June 2000
Environmental Technology
Verification Report
Monitek
TST-SC
On-Line Turbidimeter
Prepared by
Baffeiie
. . . Putting Technology To Work
Battel le
Under a cooperative agreement with
oEPA U.S. Environmental Protection Agency
ETtf ETtf ET^

-------
June 2000
Environmental Technology Verification
Report
ETV Advanced Monitoring Systems Pilot
Monitek
TST-SC On-Line Turbidimeter
By
Kenneth Cowen
Thomas Kelly
Brian Canterbury
Karen Riggs
Battelle
Columbus, Ohio 43201

-------
Notice
The U.S. Environmental Protection Agency (EPA), through its Office of Research and
Development, has financially supported and collaborated in the extramural program described
here. This document has been peer reviewed by the Agency and recommended for public release.
Mention of trade names or commercial products does not constitute endorsement or
recommendation by the EPA for use.
11

-------
Foreword
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the
Nation's air, water, and land resources. Under a mandate of national environmental laws, the
Agency strives to formulate and implement actions leading to a compatible balance between
human activities and the ability of natural systems to support and nurture life. To meet this
mandate, the EPA's Office of Research and Development (ORD) provides data and science
support that can be used to solve environmental problems and to build the scientific knowledge
base needed to manage our ecological resources wisely, to understand how pollutants affect our
health, and to prevent or reduce environmental risks.
The Environmental Technology Verification (ETV) Program has been established by the EPA to
verify the performance characteristics of innovative environmental technology across all media
and to report this objective information to permitters, buyers, and users of the technology, thus
substantially accelerating the entrance of new environmental technologies into the marketplace.
Verification Organizations oversee and report verification activities based on testing and Quality
Assurance protocols developed with input from major stakeholders and customer groups
associated with the technology area. At present, there are twelve environmental technology areas
covered by ETV. Information about each of the environmental technology areas covered by ETV
can be found on the Internet at http://www.epa.gov/etv.htm.
Effective verifications of monitoring technologies are needed to assess environmental quality and
to supply cost and performance data to select the most appropriate technology for that assess-
ment. In 1997, through a competitive cooperative agreement, Battelle was awarded EPA funding
and support to plan, coordinate, and conduct such verification tests for "Advanced Monitoring
Systems for Air, Water, and Soil" and report the results to the community at large. Information
concerning this specific environmental technology area can be found on the Internet at
http://www.epa.gov/etv/07/07_main.htm.
111

-------
Acknowledgments
The authors wish to acknowledge the support of all those who helped plan and conduct the
verification test, analyze the data, and prepare this report. In particular we would like to thank the
staff at the Dublin Road Water Plant, including Tom Camden and Terry Nichols. We also
acknowledge the participation of Mike Hurst and Chris Williams of Monitek in this verification
test. We would like to thank the Hach Company for supplying the two reference turbidimeters
used in this test.
IV

-------
Contents
Notice	 ii
Foreword 	iii
Acknowledgments	 iv
List of Abbreviations	 x
1.	Background 	1
2.	Technology Description 	2
3.	Test Design and Procedures	3
3.1	Introduction	3
3.2	Test Design Considerations	3
3.3	Experimental Apparatus 	4
3.4	Reference Instruments	5
3.5	Off-Line Testing 	6
3.5.1	Linearity 	8
3.5.2	Accuracy and Precision	8
3.5.3	Water Temperature	8
3.5.4	Flow Rate 	9
3.5.5	Color	9
3.6	On-Line Testing	10
3.6.1	Accuracy	10
3.6.2	Drift	10
4.	Quality Assurance/Quality Control 	12
4.1	Data Review and Validation 	12
4.2	Deviations from the Test/QA Plan	12
4.3	Calibration	13
4.3.1	Reference Turbidimeters 	13
4.3.2	Temperature Sensors 	14
4.3.3	Flow Meters	14
4.3.4	pHMeter 	14
v

