.03 0.04
UV-254 (1/cm) UV-254 (1 /cm)
10 -i
8 -
2~
1 6-
<
§ 4-
Q
W
2 -
0-
0.0
10
D
• D 8
°D D D ° 3-
1 6
<
• • m
w 4
D W
rf^" 2
D
D
D
- D
D° '
m
D
f
0 0.01 0.02 0.03 0.04 o.OO 0.01 0.02 0.03 0.04
UV-254 (1/cm)
UV-254 (1/cm)
Figure D-54 HAA correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 6
-------
10
8
3~
"5)
-3 6
o
9 4
Q
W
2
o
0
to
to 20-,
^5 -
^
I 10-
Q
W
Q
W
5 -
0-
1 8-1
D Single contactor effluent n
7-
• Blended effluent
6-
D 2~ '
0> 5 -
D ^
D < 4 -
D D D h-
: - .- s3:
• 2-
S J " 1-
•-•n 1 1 1 1 1 1 1 1 1 1 1 1 1 n
00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.C
UV-254 (1/cm)
20
D
D
1
• |
• o 10
m
D w
• Q
w
• 5
a
1 — i 1 1 1 1 1 1 1 1 1 1 1 1 1 n
D
D D
D
••
D
D
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07
UV-254 (1/cm)
D
D
D D
D D
• "
j—
i
0.00 0.01 0.02
0.03 0.04 0.05
UV-254 (1/cm)
0.06 0.07
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07
UV-254 (1/cm)
Figure D-55 HAA correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 7
-------
to
8 -
"
6-
^
I-
Q
W
Q
W
2 -
0
(
D Single contactor effluent
D 6-
• Blended effluent D
• D ?
D "ra
r:
D Pa-
D. Q •
D ' W 2-
o* " 1:
I ^ - - .
D
D
D
D
D
D •
•
D
100 0.01 0.02 0.03 o.oo Q.QI 0.02 0.03
UV-254 (1 /cm) UV-254 (1 /cm)
2 -|
_ _
^
;|
m
Q
w
Q
w
0 )
0.
3]
? 2 -
D)
~
o
m
Q i .
w '
30 0.01 0.02 0.03 o.
D D
D ° °
D
D
D •
D "
D
30 0.01 0.02 0.03
UV-254 (1/cm)
UV-254 (1/cm)
Figure D-56 HAA correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 8
-------
2 -i
m
o
Q
CO
Q
CO
0.00
D Single contactor effluent
• Blended effluent
0.02
0.04
UV-254(1/cm)
0.06
0.08
1 i
CO
Q
o
CO
Q
CO
Q.OO
0.02
0.04
UV-254 (1/cm)
0.06
0.08
to
^u -
15 -
"5)
n.
l£>
<; 1° -
X
CO
Q
CO
5-
0 1
2b -
D 20-
D
D -j-
D "5)
D D 3 15-
• en
^
• x
co 10 -
• O
CO
B
VD 5 -
i"
n , , , , n >
D
D
D D
rj m
D
M
•D
D
!• . . . .
0.00
0.02 0.04 0.06
UV-254 (1/cm)
0.08
0.00
0.02 0.04 0.06
UV-254(1/cm)
0.08
Figure D-57 HAA correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 1
-------
3n
m
o
Q
th .
Q 1
D Single contactor effluent
• Blended effluent
0.00
i-BD—Or—
0.01
0.02
UV-254 (1/cm)
0.03
0.04
g
O
co
Q 1
CO
0 00
D D
0 01
0.02
UV-254 (1/cm)
Q.03
0.04
to
20 -
15 -
10 -
co
Q
CO
5 -
0.00
D D
0.01 0.02 0.03
UV-254 (1/cm)
0.04
30 -i
25-
15-
5-
Q.OO
0.01 0.02 0.03
UV-254 (1/cm)
0.04
Figure D-58 HAA correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 2
-------
to
3 -
-
I'"
m
o
Q
CO
Q 1 -
0 -
0.
D Single contactor effluent
• Blended effluent
D D J2"
D ^
D D o
D D D • W 1 .
CO
• Bl B •• • • n n i n
D
D
mi mi mm \m m i-i i \ i i o -+B — •! — mj — \-i-mm — urn m u — •!_) 1 LJ — i 1
00 0.01 0.02 0.03 Q.OO 0.01 0.02 0.03
UV-254 (1/cm) UV-254 (1 /cm)
25 -i
20 -
Il5-
l£>
I 1Q_
Q
CO
5-
0 -
35 -|
30
D
D
i25
-^ 20
Q)
^f
m "^ 15
m D § 1°
¥1 D H
• 5
•• 1 , 1 , 1 , n
D
D
Q
D
,
D
f\ D B
D
. ......
0.00
0.01 0.02
UV-254 (1/cm)
0.03 Q.OO
0.01 0.02
UV-254(1/cm)
0.03
Figure D-59 HAA correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 3
-------
to
6 -
-
S 4"
m
o
Q
82-
CO
0 -
0.
D Single contactor effluent D
• Blended effluent
D D 5"
D)
n.
n HI
D Q
D 9
D • co
D Q
•• . . n
• ' i ' i ' i ' i 0 1 ••L»T—» — wn_m_i n_r« u i_n u — i u — i 1
00 0.01 0.02 0.03 0.04 Q.OO 0.01 0.02 0.03 0.04
UV-254 (1/cm) UV-254 (1/cm)
40 -|
30 -
"5)
n.
l£>
^ 20 -
X
CO
Q
CO
10 -
0.(
50
D
40
D 1 30
D
Q)
^f
5 20
D Q
D D
• J ' 1°
- D
D
D
D D
-••'
_Q
DO 0.01 0.02 0.03 0.04 Q.OO 0.01 0.02 0.03 0.04
UV-254 (1/cm) UV-254 (1/cm)
Figure D-60 HAA correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 4
-------
10
8
3~
•3 6
m
o
9 4
CO
Q
CO
2
0
0
b -
D Single contactor effluent
• Blended effluent D DD 5-
D _ •
ID 1 3-
G CO
• D
D 1 -
D
• B • . . n
D
D
D DQ
D
• •' 1 ' 1 ' 1 ' 1 U -fc-« — LJ-r-l_l»-l_|— i 1 1 1 1 1 1
00 0.01 0.02 0.03 0.04 0.00 0.01 0.02 0.03 O.C
UV-254 (1/cm) UV-254 (1/cm)
to
00 20 -|
15 -
"5)
n.
l£>
<; 10-
X
CO
Q
CO
5-
0-
0.0
40
35
DD 30
D 2~
D I25
• D | 20
D X
• • co 15
• co
D B D 10
• 5
,'3
0 0.01 0.02 0.03 0.04 n
D
D
D
D
"
• D
• l\
00 0.01 0.02 0.03 O.C
UV-254 (1/cm)
UV-254 (1/cm)
Figure D-61 HAA correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 5
-------
to
6 -
-
"^l*
H 4"
^
m
o
Q
CO _
Q 2 -
CO
n -
4 -
D Single contactor effluent D
• Blended effluent
D _ 3-
i~
"5)
n ^r
^
CO 2 •
n Q
B S •
Q
D D CO
1 -
n n
n
5 D
n
0.00 0.01 0.02 0.03 0.04 Q.OO 0.01 0.02 0.03 0.04
UV-254 (1/cm) UV-254 (1/cm)
25-i 45-1
20 -
-•— -
1
O)
in
<
co 10-
Q
CO
5-
n -
40
D
35
D —
=d 30
D)
n 5
05 25
• D 1 20
CO 15
• 10
-0 ° 5
• 1-1
o
n
n
• D
n
n n D
•
o
U ^1-1 • I • I • I • I u -Vl_l 1 1 1 1 1 1 1 1
0.00 0.01 0.02 0.03 0.04 Q.OO 0.01 0.02 0.03 0.04
UV-254 (1/cm) UV-254 (1/cm)
Figure D-62 HAA correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 6
-------
12 -
10 •
2~
• — Q
D) 0
CQ 6 -
O
9
CO .
Q 4 •
CO
2 •
0 •
D Single contactor effluent g
• Blended effluent
^
D ^
—i
I
Q
5 D %
m Q
n co
§
•_• — » . . . ... . .
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07
UV-254(1/cm)
i
to
to
0 8-,
D
D
6 - " "
^ D
=d D Q 2"
O) • O D)
-3 D ^
erf ^
m 4 " ^
1- X
CO . CO
CO CO
2- D
•
0 -i • • i • B-i 1 1 1 1 — i r-i 1 n
6 -
•
5 -
•
4 -
3 -
2 -
-
1 -
D D
D
D
D D
g
D
D
D
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07
UV-254 (1 /cm)
80
70
60
50
40
30
20
10
n
D
D
_ n
D D
% •
9
D
D •
B "
•
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07
UV-254(1/cm)
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07
UV-254 (1/cm)
Figure D-63 HAA correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 7
-------
3-1
m
o
Q
co .
Q 1
CO
D Single contactor effluent
• Blended effluent D
D „
0.00
0.01 0.02
UV-254(1/cm)
1 i
CO
o
CO
Q
CO
0.03 Q.OO
0.01
-O—iD—D-
0.02
UV-254 (1/cm)
0.03
to
to
15-,
^1 10 H
D)
| \
CO
CO 5 "
0.00
0.01 0.02
UV-254 (1/cm)
20 n
15-
D)
O5
10 -
CO
Q
0.03 Q.OO
0.01 0.02
UV-254(1/cm)
0.03
Figure D-64 HAA correlations based on GAC effluent UV-254 absorbance for single contactor and blended effluents for Water 8
-------
This page intentionally left blank.
-222-
-------
Appendix E: Logistic Function Model Curve Fits
-223-
-------
3 -
o
'•a 2
o>
o
o
O
1 -
TOC
D Single contactor effluent
Logistic function best fit (RA2 = 0.966)
O Blended effluent
- - - - Dl prediction
O
EBCT = 20 min.
c0 = 4.54 mg/L
20 40
Scaled operation time (days)
60
80
Figure E-1 Single contactor and blended effluent TOC breakthrough curves
for Water 1
0.07
0.06 -
0.05 -
o
0.04 -
0.00
UV254
D Single contactor effluent
Logistic function best fit (RA2 = 0.992)
O Blended effluent
- - - - Dl prediction
O. .
O,---'
20 40
Scaled operation time (days)
EBCT = 20 min.
c0 = 0.094 1/cm
60
80
Figure E-2 Single contactor and blended effluent UV254 breakthrough
curves for Water 1
-224-
-------
150
O
o
'-4—'
TO
O>
O
c
o
O
100 -
50 -
SDS-TOX
D Single contactor effluent
Logistic function best fit (RA2 = 0.997)
O Blended effluent
Dl prediction
EBCT = 20 min.
C0 = 224 ug/L Cl-
20 40
Scaled operation time (days)
60
80
Figure E-3 Single contactor and blended effluent SDS-TOX breakthrough
curves for Water 1
20
o>
o
c
o
O
15 -
10 H
5 -
0 +0
SDS-CF
D Single contactor effluent
Logistic function best fit (RA2 = 0.989)
O Blended effluent
- - - - Dl prediction
20
40
Scaled operation time (days)
EBCT = 20 min.
c0 = 34.2 ug/L
60
80
Figure E-4 Single contactor and blended effluent SDS-CF breakthrough
curves for Water 1
-225-
-------
30
25 -
20 -
SDS-BDCM
EBCT = 20 min
C0= 19.3 |jg/L
D Single contactor effluent
Logistic function best fit (RA2 = 0.984)
O Blended effluent
• - - - Dl prediction
20 40 60
Scaled operation time (days)
80
Figure E-5 Single contactor and blended effluent SDS-BDCM breakthrough
curves for Water 1
25
20 -
o
'-4—'
CD
t_
CD
O
c
o
O
15 -
10 -
5 -
SDS-DBCM
D Single contactor effluent
Logistic function best fit (RA2 = 0.994)
O Blended effluent
Dl prediction
40
Scaled operation time (days)
EBCT = 20 min.
c0 = 28 |jg/L
60
80
Figure E-6 Single contactor and blended effluent SDS-DBCM breakthrough
curves for Water 1
-226-
-------
20
SDS-BF
D Single contactor effluent
- Logistic function best fit (RA2 = 0.971)
Blended effluent
20 40
Scaled operation time (days)
EBCT = 20 min.
c0 = 3.7 ug/L
60
80
Figure E-7 Single contactor and blended effluent SDS-BF breakthrough
curves for Water 1
80
60 -
SDS-TTHM
D Single contactor effluent
Logistic function best fit (RA2 = 0.977)
O Blended effluent
- - - - Dl prediction
20 40
Scaled operation time (days)
EBCT = 20 min.
c0 = 85 |jg/L
60
80
Figure E-8 Single contactor and blended effluent SDS-TTHM breakthrough
curves for Water 1
-227-
-------
4 -
3 -
g
'-4—'
CD
O
O
1 -
0 +O
SDS-MCAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
- - - - Dl prediction
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 20 min.
C0 = BMRL
-rm-n-rTTi—n m—m QT-
20 40
—i-CD-
60
Scaled operation time (days)
Figure E-9 Single contactor and blended effluent SDS-MCAA breakthrough
curves for Water 1
80
12
10 -
o
|5 6
CD
O
2 -
0 +0
0
SDS-DCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.986)
O Blended effluent
Dl prediction
P'
.-O"
20 40
Scaled operation time (days)
EBCT = 20 min.
c0= 12.5 ug/L
60
80
Figure E-10 Single contactor and blended effluent SDS-DCAA breakthrough
curves for Water 1
-228-
-------
4 -
O)
•3 3 H
c
g
'-4—'
CD
O
O
1 -
SDS-TCAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
- - - - Dl prediction
Insufficient data measured above the MRL
to perform curve fit analysis
EBCT = 20 min.
c0 = 3 ug/L
0 4o i—rrn-n-TTT|—n m—en QT-
0 20 40 60 80
Scaled operation time (days)
Figure E-11 Single contactor and blended effluent SDS-TCAA breakthrough
curves for Water 1
4 -
§ 2-I
c
o
O
1 -
SDS-MBAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 20 min.
C0= BMRL
0 -K3 1—rm-n-TTTi—n en—rn cp rn i—CD-
0 20 40 60 80
Scaled operation time (days)
Figure E-12 Single contactor and blended effluent SDS-MBAA breakthrough
curves for Water 1
-229-
-------
6 -
SDS-DBAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.976)
O Blended effluent ...
Dl prediction
,O
D
TT
-n
o
EBCT = 20 min.
c0 = 4 |jg/L
20 40
Scaled operation time (days)
60
80
Figure E-13 Single contactor and blended effluent SDS-DBAA breakthrough
curves for Water 1
20
15 -
o
'•S3 10
t_
O>
o
c
o
O
5 -
SDS-HAA5
D Single contactor effluent
Logistic function best fit (RA2 = 0.977)
O Blended effluent
Dl prediction
0 40
0
-O
o.----
EBCT = 20 min.
c0 = 20 |jg/L
20 40
Scaled operation time (days)
60
80
Figure E-14 Single contactor and blended effluent SDS-HAA5 breakthrough
curves for Water 1
-230-
-------
6 -
CD
o
c
o
O
SDS-BCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.981)
O Blended effluent
Dl prediction
EBCT = 20 min.
c0= 7 |jg/L
20 40 60
Scaled operation time (days)
80
Figure E-15 Single contactor and blended effluent SDS-BCAA breakthrough
curves for Water 1
25
20 -
«=* 15 H
c
o
'-4—'
CD
t_
§ 10
c
o
O
5 -
SDS-HAA6
D Single contactor effluent
- Logistic function best fit (RA2 = 0.979)
Blended effluent
Dl prediction /D
,O
o.
20 40
Scaled operation time (days)
EBCT = 20 min.
c0 = 27 |jg/L
60
80
Figure E-16 Single contactor and blended effluent SDS-HAA6 breakthrough
curves for Water 1
-231-
-------
4 -
3 -
g
'-4—'
CD
O
O
1 -
0 +O-
0
SDS-DCBAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
Insufficient data measured above the MRL
to perform curve fit analysis
D
O
EBCT = 20 min.
c0 = 2.3 ug/L
-rm-n-rTTi—n—m—m-
20
40
60
80
Scaled operation time (days)
Figure E-17 Single contactor and blended effluent SDS-DCBAA
breakthrough curves for Water 1
4 -
3 -
O
'-4— '
CD
t_
"c
CD
O
c
o
O
1 -
SDS-CDBAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
Effluent concentrations were not detected
above the MRL for this parameter
0 40
0
-rrn-n-rTTi—n—rn—m QT-
20 40
-CD-
EBCT = 20 min.
c0= BMRL
-r-CD ,
60
80
Scaled operation time (days)
Figure E-18 Single contactor and blended effluent SDS-CDBAA
breakthrough curves for Water 1
-232-
-------
4 -
3 -
g
'-4—'
CD
O
O
1 -
0 +O
SDS-TBAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 20 min.
C0 = BMRL
-rm-n-rTTi—n m—m QT-
20 40
—i-CD-
60
80
Scaled operation time (days)
Figure E-19 Single contactor and blended effluent SDS-TBAA breakthrough
curves for Water 1
25
20 -
«=* 15 H
c
o
'-4—'
CD
t_
§ 10
c
o
O
5 -
SDS-HAA9
D Single contactor effluent
Logistic function best fit (RA2 = 0.982)^
O Blended effluent
Dl prediction
-r—O
.--6
20
40
Scaled operation time (days)
EBCT = 20 min.
c0 = 29 |jg/L
60
80
Figure E-20 Single contactor and blended effluent SDS-HAA9 breakthrough
curves for Water 1
-233-
-------
2.5
2.0 -
.g
"co
o
O
1.5 -
1.0 H
0.5 -
0.0
TOO
D Single contactor effluent
Logistic function best fit (RA2 = 0.972)
O Blended effluent
- - - - Dl prediction
50
EBCT = 20 min.
co =2.6 mg/L
100 150
Scaled operation time (days)
200
250
Figure E-21 Single contactor and blended effluent TOC breakthrough curves
for Water 2
0.040
0.030 -
o
o>
o
c
CD
.Q
O
5
.Q
0.020 -
0.010 -
0.000
UV254
EBCT = 20 min.
c0 = 0.055 1/cm
D Single contactor effluent
Logistic function best fit (RA2 = 0.982)
O Blended effluent
- - - - Dl prediction
50
100 150
Scaled operation time (days)
200
250
Figure E-22 Single contactor and blended effluent UV254 breakthrough
curves for Water 2
-234-
-------
175
150 -
-T 125 -
O
100 -
SDS-TOX
D Single contactor effluent
Logistic function best fit (RA2 = 0.991)
O Blended effluent
Dl prediction
EBCT = 20 min.
c0 =220 ug/L Cl-
50
100 150
Scaled operation time (days)
200
250
Figure E-23 Single contactor and blended effluent SDS-TOX breakthrough
curves for Water 2
25
20 -
15 -
o
'-4—'
TO
SDS-CF
D Single contactor effluent
Logistic function best fit (RA2 = 0.989)
O Blended effluent
- - - - Dl prediction
EBCT = 20 min.
c0 = 41.9 ug/L
50
100 150
Scaled operation time (days)
200
250
Figure E-24 Single contactor and blended effluent SDS-CF breakthrough
curves for Water 2
-235-
-------
30
25 -
O)
o
'•^
CD
20 -
SDS-BDCM
EBCT = 20 min.
c0 =19.8 |jg/L
50
Cr
D Single contactor effluent
Logistic function best fit (RA2 = 0.995)
O Blended effluent
- - - - Dl prediction
100 150
Scaled operation time (days)
200
250
Figure E-25 Single contactor and blended effluent SDS-BDCM breakthrough
curves for Water 2
35
30 -
25 -
20 -
SDS-DBCM
D Single contactor effluent
Logistic function best fit (RA2 = 0.995)
O Blended effluent
Dl prediction
o-
50
EBCT = 20 min.
c0= 31.7 |jg/L
100 150
Scaled operation time (days)
200
250
Figure E-26 Single contactor and blended effluent SDS-DBCM breakthrough
curves for Water 2
-236-
-------
14
12 -
10 -
g
'-4—'
CD
CD
O
c
O
O
6 -
4 -
2 -
SDS-BF
O
EBCT = 20 min.
c0 = 3.7|jg/L
D Single contactor effluent
Logistic function best fit (RA2 = 0.91)
O Blended effluent
• - - - Dl prediction
50
100 150
Scaled operation time (days)
200
250
Figure E-27 Single contactor and blended effluent SDS-BF breakthrough
curves for Water 2
100
75 -
o
'CD 50 H
t_
CD
O
c
o
O
25 -
SDS-TTHM
EBCT = 20 min
c0 = 97 |jg/L
50
D Single contactor effluent
Logistic function best fit (RA2 = 0.992)
O Blended effluent
• - - - Dl prediction
100 150
Scaled operation time (days)
200
250
Figure E-28 Single contactor and blended effluent SDS-TTHM breakthrough
curves for Water 2
-237-
-------
2.0
1.5 -
1-0 H
O>
o
c
o
O
0.5 -
o.o H
SDS-MCAA
EBCT = 20 min.
c0 = BMRL
50
Effluent concentrations were not detected
above the MRL for this parameter
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
- - - - Dl prediction
-CD—i CD 1 Oi
100 150
Scaled operation time (days)
200
250
Figure E-29 Single contactor and blended effluent SDS-MCAA breakthrough
curves for Water 2
10
6 -
SDS-DCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.975)
O Blended effluent
- - - - Dl prediction
EBCT = 20 min.
c0= 14 ug/L
50
100 150
Scaled operation time (days)
200
250
Figure E-30 Single contactor and blended effluent SDS-DCAA breakthrough
curves for Water 2
-238-
-------
3.0
2.5 -
=d 2.0 -
O)
c
g
'•S3 1-5 H
O>
o
§ 1.0 H
0.5 -
0.0 -I
SDS-TCAA
EBCT = 20 min.
c0 =5 |jg/L
50
i—O-
D Single contactor effluent
Logistic function best fit (RA2 = 0.968)
O Blended effluent
- - - - Dl prediction
100 150
Scaled operation time (days)
200
250
Figure E-31 Single contactor and blended effluent SDS-TCAA breakthrough
curves for Water 2
2.0
1.5 -
o
'•§ 1.0 H
-i—•
I
o
O
0.5 -
SDS-MBAA
EBCT = 20 min.
C0= BMRL
50
Effluent concentrations were not detected
above the MRL for this parameter
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
- - - - Dl prediction
-CD—, CE 1 Oi
100 150
Scaled operation time (days)
200
250
Figure E-32 Single contactor and blended effluent SDS-MBAA breakthrough
curves for Water 2
-239-
-------
o
D
-o-
D Single contactor effluent
Logistic function best fit (RA2 = 0.985)
O Blended effluent
Dl prediction
50
100 150
Scaled operation time (days)
200
250
Figure E-33 Single contactor and blended effluent SDS-DBAA breakthrough
curves for Water 2
25
20 -
«=* 15 H
c
o
'-4—'
CD
t_
§ 10
c
o
O
5 -
SDS-HAA5
D Single contactor effluent
Logistic function best fit (RA2 = 0.978)
O Blended effluent
- - - - Dl prediction
EBCT = 20 min.
c0 = 24 |jg/L
50
100 150
Scaled operation time (days)
200
250
Figure E-34 Single contactor and blended effluent SDS-HAA5 breakthrough
curves for Water 2
-240-
-------
10
6 -
g
'-4—'
CD
O
O
2 -
SDS-BCAA
EBCT = 20 min.
c0 =9 Mg/L
D Single contactor effluent
Logistic function best fit (RA2 = 0.977)
O Blended effluent
Dl prediction
50
100 150
Scaled operation time (days)
200
250
Figure E-35 Single contactor and blended effluent SDS-BCAA breakthrough
curves for Water 2
30
25 -
20 -
SDS-HAA6
D Single contactor effluent
Logistic function best fit (RA2 = 0.98)
O Blended effluent
- - - - Dl prediction
EBCT = 20 min.
c0 = 34
50
100 150
Scaled operation time (days)
200
250
Figure E-36 Single contactor and blended effluent SDS-HAA6 breakthrough
curves for Water 2
-241-
-------
3.0
2.5 -
2.0 -
O)
c
o
'•^ 1
CD I-
CD
O
.
0.5 -
o.o H
SDS-DCBAA
EBCT = 20 min.
co =3 |jg/L
50
D
-O
D Single contactor effluent
Logistic function best fit (RA2 = 0.952)
O Blended effluent
Dl prediction
100 150
Scaled operation time (days)
200
250
Figure E-37 Single contactor and blended effluent SDS-DCBAA
breakthrough curves for Water 2
^.u -
2.0 -
^j
co
5 1.5 -
o
'-4— '
CD
§ 1.0-
o
O
0.5 -
-
n n -
SDS-CDBAA 0
EBCT = 20 min.
n n
c0= BMRL
Insufficient data measured above the MRL
to perform curve fit analysis
O
D Single contactor effluent
1 — —j^i: — f. .-.-.i: — ,- UQI_I f\i /f?AO
Lugisiic Tunciion uesi TII ^i\ z
O Blended effluent
Dl prediction
O
- M^
- INr\j
50
100 150
Scaled operation time (days)
200
250
Figure E-38 Single contactor and blended effluent SDS-CDBAA
breakthrough curves for Water 2
-242-
-------
2.0
1.5 -
1-0 H
O>
o
c
o
O
0.5 -
o.o H
SDS-TBAA
EBCT = 20 min.
c0 = BMRL
50
Effluent concentrations were not detected
above the MRL for this parameter
-en
Single contactor effluent
- Logistic function best fit (RA2 = NA)
Blended effluent
• Dl prediction
T CD 1 Oi
100 150
Scaled operation time (days)
200
250
Figure E-39 Single contactor and blended effluent SDS-TBAA breakthrough
curves for Water 2
30
25 -
20 -
SDS-HAA9
EBCT = 20 min
c0 = 37 |jg/L
D Single contactor effluent
Logistic function best fit (RA2 = 0.988)
O Blended effluent
Dl prediction
50
100 150
Scaled operation time (days)
200
250
Figure E-40 Single contactor and blended effluent SDS-HAA9 breakthrough
curves for Water 2
-243-
-------
2.0
1.5 -
o
'•a
CD
o
c
o
O
0.5 -
0.0
TOO
D Single contactor effluent
Logistic function best fit (RA2 = 0.986)
O Blended effluent
- - - - Dl prediction
EBCT = 20 min.
c0 = 2.35 mg/L
50
100 150 200
Scaled operation time (days)
250
300
Figure E-41 Single contactor and blended effluent TOC breakthrough curves
for Water 3
0.030
0.025 -
E 0.020 -
o
o>
o
CD
.Q
0.015 -
0.010 H
0.005 -
0.000
UV254
D Single contactor effluent
Logistic function best fit (RA2 = 0.995)
O Blended effluent
- - - - Dl prediction
50
EBCT = 20 min.
c0 = 0.048 1/cm
100 150 200
Scaled operation time (days)
250
300
Figure E-42 Single contactor and blended effluent UV254 breakthrough
curves for Water 3
-244-
-------
175
150 -
-T 125 -
O
o
'-4—'
TO
100 -
75 -
o>
o
3 so H
25 -
0 -I
SDS-TOX
D Single contactor effluent
Logistic function best fit (RA2 = 0.997)
O Blended effluent
Dl prediction
EBCT = 20 min.
C0 = 255 ug/L Cl-
50
100 150 200
Scaled operation time (days)
250
300
Figure E-43 Single contactor and blended effluent SDS-TOX breakthrough
curves for Water 3
20
15 -
o
'•a 10 H
t_
o>
o
c
o
O
5 -
o ^
SDS-CF
D Single contactor effluent
Logistic function best fit (RA2 = 0.99)
O Blended effluent
- - - - Dl prediction
O'
O
50
100 150 200
Scaled operation time (days)
o
EBCT = 20 min.
c0 = 60.3 ug/L
250
300
Figure E-44 Single contactor and blended effluent SDS-CF breakthrough
curves for Water 3
-245-
-------
50
40 -
30 -
20
c
o
O
10 -
SDS-BDCM
D Single contactor effluent
Logistic function best fit (RA2 = 0.99)
O Blended effluent
Dl prediction
EBCT = 20 min.
C0 = 36.2 ug/L
50
100 150 200
Scaled operation time (days)
250
300
Figure E-45 Single contactor and blended effluent SDS-BDCM breakthrough
curves for Water 3
35
30 -
25 -
c 20
o
'
g 15
o
o
0 10 H
5 -
SDS-DBCM
D Single contactor effluent
Logistic function best fit (RA2 = 0.996)
O Blended effluent
Dl prediction
-' O
o
EBCT = 20 min.
c0 = 44.5 ug/L
50
100 150 200
Scaled operation time (days)
250
300
Figure E-46 Single contactor and blended effluent SDS-DBCM breakthrough
curves for Water 3
-246-
-------
40
30 -
g
'•£ 20 H
O>
o
c
o
O
10 -
SDS-BF
EBCT = 20 min.
C0 = 11.8 |jg/L
D Single contactor effluent
Logistic function best fit (RA2 = 0.984)
O Blended effluent
- - - - Dl prediction
50
100 150 200
Scaled operation time (days)
250
300
Figure E-47 Single contactor and blended effluent SDS-BF breakthrough
curves for Water 3
125
100 -
75 -
§ 50
c
o
O
25 -
SDS-TTHM
D Single contactor effluent
Logistic function best fit (RA2 = 0.986)
O Blended effluent
Dl prediction
EBCT = 20 min.
c0 = 154 ug/L
50
100 150 200
Scaled operation time (days)
250
300
Figure E-48 Single contactor and blended effluent SDS-TTHM breakthrough
curves for Water 3
-247-
-------
2.0
1.5 -
g
'•SS 1-0 H
O>
o
c
o
O
0.5 -
o.o -b-
0
SDS-MCAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
- - - - Dl prediction
50
Effluent concentrations were not detected
above the MRL for this parameter
100 150 200
Scaled operation time (days)
EBCT = 20 min.
C0 = BMRL
HO—
250
300
Figure E-49 Single contactor and blended effluent SDS-MCAA breakthrough
curves for Water 3
6 -
o
'•SS 4H
o>
o
c
o
O
2 -
SDS-DCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.995)
O Blended effluent
Dl prediction
O—O—r-
50
100 150 200
Scaled operation time (days)
EBCT = 20 min.
c0= 15.7ug/L
250
300
Figure E-50 Single contactor and blended effluent SDS-DCAA breakthrough
curves for Water 3
-248-
-------
3 -
O)
I
o
O
1 -
o -b-
o
SDS-TCAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
50
Insufficient data measured above the MRL
to perform curve fit analysis
D
O
EBCT = 20 min.
c0 = 5 ug/L
100 150 200
Scaled operation time (days)
250
300
Figure E-51 Single contactor and blended effluent SDS-TCAA breakthrough
curves for Water 3
2.0
1.5 -
o
'•S3 1-0 H
o>
o
c
o
O
0.5 -
0.0 -D-
0
SDS-MBAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 20 min.
c0= BMRL
50
100 150 200
Scaled operation time (days)
250
300
Figure E-52 Single contactor and blended effluent SDS-MBAA breakthrough
curves for Water 3
-249-
-------
14
12 -
10 -
.g
"co
O>
o
c
o
O
6 -
4 -
2 -
0 -
SDS-DBAA
EBCT = 20 min
c0 = 9.7 |jg/L
D Single contactor effluent
Logistic function best fit (RA2 = 0.859)
O Blended effluent
Dl prediction
50
100 150 200
Scaled operation time (days)
250
300
Figure E-53 Single contactor and blended effluent SDS-DBAA breakthrough
curves for Water 3
30
25 -
§ 20 -
3.
o
'•SS 15 H
o>
o
3 1(H
5 -
0 ^
SDS-HAA5
D Single contactor effluent
Logistic function best fit (RA2 = 0.977)
O Blended effluent
Dl prediction D
D D
-3-
EBCT = 20 min.
c0 = 30 ug/L
50
100 150 200
Scaled operation time (days)
250
300
Figure E-54 Single contactor and blended effluent SDS-HAA5 breakthrough
curves for Water 3
-250-
-------
10
o
O
6 -
4 -
2 -
0 -
SDS-BCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.92
O Blended effluent
- - - - Dl prediction
50
100 150 200
Scaled operation time (days)
EBCT = 20 min.
C0 = 12.7 ug/L
250
300
Figure E-55 Single contactor and blended effluent SDS-BCAA breakthrough
curves for Water 3
35
30 -
25 -
c 20 H
o
% 15 H
o
o
0 10 H
5 -
o ^
SDS-HAA6
D Single contactor effluent
Logistic function best fit (RA2 = 0.963P
O Blended effluent
Dl prediction
O
O
D D
50
100 150 200
Scaled operation time (days)
EBCT = 20 min.
c0 = 43 |jg/L
250
300
Figure E-56 Single contactor and blended effluent SDS-HAA6 breakthrough
curves for Water 3
-251-
-------
2.5
2.0 -
1.5 -
CD
o
c
o
O
0.5 -
0.0 -I
SDS-DCBAA
EBCT = 20 min.
C0 = 4 |jg/L
D Single contactor effluent
Logistic function best fit (RA2 = 0.531)
O Blended effluent
Dl prediction
50
100 150 200
Scaled operation time (days)
250
300
Figure E-57 Single contactor and blended effluent SDS-DCBAA
breakthrough curves for Water 3
3.0
2.5 -
2.0 -
CD 1-5 H
i_
-I—<
cz
CD
O
3 10H
0.5 -
0.0 -D-
0
SDS-CDBAA
EBCT = 20 min.
C0= 2.5|jg/L
50
Insufficient data measured above the MRL
to perform curve fit analysis D
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
100 150 200
Scaled operation time (days)
250
300
Figure E-58 Single contactor and blended effluent SDS-CDBAA
breakthrough curves for Water 3
-252-
-------
2.0
1.5 -
g
'•SS 1-0 H
O>
o
c
o
O
0.5 -
o.o -b-
0
SDS-TBAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
50
Effluent concentrations were not detected
above the MRL for this parameter
100 150 200
Scaled operation time (days)
EBCT = 20 min.
C0 = BMRL
HO—
250
300
Figure E-59 Single contactor and blended effluent SDS-TBAA breakthrough
curves for Water 3
50
40 -
30 -
§ 20
c
o
O
10 -
o ^
SDS-HAA9
D Single contactor effluent
Logistic function best fit (RA2 = 0.965)
O Blended effluent
Dl prediction
50
O
n n
EBCT = 20 min.
c0 = 49 |jg/L
100 150 200
Scaled operation time (days)
250
300
Figure E-60 Single contactor and blended effluent SDS-HAA9 breakthrough
curves for Water 3
-253-
-------
2.5
2.0 -
.g
"co
o
O
1.5 -
1.0 H
0.5 -
0.0
TOO
D Single contactor effluent
Logistic function best fit (RA2 = 0.984)
O Blended effluent
- - - - Dl prediction
EBCT = 20 min.
c0 = 2.98 mg/L
50 100 150
Scaled operation time (days)
200
Figure E-61 Single contactor and blended effluent TOC breakthrough curves
for Water 4
0.040
0.035
0.030
P 0.025
o>
o
CD
.Q
0.020 -
0.015 -
0.010
0.005
0.000
UV254
D Single contactor effluent
Logistic function best fit (RA2 = 0.998)
O Blended effluent
Dl prediction
9--
EBCT = 20 min.
c0 = 0.065 1/cm
50 100 150
Scaled operation time (days)
200
Figure E-62 Single contactor and blended effluent UV254 breakthrough
curves for Water 4
-254-
-------
175
150 -
-T 125 -
O
100 -
SDS-TOX
D Single contactor effluent
Logistic function best fit (RA2 = 0.999)
O Blended effluent
Dl prediction
EBCT = 20 min.
C0 = 288 ug/L Cl-
50 100
Scaled operation time (days)
150
200
Figure E-63 Single contactor and blended effluent SDS-TOX breakthrough
curves for Water 4
30
25 -
20 -
SDS-CF
D Single contactor effluent
Logistic function best fit (RA2 = 0.998)
O Blended effluent
- - - - Dl prediction
EBCT = 20 min.
c0 = 50.7 ug/L
50 100 150
Scaled operation time (days)
200
Figure E-64 Single contactor and blended effluent SDS-CF breakthrough
curves for Water 4
-255
-------
3 -
g
73 2H
O>
o
c
o
O
1 -
SDS-BDCM
EBCT = 20 min
c0 = 2.2 |jg/L
D Single contactor effluent
Logistic function best fit (RA2 = 0.984)
O Blended effluent
Dl prediction
100
Scaled operation time (days)
150
200
Figure E-65 Single contactor and blended effluent SDS-BDCM breakthrough
curves for Water 4
12
10 -
8 -
SDS-DBCM
D Single contactor effluent
Logistic function best fit (RA2 = 0.993)
O Blended effluent
Dl prediction
EBCT = 20 min.
c0 = 15.1 ug/L
50 100
Scaled operation time (days)
150
200
Figure E-66 Single contactor and blended effluent SDS-DBCM breakthrough
curves for Water 4
-256-
-------
20
15 -
g
'•£ 10 H
O>
o
c
o
O
5 -
SDS-BF
0 -fc>
0
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
- - - - Dl prediction
50
Effluent concentrations were not detected
above the MRL for this parameter
100
Scaled operation time (days)
EBCT = 20 min.
C0 = BMRL
150
200
Figure E-67 Single contactor and blended effluent SDS-BF breakthrough
curves for Water 4
50
40 -
30 -
SDS-TTHM
D Single contactor effluent
Logistic function best fit (RA2 = 0.995)
O Blended effluent
- - - - Dl prediction
EBCT = 20 min.
c0 = 68 ug/L
50 100
Scaled operation time (days)
150
200
Figure E-68 Single contactor and blended effluent SDS-TTHM breakthrough
curves for Water 4
-257-
-------
2.0
1.5 -
g
'•£ 1.0 H
O>
o
c
o
O
0.5 -
o.o -k>
0
SDS-MCAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
Effluent concentrations were not detected
above the MRL for this parameter
50
100
Scaled operation time (days)
EBCT = 20 min.
C0 = BMRL
150
200
Figure E-69 Single contactor and blended effluent SDS-MCAA breakthrough
curves for Water 4
15 -
10 -
SDS-DCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.99)
O Blended effluent
Dl prediction
,-O'
EBCT = 20 min.
c0 = 20.3 ug/L
50 100 150
Scaled operation time (days)
200
Figure E-70 Single contactor and blended effluent SDS-DCAA breakthrough
curves for Water 4
-258-
-------
25
20 -
O)
a 15 H
c
.0
I
§ 10
c
o
O
5 -
SDS-TCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.981)
O Blended effluent
Dl prediction
50
EBCT = 20 min.
C0 = 30.7 |jg/L
100
Scaled operation time (days)
150
200
Figure E-71 Single contactor and blended effluent SDS-TCAA breakthrough
curves for Water 4
2.0
1.5 -
o
'•S3 1-0 H
o>
o
c
o
O
0.5 -
0.0 -K>
0
SDS-MBAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
Effluent concentrations were not detected
above the MRL for this parameter
50
100
Scaled operation time (days)
EBCT = 20 min.
c0= BMRL
150
Figure E-72 Single contactor and blended effluent SDS-MBAA breakthrough
curves for Water 4
200
-259-
-------
4 -
3 -
O>
o
c
o
O
1 -
SDS-DBAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
- - - - Dl prediction
Insufficient data measured above the MRL
to perform curve fit analysis
EBCT = 20 min.
CD = 1 ug/L
-CD-
50
100
Scaled operation time (days)
150
200
Figure E-73 Single contactor and blended effluent SDS-DBAA breakthrough
curves for Water 4
40
30 -
o
'•§ 20 H
-H-<
c
O>
o
c
o
O
10 -
SDS-HAA5
D Single contactor effluent
- Logistic function best fit (RA2 = 0.982)
O Blended effluent
Dl prediction
50
EBCT = 20 min.
c0 = 52 ug/L
100 150
Scaled operation time (days)
200
Figure E-74 Single contactor and blended effluent SDS-HAA5 breakthrough
curves for Water 4
-260-
-------
4 -
3 -
O>
O
c
o
O
1 -
SDS-BCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.954)
O Blended effluent
Dl prediction
EBCT = 20 min.
C0 = 4.3 |jg/L
50 100 150
Scaled operation time (days)
200
Figure E-75 Single contactor and blended effluent SDS-BCAA breakthrough
curves for Water 4
50
40 -
30 -
SDS-HAA6
D Single contactor effluent
Logistic function best fit (RA2 = 0.98)
O Blended effluent
Dl prediction
EBCT = 20 min.
c0 = 55 pg/L
50
100
Scaled operation time (days)
150
200
Figure E-76 Single contactor and blended effluent SDS-HAA6 breakthrough
curves for Water 4
-261-
-------
4 -
.0
"co
O>
o
o 2 -
O 1
SDS-DCBAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.983)
O Blended effluent
- - - - Dl prediction
50
O
O
EBCT = 20 min.
C0 = 6.7 ug/L
100 150
Scaled operation time (days)
200
Figure E-77 Single contactor and blended effluent SDS-DCBAA
breakthrough curves for Water 4
1.0
o
'•S3 0.5
o>
o
c
o
O
SDS-CDBAA
o.o -k>
o
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
- - - - Dl prediction
Effluent concentrations were not detected
above the MRL for this parameter
50
100
Scaled operation time (days)
EBCT = 20 min.
c0= BMRL
150
200
Figure E-78 Single contactor and blended effluent SDS-CDBAA
breakthrough curves for Water 4
-262-
-------
2.0
1.5 -
g
'•SS 1-0 H
O>
o
c
o
O
0.5 -
o.o -k>
0
SDS-TBAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
Effluent concentrations were not detected
above the MRL for this parameter
50
100
Scaled operation time (days)
EBCT = 20 min.
C0 = BMRL
150
200
Figure E-79 Single contactor and blended effluent SDS-TBAA breakthrough
curves for Water 4
50
40 -
30 -
o
'-4—'
CD
8 20 H
o
O
10 -
SDS-HAA9
D Single contactor effluent
Logistic function best fit (RA2 = 0.981)
O Blended effluent
Dl prediction
9-
50
o
EBCT = 20 min.
c0 = 61 |jg/L
100
Scaled operation time (days)
150
200
Figure E-80 Single contactor and blended effluent SDS-HAA9 breakthrough
curves for Water 4
-263-
-------
3.0
2.5 -
2.0 -
o
'•a 1-
O>
o
o 1.0 -
O
0.5 -
0.0
TOO
D Single contactor effluent
Logistic function best fit (RA2 = 0.984)
O Blended effluent
- - - - Dl prediction
.O-"
. O'
o-"
-o"
.---O
EBCT = 20 min.
c0 = 3.08 mg/L
50 100 150 200 250
Scaled operation time (days)
300
350
Figure E-81 Single contactor and blended effluent TOC breakthrough curves
for Water 5
0.040
0.035 -
0.030 -
0.025 -
UV254
D Single contactor effluent
Logistic function best fit (RA2 = 0.992)
O Blended effluent
Dl prediction
EBCT = 20 min.
c0 = 0.051 1/cm
50
100 150 200 250
Scaled operation time (days)
300
350
Figure E-82 Single contactor and blended effluent UV254 breakthrough
curves for Water 5
-264-
-------
O
150
125 -
100 -
.9. 75 H
"CD
O>
O
c
o
O
50 -
25 -
SDS-TOX
D Single contactor effluent
Logistic function best fit (RA2 = 0.995)
O Blended effluent
Dl prediction
50
EBCT = 20 min.
c0 = 205 ug/L Cl-
100 150 200 250
Scaled operation time (days)
300
350
Figure E-83 Single contactor and blended effluent SDS-TOX breakthrough
curves for Water 5
12
10 -
'ro 6 H
t_
o>
o
I «H
2 -
SDS-CF
D Single contactor effluent
Logistic function best fit (RA2 = 0.987)
O Blended effluent
- - - - Dl prediction
,-O'
.--O
EBCT = 20 min.
c0 = 23.7 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure E-84 Single contactor and blended effluent SDS-CF breakthrough
curves for Water 5
-265-
-------
20
15 -
10 H
O>
o
c
o
O
5 -
SDS-BDCM
D Single contactor effluent
Logistic function best fit (RA2 = 0.97)
O Blended effluent
- - - - Dl prediction
EBCT = 20 min.
C0 = 10.8 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure E-85 Single contactor and blended effluent SDS-BDCM breakthrough
curves for Water 5
20
15 -
o
'•S3 10 H
t_
o>
o
c
o
O
5 -
SDS-DBCM
D Single contactor effluent
Logistic function best fit (RA2 = 0.988)
O Blended effluent
Dl prediction
EBCT = 20 min.
c0 = 22.7 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure E-86 Single contactor and blended effluent SDS-DBCM breakthrough
curves for Water 5
-266-
-------
2 -
SDS-BF
D Single contactor effluent
Logistic function best fit (RA2 = 0.969)
O Blended effluent
n n
Dl prediction
O
ft
D
n
O
-O
EBCT = 20 min.
C0= 1.2|jg/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure E-87 Single contactor and blended effluent SDS-BF breakthrough
curves for Water 5
50
40 -
30 -
o
'-4—'
CD
t_
§ 20
c
o
O
10 -
SDS-TTHM
D Single contactor effluent
Logistic function best fit (RA2 = 0.979)
O Blended effluent
Dl prediction
O
EBCT = 20 min.
c0 = 58 |jg/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure E-88 Single contactor and blended effluent SDS-TTHM breakthrough
curves for Water 5
-267-
-------
4 -
3 -
g
'-4—'
CD
O
O
1 -
0 -O-
0
SDS-MCAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
Insufficient data measured above the MRL
to perform curve fit analysis
D D
50
-r-O rO-
100 150 200 250
Scaled operation time (days)
EBCT = 20 min.
c0 = 2 ug/L
300
350
Figure E-89 Single contactor and blended effluent SDS-MCAA breakthrough
curves for Water 5
6 -
CD
O
c
o
O
2 -
0 -(
SDS-DCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.978)
O Blended effluent
- - - - Dl prediction
EBCT = 20 min.
c0= 10.3 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure E-90 Single contactor and blended effluent SDS-DCAA breakthrough
curves for Water 5
-268-
-------
6 -
SDS-TCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.989)
O Blended effluent
Dl prediction
EBCT = 20 min.
C0 = 12.7 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure E-91 Single contactor and blended effluent SDS-TCAA breakthrough
curves for Water 5
2.0
1.5 -
O
'•§ 1.0 H
-i—•
I
O
O
0.5 -
0.0 -O-
0
SDS-MBAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
- - - - Dl prediction
Effluent concentrations were not detected
above the MRL for this parameter
50
100 150 200 250
Scaled operation time (days)
EBCT = 20 min.
c0 = 1 ug/L
300
350
Figure E-92 Single contactor and blended effluent SDS-MBAA breakthrough
curves for Water 5
-269-
-------
3 -
o
'•^
CO
O>
O
c
o
O
1 -
SDS-DBAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.992)
O Blended effluent
Dl prediction
O
50
D
O
EBCT = 20 min.
c0 = 2 |jg/L
100 150 200 250
Scaled operation time (days)
300
350
Figure E-93 Single contactor and blended effluent SDS-DBAA breakthrough
curves for Water 5
20
EBCT = 20 min
c0 = 28 ug/L
15 -
o
'•S3 10
t_
o>
o
c
o
O
5 -
SDS-HAA5
D Single contactor effluent
Logistic function best fit (RA2 = 0.965)
O Blended effluent
- - - - Dl prediction
150 200 250
Scaled operation time (days)
300
350
Figure E-94 Single contactor and blended effluent SDS-HAA5 breakthrough
curves for Water 5
-270-
-------
6 -
CD
o
c
o
O
2 -
0 •{
SDS-BCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.987)
O Blended effluent
Dl prediction
EBCT = 20 min.