-------
4.4	Data Collection 	15
4.5	Assessments and Audits	15
4.5.1	Technical Systems Audit	15
4.5.2	Performance Evaluation Audit	15
4.5.3	Verification Test Data Audit	16
4.6	Audit Reporting	18
5.	Statistical Methods	19
5.1	Off-Line Testing 	19
5.1.1	Linearity	19
5.1.2	Accuracy 	19
5.1.3	Precision	20
5.1.4	Water Temperature Effects 	20
5.1.5	Flow Rate Sensitivity	20
5.1.6	Color Effects 	20
5.2	On-Line Testing	21
5.2.1	Accuracy 	21
5.2.2	Drift	21
6.	Test Results 	22
6.1	Off-Line Testing 	22
6.1.1	Linearity	22
6.1.2	Accuracy 	24
6.1.3	Precision	25
6.1.4	Water Temperature Effects 	27
6.1.5	Flow Rate	27
6.1.6	Color Effects 	30
6.2	On-Line Testing	32
6.2.1	Accuracy 	33
6.2.2	Drift	35
6.3	Other Performance Parameters 	38
6.3.1	Cost 	38
6.3.2	Maintenance/Operational Factors 	39
VI

-------
7.	Performance Summary	40
8.	References 	42
Appendix A: Example Data Recording Sheet	 A-l
Appendix B: Technical Systems Audit Report 	B-l
Figures
Figure 2-1. Monitek Monilog Model TST-SC On-Line Turbidimeter 	2
Figure 3-1. Schematic Representation of Recirculation System 	6
Figure 4-la. Control Chart for Performance Evaluation Calibration
Checks of the ISO 7027 Reference Turbidimeter	17
Figure 4-lb. Control Chart for Performance Evaluation Calibration
Checks of the Method 180.1 Reference Turbidimeter 	17
Figure 6-la. Linearity Plot for Monitek TST-SC Turbidimeter vs.
ISO 7027 Reference Turbidimeter 	23
Figure 6-lb. Linearity Plot for Monitek TST-SC Turbidimeter vs.
Method 180.1 Reference Turbidimeter	24
Figure 6-2a. Effect of Temperature on Monitek TST-SC Turbidity Readings
vs. ISO 7027 at Both 0.3 and 5 NTU	28
Figure 6-2b. Effect of Temperature on Monitek TST-SC Turbidity Readings
vs. Method 180.1 at Both 0.3 and 5 NTU	28
Figure 6-3. Effect of Sample Flow Rate on Monitek TST-SC
Turbidimeter Response	29
Figure 6-4a. Effect of Color on Relative Turbidity with the Monitek TST-SC
Turbidimeter vs. the ISO 7027 at Both 0.1 and 5 NTU	31
Figure 6-4b. Effect of Color on Relative Turbidity at 5 NTU with the Monitek TST-SC
Turbidimeter vs. Method 180.1 at Both 0.1 and 5 NTU	31
vii

-------
Figure 6-5. Summary of Stream Turbidity Data from On-Line Testing
of Monitek TST-SC	33
Figure 6-6. Twice-Weekly Calibration Checks During On-Line Testing
of the Monitek TST-SC Turbidimeter	36
Figure 6-7a. Final Linearity Plot for Monitek TST-SC vs. ISO 7027
Reference Turbidimeter	37
Figure 6-7b. Final Linearity Plot for Monitek TST-SC vs. Method 180.1
Reference Turbidimeter	37
Tables
Table 3-1. Performance Characteristics Evaluated and Schedule of
Verification Tests Conducted on Monitek TST-SC Turbidimeter	3
Table 3-2.	Summary of Measurements for Off-Line Testing	7
Table 3-3.	Summary of Measurements for On-Line Testing	10
Table 4-1.	Results of Linearity Check of Reference Turbidimeters	13
Table 4-2.	Summary of Flow Meter Calibration Check	14
Table 4-3. Results of Calibration Checks of Thermocouple Used
in Verification Test	18
Table 6-1. Statistical Results of Initial Linearity Test on
Monitek TST-SC Turbidimeter	23
Table 6-2. Bias of Monitek TST-SC Turbidimeter Relative to
Reference Measurements on Prepared Test Solutions 	25
Table 6-3. Adjusted Turbidity Readings for Precision Calculations
of the Monitek TST-SC Turbidimeter 	26
Table 6-4. Precision of Monitek TST-SC Turbidimeter and of the
Reference Turbidimeters	26
Table 6-5. Statistical Results of Temperature Test on the
Monitek TST-SC Turbidimeter	29
Vlll