C0 = 7.3 |jg/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure E-95 Single contactor and blended effluent SDS-BCAA breakthrough
curves for Water 5
25
20 -
«=* 15 H
c
o
'-4—'
CD
t_
§ 10 H
c
o
O
5 -
SDS-HAA6
D Single contactor effluent
Logistic function best fit (RA2 = 0.976)
O Blended effluent
- - - - Dl prediction
EBCT = 20 min.
c0 = 34 |jg/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure E-96 Single contactor and blended effluent SDS-HAA6 breakthrough
curves for Water 5
-271-
-------
10
g
'-4—'
CD
CD
O
c
O
O
6 -
4 -
2 -
SDS-DCBAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.987)
O Blended effluent
Dl prediction
EBCT = 20 min.
C0 = 10.7 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure E-97 Single contactor and blended effluent SDS-DCBAA
breakthrough curves for Water 5
5 -
O
•^
CD
t_
CD
O
1 -
0 -{
SDS-CDBAA
EBCT = 20 min.
c0 = 3.7 |jg/L
50
D Single contactor effluent
Logistic function best fit (RA2 = 0.95)
O Blended effluent
- - - - Dl prediction
100 150 200 250
Scaled operation time (days)
300
350
Figure E-98 Single contactor and blended effluent SDS-CDBAA
breakthrough curves for Water 5
-272-
-------
2.0
1.5 -
1.0 -
O>
o
c
o
O
0.5 -
0.0 -O-
0
SDS-TBAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
50
Effluent concentrations were not detected
above the MRL for this parameter
100 150 200 250
Scaled operation time (days)
EBCT = 20 min.
C0 = BMRL
300
350
Figure E-99 Single contactor and blended effluent SDS-TBAA breakthrough
curves for Water 5
40
30 -
o
'•S3 20 H
t_
o>
o
c
o
O
10 -
0 -O-
0
SDS-HAA9
EBCT = 20 min
c0 = 48 |jg/L
50
D Single contactor effluent
Logistic function best fit (RA2 = 0.99)
O Blended effluent
- - - - Dl prediction
100 150 200 250
Scaled operation time (days)
300
350
Figure E-100 Single contactor and blended effluent SDS-HAA9 breakthrough
curves for Water 5
-273-
-------
2.5
2.0 -
.g
"co
o
O
1.5 -
1.0 H
0.5 -
0.0
TOO
D Single contactor effluent
Logistic function best fit (RA2 = 0.992)
O Blended effluent
- - - - Dl prediction
50
100 150 200
Scaled operation time (days)
EBCT = 20 min.
c0 = 2.64 mg/L
250
300
Figure E-101 Single contactor and blended effluent TOC breakthrough
curves for Water 6
0.040
0.035 -
0.030 -
0.025 -
o>
o
UV254
D Single contactor effluent
Logistic function best fit (RA2 = 0.998)
O Blended effluent
- - - - Dl prediction
50
O
o---
O-'
EBCT = 20 min.
c0 = 0.06 1/cm
100 150 200
Scaled operation time (days)
250
300
Figure E-102 Single contactor and blended effluent UV254 breakthrough
curves for Water 6
-274-
-------
225
200
175
O
150 -
O)
^ 125
c
o
E 100
O>
o
c
o
O
75 -
50
25
0
SDS-TOX
D Single contactor effluent
Logistic function best fit (RA2 = 0.997)
O Blended effluent
Dl prediction
50
.0-'
100 150 200
Scaled operation time (days)
.o
EBCT = 20 min.
C0 = 305 ug/L Cl-
250
300
Figure E-103 Single contactor and blended effluent SDS-TOX breakthrough
curves for Water 6
25
20 -
^ 15 H
c
o
'-4—'
TO
t_
§ 10 H
c
o
O
5 -
o ^
SDS-CF
D Single contactor effluent
Logistic function best fit (RA2 = 0.991)
O Blended effluent
- - - - Dl prediction
EBCT = 20 min.
c0 = 55.3 ug/L
50
100 150 200
Scaled operation time (days)
250
300
Figure E-104 Single contactor and blended effluent SDS-CF breakthrough
curves for Water 6
-275-
-------
35
30 -
25 -
c 20 H
o
g 15 H
o
10 -
5 -
o ^
SDS-BDCM
EBCT = 20 min.
C0 = 27.4 |jg/L
50
O
O .'
o-
o-
D Single contactor effluent
Logistic function best fit (RA2 = 0.979)
O Blended effluent
Dl prediction
100 150 200
Scaled operation time (days)
250
300
Figure E-105 Single contactor and blended effluent SDS-BDCM
breakthrough curves for Water 6
35
30 -
25 -
c 20 H
o
o
o
0 10 H
5 -
o ^
SDS-DBCM
D Single contactor effluent
Logistic function best fit (RA2 = 0.993)
O Blended effluent
Dl prediction
EBCT = 20 min.
c0= 41.6ug/L
50
100 150 200
Scaled operation time (days)
250
300
Figure E-106 Single contactor and blended effluent SDS-DBCM
breakthrough curves for Water 6
-276-
-------
20
15 -
g
'• 10
O>
o
c
o
O
5 -
SDS-BF
EBCT = 20 min.
C0 = 3.3 |jg/L
0 H
0
50
O
D Single contactor effluent
Logistic function best fit (RA2 = 0.978)
O Blended effluent
Dl prediction
100 150 200
Scaled operation time (days)
250
300
Figure E-107 Single contactor and blended effluent SDS-BF breakthrough
curves for Water 6
100
75 -
o
'CD 50 H
o>
o
c
o
O
25 -
0 -I
SDS-TTHM
D Single contactor effluent
Logistic function best fit (RA2 = 0.98)
O Blended effluent
Dl prediction
50
EBCT = 20 min.
c0 = 128 ug/L
100 150 200
Scaled operation time (days)
250
300
Figure E-108 Single contactor and blended effluent SDS-TTHM breakthrough
curves for Water 6
-277-
-------
2.0
1.5 -
g
'•S3 1-0 H
CD
o
c
o
O
0.5 -
SDS-MCAA
D Single contactor effluent
0.0 -t>
0
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
50
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 20 min.
C0 = BMRL
100 150 200
Scaled operation time (days)
250
300
Figure E-109 Single contactor and blended effluent SDS-MCAA
breakthrough curves for Water 6
10
O)
3.
c
o
'-4—'
CD
t_
"£
CD
o
C
O
O
6 -
4 -
2 -
0 -I
SDS-DCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.976)
O Blended effluent
Dl prediction
EBCT = 20 min.
c0= 17.3ug/L
T-O-
50
100 150 200
Scaled operation time (days)
250
300
Figure E-110 Single contactor and blended effluent SDS-DCAA breakthrough
curves for Water 6
-278-
-------
3.0
2.5 -
2.0 -
g
'•SS 1-5 H
O>
o
O
1 0 -
LU 1
0.5 -
0.0 -I
SDS-TCAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
Insufficient data measured above the MRL
to perform curve fit analysis
O
EBCT = 20 min.
C0 = 13.3 ug/L
50
-D Ol—D—D-Oh-ID—Oh—
100 150
200
250
300
Scaled operation time (days)
Figure E-111 Single contactor and blended effluent SDS-TCAA breakthrough
curves for Water 6
2.0
1.5 -
o
'•§ 1.0
-i—•
I
o
O
0.5 -
o.o -b-
0
SDS-MBAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 20 min.
c0= BMRL
50
100 150 200
Scaled operation time (days)
250
300
Figure E-112 Single contactor and blended effluent SDS-MBAA
breakthrough curves for Water 6
-279-
-------
10
c
CD
O
c
o
O
6 -
4 -
2 -
0 -
SDS-DBAA
EBCT = 20 min.
c0 = 5.7 |jg/L
O
o 9,-'
D Single contactor effluent
Logistic function best fit (RA2 = 0.952)
O Blended effluent
- - - - Dl prediction
50
100 150 200
Scaled operation time (days)
250
300
Figure E-113 Single contactor and blended effluent SDS-DBAA breakthrough
curves for Water 6
25
20 -
15 -
o
'-4—'
CD
8 10
c
o
O
5 -
0 -I
SDS-HAA5
D Single contactor effluent
Logistic function best fit (RA2 = 0.969)
O Blended effluent
Dl prediction
D D
"Q O.O-D-"
50
.-O"
EBCT = 20 min.
c0 = 36 ug/L
100 150 200
Scaled operation time (days)
250
300
Figure E-114 Single contactor and blended effluent SDS-HAA5 breakthrough
curves for Water 6
-280-
-------
10
6 -
I 4
o
O
2 -
0 -t>
0
SDS-BCAA
D Single contactor effluent
- Logistic function best fit (RA2 = 0.967)
Blended effluent
Dl prediction
o-
50
100 150 200
Scaled operation time (days)
EBCT = 20 min.
c0 = 12.3 ug/L
250
300
Figure E-115 Single contactor and blended effluent SDS-BCAA breakthrough
curves for Water 6
35
30 -
25 -
20 -
o
'-4—'
CD
CD
O
C
O
O
15 -
10 -
5 -
0 -I
SDS-HAA6
D Single contactor effluent
Logistic function best fit (RA2 = 0.97)
O Blended effluent
Dl prediction
EBCT = 20 min.
CD = 49 ug/L
50
100 150 200
Scaled operation time (days)
250
300
Figure E-116 Single contactor and blended effluent SDS-HAA6 breakthrough
curves for Water 6
-281-
-------
5 -
4 -
O)
g
I 3
"c
o>
o
c
o 2
O ^
1 -
SDS-DCBAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.975)
O Blended effluent
Dl prediction
0 -t>
0
i-O
50
100 150 200
Scaled operation time (days)
EBCT = 20 min.
c0 = 9 |jg/L
250
300
Figure E-117 Single contactor and blended effluent SDS-DCBAA
breakthrough curves for Water 6
3.5
3.0 -
2.5 -
2.0 -
o
'-4—'
CD
CD
O
C
O
O
1.5 -
1.0 -
0.5 -
0.0 -I
SDS-CDBAA
EBCT = 20 min.
c0 = 3 |jg/L
50
D D
D Single contactor effluent
Logistic function best fit (RA2 = 0.913)
O Blended effluent
Dl prediction
100 150 200
Scaled operation time (days)
250
300
Figure E-118 Single contactor and blended effluent SDS-CDBAA
breakthrough curves for Water 6
-282-
-------
2.0
1.5 -
g
'•S3 1-0 H
CD
o
c
o
O
0.5 -
SDS-TBAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
0.0 -t>
0
50
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 20 min.
C0 = BMRL
100 150 200
Scaled operation time (days)
250
300
Figure E-119 Single contactor and blended effluent SDS-TBAA breakthrough
curves for Water 6
45
40 -
35 -
§ 30 -
3.
I 25-
H—<
CD
£ 20 H
CD
O
O 1£
10 -
5 -
0 -
SDS-HAA9
D Single contactor effluent
- Logistic function best fit (RA2 = 0.966)
Blended effluent
• Dl prediction
D D
O
EBCT = 20 min.
c0 = 61 |jg/L
50
100 150 200
Scaled operation time (days)
250
300
Figure E-120 Single contactor and blended effluent SDS-HAA9 breakthrough
curves for Water 6
-283-
-------
4 -
3 -
.g
"co
o> 9
o *-
c
o
O
1 -
TOO
D Single contactor effluent
Logistic function best fit (RA2 = 0.979)
O Blended effluent
- - - - Dl prediction
EBCT = 20 min.
c0 = 5.58 mg/L
25
50 75 100
Scaled operation time (days)
125
150
Figure E-121 Single contactor and blended effluent TOC breakthrough
curves for Water 7
0.07
0.06 -
0.05 -
.o
^ 0.04 -
o>
o
c
CD
•9 0.03 -
o
0.02 H
0.01 -
0.00
UV254
D Single contactor effluent
Logistic function best fit (RA2 = 0.994)
O Blended effluent
- - - - Dl prediction
O.-
25
EBCT = 20 min.
c0 = 0.109 1/cm
50 75 100
Scaled operation time (days)
125
150
Figure E-122 Single contactor and blended effluent UV254 breakthrough
curves for Water 7
-284-
-------
300
250 -
O
200 -
O)
3.
.1 150 H
-I—'
TO
O>
o
c
o
O
100 -
50 -
o H
SDS-TOX
D Single contactor effluent
Logistic function best fit (RA2 = 0.994)
O Blended effluent
Dl prediction
O
25
O,---
50 75 100
Scaled operation time (days)
EBCT = 20 min.
C0 = 486 ug/L Cl-
125
150
Figure E-123 Single contactor and blended effluent SDS-TOX breakthrough
curves for Water 7
20
15 -
o
'•a 10 H
t_
o>
o
c
o
O
5 -
o ^
SDS-CF
D Single contactor effluent
Logistic function best fit (RA2 = 0.996)
O Blended effluent
- - - - Dl prediction
EBCT = 20 min
c0 = 52.4 ug/L
25
50 75 100
Scaled operation time (days)
125
150
Figure E-124 Single contactor and blended effluent SDS-CF breakthrough
curves for Water 7
-285-
-------
80
70
60
50
40 -
CD
c
o
O
30 -
20
10
0
SDS-BDCM
D Single contactor effluent
Logistic function best fit (RA2 = 0.991)
O Blended effluent
Dl prediction
25
50 75 100
Scaled operation time (days)
EBCT = 20 min.
C0 = 65.3 |jg/L
125
150
Figure E-125 Single contactor and blended effluent SDS-BDCM
breakthrough curves for Water 7
50
40 -
«=* 30 H
c
o
'-4—'
CD
t_
§ 20 H
c
o
O
10 -
o ^
SDS-DBCM
D
Single contactor effluent
Logistic function best fit (RA2 = 0.996)
O Blended effluent
- - - - Dl prediction
25
o-'
50 75 100
Scaled operation time (days)
' o
EBCT = 20 min.
c0 = 66.2 |jg/L
125
150
Figure E-126 Single contactor and blended effluent SDS-DBCM
breakthrough curves for Water 7
-286-
-------
60
50 -
O)
40 -
30 -
CD
O
o 20 J
10 -
SDS-BF
D Single contactor effluent
Logistic function best fit (RA2 = 0.971)
O Blended effluent
Dl prediction
O
O
O
O
EBCT = 20 min.
c0 = 16 ug/L
25
50 75 100
Scaled operation time (days)
125
150
Figure E-127 Single contactor and blended effluent SDS-BF breakthrough
curves for Water 7
200
o
'-4—'
CD
t_
CD
O
c
o
O
150 -
100 -
50 -
0 ^
SDS-TTHM
D Single contactor effluent
Logistic function best fit (RA2 = 0.987)
O Blended effluent
- - - - Dl prediction
25
50 75 100
Scaled operation time (days)
EBCT = 20 min.
c0 = 200 |jg/L
125
150
Figure E-128 Single contactor and blended effluent SDS-TTHM breakthrough
curves for Water 7
-287-
-------
2.0
1.5 -
1-0 H
O>
o
c
o
O
0.5 -
0.0 -t>
0
SDS-MCAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
- - - - Dl prediction
25
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 20 min.
C0 = 5 ug/L
-O—i
50 75 100
Scaled operation time (days)
125
150
Figure E-129 Single contactor and blended effluent SDS-MCAA
breakthrough curves for Water 7
12
10 -
o
'CD 6 H
t_
o>
o
2 -
o ^
SDS-DCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.922)
O Blended effluent
- - - - Dl prediction
.o-
. - - -' "o
-D—O—r-
25
50 75 100
Scaled operation time (days)
EBCT = 20 min.
c0 = 25.3 ug/L
125
150
Figure E-130 Single contactor and blended effluent SDS-DCAA breakthrough
curves for Water 7
-288-
-------
6 -
g
'•SS 4H
O>
o
c
o
O
SDS-TCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.999)
O Blended effluent
Dl prediction
25
D
50 75 100
Scaled operation time (days)
o
.O
EBCT = 20 min.
C0 = 18 |jg/L
125
150
Figure E-131 Single contactor and blended effluent SDS-TCAA breakthrough
curves for Water 7
2.0
1.5 -
g
'•§ 1.0 H
-I—•
I
o
o
0.5 -
o.o -b-
0
SDS-MBAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
Effluent concentrations were not detected
above the MRL for this parameter
25
50 75 100
Scaled operation time (days)
125
150
Figure E-132 Single contactor and blended effluent SDS-MBAA
breakthrough curves for Water 7
-289-
-------
20
15 -
10 H
O>
o
c
o
O
5 -
o H
SDS-DBAA
EBCT = 20 min
C0 = 16.3 |jg/L
25
D Single contactor effluent
Logistic function best fit (RA2 = 0.976)
O Blended effluent
- - - - Dl prediction
50 75 100
Scaled operation time (days)
125
150
Figure E-133 Single contactor and blended effluent SDS-DBAA breakthrough
curves for Water 7
40
30 -
o
'•S3 20
t_
o>
o
c
o
O
10 -
o ^
SDS-HAA5
D Single contactor effluent
Logistic function best fit (RA2 = 0.994)
O Blended effluent
- - - - Dl prediction
25
50 75 100
Scaled operation time (days)
EBCT = 20 min.
c0 = 65 ug/L
125
150
Figure E-134 Single contactor and blended effluent SDS-HAA5 breakthrough
curves for Water 7
-290-
-------
16
14
12
13"
1 10
c
o
'•^ ft
CD o
CD
O
c
O
O
6 -
4
2
0
SDS-BCAA
D Single contactor effluent
- Logistic function best fit (RA2 = 0.952)
O Blended effluent
Dl prediction
25
50 75 100
Scaled operation time (days)
.--o
EBCT = 20 min.
C0 = 23.7 |jg/L
125
150
Figure E-135 Single contactor and blended effluent SDS-BCAA breakthrough
curves for Water 7
50
40 -
30 -
o
'-4—'
CD
t_
§ 20
c
o
O
10 -
o ^
SDS-HAA6
D Single contactor effluent
Logistic function best fit (RA2 = 0.988)
O Blended effluent
Dl prediction
25
50 75 100
Scaled operation time (days)
EBCT = 20 min.
c0 = 85 |jg/L
125
150
Figure E-136 Single contactor and blended effluent SDS-HAA6 breakthrough
curves for Water 7
-291-
-------
12
10 -
8 -
o
'•^ R -
CD D 1
CD
O
2 -
o H
SDS-DCBAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.982)
O Blended effluent
Dl prediction
25
50 75 100
Scaled operation time (days)
EBCT = 20 min.
c0 = 26.7 ug/L
125
150
Figure E-137 Single contactor and blended effluent SDS-DCBAA
breakthrough curves for Water 7
6 -
5 -
4 -
SDS-CDBAA
EBCT = 20 min
c0= 11.7 |jg/L
25
D Single contactor effluent
Logistic function best fit (RA2 = 0.976)
O Blended effluent
• - - - Dl prediction
50 75 100
Scaled operation time (days)
125
150
Figure E-138 Single contactor and blended effluent SDS-CDBAA
breakthrough curves for Water 7
-292-
-------
6 -
SDS-TBAA
EBCT = 20 min
c0 = BMRL
25
D Single contactor effluent
Logistic function best fit (RA2 = 0.976)
O Blended effluent
- - - - Dl prediction
50 75 100
Scaled operation time (days)
125
150
Figure E-139 Single contactor and blended effluent SDS-TBAA breakthrough
curves for Water 7
80
60 -
o
'•S3 40 H
t_
o>
o
c
o
O
20 -
o ^
SDS-HAA9
D Single contactor effluent
Logistic function best fit (RA2 = 0.98)
O Blended effluent
Dl prediction
25
50 75 100
Scaled operation time (days)
EBCT = 20 min.
c0= 124 ug/L
125
150
Figure E-140 Single contactor and blended effluent SDS-HAA9 breakthrough
curves for Water 7
-293
-------
2.0
1.5 -
o
'•a
CD
o
c
o
O
0.5 -
0.0
TOO
D Single contactor effluent
Logistic function best fit (RA2 = 0.974)
O Blended effluent
- - - - Dl prediction
EBCT = 7.2min.
c0 = 2.02 mg/L
50 100
Scaled operation time (days)
150
200
Figure E-141 Single contactor and blended effluent TOC breakthrough
curves for Water 8
0.025
0.020 -
o
^ 0.015 H
CD
O
CD
.Q
O 0.010 H
.a
0.005 -
0.000
UV254
D Single contactor effluent
Logistic function best fit (RA2 = 0.994)
O Blended effluent
- - - - Dl prediction
EBCT = 7.2 min.
c0 = 0.033 1/cm
50 100 150
Scaled operation time (days)
200
Figure E-142 Single contactor and blended effluent UV254 breakthrough
curves for Water 8
-294-
-------
O
o
'-4—'
TO
O>
O
c
o
O
125
100 -
75 -
50 -
25 -
0 -to
SDS-TOX
D Single contactor effluent
Logistic function best fit (RA2 = 0.99)
O Blended effluent
- - - - Dl prediction
EBCT = 7.2min.
c0= 156 ug/LCI-
50 100
Scaled operation time (days)
150
200
Figure E-143 Single contactor and blended effluent SDS-TOX breakthrough
curves for Water 8
25
20 -
15 -
o
'-4—'
TO
t_
O>
O
c
o
O
10 -
SDS-CF
D Single contactor effluent
Logistic function best fit (RA2 = 0.985)
O Blended effluent
- - - - Dl prediction
P--
EBCT = 7.2min.
c0 = 29.1Mg/L
50 100
Scaled operation time (days)
150
200
Figure E-144 Single contactor and blended effluent SDS-CF breakthrough
curves for Water 8
-295-
-------
4 -
3 -
g
'-4—'
CD
O
O
1 -
0 -O
SDS-BDCM
EBCT = 7.2min.
C0 = 2.6 |jg/L
D D
O
O
CD Or—
50
O
D
O
D Single contactor effluent
Logistic function best fit (RA2 = 0.946)
O Blended effluent
- - - - Dl prediction
100
Scaled operation time (days)
150
200
Figure E-145 Single contactor and blended effluent SDS-BDCM
breakthrough curves for Water 8
12
10 -
o
'CD 6
SDS-DBCM
D Single contactor effluent
Logistic function best fit (RA2 = 0.98)
O Blended effluent
- - - - Dl prediction
-ID
D
O
EBCT = 7.2min.
c0 = 10.3 |jg/L
50 100
Scaled operation time (days)
150
200
Figure E-146 Single contactor and blended effluent SDS-DBCM
breakthrough curves for Water 8
-296-
-------
4 -
3 -
g
'-4—'
CD
O
O
1 -
0 -O
SDS-BF
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
Insufficient data measured above the MRL
to perform curve fit analysis
-ii ii i k-ry-n—<"n Oh QI-
50
-CD-
EBCT = 7.2min.
C0 = BMRL
100
-Ch—
150
200
Scaled operation time (days)
Figure E-147 Single contactor and blended effluent SDS-BF breakthrough
curves for Water 8
40
30 -
SDS-TTHM
D Single contactor effluent
Logistic function best fit (RA2 = 0.98)
O Blended effluent
- - - - Dl prediction
EBCT = 7.2min.
C0 = 42 jjg/L.
50 100
Scaled operation time (days)
150
200
Figure E-148 Single contactor and blended effluent SDS-TTHM breakthrough
curves for Water 8
-297-
-------
2.0
1.5 -
1.0 -
O>
o
c
o
O
0.5 -
SDS-MCAA
o.o -lo-
0
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
- - - - Dl prediction
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 7.2min.
C0 = BMRL
-ii ii i k-ry-n—m—Oh QJ-
50
-CD-
100
-OH—
150
Scaled operation time (days)
Figure E-149 Single contactor and blended effluent SDS-MCAA
breakthrough curves for Water 8
200
6 -
4 -
o>
o
c
o
O
SDS-DCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.974)
O Blended effluent
Dl prediction
50
EBCT = 7.2min.
c0= 10.7ug/L
100
Scaled operation time (days)
150
200
Figure E-150 Single contactor and blended effluent SDS-DCAA breakthrough
curves for Water 8
-298-
-------
6 -
5 -
4 -
g
'-4—'
CD
CD
O
c
O
O
3 -
SDS-TCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.93)
O Blended effluent
- - - - Dl prediction
EBCT = 7.2 min.
C0 = 12.7 |jg/L
50 100
Scaled operation time (days)
150
200
Figure E-151 Single contactor and blended effluent SDS-TCAA breakthrough
curves for Water 8
2.0
1.5 -
O
'•§ 1.0 H
-i—•
I
O
O
0.5 -
SDS-MBAA
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
0.0 Hd-
0
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 7.2 min.
C0= BMRL
-ii ii i K-ry-n—m Oh QI-
50
-CD-
100
-CD-,—
150
200
Scaled operation time (days)
Figure E-152 Single contactor and blended effluent SDS-MBAA
breakthrough curves for Water 8
-299-
-------
10
6 -
g
'-4—'
CD
O
O
2 -
SDS-DBAA
o -lo
o
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 7.2min.
C0 = BMRL
-03EEKJJ-D—O\ Oh Ol-
50
-CD-
100
-CDn—
150
200
Scaled operation time (days)
Figure E-153 Single contactor and blended effluent SDS-DBAA breakthrough
curves for Water 8
15
10 -
o
'-4—'
CD
SDS-HAA5
D Single contactor effluent
Logistic function best fit (RA2 = 0.955)
O Blended effluent
Dl prediction
50
EBCT = 7.2min.
c0 = 23 |jg/L
100
Scaled operation time (days)
150
200
Figure E-154 Single contactor and blended effluent SDS-HAA5 breakthrough
curves for Water 8
-300-
-------
3 -
o
'•^
CO
O>
O
c
o
O
1 -
0 -O
SDS-BCAA
D Single contactor effluent
Logistic function best fit (RA2 = 0.971)
O Blended effluent
Dl prediction
-D-
O
O
D
EBCT = 7.2min.
c0 = 3 |jg/L
100
Scaled operation time (days)
150
200
Figure E-155 Single contactor and blended effluent SDS-BCAA breakthrough
curves for Water 8
20
15 -
SDS-HAA6
D Single contactor effluent
Logistic function best fit (RA2 = 0.956)
O Blended effluent
EBCT = 7.2min.
c0 = 27 |jg/L
50 100 150
Scaled operation time (days)
200
Figure E-156 Single contactor and blended effluent SDS-HAA6 breakthrough
curves for Water 8
-301-
-------
3.0
2.5 -
2.0 -
O)
c
o
'•^ 1
CD I-
CD
O
.
0.5 -
SDS-DCBAA
EBCT = 7.2min.
c0 = 3 |jg/L
0.0 -lo-
0
D
O
-D-
D Single contactor effluent
Logistic function best fit (RA2 = 0.939)
O Blended effluent
Dl prediction
50 100 150
Scaled operation time (days)
200
Figure E-157 Single contactor and blended effluent SDS-DCBAA
breakthrough curves for Water 8
2.0
1.5 -
o
'•CD 1.0 H
t_
CD
O
c
o
O
0.5 -
SDS-CDBAA
o.o Ho-
o
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 7.2 min.
c0= BMRL
-ini i KTy-n—m—Qh QJ-
50
-CD-
-OH-
100
150
200
Scaled operation time (days)
Figure E-158 Single contactor and blended effluent SDS-CDBAA
breakthrough curves for Water 8
-302-
-------
2.0
1.5 -
1-0 H
O>
o
c
o
O
0.5 -
SDS-TBAA
o.o -lo-
0
D Single contactor effluent
Logistic function best fit (RA2 = NA)
O Blended effluent
Dl prediction
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 7.2min.
C0 = BMRL
-ii ii i k-ry-n—m—Oh QJ-
50
-CD-
100
-OH—
150
200
Scaled operation time (days)
Figure E-159 Single contactor and blended effluent SDS-TBAA breakthrough
curves for Water 8
20
15 -
SDS-HAA9
D Single contactor effluent
Logistic function best fit (RA2 = 0.952)
O Blended effluent
- - - - Dl prediction
EBCT = 7.2min.
c0 = 30 |jg/L
50 100 150
Scaled operation time (days)
200
Figure E-160 Single contactor and blended effluent SDS-HAA9 breakthrough
curves for Water 8
-303-
-------
This page intentionally left blank.
-304-
-------
Appendix F: Comparison of SCA Method to Dl Approach for Integral
Breakthrough Curve Prediction
-305-
-------
2.5
2.0 -
O)
o
"co
o
O
1.5 4
1.0 H
0.5 -
0.0
TOO
EBCT = 15 min.
c0 = 4.54 mg/L
X •'
X '
X
X
• Observed data
Best fit
Dl prediction (RSS = 0.152)
25 50
Scaled operation time (days)
75
Figure F-1 Dl method prediction of the TOC integral breakthrough curve for
Water 1
0.035
0.000
• Observed data
Best fit
Dl prediction (RSS = 0.0011)
SCA prediction (RSS = 0.0017)
0 25 50
Scaled operation time (days)
Figure F-2 Comparison of Dl and SCA methods for predicting the UV254
integral breakthrough curve for Water 1
75
-306-
-------
200
150 -
O
o 100 H
o>
o
c
o
0 50 H
SDS-TOX
EBCT = 15 min.
CD = 224 |jg/L Cl-
SDS-TOX not analyzed in
the blended effluent
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
25 50
Scaled operation time (days)
75
Figure F-3 Comparison of Dl and SCA methods for predicting the SDS-TOX
integral breakthrough curve for Water 1
5 -
4 -
o
'•a
o>
o
8 2
1 -
SDS-CF
EBCT = 15 min.
CD = 34.2 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.39)
SCA prediction (RSS = 0.65)
25 50
Scaled operation time (days)
75
Figure F-4 Comparison of Dl and SCA methods for predicting the SDS-CF
integral breakthrough curve for Water 1
-307-
-------
20
16 -
O)
a 12 H
c
o
'-4— '
S
§ 8
c
o
O
4 -
SDS-BDCM
EBCT = 15 min.
c0 = 19.3|jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.52)
SCA prediction (RSS = 1.03)
0 25 50
Scaled operation time (days)
Figure F-5 Comparison of Dl and SCA methods for predicting the SDS-
BDCM integral breakthrough curve for Water 1
75
12
10 -
CD 6 -\
1_
-I—•
§
I 4
2 -
SDS-DBCM
EBCT = 15 min
co = 28 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.4)
SCA prediction (RSS = 0.73)
0 25 50
Scaled operation time (days)
Figure F-6 Comparison of Dl and SCA methods for predicting the SDS-
DBCM integral breakthrough curve for Water 1
75
-308-
-------
16
14 -
12 -
10 -
o
'-4—'
CD
CD
O
c
o
O
SDS-BF
EBCT = 15 min
c0= 3.7|jg/L
• Observed data
Best fit
Dl prediction (RSS = 2.76)
SCA prediction (RSS = 1.31)
0 25 50
Scaled operation time (days)
Figure F-7 Comparison of Dl and SCA methods for predicting the SDS-BF
integral breakthrough curve for Water 1
75
Observed data
Best fit
Dl prediction (RSS = 3.93)
SCA prediction (RSS = 2.55)
0 25 50
Scaled operation time (days)
Figure F-8 Comparison of Dl and SCA methods for predicting the SDS-TTHM
integral breakthrough curve for Water 1
75
-309-
-------
2.0
1.5 -
g
'•S3 1-0 H
O>
o
c
o
O
0.5 -
SDS-MCAA
EBCT = 15 min.
c0 = BMRL
0.0 +•-
0
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
25 50
Scaled operation time (days)
75
Figure F-9 Comparison of Dl and SCA methods for predicting the SDS-MCAA
integral breakthrough curve for Water 1
10
6 -
§ 4H
c
o
O
2 -
SDS-DCAA
EBCT = 15 min.
c0= 12.5 pg/L
• Observed data
Best fit
Dl prediction (RSS = 0.35)
SCA prediction (RSS = 1.46)
0 25 50
Scaled operation time (days)
Figure F-10 Comparison of Dl and SCA methods for predicting the SDS-
DCAA integral breakthrough curve for Water 1
75
-310-
-------
2.0
1.5 -
g
'•S3 1-0 H
O>
o
c
o
O
0.5 -
SDS-TCAA
EBCT = 15 min.
c0 = 3 |jg/L
0.0 +•-
0
Blended effluent concentrations were not
detected above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
25 50
Scaled operation time (days)
75
Figure F-11 Comparison of Dl and SCA methods for predicting the SDS-
TCAA integral breakthrough curve for Water 1
2.0
1.5 -
O
'•§ 1.0 H
-i—'
O>
o
c
o
O
0.5 -
SDS-MBAA
EBCT = 15 min.
CO = BMRL
o.o 4*-
o
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
25 50
Scaled operation time (days)
75
Figure F-12 Comparison of Dl and SCA methods for predicting the SDS-
MBAA integral breakthrough curve for Water 1
-311-
-------
5 -
4 -
3 -
CD
O
O o J
O ^ 1
1 -
SDS-DBAA
EBCT = 15 min
co = 4 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.46)
SCA prediction (RSS = 0.39)
0 25 50
Scaled operation time (days)
Figure F-13 Comparison of Dl and SCA methods for predicting the SDS-
DBAA integral breakthrough curve for Water 1
75
15
10 -
o
"CD
CD
o
c
o
O
5 -
SDS-HAA5
EBCT = 15 min
co=20 |jg/L
Observed data
Best fit
Dl prediction (RSS = 0.38)
SCA prediction (RSS = 1.58)
0 25 50
Scaled operation time (days)
Figure F-14 Comparison of Dl and SCA methods for predicting the SDS-
HAA5 integral breakthrough curve for Water 1
75
-312-
-------
4 -
O)
a 3 -
c
o
I
12
o
O
1 -
SDS-BCAA
EBCT = 15 min.
c0 = 7 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.38)
SCA prediction (RSS = 0.29)
0 25 50
Scaled operation time (days)
Figure F-15 Comparison of Dl and SCA methods for predicting the SDS-
BCAA integral breakthrough curve for Water 1
75
20
15 -
o
'•S3 10 H
o>
o
c
o
O
5 -
SDS-HAA6
EBCT = 15 min.
c0 = 27 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.63)
SCA prediction (RSS = 1.26)
0 25 50
Scaled operation time (days)
Figure F-16 Comparison of Dl and SCA methods for predicting the SDS-
HAA6 integral breakthrough curve for Water 1
75
-313-
-------
2.0
1.5 -
o
'OB 1.0 H
O>
o
c
o
O
0.5 -
0.0
SDS-DCBAA
EBCT = 15 min.
c0 = 2.3 ug/L
Insufficient data measured above the MRL to
compare Dl and SCA predictions
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
0 25 50
Scaled operation time (days)
Figure F-17 Comparison of Dl and SCA methods for predicting the SDS-
DCBAA integral breakthrough curve for Water 1
75
2.0
1.5 -
o
'•g 1.0 H
-I—I
§
c
o
O
0.5 -
0.0
SDS-CDBAA
EBCT = 15 min.
rv= BMRL
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
0 25 50
Scaled operation time (days)
Figure F-18 Comparison of Dl and SCA methods for predicting the SDS-
CDBAA integral breakthrough curve for Water 1
75
-314-
-------
2.0
1.5 -
o
'7J 1-0 H
-i—'
I
o
O
0.5 -
0.0
SDS-TBAA
EBCT = 15 min.
rv= BMRL
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
0 25 50
Scaled operation time (days)
Figure F-19 Comparison of Dl and SCA methods for predicting the SDS-
TBAA integral breakthrough curve for Water 1
75
20
15 -
O)
o
73 10
-I— i
§
c
o
O
5 -
SDS-HAA9
EBCT = 15 min.
c0= 29
• Observed data
Best fit
Dl prediction (RSS = 0.78)
SCA prediction (RSS = 1.04)
0 25 50
Scaled operation time (days)
Figure F-20 Comparison of Dl and SCA methods for predicting the SDS-
HAA9 integral breakthrough curve for Water 1
75
-315-
-------
1.50
1.25 -
1.00 -
o
'•a °-75 H
o>
o
o 0.50 -
O
0.25 -
0.00
TOO
EBCT = 20min.
c0 = 2.6 mg/L
• Observed data
Best fit
Dl prediction (RSS = 0.057)
50
100 150
Scaled operation time (days)
200
250
Figure F-21 Dl method prediction of the TOC integral breakthrough curve for
Water 2
0.025
0.020 -
o
^ 0.015 H
CD
O
CD
.Q
O 0.010 H
to
0.005 -
0.000
UV254
EBCT = 20 min.
co = 0.055 1/cm
• Observed data
Best fit
Dl prediction (RSS = 0.0012)
SCA prediction (RSS = 0.0015)
50
100 150
Scaled operation time (days)
200
250
Figure F-22 Comparison of Dl and SCA methods for predicting the UV254
integral breakthrough curve for Water 2
-316-
-------
200
150 -
O
. 100 H
-i—'
CD
CD
O
50 H
SDS-TOX
EBCT = 20 min.
CD = 220 |jg/L Cl-
50
SDS-TOX not analyzed in
the blended effluent
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
100 150
Scaled operation time (days)
200
250
Figure F-23 Comparison of Dl and SCA methods for predicting the SDS-TOX
integral breakthrough curve for Water 2
15
10 -
O
IE
-i—'
CD
O
SDS-CF
EBCT = 20 min
c0= 41.9|jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.57)
SCA prediction (RSS = 1.59)
50
100 150
Scaled operation time (days)
200
250
Figure F-24 Comparison of Dl and SCA methods for predicting the SDS-CF
integral breakthrough curve for Water 2
-317-
-------
30
25 -
SDS-BDCM
EBCT = 20 min.
c0 = 19.8 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 4.35)
SCA prediction (RSS = 1.98)
50
100 150
Scaled operation time (days)
200
250
Figure F-25 Comparison of Dl and SCA methods for predicting the
SDS-BDCM integral breakthrough curve for Water 2
25
20 -
15 -
O
c
o
O
-in J
ID -
5 -
SDS-DBCM
EBCT = 20 min.
Co= 31.7M9/L
50
• Observed data
Best fit
Dl prediction (RSS = 2.18)
SCA prediction (RSS = 1.72)
100 150
Scaled operation time (days)
200
250
Figure F-26 Comparison of Dl and SCA methods for predicting the SDS-
DBCM integral breakthrough curve for Water 2
-318-
-------
14
12 -
10 -
o
"-i—'
2
| 6H
o
° 4J
2 -
0
0
SDS-BF
EBCT = 20 min.
c0 = 3.7 |jg/L |
50
• Observed data
Best fit
Dl prediction (RSS = 3.73)
SCA prediction (RSS = 2.18)
200
100 150
Scaled operation time (days)
Figure F-27 Comparison of Dl and SCA methods for predicting the SDS-BF
integral breakthrough curve for Water 2
250
80
70 -
60 -
50 -
o
IS 40 H
o>
g
o
O
30 -
20 -
10 -
0
SDS-TTHM
EBCT = 20 min
c0 = 97 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 8.49)
SCA prediction (RSS = 4.38)
50
100 150
Scaled operation time (days)
200
250
Figure F-28 Comparison of Dl and SCA methods for predicting the SDS-
TTHM integral breakthrough curve for Water 2
-319-
-------
2.0
1.5 -
o
'•S 1-0 H
o
O
0.5 -
0.0
SDS-MCAA
EBCT = 20 min.
co= BMRL
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50
100 150
Scaled operation time (days)
200
250
Figure F-29 Comparison of Dl and SCA methods for predicting the SDS-
MCAA integral breakthrough curve for Water 2
5 -
4 -
o
*= o
CD J
o
<§2
1 -
SDS-DCAA
EBCT = 20 min.
c0= 14 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.49)
SCA prediction (RSS = 0.68)
50
100 150
Scaled operation time (days)
200
250
Figure F-30 Comparison of Dl and SCA methods for predicting the SDS-
DCAA integral breakthrough curve for Water 2
-320-
-------
2.5
2.0 -
O)
1.5 -
o
-i—•
CD
8 1-0-1
o
O -I
0.5 -
0.0
SDS-TCAA
EBCT = 20 min.
c0 = 5 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.27)
SCA prediction (RSS = 0.34)
50
100 150
Scaled operation time (days)
200
250
Figure F-31 Comparison of Dl and SCA methods for predicting the SDS-
TCAA integral breakthrough curve for Water 2
2.0
1.5 -
1-0 H
O>
o
c
o
O
0.5 -
0.0
SDS-MBAA
EBCT = 20 min.
c°= BMRL
Blended effluent concentrations were not
detected above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50
100 150
Scaled operation time (days)
200
250
Figure F-32 Comparison of Dl and SCA methods for predicting the SDS-
MBAA integral breakthrough curve for Water 2
-321-
-------
6 -
o
14H
I
o
O
SDS-DBAA
EBCT = 20 min
c0 = 5 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.1)
SCA prediction (RSS = 0.6)
100 150
Scaled operation time (days)
200
250
Figure F-33 Comparison of Dl and SCA methods for predicting the SDS-
DBAA integral breakthrough curve for Water 2
15
10 -
o
'-4—'
CD
CD
O
0 «; J
O 5 H
SDS-HAA5
EBCT = 20 min.
co = 24 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.69)
SCA prediction (RSS = 1.48)
50
100 150
Scaled operation time (days)
200
250
Figure F-34 Comparison of Dl and SCA methods for predicting the SDS-
HAA5 integral breakthrough curve for Water 2
-322-
-------
6 -
o
'•SS 4H
O>
o
c
o
O
2 -
SDS-BCAA
EBCT = 20 min.
c0= 9|jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.82)
SCA prediction (RSS = 0.78)
50
100 150
Scaled operation time (days)
200
250
Figure F-35 Comparison of Dl and SCA methods for predicting the SDS-
BCAA integral breakthrough curve for Water 2
25
20 -
O)
a 15 H
c
o
'-4—'
S
§ 10
c
o
O
5 -
SDS-HAA6
EBCT = 20 min.
c0= 34
Observed data
Best fit
Dl prediction (RSS = 2.5)
SCA prediction (RSS = 2.25)
50
100 150
Scaled operation time (days)
200
250
Figure F-36 Comparison of Dl and SCA methods for predicting the SDS-
HAA6 integral breakthrough curve for Water 2
-323-
-------
2.5
2.0 -
O)
1.5 -
o
•*=
£5
-i—i
§ 1.0
c
o
O
0.5 -
SDS-DCBAA
EBCT = 20 min.
CD = 3 |jg/L
0.0 -(•-
0
• Observed data
Best fit
Dl prediction (RSS = 0.48)
SCA prediction (RSS = 0.35)
100 150
Scaled operation time (days)
200
250
Figure F-37 Comparison of Dl and SCA methods for predicting the SDS-
DCBAA integral breakthrough curve for Water 2
3.0
2.5 -
2.0 -
O)
1-5 H
0.5 -
0.0
SDS-CDBAA
EBCT = 20 min.
c0= BMRL
Insufficient data measured above the MRL
to compare Dl and SCA predictions
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50
100 150
Scaled operation time (days)
200
250
Figure F-38 Comparison of Dl and SCA methods for predicting the SDS-
CDBAA integral breakthrough curve for Water 2
-324-
-------
2.0
1.5 -
o
'•S 1-0
o
o
o
0.5 -
SDS-TBAA
EBCT = 20 min.
co= BMRL
0.0 -(•-
0
Blended effluent concentrations were not
detected above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50
100 150
Scaled operation time (days)
200
250
Figure F-39 Comparison of Dl and SCA methods for predicting the SDS-
TBAA integral breakthrough curve for Water 2
30
25 -
20 -
O)
15 H
5 -
SDS-HAA9
EBCT = 20 min.
c0 = 37
• Observed data
Best fit
Dl prediction (RSS = 3.64)
SCA prediction (RSS = 3.43)
50
100 150
Scaled operation time (days)
200
250
Figure F-40 Comparison of Dl and SCA methods for predicting the SDS-
HAA9 integral breakthrough curve for Water 2
-325-
-------
1.25
1.00 -
O)
-§ 0.75 -
.g
"co
g 0.50
c
o
O
0.25 -
0.00
TOO
EBCT = 20 min.
c0 = 2.35 mg/L
X
X
X
X , •
xx<
x , -'
• Observed data
Best fit
Dl prediction (RSS = 0.044)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-41 Dl method prediction of the TOC integral breakthrough curve for
Water 3
0.020
0.015 -
o
o>
o
c
CD
.Q
O
5
.Q
0.010 -
0.005 -
0.000
UV254
EBCT = 20 min.
co= 0.048 1/cm
• Observed data
Best fit
Dl prediction (RSS = 0.001)
SCA prediction (RSS = 0.001)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-42 Comparison of Dl and SCA methods for predicting the UV254
integral breakthrough curve for Water 3
-326-
-------
200
150 -
O
.2 100 -
"en
O>
o
c
o
0 50 H
SDS-TOX
EBCT = 20 min.
c0 = 255 |jg/L Cl-
50
SDS-TOX not analyzed in
the blended effluent
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
100 150 200
Scaled operation time (days)
250
300
Figure F-43 Comparison of Dl and SCA methods for predicting the SDS-TOX
integral breakthrough curve for Water 3
6 -
5 -
4 H
o
"on
o
O
3 -
2 -
1 -
SDS-CF
EBCT = 20 min.
co = 60.3 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.85)
SCA prediction (RSS = 0.4)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-44 Comparison of Dl and SCA methods for predicting the SDS-CF
integral breakthrough curve for Water 3
-327-
-------
35
30 -
25 -
O)
20 -
o
0
10 -
5 -
SDS-BDCM
EBCT = 20 min.
c0 = 36.2 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.05)
SCA prediction (RSS = 2.57)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-45 Comparison of Dl and SCA methods for predicting the SDS-
BDCM integral breakthrough curve for Water 3
16
14 -
12 -
10 -
I 6H
O
O
4 -
2 -
SDS-DBCM
EBCT = 20 min
co = 44.5 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.04)
SCA prediction (RSS = 0.68)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-46 Comparison of Dl and SCA methods for predicting the SDS-
DBCM integral breakthrough curve for Water 3
-328-
-------
SDS-BF
EBCT = 20 min
11.8 |jg/L
Observed data
Best fit
Dl prediction (RSS = 5.67)
SCA prediction (RSS = 3.14)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-47 Comparison of Dl and SCA methods for predicting the SDS-BF
integral breakthrough curve for Water 3
o
O
80
70 -
60 -
50 -
40 -
30 -
20 -
10 -
0
SDS-TTHM
EBCT = 20 min.