-------
Table 6-6. Statistical Results of Flow Rate Test for the
Monitek TST-SC Turbidimeter	30
Table 6-7. Statistical Results of the Color Test with the
Monitek TST-SC Turbidimeter	32
Table 6-8. On-Line Daily Accuracy Check Results	34
Table 6-9. Results of Calibration Checks Performed During On-Line Testing	36
Table 6-10. Statistical Results of Final Linearity Test	38
Table 6-11. Comparison of Results from Linearity Tests at Beginning and
End of the Verification Test	38
IX

-------
List of Abbreviations
AC
alternating current
AMS
Advanced Monitoring Systems
CU
color unit
DC
direct current
EPA
U.S. Environmental Protection Agency
ETV
Environmental Technology Verification
gpm
gallons per minute
LOD
limit of detection
NIST
National Institute of Standards and Technology
NPT
normal pipe thread
NTU
nephelometric turbidity unit
QA
quality assurance
QC
quality control
QMP
Quality Management Plan
RSD
relative standard deviation
x

-------
Chapter 1
Background
The U.S. Environmental Protection Agency (EPA) has created the Environmental Technology
Verification (ETV) Program to facilitate the deployment of innovative environmental tech-
nologies through performance verification and dissemination of information. The goal of the
ETV Program is to further environmental protection by substantially accelerating the acceptance
and use of improved and cost-effective technologies. ETV seeks to achieve this goal by provid-
ing high quality, peer-reviewed data on technology performance to those involved in the design,
distribution, permitting, purchase, and use of environmental technologies.
ETV works in partnership with recognized testing organizations; with stakeholder groups
consisting of regulators, buyers and vendor organizations; and with the full participation of
individual technology developers. The program evaluates the performance of innovative tech-
nologies by developing test plans that are responsive to the needs of stakeholders, conducting
field or laboratory tests (as appropriate), collecting and analyzing data, and preparing peer-
reviewed reports. All evaluations are conducted in accordance with rigorous quality assurance
protocols to ensure that data of known and adequate quality are generated and that the results are
defensible.
The EPA's National Exposure Research Laboratory and its verification organization partner,
Battelle, operate the Advanced Monitoring Systems (AMS) pilot under ETV. The AMS pilot
recently evaluated the performance of on-line turbidimeters for use in water treatment facilities.
This verification report presents the procedures and results of the verification test for the Monitek
TST-SC on-line turbidimeter.
1

-------
Chapter 2
Technology Description
The following description of the Monitek TST-SC turbidimeter is based on information provided
by the vendor.
The Monitek Technologies Monilog Model TST-SC uses on-line turbidimeter alternating four-
beam technology to measure turbidity and suspended solids. This method uses an array of two
light-emitting diodes (LEDs) and two detectors. The LEDs are oriented 90 degrees from each
other and are pulsed alternately. This causes the two detectors (located across the process stream
and 90 degrees apart) to alternate between their functions as the scatter and direct beam detectors.
Since each detector element alternates between functions as the measurement and reference
channel, the system compensates for variations in the LED output, window coatings, and other
effects. The 90 degree angle, the correction of the ray path with lenses, and the infrared light
wavelength of 875 nm meet the requirements of ISO 7027.(1) The lower limit of the TST-SC's
response range is 0.1 NTU.