Co= 154|jg/L
• Observed data
Best fit
Dl prediction (RSS = 2.87)
SCA prediction (RSS = 2.64)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-48 Comparison of Dl and SCA methods for predicting the SDS-
TTHM integral breakthrough curve for Water 3
-329-
-------
2.0
1.5 -
g
'•SS 1-0 H
O>
o
o
O
0.5 -
0.0
SDS-MCAA
EBCT = 20 min.
co= BMRL
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-49 Comparison of Dl and SCA methods for predicting the SDS-
MCAA integral breakthrough curve for Water 3
3.0
2.5 -
2.0 -
o
]3 1.5 H
-i—'
o>
o
o 1.0 H
0.5 -
0.0
SDS-DCAA
EBCT = 20 min.
c0= 15.7
Observed data
• Best fit
Dl prediction (RSS = 0.4)
SCA prediction (RSS = 0.37)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-50 Comparison of Dl and SCA methods for predicting the SDS-
DCAA integral breakthrough curve for Water 3
-330-
-------
2.0
1.5 -
o
'•SS 1-0 H
O>
o
o
O
0.5 -
0.0
SDS-TCAA
EBCT = 20 min.
co -i
'5 ug/L
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-51 Comparison of Dl and SCA methods for predicting the SDS-
TCAA integral breakthrough curve for Water 3
2.0
1.5 -
'•§ 1.0 H
-I—'
o>
o
c
o
O
0.5 -
0.0
SDS-MBAA
EBCT = 20 min.
co= BMRL
Blended effluent concentrations were not
detected above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-52 Comparison of Dl and SCA methods for predicting the SDS-
MBAA integral breakthrough curve for Water 3
-331-
-------
12
10 -
CD
O
§ 4H
2 -
SDS-DBAA
EBCT = 20 min
c0 = 9.7 pg/L
• Observed data
Best fit
Dl prediction (RSS = 1.02)
SCA prediction (RSS = 1.56)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-53 Comparison of Dl and SCA methods for predicting the SDS-
DBAA integral breakthrough curve for Water 3
20
15 -
o
'•SS 10 H
CD
O
c
o
O
5 -
SDS-HAA5
EBCT = 20 min.
co = 30 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 2.48)
SCA prediction (RSS = 1.13)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-54 Comparison of Dl and SCA methods for predicting the SDS-
HAA5 integral breakthrough curve for Water 3
-332-
-------
5 -
4 -
o
73 3
O>
o
1 -
SDS-BCAA
EBCT = 20 min
c0=12.7|jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.63)
SCA prediction (RSS = 0.63)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-55 Comparison of Dl and SCA methods for predicting the SDS-
BCAA integral breakthrough curve for Water 3
20
15 -
o
73 1°
CD
O
c
o
O
5 -
SDS-HAA6
EBCT = 20 min
co = 43 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 2.29)
SCA prediction (RSS = 2.3)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-56 Comparison of Dl and SCA methods for predicting the SDS-
HAA6 integral breakthrough curve for Water 3
-333-
-------
2.0
1.5 -
o
'"S 1.0 -\
(D
O
O
O
0.5 -
0.0
SDS-DCBAA
EBCT = 20 min.
c0= 4|jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.68)
SCA prediction (RSS = 0.72)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-57 Comparison of Dl and SCA methods for predicting the SDS-
DCBAA integral breakthrough curve for Water 3
2.0
1.5 -
o
'•§ 1.0 H
-i—'
o>
o
c
o
O
0.5 -
0.0
SDS-CDBAA
EBCT = 20 min.
c0=2.5|jg/L
50
Blended effluent concentrations were not
detected above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
100 150 200
Scaled operation time (days)
250
300
Figure F-58 Comparison of Dl and SCA methods for predicting the SDS-
CDBAA integral breakthrough curve for Water 3
-334-
-------
2.0
1.5 -
g
'•SS 1-0 H
CD
o
c
o
O
0.5 -
SDS-TBAA
EBCT = 20 min.
co= BMRL
0.0 -!•-
0
Blended effluent concentrations were not detected above the
MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-59 Comparison of Dl and SCA methods for predicting the SDS-
TBAA integral breakthrough curve for Water 3
25
20 -
o
'-4—'
CD
o
O
15 -
-in J
10 -
5 -
o -k
SDS-HAA9
EBCT = 20 min
c0= 49
• Observed data
Best fit
Dl prediction (RSS = 4.36)
SCA prediction (RSS = 2.2)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-60 Comparison of Dl and SCA methods for predicting the SDS-
HAA9 integral breakthrough curve for Water 3
-335-
-------
1.50
1.25 -
1.00 -
c
o
0.75 -
CD
1 0.50 H
0.25 -
0.00
TOC
EBCT = 20 min.
c0 = 2.98 mg/L
• Observed data
Best fit
PI prediction (RSS = 0.05)
50 100
Scaled operation time (days)
150
200
Figure F-61 Dl method prediction of the TOC integral breakthrough curve for
Water 4
0.025
0.020 -
o
^ 0.015
CD
o 0.010
0.005 -
0.000
UV254
EBCT = 20 min.
co= 0.065 1/cm
• Observed data
Best fit
Dl prediction (RSS = 0.0015)
SCA prediction (RSS = 0.0023)
50 100
Scaled operation time (days)
150
200
Figure F-62 Comparison of Dl and SCA methods for predicting the UV254
integral breakthrough curve for Water 4
-336-
-------
200
150 -
O
o 100 -
O>
o
50 4
SDS-TOX
EBCT = 20 min.
c0 =288 |jg/L Cl-
SDS-TOX not analyzed in the
blended effluent
Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50 100
Scaled operation time (days)
150
200
Figure F-63 Comparison of Dl and SCA methods for predicting the SDS-TOX
integral breakthrough curve for Water 4
14
12 -
10 -
o
"co
o>
o
c
o
O
6 -
4 -
2 -
SDS-CF
EBCT = 20 min.
• Observed data
Best fit
Dl prediction (RSS = 0.41)
SCA prediction (RSS = 1.33)
50 100
Scaled operation time (days)
150
200
Figure F-64 Comparison of Dl and SCA methods for predicting the SDS-CF
integral breakthrough curve for Water 4
-337-
-------
3 -
g
•^
CD
CD
O
c
O
O
1 -
SDS-BDCM
EBCT = 20 mm.
c0 = 2.2Mg/L
Observed data
Best fit
Dl prediction (RSS = 0.58)
SCA prediction (RSS = 0.55)
50 100
Scaled operation time (days)
150
200
Figure F-65 Comparison of Dl and SCA methods for predicting the SDS-
BDCM integral breakthrough curve for Water 4
6 -
o
I 4H
I
o
O
2 -
SDS-DBCM
EBCT = 20 min.
c0= 15.1M9/L
• Observed data
Best fit
Dl prediction (RSS = 0.59)
SCA prediction (RSS = 0.55)
50 100
Scaled operation time (days)
150
200
Figure F-66 Comparison of Dl and SCA methods for predicting the SDS-
DBCM integral breakthrough curve for Water 4
-338-
-------
2.0
1.5 -
g
'•SS 1-0 H
O>
o
c
o
O
0.5 -
0.0
SDS-BF
EBCT = 20 min.
c0= BMRL
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50 100
Scaled operation time (days)
150
200
Figure F-67 Comparison of Dl and SCA methods for predicting the SDS-BF
integral breakthrough curve for Water 4
25
20 -
O)
15 H
o
'-4— '
§ 10
c
o
O
5 -
SDS-TTHM
EBCT = 20 min.
c0 = 68 |jg/L
Observed data
Best fit
Dl prediction (RSS = 1.39)
SCA prediction (RSS = 2.03)
50 100
Scaled operation time (days)
150
200
Figure F-68 Comparison of Dl and SCA methods for predicting the SDS-
TTHM integral breakthrough curve for Water 4
-339-
-------
2.0
1.5 -
c
o
ro
o
c
o
O
1.0 -
0.5 -
SDS-MCAA
EBCT = 20 min.
CD= BMRL
Blended effluent concentrations were not detected above the
MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
0.0 4«-
0
50 100
Scaled operation time (days)
150
200
Figure F-69 Comparison of Dl and SCA methods for predicting the SDS-
MCAA integral breakthrough curve for Water 4
6 -
5 -
SDS-DCAA
EBCT = 20 min.
co = 20.3 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.32)
SCA prediction (RSS = 0.47)
50
100
Scaled operation time (days)
150
200
Figure F-70 Comparison of Dl and SCA methods for predicting the SDS-
DCAA integral breakthrough curve for Water 4
-340-
-------
10
O)
6 -
c
o
•*=
£5
-H-
C
0> „
o 4
c
o
O
2 -
SDS-TCAA
EBCT = 20 min.
c0 = 30.7 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.96)
SCA prediction (RSS = 0.64)
50 100
Scaled operation time (days)
150
200
Figure F-71 Comparison of Dl and SCA methods for predicting the SDS-
TCAA integral breakthrough curve for Water 4
2.0
1.5 -
c
o
jo 1-0 H
-i—<
c
o>
o
c
o
O
0.5 -
SDS-MBAA
EBCT = 20 min.
co= BMRL
0.0 -k
0
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50 100
Scaled operation time (days)
150
200
Figure F-72 Comparison of Dl and SCA methods for predicting the SDS-
MBAA integral breakthrough curve for Water 4
-341-
-------
2.0
1.5 -
1.0 H
o
O
0.5 -
0.0
SDS-DBAA
EBCT = 20 min.
c0 = 1 H9/L
Blended effluent concentrations were not detected above the
MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50 100
Scaled operation time (days)
150
200
Figure F-73 Comparison of Dl and SCA methods for predicting the SDS-
DBAA integral breakthrough curve for Water 4
Observed data
Best fit
Dl prediction (RSS = 1.24)
SCA prediction (RSS = 0.87)
50 100
Scaled operation time (days)
150
200
Figure F-74 Comparison of Dl and SCA methods for predicting the SDS-
HAA5 integral breakthrough curve for Water 4
-342-
-------
3.0
2.5 -
2.0 -
O)
o
<§
1-5 H
1.0 H
0.5 -
0.0
SDS-BCAA
EBCT = 20 min.
co = 4.3 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.31)
SCA prediction (RSS = 0.22)
50 100
Scaled operation time (days)
150
200
Figure F-75 Comparison of Dl and SCA methods for predicting the SDS-
BCAA integral breakthrough curve for Water 4
20
15 -
o
jo 10 H
-I—I
I
o
O
5 -
SDS-HAA6
EBCT = 20 min.
co = 55 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.34)
SCA prediction (RSS = 1.07)
50 100
Scaled operation time (days)
150
200
Figure F-76 Comparison of Dl and SCA methods for predicting the SDS-
HAA6 integral breakthrough curve for Water 4
-343-
-------
3.0
2.5-
r2.o-
c
o
;i.5 -
o
c
O1.0 -
0.5-
0.0
SDS-DCBAA
EBCT = 20 min.
c0= 6.7|jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.49)
SCA prediction (RSS = 0.5)
50 100
Scaled operation time (days)
150
200
Figure F-77 Comparison of Dl and SCA methods for predicting the SDS-
DCBAA integral breakthrough curve for Water 4
2.0
1.5-
O)
o
't5 1.0 H
8
c
o
O
0.5 -
0.0
SDS-CDBAA
EBCT = 20 min.
CD= BMRL
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50 100
Scaled operation time (days)
150
200
Figure F-78 Comparison of Dl and SCA methods for predicting the SDS-
CDBAA integral breakthrough curve for Water 4
-344-
-------
2.0
1.5-
o
M.O-
0)
o
o
O
0.5 -
0.0
SDS-TBAA
EBCT = 20 m\n.
co= BMRL
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50 100
Scaled operation time (days)
150
200
Figure F-79 Comparison of Dl and SCA methods for predicting the SDS-
TBAA integral breakthrough curve for Water 4
25
20-
15 -
c
o
<& in J
O 10 -
C
o
O
5 -
SDS-HAA9
EBCT = 20 min.
cn=61 M9/L
• Observed data
Best fit
Dl prediction (RSS = 1.7)
SCA prediction (RSS = 1.12)
50 100
Scaled operation time (days)
150
200
Figure F-80 Comparison of Dl and SCA methods for predicting the SDS-
HAA9 integral breakthrough curve for Water 4
-345-
-------
1.75
1.50 -
1.25 H
^ 1.00 H
o
u-»
TO
c 0.75 -
-------
84
72 -
O 60 -
CD
O
I
48 -
36 -
12 -
0
0
SDS-TOX
EBCT = 20 min.
c0 = 205 |jg/L Cl-
SDS-TOX not analyzed in
the blended effluent
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50 100 150 200 250
Scaled operation time (days)
300
350
Figure F-83 Comparison of Dl and SCA methods for predicting the SDS-TOX
integral breakthrough curve for Water 5
4 -
3 -
c
o
CD
O
c
o
O
1 -
SDS-CF
EBCT = 20 min.
c0 =23.7 pg/L
• Observed data
Best fit
Dl prediction (RSS = 0.31)
SCA prediction (RSS = 0.65)
50
100 150 200 250
Scaled operation time (days)
300
350
Figure F-84 Comparison of Dl and SCA methods for predicting the SDS-CF
integral breakthrough curve for Water 5
-347-
-------
12
10 -
O)
c
g
'CD 6
O>
o
3 4
2 -
SDS-BDCM
EBCT = 20 min
c0 = 10.8 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.85)
SCA prediction (RSS = 0.34)
50
100 150 200 250
Scaled operation time (days)
300
350
Figure F-85 Comparison of Dl and SCA methods for predicting the SDS-
BDCM integral breakthrough curve for Water 5
10
6 -
o
"co
o
O
2 -
SDS-DBCM
EBCT = 20 min.
C0 = 22.7 M9/L
• Observed data
Best fit
Dl prediction (RSS = 0.54)
SCA prediction (RSS = 0.72)
50
100 150 200 250
Scaled operation time (days)
300
350
Figure F-86 Comparison of Dl and SCA methods for predicting the SDS-
DBCM integral breakthrough curve for Water 5
-348-
-------
3.0
2.5 -
2.0 -
g
'•a 1.5 H
O>
0.5 -
0.0
SDS-BF
EBCT = 20 min.
c0= 1-2M9/L
• Observed data
Best fit
Dl prediction (RSS = 0.55)
SCA prediction (RSS = 0.26)
50
100 150 200 250
Scaled operation time (days)
300
350
Figure F-87 Comparison of Dl and SCA methods for predicting the SDS-BF
integral breakthrough curve for Water 5
30
25 -
20 -
SDS-TTHM
EBCT = 20 min.
c0 = 58 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.89)
SCA prediction (RSS = 1.3)
50
100 150 200 250
Scaled operation time (days)
300
350
Figure F-88 Comparison of Dl and SCA methods for predicting the SDS-TTHM
integral breakthrough curve for Water 5
-349-
-------
2.0
1.5 -
g
'•a 1.0 H
CD
o
c
o
O
0.5 -
0.0
SDS-MCAA
EBCT = 20 min.
c0 = 2 |jg/L
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50
100 150 200 250
Scaled operation time (days)
300
350
Figure F-89 Comparison of Dl and SCA methods for predicting the SDS-
MCAA integral breakthrough curve for Water 5
3 -
o
•^
CD
CD
O
c
o
O
1 -
SDS-DCAA
EBCT = 20 min.
co= 10.3|jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.49)
SCA prediction (RSS = 0.59)
50
100 150 200 250
Scaled operation time (days)
300
350
Figure F-90 Comparison of Dl and SCA methods for predicting the SDS-
DCAA integral breakthrough curve for Water 5
-350-
-------
3 -
O)
o
*J O
s
I
o
O
1 -
0 -li
SDS-TCAA
EBCT = 20 min.
co= 12.7|jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.28)
SCA prediction (RSS = 0.39)
50
100 150 200 250
Scaled operation time (days)
300
350
Figure F-91 Comparison of Dl and SCA methods for predicting the SDS-
TCAA integral breakthrough curve for Water 5
2.0
1.5 -
o
'•SS 1-0 H
o>
o
c
o
O
0.5 -
0.0
SDS-MBAA
EBCT = 20 min.
c0 = 1 H9/L
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50
100 150 200 250
Scaled operation time (days)
300
350
Figure F-92 Comparison of Dl and SCA methods for predicting the SDS-
MBAA integral breakthrough curve for Water 5
-351-
-------
3.5
3.0 -
2.5 -
O)
o
O
1.5 H
1.0 H
0.5 -
0.0
SDS-DBAA
EBCT = 20 min
co = 2 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.47)
SCA prediction (RSS = 0.2)
50
100 150 200 250
Scaled operation time (days)
300
350
Figure F-93 Comparison of Dl and SCA methods for predicting the SDS-
DBAA integral breakthrough curve for Water 5
12
10 -
8 -
o
'•a
o>
o
o 4 .
O 1
2 -
SDS-HAA5
EBCT = 20 min.
c0 = 28 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.56)
SCA prediction (RSS = 0.52)
50
100 150 200 250
Scaled operation time (days)
300
350
Figure F-94 Comparison of Dl and SCA methods for predicting the SDS-
HAA5 integral breakthrough curve for Water 5
-352-
-------
4 -
3 -
.o
"co
o>
o
c
o
O
1 -
SDS-BCAA
EBCT = 20 min
c° = 7.3 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.4)
SCA prediction (RSS = 0.21)
50
100 150 200 250
Scaled operation time (days)
300
350
Figure F-95 Comparison of Dl and SCA methods for predicting the SDS-
BCAA integral breakthrough curve for Water 5
16
14 -
12 -
Q"
1 1°-
o
'•S3 8
o>
c 6
o
O
4 -
2 -
SDS-HAA6
EBCT = 20 min.
c0 = 34 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.82)
SCA prediction (RSS = 0.55)
50
100 150 200 250
Scaled operation time (days)
300
350
Figure F-96 Comparison of Dl and SCA methods for predicting the SDS-
HAA6 integral breakthrough curve for Water 5
-353-
-------
5 -
4 •
O)
o
*J
CD
CD
O
c
0 o
O 2
1 -
SDS-DCBAA
EBCT = 20 min.
c0= 10.7|jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.4)
SCA prediction (RSS = 0.34)
50
100 150 200
Scaled operation time (days)
250
300
350
Figure F-97 Comparison of Dl and SCA methods for predicting the SDS-
DCBAA integral breakthrough curve for Water 5
4 -
3 -
CD
O
c
o
O
1 -
SDS-CDBAA
EBCT = 20 min.
co = 3.7 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.88)
SCA prediction (RSS = 1.1]
50
100 150 200 250
Scaled operation time (days)
300
350
Figure F-98 Comparison of Dl and SCA methods for predicting the SDS-
CDBAA integral breakthrough curve for Water 5
-354-
-------
2.0
1.5 -
o
'•S 1.0 H
o>
o
c
o
o
0.5 -
0.0
SDS-TBAA
EBCT = 20 min.
c0= BMRL
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50
100 150 200
Scaled operation time (days)
250
300
350
Figure F-99 Comparison of Dl and SCA methods for predicting the SDS-
TBAA integral breakthrough curve for Water 5
25
20 -
a 15 -\
o
'-4—I
E
-i—i
| 10-|
o
O
5 -
SDS-HAA9
EBCT = 20 min
c0 = 48 pg/L
• Observed data
Best fit
Dl prediction (RSS = 1.19)
SCA prediction (RSS = 1.13)
50
100 150 200 250
Scaled operation time (days)
300
350
Figure F-100 Comparison of Dl and SCA methods for predicting the SDS-
HAA9 integral breakthrough curve for Water 5
-355-
-------
1.25
1.00 -
O)
0.75 -
g 0.50 H
c
o
O
0.25 -
0.00
TOO
EBCT = 20 min.
c0 = 2.64 mg/L
• Observed data
Best fit
Dl prediction (RSS = 0.015)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-101 Dl method prediction of the TOC integral breakthrough curve
for Water 6
0.025
0.020 -
o
^ 0.015
o>
o
c
CD
.Q
o 0.010
5
.Q
0.005 -
0.000
UV254
EBCT = 20 min.
co = 0.06 1/cm
• Observed data
Best fit
Dl prediction (RSS = 0.0017)
SCA prediction (RSS = 0.0026)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-102 Comparison of Dl and SCA methods for predicting the UV254
integral breakthrough curve for Water 6
-356-
-------
o
o
'-4—'
CD
CD
O
C
O
O
100
80 -
60 H
40 -
20 -
SDS-TOX
EBCT = 20 min.
co = 305 |jg/L Cl-
• Observed data
Best fit
Dl prediction (RSS = 2.39)
SCA prediction (RSS = 6.26)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-103 Comparison of Dl and SCA methods for predicting the SDS-TOX
integral breakthrough curve for Water 6
6 -
O
'•£ 4
CD
O
c
o
O
SDS-CF
EBCT = 20 min
c0 = 55.3 pg/L
50
• Observed data
Best fit
Dl prediction (RSS = 1.37)
SCA prediction (RSS = 0.74)
100 150 200
Scaled operation time (days)
250
300
Figure F-104 Comparison of Dl and SCA methods for predicting the SDS-CF
integral breakthrough curve for Water 6
-357-
-------
30
25 -
20 -
15 -
CD
O
c
° m
O 1U
5 -
SDS-BDCM
EBCT = 20 min.
c0 = 27.4 |jg/L
Observed data
Best fit
Dl prediction (RSS = 3.63)
SCA prediction (RSS = 2.31)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-105 Comparison of Dl and SCA methods for predicting the SDS-
BDCM integral breakthrough curve for Water 6
16
14
12
10
o
"CD
CD
O
c
o
O
SDS-DBCM
EBCT = 20 min.
C0= 41.6|jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.02)
SCA prediction (RSS = 1.84)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-106 Comparison of Dl and SCA methods for predicting the SDS-
DBCM integral breakthrough curve for Water 6
-358-
-------
20
15 -
g
'•S3 10 H
O>
o
c
o
O
5 -
SDS-BF
EBCT = 20 min.
c0 = 3.3 |jg/L
I Observed data
- - - Best fit
— Dl prediction (RSS = 3.47)
— SCA prediction (RSS = 1.86)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-107 Comparison of Dl and SCA methods for predicting the SDS-BF
integral breakthrough curve for Water 6
70
60 -
50 -
40 -
o
"co
o>
o
c
o
O
30 -
20 -
10 -
SDS-TTHM
EBCT = 20 min.
c0= 128|jg/L
• Observed data
Best fit
Dl prediction (RSS = 6.43)
SCA prediction (RSS = 5.91)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-108 Comparison of Dl and SCA methods for predicting the SDS-
TTHM integral breakthrough curve for Water 6
-359-
-------
2.0
1.5 -
o
1.0 -
o
c
o
O
0.5 -
0.0
SDS-MCAA
EBCT = 20 min.
Co= BMRL
Blended effluent concentrations were not detected above the
MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
-SCA prediction (RSS = NA)
• •
50
100 150 200
Scaled operation time (days)
250
300
Figure F-109 Comparison of Dl and SCA methods for predicting the SDS-
MCAA integral breakthrough curve for Water 6
3.5
3.0 -
2.5 -
SDS-DCAA
EBCT = 20 min
co= 17.3|jg/L
0
50
Observed data
Best fit
Dl prediction (RSS = 0.47)
SCA prediction (RSS = 0.31) /
250
100 150 200
Scaled operation time (days)
Figure F-110 Comparison of Dl and SCA methods for predicting the SDS-
DCAA integral breakthrough curve for Water 6
300
-360-
-------
2.0
1.5 -
o
IS 1.0
o
O
0.5 -
o.o
SDS-TCAA
EBCT = 20 min.
c0 = 13.3|jg/L
Insufficient data measured above the MRL to compare
Dl and SCA predictions
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-111 Comparison of Dl and SCA methods for predicting the SDS-
TCAA integral breakthrough curve for Water 6
2.0
1.5 -
o
'•& 1.0 H
o>
o
c
o
O
0.5 -
SDS-MBAA
EBCT = 20 min.
c0= BMRL
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
0.0 -!•-
0
50
100 150 200
Scaled operation time (days)
250
300
Figure F-112 Comparison of Dl and SCA methods for predicting the SDS-
MBAA integral breakthrough curve for Water 6
-361-
-------
10
6 -
o
'-4—'
CD
CD ..
o 4
o
o
2 -
SDS-DBAA
EBCT = 20 min.
c0 = 5.7 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.31)
SCA prediction (RSS = 1.1
50
100 150 200
Scaled operation time (days)
250
300
Figure F-113 Comparison of Dl and SCA methods for predicting the SDS-
DBAA integral breakthrough curve for Water 6
14
12 -
10 -
o
'-4—»
CD
CD
O
C
O
O
6 -
4 -
2 -
SDS-HAA5
EBCT = 20 min.
c0 = 36 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.02)
SCA prediction (RSS = 1.54)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-114 Comparison of Dl and SCA methods for predicting the SDS-
HAA5 integral breakthrough curve for Water 6
-362-
-------
5 -
4 -
O)
3 -
1 -
o -k
SDS-BCAA
EBCT = 20 min.
c0 12.3 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.43)
SCA prediction (RSS = 0.44)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-115 Comparison of Dl and SCA methods for predicting the SDS-
BCAA integral breakthrough curve for Water 6
20
15 -
o
'•& 10
o>
o
c
o
O
5 -
SDS-HAA6
EBCT = 20 min.
c0 = 49 pg/L
• Observed data
Best fit
Dl prediction (RSS = 1.3)
SCA prediction (RSS = 1.92)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-116 Comparison of Dl and SCA methods for predicting the SDS-
HAA6 integral breakthrough curve for Water 6
-363-
-------
2.5
2.0 -
1.5 -
SDS-DCBAA
EBCT = 20 min.
co = 9 pg/L
§ 1.0 H
o
O
0.5 -
50
• Observed data
Best fit
Dl prediction (RSS = 0.34)
SCA prediction (RSS = 0.34)
100 150 200
Scaled operation time (days)
250
300
Figure F-117 Comparison of Dl and SCA methods for predicting the SDS-
DCBAA integral breakthrough curve for Water 6
3.0
2.5 -
2.0 -
O)
1-5 H
0.5 -
0.0
SDS-CDBAA
EBCT = 20 min.
c0 = 3 |jg/L
50
• Observed data
Best fit
Dl prediction (RSS = 0.72)
SCA prediction (RSS = 0.67)
100 150 200
Scaled operation time (days)
250
Figure F-118 Comparison of Dl and SCA methods for predicting the SDS-
CDBAA integral breakthrough curve for Water 6
300
-364-
-------
2.0
1.5 -
O)
g
jo 1-0 H
-i— •
I
o
O
0.5 -
0.0
SDS-TBAA
EBCT = 20 min.
c0= BMRL
Blended effluent concentrations were not detected above the
MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-119 Comparison of Dl and SCA methods for predicting the SDS-
TBAA integral breakthrough curve for Water 6
25
20 -
O)
o
'-4— '
o
O
15 H
10 -
5 -
SDS-HAA9
EBCT = 20 min.
Co = 61 |jg/L
Observed data
Best fit
Dl prediction (RSS = 1.88)
SCA prediction (RSS = 2.41)
50
100 150 200
Scaled operation time (days)
250
300
Figure F-120 Comparison of Dl and SCA methods for predicting the SDS-
HAA9 integral breakthrough curve for Water 6
-365-
-------
2.5
2.0 -
O)
g
'-4— '
CD
8 to H
o
O
0.5 -
0.0
TOC
EBCT = 20 min.
co = 5.58 mg/L
X-
X. '
x .
X , '
X , '
• Observed data
Best fit
Dl prediction (RSS = 0.065)
25
50 75 100
Scaled operation time (days)
125
150
Figure F-121 Dl method prediction of the TOC integral breakthrough curve
for Water 7
0.04
0.03 -
o
CD
CD
O
to
.a
0.02 H
0.01 -
0.00
UV254
EBCT = 20 min.
CD =0.109 1/cm
• Observed data
Best fit
Dl prediction (RSS = 0.0015)
SCA prediction (RSS = 0.0021]
25
50 75 100
Scaled operation time (days)
125
150
Figure F-122 Comparison of Dl and SCA methods for predicting the UV254
integral breakthrough curve for Water 7
-366-
-------
200
150 -
O
.2 100 -
"en
O>
o
c
o
0 50 H
SDS-TOX
EBCT = 20 min.
co = 486 |jg/L Cl
• Observed data
Best fit
Dl prediction (RSS = 6.92)
SCA prediction (RSS = 9.3)
25
50 75 100
Scaled operation time (days)
125
150
Figure F-123 Comparison of Dl and SCA methods for predicting the SDS-
TOX integral breakthrough curve for Water 7
5 -
4 -
o>
o
1 -
SDS-CF
EBCT = 20 min.
C0 = 52.4 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.85)
SCA prediction (RSS = 0.46)
25
50 75 100
Scaled operation time (days)
125
150
Figure F-124 Comparison of Dl and SCA methods for predicting the SDS-CF
integral breakthrough curve for Water 7
-367-
-------
60
50 -
40 -
O)
o
8 30
CD
o
5 20
10 -
SDS-BDCM
EBCT = 20 min.
Co = 65.3 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 5.57)
SCA prediction (RSS = 3.38)
25
50 75 100
Scaled operation time (days)
125
150
Figure
BDCM
F-125 Comparison of Dl and SCA methods for predicting the SDS-
integral breakthrough curve for Water 7
25
20 -
15 -
o
'-4—'
CD
i_
CD
O
c
o
O
5 -
SDS-DBCM
EBCT = 20 min.
c0 = 66.2 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.99)
SCA prediction (RSS = 1.1]
25
50 75 100
Scaled operation time (days)
125
150
Figure F-126 Comparison of Dl and SCA methods for predicting the SDS-
DBCM integral breakthrough curve for Water 7
-368-
-------
50
40 -
30 -
.o
"co
o>
o
c
o
O
10 -
SDS-BF
EBCT = 20 min.
c°= 16 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 11.53)
SCA prediction (RSS = 5.39)
25
50 75 100
Scaled operation time (days)
125
150
Figure F-127 Comparison of Dl and SCA methods for predicting the SDS-BF
integral breakthrough curve for Water 7
125
100 -
a 75 H
o
'-4—'
CD
£
8 50
o
O
25 -
SDS-TTHM
EBCT = 20 min.
co = 200 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 15.22)
SCA prediction (RSS = 9.72)
25
50 75 100
Scaled operation time (days)
125
150
Figure F-128 Comparison of Dl and SCA methods for predicting the SDS-
TTHM integral breakthrough curve for Water 7
-369-
-------
2.0
1.5 -
O)
o
'•a 1.0
O>
o
c
o
O
0.5 -
0.0
Figure
MCA A
SDS-MCAA
EBCT = 20 min.
Blended effluent concentrations were not detected above the
MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
25
50 75 100
Scaled operation time (days)
125
150
F-129 Comparison of Dl and SCA methods for predicting the SDS-
integral breakthrough curve for Water 7
4 -
O)
3 -
c
o
'-4— <
cc
l_
-I— «
§ 2 -I
c
o
O
1 -
0
SDS-DCAA
EBCT = 20 min
co = 25.3 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.59)
SCA prediction (RSS = 0.63)
0
25
125
50 75 100
Scaled operation time (days)
Figure F-130 Comparison of Dl and SCA methods for predicting the SDS-
DCAA integral breakthrough curve for Water 7
150
-370-
-------
3.5
3.0 -
2.5 -
2.0-
SDS-TCAA
EBCT = 20 min.
co= 18 |jg/L
• Observed data
Dl prediction (RSS = 0.4)
. prediction (RSS = 1.34)
8
o
0
1.0 -
0.5 -
25
50 75 100
Scaled operation time (days)
125
150
Figure F-131 Comparison of Dl and SCA methods for predicting the SDS-
TCAA integral breakthrough curve for Water 7
2.0
1.5 -
O)
o
IS 1.0 H
8
c
o
o
0.5 -
0.0
SDS-MBAA
EBCT = 20 min.
C0 = BMRL
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
prediction (RSS = NA)
25
50 75 100
Scaled operation time (days)
125
150
Figure F-132 Comparison of Dl and SCA methods for predicting the SDS-
MBAA integral breakthrough curve for Water 7
-371-
-------
Observed data
Best fit
Dl prediction (RSS = 2.21)
SCA prediction (RSS = 0.97)
25
50 75 100
Scaled operation time (days)
125
150
Figure F-133 Comparison of Dl and SCA methods for predicting the SDS-
DBAA integral breakthrough curve for Water 7
25
20 -
15 -
CD
O
c
o
O
5 -
SDS-HAA5
EBCT = 20 min.
c0 = 65 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 2.02)
SCA prediction (RSS = 1.52)
25
50 75 100
Scaled operation time (days)
125
150
Figure F-134 Comparison of Dl and SCA methods for predicting the SDS-
HAA5 integral breakthrough curve for Water 7
-372-
-------
10
SDS-BCAA
EBCT = 20 min.
c0 = 23.7 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.35)
SCA prediction (RSS = 1.1)
25
50 75 100
Scaled operation time (days)
125
150
Figure F-135 Comparison of Dl and SCA methods for predicting the SDS-
BCAA integral breakthrough curve for Water 7
35
30 -
25 -
20 -
CD
O
O
O
15 -
10 -
5 -
SDS-HAA6
EBCT = 20 min.
c0 = 85 pg/L
• Observed data
Best fit
Dl prediction (RSS = 2.09)
SCA prediction (RSS = 2.25)
25
50 75 100
Scaled operation time (days)
125
150
Figure F-136 Comparison of Dl and SCA methods for predicting the SDS-
HAA6 integral breakthrough curve for Water 7
-373-
-------
4 -
O)
3 -
-—
§ 2
c
o
O
1 -
SDS-DCBAA
EBCT = 20 min.
c° = 26.7 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.13)
SCA prediction (RSS = 0.85)
25
50 75 100
Scaled operation time (days)
125
150
Figure F-137 Comparison of Dl and SCA methods for predicting the SDS-
DCBAA integral breakthrough curve for Water 7
5 -
4 -
O)
g
't-> O _
CO
-I—'
I 1
o ^
O "
1 -
SDS-CDBAA
EBCT = 20 min.
• Observed data
Best fit
Dl prediction (RSS = 0.25)
SCA prediction (RSS = 1.02)
25
50 75 100
Scaled operation time (days)
125
150
Figure F-138 Comparison of Dl and SCA methods for predicting the SDS-
CDBAA integral breakthrough curve for Water 7
-374-
-------
6 -
5 -
O)
3 -
o
O
2 -
1 -
SDS-TBAA
EBCT = 20 min.
c0= BMRL
• Observed data
Best fit
Dl prediction (RSS = 1.92)
SCA prediction (RSS = 1.84)
25
50 75 100
Scaled operation time (days)
125
150
Figure F-139 Comparison of Dl and SCA methods for predicting the SDS-
TBAA integral breakthrough curve for Water 7
60
50 -
40 -
o
|8 30 H
-I—'
I 1
5 2°H
10 -
SDS-HAA9
EBCT = 20 min.
c0= 124|jg/L
• Observed data
Best fit
Dl prediction (RSS = 4.52)
SCA prediction (RSS = 4.7)
25
50 75 100
Scaled operation time (days)
125
150
Figure F-140 Comparison of Dl and SCA methods for predicting the SDS-
HAA9 integral breakthrough curve for Water 7
-375-
-------
1.25
1.00 -
O)
o
0.75 -
§ 0.50 -
o
O
0.25 -
0.00
TOO
EBCT = 7.2 min.
co = 2.02 mg/L
• Observed data
Best fit
Dl prediction (RSS = 0.033)
50 100 150
Scaled operation time (days)
200
Figure F-141 Dl method prediction of the TOC integral breakthrough curve
for Water 8
0.015
E 0.010 -
o
o>
o
0.005 -
0.000
UV254
EBCT = 7.2 min.
co= 0.033 1/cm
• Observed data
Best fit
Dl prediction (RSS = 0.0017)
SCA prediction (RSS = 0.0017)
50 100
Scaled operation time (days)
150
200
Figure F-142 Comparison of Dl and SCA methods for predicting the UV254
integral breakthrough curve for Water 8
-376-
-------
200
:r 150 -
o
o 100 -
-i—'
CD
CD
O
c
O 50 -
SDS-TOX
EBCT = 7.2min.
co= 156|jg/LCI-
SDS-TOX not analyzed in
blended effluent
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50 100
Scaled operation time (days)
150
200
Figure F-143 Comparison of Dl and SCA methods for predicting the SDS-
TOX integral breakthrough curve for Water 8
14
12
10 -
o
'-4— '
CD
o
O 4
2 -
SDS-CF
EBCT = 7.2min.
co= 29.1 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.58)
SCA prediction (RSS = 1.62)
50 100
Scaled operation time (days)
150
200
Figure F-144 Comparison of Dl and SCA methods for predicting the SDS-CF
integral breakthrough curve for Water 8
-377-
-------
4 -
8
a 3.
01 2 -
o n
o
O
1 -
SDS-BDCM
EBCT = 7.2min.
c0 = 2.6 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.2)
SCA prediction (RSS = 0.81)
50 100
Scaled operation time (days)
150
200
Figure F-145 Comparison of Dl and SCA methods for predicting the SDS-
BDCM integral breakthrough curve for Water 8
10
c
o
is
-t—>
§
c
o
O
6 -
4 -
2 -
SDS-DBCM
EBCT = 7.2 min.
co= 10.3|jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.15)
SCA prediction (RSS = 1.19)
50 100
Scaled operation time (days)
150
200
Figure F-146 Comparison of Dl and SCA methods for predicting the SDS-
DBCM integral breakthrough curve for Water 8
-378-
-------
2.0
1.5 -
1.0 H
o>
o
c
o
O
0.5 -
0.0
SDS-BF
EBCT = 7.2min.
co= BMRL
Insufficient data measured above the MRL to
compare Dl and SCA predictions
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50
100
150
200
Scaled operation time (days)
Figure F-147 Comparison of Dl and SCA methods for predicting the SDS-BF
integral breakthrough curve for Water 8
30
25 -
20 -
o
'•§ 15 H
-i—'
o>
o
o 10 -
O
5 -
SDS-TTHM
EBCT = 7.2min.
co = 42 ug/L
• Observed data
Best fit
Dl prediction (RSS = 2.59)
SCA prediction (RSS = 3.87)
50
100
150
200
Scaled operation time (days)
Figure F-148 Comparison of Dl and SCA methods for predicting the SDS-
TTHM integral breakthrough curve for Water 8
-379-
-------
2.0
1.5 -
o
'•a 1.0 H
CD
O
O
O
0.5 -
SDS-MCAA
EBCT = 7.2min.
co= BMRL
0.0 -!•-
0
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50
100
150
200
Scaled operation time (days)
Figure F-149 Comparison of Dl and SCA methods for predicting the SDS-
MCAA integral breakthrough curve for Water 8
5 -
4 -
c
o
3 -
§
§ 2
O
1 -
SDS-DCAA
EBCT = 7.2min.
co= 10.7|jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.68)
SCA prediction (RSS = 0.75)
50 100
Scaled operation time (days)
150
200
Figure F-150 Comparison of Dl and SCA methods for predicting the SDS-
DCAA integral breakthrough curve for Water 8
-380-
-------
4 -
3 -
CD 9 -
o ^ n
o
o
1 -
SDS-TCAA
EBCT = 7.2min.
c0 = 12.7 |jg/L
Observed data
Best fit
Dl prediction (RSS = 0.73)
SCA prediction (RSS = 1)
50 100
Scaled operation time (days)
150
200
Figure F-151 Comparison of Dl and SCA methods for predicting the SDS-
TCAA integral breakthrough curve for Water 8
2.0
c
o
c
CD
O
c
o
O
1.5 -
1.0 -
0.5 -
0.0
SDS-MBAA
EBCT = 7.2 min.
co= BMRL
Blended effluent concentrations were not detected above the
MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50
100
150
200
Scaled operation time (days)
Figure F-152 Comparison of Dl and SCA methods for predicting the SDS-
MBAA integral breakthrough curve for Water 8
-381-
-------
2.0
1.5 -
o
'•S3 1-0 H
CD
O
c
o
O
0.5 -
o.o -k-
0
SDS-DBAA
EBCT = 7.2min.
co= BMRL
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50 100
Scaled operation time (days)
150
200
Figure F-153 Comparison of Dl and SCA methods for predicting the SDS-
DBAA integral breakthrough curve for Water 8
12
10 -
c
o
CO
6 -
c
CD
O
§ 4
O
2 -
0 -!•-
0
SDS-HAA5
EBCT = 7.2min.
co = 23 |jg/L
Observed data
Best fit
Dl prediction (RSS = 1.35)
SCA prediction (RSS = 1.84)
50 100
Scaled operation time (days)
150
200
Figure F-154 Comparison of Dl and SCA methods for predicting the SDS-
HAA5 integral breakthrough curve for Water 8
-382-
-------
3.5
3.0 -
o>
2.0 H
O
0 1.0 H
0.5 -
0.0
SDS-BCAA
EBCT = 7.2min.
CD = 3 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 0.59)
SCA prediction (RSS = 0.31)
50 100
Scaled operation time (days)
150
200
Figure F-155 Comparison of Dl and SCA methods for predicting the SDS-
BCAA integral breakthrough curve for Water 8
14
12 -
10 -
o
'-4—'
CD
CD
O
C
O
O
6 -
4 -
2 -
SDS-HAA6
EBCT = 7.2min.
co = 27 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 1.83)
SCA prediction (RSS = 1.86)
50 100
Scaled operation time (days)
150
200
Figure F-156 Comparison of Dl and SCA methods for predicting the SDS-
HAA6 integral breakthrough curve for Water 8
-383-
-------
2.5
2.0 -
1.5 -
o
"-i—'
s
"c
o>
o
o
O
0.5 -
SDS-DCBAA
EBCT = 7.2min.
C° ~ 3 |jg/L
0.0 -\»-
0
• Observed data
Best fit
Dl prediction (RSS = 0.46)
SCA prediction (RSS = 0.29)
50 100
Scaled operation time (days)
150
200
Figure F-157 Comparison of Dl and SCA methods for predicting the SDS-
DCBAA integral breakthrough curve for Water 8
2.0
1.5 -
1.0 H
o>
o
c
o
O
0.5 -
SDS-CDBAA
EBCT = 7.2 min.
c°= BMRL
Blended effluent concentrations were not detected above
the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
0.0 -{•-
0
50 100
Scaled operation time (days)
150
200
Figure F-158 Comparison of Dl and SCA methods for predicting the SDS-
CDBAA integral breakthrough curve for Water 8
-384-
-------
2.0
1.5 -
1-0 H
O>
o
c
o
O
0.5 -
0.0
SDS-TBAA
EBCT = 7.2min.
CD = BMRL
Blended effluent concentrations were not detected
above the MRL for this parameter
• Observed data
Best fit
Dl prediction (RSS = NA)
SCA prediction (RSS = NA)
50 100
Scaled operation time (days)
150
200
Figure F-159 Comparison of Dl and SCA methods for predicting the SDS-
TBAA integral breakthrough curve for Water 8
16
14 -
12 -
Q"
1 1°-
o
'•^ ft -
CD O 1
CD
£ e ^
o
O
4 -
2 -
0
SDS-HAA9
EBCT = 7.2min.
co = 30 |jg/L
• Observed data
Best fit
Dl prediction (RSS = 2.2)
SCA prediction (RSS = 1.85)
0
50 100
Scaled operation time (days)
150
200
Figure F-160 Comparison of Dl and SCA methods for predicting the SDS-
HAA9 integral breakthrough curve for Water 8
-385-
-------
This page intentionally left blank.