Figure 2-1. Monitek Monilog Model TST-SC Qn-Line
Turbidimeter
2

-------
Chapter 3
Test Design and Procedures
3.1 Introduction
The verification test was conducted according to procedures specified in the Test/QA Plan for
Verification of On-Line Turbidimeters,(2) Performance characteristics evaluated in the verifica-
tion test are listed in Table 3-1, along with the dates that data were collected for these evalua-
tions. The test was conducted at a full-scale municipal water treatment facility in Columbus,
Ohio. The verification test described in this report was conducted from September 9 through
October 26, 1999, as indicated in Table 3-1.
Table 3-1. Performance Characteristics Evaluated and Schedule of Verification Tests
Conducted on the Monitek TST-SC Turbidimeter
Performance Characteristic
Date Data Collected
Off-Line Phase
Linearity
Accuracy
Precision
Water temperature effects
Flow rate sensitivity
Color effects
On-Line Phase
Accuracy
Calibration checks
September 9 to 10; October 20-21
September 9 to 10
September 9 to 10
September 10, 14 to 15
September 15 to 16
October 25 to 26
September 17 to October 18
September 23, 24, 27, 30 and October 6, 8, 12, 18
3.2 Test Design Considerations
Since turbidity is a measurement of light scattering, a number of factors can influence the
measurement of turbidity in a given sample solution. Instrument design, including light source
selection and geometric differences, may result in significant differences between the responses
of different turbidimeters. Further differences may result from the variable nature of both the size
and composition of particles typically found in water streams, relative to those in standard
3

-------
solutions made with formazin or polymer beads. These issues were addressed in this verification
test in two ways: (1) by using different instrumental designs for reference turbidimeters, and (2)
by evaluating a variety of samples.
To avoid potential bias associated with a single method of comparison, the verification test used
two reference methods for data comparisons: ISO 7027, "Water Quality—Determination of
Turbidity"(1) and EPA Method 180.1, "Determination of Turbidity by Nephelometry."(3) Both of
these methods measure turbidity using a nephelometric turbidimeter, but they differ in the type of
light source and the wavelength used. ISO 7027 calls for an infrared light source, whereas
Method 180.1 calls for a visible light source. The Monitek TST-SC is designed to conform to the
requirements of ISO 7027, and thus that method is the appropriate reference for verification of
the TST-SC performance. Verification results presented in this report, and summarized in the
Verification Statement, are based on comparisons with the ISO 7027 data. However, secondary
comparisons also are shown in this report, based on data from the TST-SC and Method 180.1.
The secondary comparisons are of interest because Method 180.1 is widely recognized in the
U.S. and is designated as the required method for drinking water compliance measurements. The
secondary comparisons are shown only to illustrate the performance capabilities of the TST-SC
and should not be taken as having equal weight as the comparisons with ISO 7027.
Additionally, to assess the response of the Monitek TST-SC turbidimeter to both prepared
solutions and real-world water samples, verification involved both off-line and on-line phases.
The off-line phase challenged the turbidimeter with a series of prepared standards and other test
solutions to verify performance under controlled conditions. The on-line phase assessed long-
term performance under realistic operating conditions by monitoring a sample stream in a
municipal water treatment plant under normal operation. With the cooperation of the City of
Columbus Water Division, both off-line and on-line phases were performed at the Dublin Road
Water Plant in Columbus, Ohio.
3.3 Experimental Apparatus
On-line turbidimeters measure turbidity continuously on flowing sample streams as opposed to
the static grab samples analyzed by the bench-top reference turbidimeters. Consequently, great
care was taken to ensure that the samples collected for reference analysis were representative of
the sample flow measured by the Monitek TST-SC turbidimeter. A cylindrical distribution mani-
fold provided identical sample streams to sample ports spaced equally around the circumference
of the manifold. Throughout the verification test, three ports were used for the turbidimeters
being verified, and one port provided a stream for the grab samples. A single port centered in the
bottom of the manifold introduced the sample stream to the distribution manifold. All the ports
were tapped for V2" male normal pipe thread (NPT) fittings, and hard plastic compression fittings
were used to connect the distribution manifold to the tubing (V2" OD polyethylene) used in the
recirculation system. Using a consistent tubing size and fitting style enabled rapid switching of
the turbidimeters on a scheduled basis among the ports on the distribution manifold. Providing
identical samples to each of the manifold ports minimized biases arising from water quality or
turbulence issues; rotation of the technologies to each of these ports was used to identify if biases
existed.
4