-386-
-------
Appendix G: Logistic Function Model Best-Fit Parameters
-387-
-------
Analyte
TOC
UV-254
SDS-TOX
SDS-TTHM
SDS-HAA5
SDS-HAA6
SDS-HAA9
SDS-CF
SDS-BDCM
SDS-DBCM
SDS-BF
SDS-MCAA
SDS-DCAA
SDS-TCAA
SDS-MBAA
SDS-DBAA
SDS-BCAA
SDS-CDBAA
SDS-DCBAA
SDS-TBAA
Type of curve
fit
Step
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag-peak
No fit
Step-lag
No fit
No fit
Step-lag
Step-lag
No fit
No fit
No fit
Ao
0.00
-0.03
-55.82
-33.46
-1.25
-10.79
-11.49
-2.12
-12.01
-7.70
-0.40
NA
-0.36
NA
NA
-0.05
-2.94
NA
NA
NA
A
3.28
0.09
186.75
100.37
15.59
30.90
33.13
20.37
35.33
31.82
13.59
NA
8.52
NA
NA
5.35
8.66
NA
NA
NA
B
11.11
3.21
4.02
3.45
180.13
12.40
9.47
19.81
5.74
5.91
200.86
NA
1 .39E+04
NA
NA
124.71
6.86
NA
NA
NA
D
0.104
0.058
0.070
0.083
0.273
0.157
0.137
0.068
0.117
0.066
0.388
NA
0.470
NA
NA
0.302
0.103
NA
NA
NA
S
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
-0.148
NA
NA
NA
NA
NA
NA
NA
NA
NA
*P
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
28
NA
NA
NA
NA
NA
NA
NA
NA
NA
n*
13
13
12
11
11
11
11
8
11
11
11
0
9
2
0
11
9
0
3
0
NA: not applicable
'Number of observations above the MRL
Table G-1 Summary of single contactor best-fit logistic function model parameters for Water 1
Analyte
TOC
UV-254
SDS-TOX
SDS-TTHM
SDS-HAA5
SDS-HAA6
SDS-HAA9
SDS-CF
SDS-BDCM
SDS-DBCM
SDS-BF
SDS-MCAA
SDS-DCAA
SDS-TCAA
SDS-MBAA
SDS-DBAA
SDS-BCAA
SDS-CDBAA
SDS-DCBAA
SDS-TBAA
Type of curve
fit
Step
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag-peak
No fit
Step-lag
Step-lag
No fit
Step-lag
Step-lag
No fit
Step-lag
No fit
Ao
0.00
-0.02
-75.03
-38.87
-8.74
-12.84
-14.05
-9.99
-7.34
-12.70
0.06
NA
-4.27
0.00
NA
-0.56
-4.10
NA
-1.21
NA
A
1.78
0.05
218.15
113.39
24.47
35.99
40.85
35.27
31.41
37.33
11.27
NA
12.29
2.10
NA
6.65
11.54
NA
3.42
NA
B
7.48
2.03
2.32
2.50
2.89
2.89
3.06
3.20
8.73
3.51
11.03
NA
2.84
3.51 E+06
NA
11.45
2.86
NA
15.85
NA
D
0.037
0.022
0.025
0.037
0.032
0.032
0.031
0.016
0.071
0.036
0.127
NA
0.021
0.245
NA
0.073
0.030
NA
0.067
NA
S
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
-0.036
NA
NA
NA
NA
NA
NA
NA
NA
NA
*P
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
39
NA
NA
NA
NA
NA
NA
NA
NA
NA
n*
13
13
12
12
11
11
11
10
11
11
12
0
9
6
0
11
10
2
8
0
NA: not applicable
'Number of observations above the MRL
Table G-2 Summary of single contactor best-fit logistic function model parameters for Water 2
-388-
-------
Analyte
TOC
UV-254
SDS-TOX
SDS-TTHM
SDS-HAA5
SDS-HAA6
SDS-HAA9
SDS-CF
SDS-BDCM
SDS-DBCM
SDS-BF
SDS-MCAA
SDS-DCAA
SDS-TCAA
SDS-MBAA
SDS-DBAA
SDS-BCAA
SDS-CDBAA
SDS-DCBAA
SDS-TBAA
Type of curve
fit
Step
Step-lag
Step-lag
Step-lag
Step-lag-peak
Step-lag
Step-lag-peak
Step-lag
Step-lag
Step-lag
Step-lag-peak
No fit
Step-lag
No fit
No fit
Step-lag
Step-lag
No fit
Step-lag
No fit
Ao
0.00
-0.01
-24.55
-45.77
-0.33
2.79
-1.91
-0.91
-1.52
-7.80
3.92
NA
-2.92
NA
NA
0.37
0.44
NA
-0.96
NA
A
1.50
0.04
161.23
144.99
30.45
22.97
48.87
18.88
35.95
36.72
27.13
NA
9.26
NA
NA
10.77
6.91
NA
2.87
NA
B
9.36
3.15
4.95
1.92
24.34
228.87
17.53
36.59
12.00
5.97
32.31
NA
12.56
NA
NA
25.97
85.06
NA
1.75
NA
D
0.027
0.019
0.019
0.020
0.026
0.057
0.024
0.022
0.041
0.018
0.080
NA
0.024
NA
NA
0.048
0.047
NA
0.016
NA
S
NA
NA
NA
NA
0.000
NA
0.000
NA
NA
NA
-0.050
NA
NA
NA
NA
NA
NA
NA
NA
NA
*P
NA
NA
NA
NA
140
NA
140
NA
NA
NA
78
NA
NA
NA
NA
NA
NA
NA
NA
NA
n*
12
12
12
12
11
11
11
9
12
11
12
0
6
5
0
11
11
2
9
0
NA: not applicable
'Number of observations above the MRL
Table G-3 Summary of single contactor best-fit logistic function model parameters for Water 3
Analyte
TOC
UV-254
SDS-TOX
SDS-TTHM
SDS-HAA5
SDS-HAA6
SDS-HAA9
SDS-CF
SDS-BDCM
SDS-DBCM
SDS-BF
SDS-MCAA
SDS-DCAA
SDS-TCAA
SDS-MBAA
SDS-DBAA
SDS-BCAA
SDS-CDBAA
SDS-DCBAA
SDS-TBAA
Type of curve
fit
Step
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag-peak
Step-lag
No fit
No fit
Step-lag
Step-lag
No fit
No fit
Step-lag
No fit
Step-lag
No fit
Ao
0.00
-0.01
-75.14
-19.39
-10.44
-12.99
-14.39
-10.41
-1.62
-5.61
NA
NA
-5.31
-4.35
NA
NA
-2.05
NA
-1.52
NA
A
2.10
0.05
246.00
61.91
51.98
58.13
66.66
41.38
4.69
16.64
NA
NA
20.22
31.76
NA
NA
5.78
NA
8.53
NA
B
27.87
6.52
7.02
5.76
9.16
7.66
7.19
7.49
3.64E+03
6.38
NA
NA
6.76
15.48
NA
NA
8.18
NA
4.59
NA
D
0.044
0.025
0.028
0.026
0.021
0.021
0.020
0.023
0.162
0.032
NA
NA
0.022
0.022
NA
NA
0.040
NA
0.017
NA
S
NA
NA
NA
NA
NA
NA
NA
NA
-0.006
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
*P
NA
NA
NA
NA
NA
NA
NA
NA
91
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
n*
13
13
12
11
10
10
11
11
10
11
0
0
10
10
0
4
10
0
11
0
NA: not applicable
'Number of observations above the MRL
Table G-4 Summary of single contactor best-fit logistic function model parameters for Water 4
-389-
-------
Analyte
TOC
UV-254
SDS-TOX
SDS-TTHM
SDS-HAA5
SDS-HAA6
SDS-HAA9
SDS-CF
SDS-BDCM
SDS-DBCM
SDS-BF
SDS-MCAA
SDS-DCAA
SDS-TCAA
SDS-MBAA
SDS-DBAA
SDS-BCAA
SDS-CDBAA
SDS-DCBAA
SDS-TBAA
Type of curve
fit
Step
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
No fit
Step-lag
Step-lag
No fit
Step-lag
Step-lag
Step-lag-peak
Step-lag
No fit
Ao
0.00
-0.01
-53.67
-22.75
-7.50
-10.90
-17.05
-2.28
-7.19
-6.98
-1.20
NA
-2.40
-3.00
NA
-1.60
-3.40
-2.60
-4.25
NA
A
2.33
0.05
196.80
68.25
22.06
32.01
50.40
16.34
20.65
27.30
3.34
NA
8.70
9.30
NA
4.65
9.96
7.06
12.75
NA
B
12.66
3.83
4.28
3.01
4.85
4.84
6.20
9.86
3.14
4.02
116.01
NA
4.51
7.63
NA
62.79
4.82
148.78
4.99
NA
D
0.023
0.013
0.014
0.015
0.021
0.021
0.023
0.013
0.021
0.013
0.090
NA
0.014
0.019
NA
0.064
0.022
0.060
0.019
NA
S
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
-0.004
NA
NA
*P
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
158
NA
NA
n*
13
13
13
12
11
11
11
10
12
12
11
3
10
9
0
11
11
8
11
0
NA: not applicable
'Number of observations above the MRL
Table G-5 Summary of single contactor best-fit logistic function model parameters for Water 5
Analyte
TOC
UV-254
SDS-TOX
SDS-TTHM
SDS-HAA5
SDS-HAA6
SDS-HAA9
SDS-CF
SDS-BDCM
SDS-DBCM
SDS-BF
SDS-MCAA
SDS-DCAA
SDS-TCAA
SDS-MBAA
SDS-DBAA
SDS-BCAA
SDS-CDBAA
SDS-DCBAA
SDS-TBAA
Type of curve
fit
Step
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step
Step-lag
Step-lag
Step-lag-peak
No fit
Step-lag
No fit
No fit
Step-lag
Step-lag
Step-lag
Step-lag
No fit
Ao
0.00
-0.01
-77.38
-44.21
-9.22
-14.16
-18.87
0.00
-2.82
-13.11
-0.13
NA
-2.97
NA
NA
0.18
-4.62
0.04
-2.59
NA
A
1.98
0.05
283.38
131.09
30.95
44.81
57.94
24.13
31.63
44.11
15.07
NA
12.66
NA
NA
7.36
13.86
2.64
8.53
NA
B
69.54
9.28
9.39
12.29
10.99
12.39
14.38
372.30
605.70
14.75
8.23E+03
NA
17.57
NA
NA
1 .73E+05
18.59
7.20E+12
17.66
NA
D
0.033
0.017
0.018
0.027
0.020
0.022
0.023
0.031
0.062
0.022
0.099
NA
0.018
NA
NA
0.110
0.026
0.263
0.021
NA
S
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
-0.057
NA
NA
NA
NA
NA
NA
NA
NA
NA
*P
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
131
NA
NA
NA
NA
NA
NA
NA
NA
NA
n*
13
13
11
11
10
10
10
8
11
10
11
0
7
4
0
10
9
7
7
0
NA: not applicable
'Number of observations above the MRL
Table G-6 Summary of single contactor best-fit logistic function model parameters for Water 6
-390-
-------
Analyte
TOO
UV-254
SDS-TOX
SDS-TTHM
SDS-HAA5
SDS-HAA6
SDS-HAA9
SDS-CF
SDS-BDCM
SDS-DBCM
SDS-BF
SDS-MCAA
SDS-DCAA
SDS-TCAA
SDS-MBAA
SDS-DBAA
SDS-BCAA
SDS-CDBAA
SDS-DCBAA
SDS-TBAA
Type of curve
fit
Step
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
No fit
Step-lag
Step-lag
No fit
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag-peak
Ao
0.00
-0.03
-134.27
-81.69
-16.06
-23.26
-34.18
-2.35
-34.72
-15.55
1.39
NA
-4.10
-0.03
NA
-0.06
-7.20
-1.78
-2.77
-3.67
A
4.01
0.09
420.30
239.78
47.02
67.10
96.81
25.39
103.28
64.36
35.62
NA
13.83
6.60
NA
16.06
20.05
7.56
16.70
10.33
B
15.49
4.57
4.63
4.18
5.56
6.08
8.04
33.37
5.34
7.25
459.64
NA
3.28
6.68E+05
NA
136.54
7.76
262.52
9.30
471 .50
D
0.049
0.032
0.036
0.042
0.051
0.052
0.060
0.034
0.045
0.031
0.187
NA
0.026
0.193
NA
0.127
0.058
0.138
0.027
0.174
S
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
-0.034
*P
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
65
n*
13
13
12
12
11
11
11
7
11
10
12
0
10
6
0
11
10
9
9
9
NA: not applicable
'Number of observations above the MRL
Table G-7 Summary of single contactor best-fit logistic function model parameters for Water 7
Analyte
TOC
UV-254
SDS-TOX
SDS-TTHM
SDS-HAA5
SDS-HAA6
SDS-HAA9
SDS-CF
SDS-BDCM
SDS-DBCM
SDS-BF
SDS-MCAA
SDS-DCAA
SDS-TCAA
SDS-MBAA
SDS-DBAA
SDS-BCAA
SDS-CDBAA
SDS-DCBAA
SDS-TBAA
Type of curve
fit
Step
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag
Step-lag-peak
Step-lag
No fit
No fit
Step-lag
Step-lag
No fit
No fit
Step-lag
No fit
Step-lag
No fit
Ao
0.00
-0.01
-56.70
-18.32
-6.73
-8.03
-9.14
-11.78
-1.91
-5.12
NA
NA
-3.55
-3.19
NA
NA
-1.30
NA
-0.97
NA
A
1.53
0.03
162.37
52.36
20.20
22.41
25.34
35.35
4.97
14.78
NA
NA
10.03
9.56
NA
NA
3.68
NA
3.05
NA
B
27.83
5.40
6.31
5.00
5.50
9.66
14.18
4.06
1 .26E+03
9.93
NA
NA
8.52
5.04
NA
NA
2.49E+03
NA
1 .44E+04
NA
D
0.064
0.037
0.042
0.036
0.034
0.052
0.062
0.024
0.196
0.055
NA
NA
0.048
0.030
NA
NA
0.199
NA
0.236
NA
S
NA
NA
NA
NA
NA
NA
NA
NA
-0.004
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
*P
NA
NA
NA
NA
NA
NA
NA
NA
113
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
n*
13
13
13
11
10
10
10
11
11
11
0
0
10
9
0
0
10
0
9
0
NA: not applicable
'Number of observations above the MRL
Table G-8 Summary of single contactor best-fit logistic function model parameters for Water 8
-391-
-------
This page intentionally left blank.
-392-
-------
Appendix H: Impact of Extrapolation on SCA Prediction of the
Integral Breakthrough Curve
-393-
-------
3.0
2.5 -
2.0 -
o
•^= -I c
CD '•"
CD
o
o 1.0 H
O
0.5 -
0.0
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.984)
- - Extrapolated logistic function best fit (RA2 = 0.979)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
TOO
EBCT = 20 min.
c0 = 3.08 mg/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure H-1 Impact of extrapolation on Dl prediction of the TOC integral
breakthrough curve for Water 5
0.035
0.030 -
0.025 -
.o
^ 0.020 -
o>
o
c
CD
£ 0.015 -
o
0.010 H
0.005 -
0.000
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.992)
- - Extrapolated logistic function best fit (RA2 = 0.984)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
UV254
EBCT = 20 min.
c0 = 0.051 1/cm
50 100 150 200 250
Scaled operation time (days)
300
350
Figure H-2 Impact of extrapolation on SCA prediction of the UV254 integral
breakthrough curve for Water 5
-394-
-------
O
150
125 -
100 -
.2 75 -
"oo
O>
£ 50 H
o
O
25 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.995)
- - - Extrapolated logistic function best fit (RA2 = 0.99)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-TOX
EBCT = 20 min.
C0 = 205 ug/L Cl-
50 100 150 200 250
Scaled operation time (days)
300
350
Figure H-3 Impact of extrapolation on SCA prediction of the SDS-TOX
integral breakthrough curve for Water 5
12
10 -
o
'« 6 H
o>
o
O
2 -
0 \
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.987)
- - Extrapolated logistic function best fit (RA2 = 0.982)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-CF
EBCT = 20 min.
C0 = 23.7 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure H-4 Impact of extrapolation on SCA prediction of the SDS-CF integral
breakthrough curve for Water 5
-395-
-------
16
14 -
12 -
Q"
| 10-
g
'•£ 8 H
o
O
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.97)
- - Extrapolated logistic function best fit (RA2 = 0.967)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-BDCM
---X)
EBCT = 20 min.
C0 = 10.8 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure H-5 Impact of extrapolation on SCA prediction of the SDS-BDCM
integral breakthrough curve for Water 5
20
o>
o
c
o
o
15 -
10 -
5 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.988)
- - Extrapolated logistic function best fit (RA2 = 0.986)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-DBCM
EBCT = 20 min.
C0 = 22.7 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure H-6 Impact of extrapolation on SCA prediction of the SDS-DBCM
integral breakthrough curve for Water 5
-396-
-------
3.0
2.5 -
§2.0-
3.
c
g
'•S3 1-5 H
CD
o
§ 1.0 -
O
0.5 -
0.0 4
EBCT = 20 min.
C0= 1.2|jg/L
SDS-BF
50
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.969)
- - - Extrapolated logistic function best fit (RA2 = 0.969)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
100 150 200 250
Scaled operation time (days)
300
350
Figure H-7 Impact of extrapolation on SCA prediction of the SDS-BF integral
breakthrough curve for Water 5
50
40 -
30 -
o
'-4—'
CD
20 -
o
O
10 -
o 4
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.979)
- - Extrapolated logistic function best fit (RA2 = 0.975)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-TTHM
EBCT = 20 min.
C0 = 58 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure H-8 Impact of extrapolation on SCA prediction of the SDS-TTHM
integral breakthrough curve for Water 5
-397-
-------
4 -
3 -
g
'-4—'
CD
CD 2 -
o ^ ~
o
o
1 -
0 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = NA)
• • Extrapolated logistic function best fit
O Blended effluent
SCA prediction
SCA prediction - extrapolated
Insufficient data measured above the MRL
to perform curve fit and extrapolation analysis
-Q—Ol—i-O—Oh-
50 100
150 200 250
Scaled operation time (days)
SDS-MCAA
EBCT = 20 min.
c0 = 2 ug/L
300
350
Figure H-9 Impact of extrapolation on SCA prediction of the SDS-MCAA
integral breakthrough curve for Water 5
7
O
'-4—'
CD
6 -
5 -
4 -
3 -
o
O 2 4
1 -
o 4
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.978)
- - Extrapolated logistic function best fit (RA2 = 0.963)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-DCAA
50
EBCT = 20 min.
C0 = 10.3 |jg/L
100 150 200 250
Scaled operation time (days)
300
350
Figure H-10 Impact of extrapolation on SCA prediction of the SDS-DCAA
integral breakthrough curve for Water 5
-398-
-------
6 -
5 -
.g
"co
4 -
3 -
CD
O
O
O 2H
1 -
o 4
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = 0.989)
Extrapolated logistic function best fit (RA2 = 0.987)
Blended effluent
-SCA prediction
SCA prediction - extrapolated
SDS-TCAA
EBCT = 20 min.
C0 = 12.7 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure H-11 Impact of extrapolation on SCA prediction of the SDS-TCAA
integral breakthrough curve for Water 5
4 -
3 -
o
'-4—'
CD
CD
O
c
o
O
1 -
0 -O-
0
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = NA)
Extrapolated logistic function best fit
Blended effluent
-SCA prediction
• SCA prediction - extrapolated
SDS-MBAA
Effluent concentrations were not detected
above the MRL for this parameter
50
100 150 200 250
Scaled operation time (days)
EBCT = 20 min.
c0 = 1 ug/L
300
350
Figure H-12 Impact of extrapolation on SCA prediction of the SDS-MBAA
integral breakthrough curve for Water 5
-399-
-------
3.5
3.0 -
2.5 -
.g
"co
2.0 -
1.5 -
o>
o
c
o
0 1.0 H
0.5 -
o.o 4
EBCT = 20 min.
c0 = 2 ug/L
50
SDS-DBAA
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.992)
- - - Extrapolated logistic function best fit (RA2 = 0.992)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
100 150 200 250
Scaled operation time (days)
300
350
Figure H-13 Impact of extrapolation on SCA prediction of the SDS-DBAA
integral breakthrough curve for Water 5
20
EBCT = 20 mm.
C0 = 28 ug/L
15 -
o
'•S3 10
o>
o
c
o
O
5 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.965)
- - Extrapolated logistic function best fit (RA2 = 0.959)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-HAA5
100
150 200
Scaled operation time (days)
250
300
350
Figure H-14 Impact of extrapolation on SCA prediction of the SDS-HAA5
integral breakthrough curve for Water 5
-400-
-------
6 -
g
'•S3 4H
CD
o
c
o
O
2 -
0 -<
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.987)
- - - Extrapolated logistic function best fit (RA2 = 0.982)
O Blended effluent
SCA prediction
SCA prediction - extrapolated _^^^". D
SDS-BCAA
EBCT = 20 min.
C0 = 7.3 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure H-15 Impact of extrapolation on SCA prediction of the SDS-BCAA
integral breakthrough curve for Water 5
25
20 -
o
'-4—'
CD
o
o
15 -
10 -
5 -
0 4
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.976)
- - Extrapolated logistic function best fit (RA2 = 0.971)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-HAA6
EBCT = 20 min.
c0 = 34 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure H-16 Impact of extrapolation on SCA prediction of the SDS-HAA6
integral breakthrough curve for Water 5
-401-
-------
10
O)
.0
"co
o
O
6-\
2 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.987)
- - Extrapolated logistic function best fit (RA2 = 0.987)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-DCBAA
EBCT = 20 min.
c0= 10.7 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure H-17 Impact of extrapolation on SCA prediction of the SDS-DCBAA
integral breakthrough curve for Water 5
5 -
o
•^
CD
CD
O
o -2\
O
1 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (R^2 = 0.95)
- - Extrapolated logistic function best fit (RA2 = 0.947)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-CDBAA
EBCT = 20 min.
c0 = 3.7 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure H-18 Impact of extrapolation on SCA prediction of the SDS-CDBAA
integral breakthrough curve for Water 5
-402-
-------
4 -
3 -
.g
"co
o>
o
c
o
O
1 -
o -b-
o
SDS-TBAA
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = NA)
Extrapolated logistic function best fit
Blended effluent
-SCA prediction
SCA prediction - extrapolated
Effluent concentrations were not detected
above the MRL for this parameter
50
100 150 200 250
Scaled operation time (days)
EBCT = 20 min.
C0 = BMRL
300
350
Figure H-19 Impact of extrapolation on SCA prediction of the SDS-TBAA
integral breakthrough curve for Water 5
40
o>
£
o
O
35 -
30 -
25 -
20 -
15 -
10 -
5 -
0 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.99)
- - Extrapolated logistic function best fit (RA2 = 0.988)
O Blended effluent
SCA prediction
SCA prediction - extrapolated __^j»*» " n
SDS-HAA9
EBCT = 20 min.
C0 = 48 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure H-20 Impact of extrapolation on SCA prediction of the SDS-HAA9
integral breakthrough curve for Water 5
-403-
-------
2.0
1.5 -
'•a 1-0 H
CD
o
c
o
O
0.5 -
0.0
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.974)
- - Extrapolated logistic function best fit (RA2 = 0.961)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
TOO
EBCT = 7.2 min.
c0 = 2.02 mg/L
50 100
Scaled operation time (days)
150
200
Figure H-21 Impact of extrapolation on Dl prediction of the TOC integral
breakthrough curve for Water 8
0.025
0.020 -
o
^ 0.015
CD
O
CD
o 0.010
0.005 -
0.000 -K>
0
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.994)
- - Extrapolated logistic function best fit (RA2 = 0.98
O Blended effluent
SCA prediction
SCA prediction - extrapolated
UV254
EBCT = 7.2 min.
C0 = 0.033 1/cm
50 100
Scaled operation time (days)
150
200
Figure H-22 Impact of extrapolation on SCA prediction of the UV254 integral
breakthrough curve for Water 8
-404-
-------
O
o
"co
o>
o
c
o
O
125
100 -
75 -
50 -
25 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.99)
- - Extrapolated logistic function best fit (RA2 = 0.98)_
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-TOX
EBCT = 7.2 min.
c0= 156 ug/LCI-
50 100
Scaled operation time (days)
150
200
Figure H-23 Impact of extrapolation on SCA prediction of the SDS-TOX
integral breakthrough curve for Water 8
25
20 -
15 -
o
'-4—'
CD
o
O
10 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.985)
- - Extrapolated logistic function best fit (RA2 = 0.917)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-CF
EBCT = 7.2 min.
C0 = 29.1 ug/L
50 100
Scaled operation time (days)
150
200
Figure H-24 Impact of extrapolation on SCA prediction of the SDS-CF
integral breakthrough curve for Water 8
-405-
-------
4 -
3 -
.g
"co
o>
o
c
o
O
1 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.946)
- - Extrapolated logistic function best fit (RA2 = 0.942)
O Blended effluent
SCA prediction u
• • • • SCA prediction - extrapolated
O
SDS-BDCM
50 100
Scaled operation time (days)
150
200
Figure H-25 Impact of extrapolation on SCA prediction of the SDS-BDCM
integral breakthrough curve for Water 8
12
10 -
^ 8-
o
73
o>
o
c
o
O
6H
4 -
2 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.98)
- - - Extrapolated logistic function best fit (RA2 = 0.978}
O Blended effluent D
SCA prediction
SCA prediction - extrapolated
SDS-DBCM
EBCT = 7.2 min.
C0 = 10.3 ug/L
50 100
Scaled operation time (days)
150
200
Figure H-26 Impact of extrapolation on SCA prediction of the SDS-DBCM
integral breakthrough curve for Water 8
-406-
-------
2.0
1.5 -
g
'•SS 1-0 H
CD
O
O
O
0.5 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = NA)
- - - Extrapolated logistic function best fit
O Blended effluent
SCA prediction
SCA prediction - extrapolated
0.0 -K>
0
O
Insufficient data measured above the MRL
to perform curve fit and extrapolation analysis
JQHXD——CBi CD-
50
-CD-
SDS-BF
EBCT = 7.2min.
C0 = BMRL
100
150
200
Scaled operation time (days)
Figure H-27 Impact of extrapolation on SCA prediction of the SDS-BF
integral breakthrough curve for Water 8
40
30 -
o
'•S3 20 H
CD
O
c
o
O
10 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.98)
- - - Extrapolated logistic function best fit (RA2 = 0.966)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-TTHM
EBCT = 7.2min.
c0 = 42 ug/L
50 100
Scaled operation time (days)
150
200
Figure H-28 Impact of extrapolation on SCA prediction of the SDS-TTHM
integral breakthrough curve for Water 8
-407-
-------
4 -
3 -
g
'-4—'
CD
CD 2
O ^
c
o
O
1 -
o -ta-
o
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = NA)
Extrapolated logistic function best fit
Blended effluent
-SCA prediction
SCA prediction - extrapolated
SDS-MCAA
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 7.2min.
C0 = BMRL
-CD-
50
100
Scaled operation time (days)
-On—
150
200
Figure H-29 Impact of extrapolation on SCA prediction of the SDS-MCAA
integral breakthrough curve for Water 8
6 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.974)
- - Extrapolated logistic function best fit (RA2 = 0.953)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-DCAA
EBCT = 7.2min.
C0 = 10.7 ug/L
50
100
Scaled operation time (days)
150
200
Figure H-30 Impact of extrapolation on SCA prediction of the SDS-DCAA
integral breakthrough curve for Water 8
-408-
-------
6 -
5 -
4 -
g
'-4—'
CD
3 -
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = 0.93)
Extrapolated logistic function best fit (RA2 = 0.835)
Blended effluent
-SCA prediction
SCA prediction - extrapolated
SDS-TCAA
EBCT = 7.2 min.
c0= 12.7 ug/L
50
100
Scaled operation time (days)
150
200
Figure H-31 Impact of extrapolation on SCA prediction of the SDS-TCAA
integral breakthrough curve for Water 8
4 -
3 -
o
'-4—'
CD
CD
O
c
o
O
1 -
0 -K>
0
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = NA)
• • Extrapolated logistic function best fit
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-MBAA
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 7.2 min.
C0= BMRL
-CD-
50 100
Scaled operation time (days)
•Or—
150
200
Figure H-32 Impact of extrapolation on SCA prediction of the SDS-MBAA
integral breakthrough curve for Water 8
-409-
-------
2.0
1.5 -
g
'•S3 1-0 H
o>
o
o
O
0.5 -
0.0 HO-
0
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = NA)
Extrapolated logistic function best fit
Blended effluent
-SCA prediction
SCA prediction - extrapolated
SDS-DBAA
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 7.2 min.
C0 = BMRL
-ii ii i urryjT—rn Oh O>
50
-CD-
•OT-
100
150
200
Scaled operation time (days)
Figure H-33 Impact of extrapolation on SCA prediction of the SDS-DBAA
integral breakthrough curve for Water 8
16
14 -
12 -
10 -
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = 0.955)
Extrapolated logistic function best fit (RA2 = 0.911)
Blended effluent
-SCA prediction
• SCA prediction - extrapolated
SDS-HAA5
EBCT = 7.2 min.
c0 = 23 ug/L
50
100
Scaled operation time (days)
150
200
Figure H-34 Impact of extrapolation on SCA prediction of the SDS-HAA5
integral breakthrough curve for Water 8
-410-
-------
3.5
3.0 -
2.5 -
2.0 -
.o
"co
o>
o
c
o
O
1.5 -
1.0 -
0.5 -
0.0 -K>
0
EBCT = 7.2min.
C0 = 3 |jg/L
50
SDS-BCAA
O
Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.971)
- - - Extrapolated logistic function best fit (RA2 = 0.97)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
100
Scaled operation time (days)
150
200
Figure H-35 Impact of extrapolation on SCA prediction of the SDS-BCAA
integral breakthrough curve for Water 8
20
15 -
o
73 10
o>
o
c
o
O
5 H
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.956)
- - - Extrapolated logistic function best fit (RA2 = 0.923)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-HAA6
50 100
Scaled operation time (days)
150
200
Figure H-36 Impact of extrapolation on SCA prediction of the SDS-HAA6
integral breakthrough curve for Water 8
-411-
-------
3.0
2.5 -
2.0 -
o
73 1-5 H
O>
o
o 1.0 H
0.5 -
0.0 -O
0
EBCT = 7.2min.
C0 = 3 |jg/L
SDS-DCBAA
D Single contactor effluent
• Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = 0.939)
- - - Extrapolated logistic function best fit (RA2 = 0.914)
O Blended effluent
SCA prediction
SCA prediction - extrapolated
50
100
Scaled operation time (days)
150
200
Figure H-37 Impact of extrapolation on SCA prediction of the SDS-DCBAA
integral breakthrough curve for Water 8
2.0
1.5 -
o
73 1-0 H
o>
o
c
o
O
0.5 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = NA)
• • • Extrapolated logistic function best fit
O Blended effluent
SCA prediction
SCA prediction - extrapolated
o.o -k>
o
50
Effluent concentrations were not detected
above the MRL for this parameter
-CT-
SDS-CDBAA
EBCT = 7.2min.
c0 = BMRL
100
Scaled operation time (days)
•Or—
150
200
Figure H-38 Impact of extrapolation on SCA prediction of the SDS-CDBAA
integral breakthrough curve for Water 8
-412-
-------
4 -
O)
.0
I
I 2
o
O
1 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = NA)
• • Extrapolated logistic function best fit
O Blended effluent
SCA prediction
SCA prediction - extrapolated
SDS-TBAA
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 7.2 min.
C0= BMRL
-iMI i urryjT—rn Qh CD-
50
-CD-
100
150
200
Scaled operation time (days)
Figure H-39 Impact of extrapolation on SCA prediction of the SDS-TBAA
integral breakthrough curve for Water 8
20
15 -
o
'•S3 10
o>
o
c
o
O
5 -
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = 0.952)
Extrapolated logistic function best fit (RA2 = 0.924)
Blended effluent
-SCA prediction
• SCA prediction - extrapolated
SDS-HAA9
EBCT = 7.2 min.
c0 = 30 ug/L
50
100
Scaled operation time (days)
150
200
Figure H-40 Impact of extrapolation on SCA prediction of the SDS-HAA9
integral breakthrough curve for Water 8
-413-
-------
This page intentionally left blank.
-414-
-------
Appendix I: Impact of Extrapolation on Dl Prediction of the Integral
Breakthrough Curve
-415-
-------
3.0
2.5 -
2.0 -
o
•^= -I c
CO '•*•
O>
o
o 1.0 H
O
0.5 -
0.0
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.984)
- - Extrapolated logistic function best fit (RA2 = 0.979)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
TOO
EBCT = 20 min.
c0 = 3.08 mg/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure 1-1 Impact of extrapolation on the Dl prediction of the TOC integral
breakthrough curve for Water 5
0.035
0.030 -
0.025 -
o
o>
o
0.020 -
0.000
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.992)
- - Extrapolated logistic function best fit (RA2 = 0.984)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
UV254
EBCT = 20 min.
c0 = 0.051 1/cm
50 100 150 200 250
Scaled operation time (days)
300
350
Figure I-2 Impact of extrapolation on the Dl prediction of the UV254 integral
breakthrough curve for Water 5
-416-
-------
O
150
125 -
100 -
o 75 -
o>
£ 50 H
o
O
25 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.995)
- - - Extrapolated logistic function best fit (RA2 = 0.99)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
SDS-TOX
EBCT = 20 min.
C0 = 205 ug/L Cl-
50 100 150 200 250
Scaled operation time (days)
300
350
Figure I-3 Impact of extrapolation on the Dl prediction of the SDS-TOX
integral breakthrough curve for Water 5
12
10 -
o
'« 6 H
o>
o
O
2 -
0 \
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.987)
- - Extrapolated logistic function best fit (RA2 = 0.982)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
SDS-CF
EBCT = 20 min.
C0 = 23.7 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure I-4 Impact of extrapolation on the Dl prediction of the SDS-CF integral
breakthrough curve for Water 5
-417-
-------
16
14 -
12 -
Q"
| 10-
g
'•£ 8 H
o
O
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.97)
- - Extrapolated logistic function best fit (RA2 = 0.967)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
SDS-BDCM
EBCT = 20 min.
C0 = 10.8 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure I-5 Impact of extrapolation on the Dl prediction of the SDS-BDCM
integral breakthrough curve for Water 5
20
o>
o
c
o
o
15 -
10 -
5 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.988)
- - Extrapolated logistic function best fit (RA2 = 0.986)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
SDS-DBCM
EBCT = 20 min.
C0 = 22.7 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure I-6 Impact of extrapolation on the Dl prediction of the SDS-DBCM
integral breakthrough curve for Water 5
-418-
-------
3.0
2.5 -
§2.0-
3.
c
g
'•S3 1-5 H
CD
o
§ 1.0 -
O
0.5 -
0.0 4
EBCT = 20 min.
C0= 1.2|jg/L
SDS-BF
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.969)
- - - Extrapolated logistic function best fit (RA2 = 0.969)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
50
100 150 200 250
Scaled operation time (days)
300
350
Figure I-7 Impact of extrapolation on the Dl prediction of the SDS-BF integral
breakthrough curve for Water 5
50
40 -
30 -
o
'-4—'
CD
20 -
o
O
10 -
o 4
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.979)
- - Extrapolated logistic function best fit (RA2 = 0.975)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
SDS-TTHM
EBCT = 20 min.
C0 = 58 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure I-8 Impact of extrapolation on the Dl prediction of the SDS-TTHM
integral breakthrough curve for Water 5
-419-
-------
4 -
3 -
g
'-4—'
CD
CD 2 -
o ^ ~
o
o
1 -
0 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = NA)
• • Extrapolated logistic function best fit
O Blended effluent
Dl prediction
Dl prediction - extrapolated
Insufficient data measured above the MRL
to perform curve fit and extrapolation analysis
-Di—Ol—i-O—Oh-
50 100
150 200 250
Scaled operation time (days)
SDS-MCAA
EBCT = 20 min.
c0 = 2 ug/L
300
350
Figure I-9 Impact of extrapolation on the Dl prediction of the SDS-MCAA
integral breakthrough curve for Water 5
7
O
'-4—'
CD
6 -
5 -
4 -
3 -
o
O 2 4
1 -
o 4
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.978)
- - Extrapolated logistic function best fit (RA2 = 0.963)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
SDS-DCAA
50
100 150 200 250
Scaled operation time (days)
EBCT = 20 min.
C0 = 10.3 |jg/L
300
350
Figure 1-10 Impact of extrapolation on the Dl prediction of the SDS-DCAA
integral breakthrough curve for Water 5
-420-
-------
6 -
5 -
.g
"co
4 -
3 -
CD
O
O
O 2H
1 -
o 4
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = 0.989)
Extrapolated logistic function best fit (RA2 = 0.987)
Blended effluent
-Dl prediction
Dl prediction - extrapolated
SDS-TCAA
EBCT = 20 min.
C0 = 12.7 ug/L
50
100 150 200
Scaled operation time (days)
250
300
350
Figure 1-11 Impact of extrapolation on the Dl prediction of the SDS-TCAA
integral breakthrough curve for Water 5
4 -
3 -
o
'-4—'
CD
CD
O
c
o
O
1 -
0 -O-
0
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = NA)
Extrapolated logistic function best fit
Blended effluent
-Dl prediction
• Dl prediction - extrapolated
SDS-MBAA
Effluent concentrations were not detected
above the MRL for this parameter
50
100 150 200 250
Scaled operation time (days)
EBCT = 20 min.
c0 = 1 ug/L
300
350
Figure 1-12 Impact of extrapolation on the Dl prediction of the SDS-MBAA
integral breakthrough curve for Water 5
-421-
-------
3.5
3.0 -
2.5 -
.g
"co
2.0 -
1.5 -
o>
o
c
o
0 1.0 H
0.5 -
o.o 4
EBCT = 20 min
c0 = 2 |jg/L
50
SDS-DBAA
D Single contactor effluent
• Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = 0.992)
- - - Extrapolated logistic function best fit (RA2 = 0.992)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
100 150 200 250
Scaled operation time (days)
300
350
Figure 1-13 Impact of extrapolation on the Dl prediction of the SDS-DBAA
integral breakthrough curve for Water 5
20
EBCT = 20 mm.
C0 = 28 ug/L
15 -
o
'•S3 10
o>
o
c
o
O
5 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.965)
- - Extrapolated logistic function best fit (RA2 = 0.959)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
SDS-HAA5
100
150 200
Scaled operation time (days)
250
300
350
Figure 1-14 Impact of extrapolation on the Dl prediction of the SDS-HAA5
integral breakthrough curve for Water 5
-422-
-------
6 -
g
'•S3 4
CD
o
c
o
O
2 -
0 -<
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.987)
- - Extrapolated logistic function best fit (RA2 = 0.982)
O Blended effluent
Dl prediction
Dl prediction - extrapolated _^^^". D
SDS-BCAA
EBCT = 20 min.
C0 = 7.3 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure 1-15 Impact of extrapolation on the Dl prediction of the SDS-BCAA
integral breakthrough curve for Water 5
25
20 -
o
'-4—'
CD
o
o
15 -
10 -
5 -
0 4
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.976)
- - Extrapolated logistic function best fit (RA2 = 0.971)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
SDS-HAA6
EBCT = 20 min.
c0 = 34 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure 1-16 Impact of extrapolation on the Dl prediction of the SDS-HAA6
integral breakthrough curve for Water 5
-423-
-------
10
O)
.0
"co
o
O
6-\
2 -
0 -i
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.987)
- - Extrapolated logistic function best fit (RA2 = 0.987)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
SDS-DCBAA
EBCT = 20 min.
c0= 10.7 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure 1-17 Impact of extrapolation on the Dl prediction of the SDS-DCBAA
integral breakthrough curve for Water 5
5 -
o
•^
CD
CD
O
o -2\
O
1 -
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (R^2 = 0.95)
Extrapolated logistic function best fit (RA2 = 0.947)
Blended effluent
-Dl prediction
• Dl prediction - extrapolated
SDS-CDBAA
--O
EBCT = 20 min.
c0 = 3.7 ug/L
-O
50
100
150 200 250
Scaled operation time (days)
300
350
Figure 1-18 Impact of extrapolation on the Dl prediction of the SDS-CDBAA
integral breakthrough curve for Water 5
-424-
-------
4 -
3 -
.g
"co
o>
o
c
o
O
1 -
o -b-
o
SDS-TBAA
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = NA)
Extrapolated logistic function best fit
Blended effluent
-Dl prediction
Dl prediction - extrapolated
Effluent concentrations were not detected
above the MRL for this parameter
50
100 150 200 250
Scaled operation time (days)
EBCT = 20 min.
C0 = BMRL
300
350
Figure 1-19 Impact of extrapolation on the Dl prediction of the SDS-TBAA
integral breakthrough curve for Water 5
40
o>
£
o
O
35 -
30 -
25 -
20 -
15 -
10 -
5 -
0 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.99)
- - Extrapolated logistic function best fit (RA2 = 0.988)
O Blended effluent
Dl prediction
Dl prediction - extrapolated __^j»*» " n
SDS-HAA9
EBCT = 20 min.
C0 = 48 ug/L
50
100 150 200 250
Scaled operation time (days)
300
350
Figure I-20 Impact of extrapolation on the Dl prediction of the SDS-HAA9
integral breakthrough curve for Water 5
-425-
-------
2.0
1.5 -
'•a 1-0 H
CD
o
c
o
O
0.5 -
0.0
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = 0.974)
Extrapolated logistic function best fit (RA2 = 0.961)
Blended effluent
-Dl prediction
Dl prediction - extrapolated
TOO
EBCT = 7.2 min.
c0 = 2.02 mg/L
50 100
Scaled operation time (days)
150
200
Figure 1-21 Impact of extrapolation on the Dl prediction of the TOC integral
breakthrough curve for Water 8
0.025
0.020 -
o
^ 0.015
CD
O
CD
o 0.010
0.005 -
0.000 -K>
0
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.994)
- - Extrapolated logistic function best fit (RA2 = 0.98
O Blended effluent
Dl prediction
Dl prediction - extrapolated
UV254
EBCT = 7.2 min.
C0 = 0.033 1/cm
50 100
Scaled operation time (days)
150
200
Figure I-22 Impact of extrapolation on the Dl prediction of the UV254 integral
breakthrough curve for Water 8
-426-
-------
O
o
"co
O>
o
c
o
O
125
100 -
75 -
50 -
25 -
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = 0.99)
Extrapolated logistic function best fit (RA2 = 0.98)_
Blended effluent
-Dl prediction
Dl prediction - extrapolated
SDS-TOX
EBCT = 7.2 min.
c0= 156 ug/LCI-
50 100
Scaled operation time (days)
150
200
Figure I-23 Impact of extrapolation on the Dl prediction of the SDS-TOX
integral breakthrough curve for Water 8
25
20 -
15 -
o
'-4—'
CD
o
O
10 -
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = 0.985)
Extrapolated logistic function best fit (RA2 = 0.917)
Blended effluent
-Dl prediction
• Dl prediction - extrapolated
SDS-CF
EBCT = 7.2 min.
C0 = 29.1 ug/L
50 100
Scaled operation time (days)
150
200
Figure I-24 Impact of extrapolation on the Dl prediction of the SDS-CF
integral breakthrough curve for Water 8
-427-
-------
4 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.946)
- - Extrapolated logistic function best fit (RA2 = 0.942)
O Blended effluent
Dl prediction D
Dl prediction - extrapolated
O
SDS-BDCM
50 100
Scaled operation time (days)
150
200
Figure I-25 Impact of extrapolation on the Dl prediction of the SDS-BDCM
integral breakthrough curve for Water 8
12
10 -
^ 8-
o
13 6
*-t
O>
o
o 4
O
2 -
0 -lO
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.98)
- - Extrapolated logistic function best fit (RA2 = 0.978)
O Blended effluent D
Dl prediction
Dl prediction - extrapolated
SDS-DBCM
EBCT = 7.2 min.
C0 = 10.3 ug/L
50 100
Scaled operation time (days)
150
200
Figure I-26 Impact of extrapolation on the Dl prediction of the SDS-DBCM
integral breakthrough curve for Water 8
-428-
-------
2.0
1.5 -
g
'•S3 1-0 H
o>
o
o
O
0.5 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = NA)
- - - Extrapolated logistic function best fit
O Blended effluent
Dl prediction
Dl prediction - extrapolated
o.o -lo-
o
O
Insufficient data measured above the MRL
to perform curve fit and extrapolation analysis
-ii n i urr>-n—rn—Oh CD-
50
-CD-
SDS-BF
EBCT = 7.2min.
C0 = BMRL
100
150
200
Scaled operation time (days)
Figure I-27 Impact of extrapolation on the Dl prediction of the SDS-BF
integral breakthrough curve for Water 8
40
30 -
o
'•S3 20
o>
o
o
O
10 -
o Ho
o
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.98)
- - Extrapolated logistic function best fit (RA2 = 0.966)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
SOS-^HM
EBCT = 7.2min.
c0 = 42 ug/L
50 100
Scaled operation time (days)
150
200
Figure I-28 Impact of extrapolation on the Dl prediction of the SDS-TTHM
integral breakthrough curve for Water 8
-429-
-------
4 -
3 -
g
'-4—'
CD
CD 9 -
o ^ ~
o
o
1 -
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = NA)
Extrapolated logistic function best fit
Blended effluent
-Dl prediction
Dl prediction - extrapolated
SDS-MCAA
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 7.2min.
C0 = BMRL
-dEEXB-O—CD CDi CD-
50
-CD-
•OT-
100
150
200
Scaled operation time (days)
Figure I-29 Impact of extrapolation on the Dl prediction of the SDS-MCAA
integral breakthrough curve for Water 8
6 -
o
'•S3 4
CD
O
c
o
O
Single contactor effluent
Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.974)
Extrapolated logistic function best fit (RA2 = 0.953)
Blended effluent
Dl prediction
• Dl prediction - extrapolated
SDS-DCAA
EBCT = 7.2min.
C0 = 10.7 ug/L
50
100
Scaled operation time (days)
150
200
Figure I-30 Impact of extrapolation on the Dl prediction of the SDS-DCAA
integral breakthrough curve for Water 8
-430-
-------
6 -
5 -
.g
"co
4 -
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = 0.93)
Extrapolated logistic function best fit (RA2 = 0.835)
Blended effluent
-Dl prediction
Dl prediction - extrapolated
SDS-TCAA
EBCT = 7.2 min.
C0 = 12.7 ug/L
50
100
Scaled operation time (days)
150
200
Figure 1-31 Impact of extrapolation on the Dl prediction of the SDS-TCAA
integral breakthrough curve for Water 8
4 -
3 -
o
'-4—'
CD
CD
O
c
o
O
1 -
0 -K>
0
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = NA)
Extrapolated logistic function best fit
Blended effluent
-Dl prediction
• Dl prediction - extrapolated
SDS-MBAA
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 7.2 min.
C0= BMRL
-CD-
50 100
Scaled operation time (days)
•Or—
150
200
Figure I-32 Impact of extrapolation on the Dl prediction of the SDS-MBAA
integral breakthrough curve for Water 8
-431-
-------
2.0
1.5 -
g
'•S3 1-0 H
CD
o
c
o
O
0.5 -
o.o -k>
0
Single contactor effluent
Extrapolation experimental data points
•Logistic function best fit - all data (RA2 = NA)
Extrapolated logistic function best fit
Blended effluent
-Dl prediction
Dl prediction - extrapolated
SDS-DBAA
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 7.2 min.
C0 = BMRL
-ii n i uoyjT—rn Oh O>
50
-CD-
•OT-
100
150
200
Scaled operation time (days)
Figure I-33 Impact of extrapolation on the Dl prediction of the SDS-DBAA
integral breakthrough curve for Water 8
16
14 -
12 -
10 -
o
'-4—'
CD
Single contactor effluent
Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.955)
Extrapolated logistic function best fit (RA2 = 0.911)
Blended effluent
Dl prediction
Dl prediction - extrapolated
SDS-HAA5
EBCT = 7.2 min.
c0 = 23 ug/L
50
100
Scaled operation time (days)
150
200
Figure I-34 Impact of extrapolation on the Dl prediction of the SDS-HAA5
integral breakthrough curve for Water 8
-432-
-------
3.0
2.5 -
2.0 -
EBCT = 7.2min.
C0 = 3 |jg/L
SDS-BCAA
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.971)
- - - Extrapolated logistic function best fit (RA2 = 0.97)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
50
100
Scaled operation time (days)
150
200
Figure I-35 Impact of extrapolation on the Dl prediction of the SDS-BCAA
integral breakthrough curve for Water 8
20
15 -
o
'•S3 10
o>
o
c
o
O
5 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.956)
- - Extrapolated logistic function best fit (RA2 = 0.923)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
50
SDS-HAA6
EBCT = 7.2min.
c0 = 27 ug/L
100
Scaled operation time (days)
150
200
Figure I-36 Impact of extrapolation on the Dl prediction of the SDS-HAA6
integral breakthrough curve for Water 8
-433-
-------
3.0
2.5 -
2.0 -
g
'•§ 1.5 H
"c
O>
o
o 1.0 -
O
0.5 -
0.0
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.939)
- - Extrapolated logistic function best fit (RA2 = 0.914)
O Blended effluent
Dl prediction ...
Dl prediction - extrapolated
50
SDS-DCBAA
O
EBCT = 7.2 min.
c0 = 3 ug/L
100
Scaled operation time (days)
150
200
Figure I-37 Impact of extrapolation on the Dl prediction of the SDS-DCBAA
integral breakthrough curve for Water 8
2.0
1.5 -
o
'•S3 1-0 H
o>
o
c
o
O
0.5 -
0.0 -K>
0
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = NA)
• • • Extrapolated logistic function best fit
O Blended effluent
Dl prediction
Dl prediction - extrapolated
SDS-CDBAA
Effluent concentrations were not detected
above the MRL for this parameter
EBCT = 7.2 min.
C0 = BMRL
-in11 M"r>-n-^"n——m, rn-
50
-CD-
100
•Or—
150
200
Scaled operation time (days)
Figure I-38 Impact of extrapolation on the Dl prediction of the SDS-CDBAA
integral breakthrough curve for Water 8
-434-
-------
4 -
3 -
g
'-4—'
CD
CD 9 -
o ^ ~
o
o
1 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = NA)
- - Extrapolated logistic function best fit
O Blended effluent
Dl prediction
Dl prediction - extrapolated
Effluent concentrations were not detected
above the MRL for this parameter
JOHXB——Oli CT-
50
-CT-
SDS-TBAA
EBCT = 7.2min.
C0= BMRL
100
-On—
150
200
Scaled operation time (days)
Figure I-39 Impact of extrapolation on the Dl prediction of the SDS-TBAA
integral breakthrough curve for Water 8
20
15 -
o
'•SS 10
CD
O
c
o
O
5 -
D Single contactor effluent
• Extrapolation experimental data points
Logistic function best fit - all data (RA2 = 0.952)
- - Extrapolated logistic function best fit (RA2 = 0.924)
O Blended effluent
Dl prediction
Dl prediction - extrapolated
50
SDS-HAA9
EBCT = 7.2min.
c0 = 30 ug/L
100
Scaled operation time (days)
150
200
Figure I-40 Impact of extrapolation on the Dl prediction of the SDS-HAA9
integral breakthrough curve for Water 8
-435-
------- |