-------
A schematic representation of this recirculation system is provided in Figure 3-1, where T1
through T3 represent the three on-line turbidimeters undergoing verification testing. T1 repre-
sents the Monitek TST-SC turbidimeter. Prepared solutions were supplied to all three turbidi-
meters simultaneously in a closed-loop recirculation system that used a 40-L reservoir and a
centrifugal pump. Stream water from the plant was sampled from a pressurized source in a once-
through configuration (i.e., without use of the pump or reservoir). In-line particle filters were
inserted into the water flow using appropriate valving when reduction of turbidity levels was
needed.
Before verification testing began, a series of five grab samples was collected from each port on
the cylindrical manifold while recirculating a formazin solution with a nominal turbidity of
0.5 nephelometric turbidity unit (NTU). These samples were analyzed with the reference
turbidimeters and compared to ensure uniformity of the turbidity of the solution. Comparison of
the sample analyses indicated agreement in turbidity readings within ± 5% among all of the ports.
Before verification testing began, the on-line turbidimeters verified in this test were installed in
the test apparatus at the Dublin Road Water Plant. The Monitek TST-SC turbidimeter (Serial
Number 0618) was installed by Battelle and inspected by a field representative of Monitek. As
supplied, the Monitek turbidimeter included I-V2" female NPT fittings for direct in-line pipe
mounting. Reducing fittings were used to couple the turbidimeter to the recirculation system,
which was built with V2" OD tubing and V2" NPT plastic compression fittings. Much of the
recirculation system, including the flow meters and the distribution manifold, was mounted to a
'/V-thick aluminum panel installed in the water plant specifically for this verification test. The
TST-SC turbidimeter was bolted to the panel using two L brackets that sandwiched the
turbidimeter housing and were held in place using two of the six bolts around the housing
circumference. Since the diameter of the tubing was smaller than the inside diameter of the
Monitek TST-SC turbidimeter, the system was mounted vertically to ensure that no pockets of air
were present in the turbidimeter during testing.
No control unit was provided with the Monitek TST-SC turbidimeter for verification testing,
although control units are available from Monitek. Power was supplied to the sensor using a
Lambda LQ-415 power supply (0 to 250 V, 80 mA, Battelle Asset Number LN 426441) provided
by Battelle. The Monitek turbidimeter sensor was provided with 9 to 10V power from this power
supply throughout the verification test. The sensor output was converted from a 4 to 20 mA
signal to a direct current (DC) voltage using a precision resistor, and was recorded every
10 seconds throughout the test using LabTech Notebook software, which was run on a personal
computer at the test site.
3.4 Reference Instruments
Owing to the nature of turbidity measurement and the inherent differences in response arising
from different instrumental designs, separate bench-top turbidimeters meeting the design criteria
detailed in ISO 7027(1) and EPA Method 180.1(3) were used as reference instruments in this test.
Both methods describe procedures to measure the nephelometric light scattering of a formazin
solution, albeit with different prescribed instrumental design parameters. The primary difference
5

-------
Flow Meter
Ball Valve
Grab Sample Port
A - Distribution Manifold
Exit Port for On-
Line Tests
Off-line
Sample
Introduction
Port
rfxH


T3
By-Pass Channel
51
H
In-Line Filters
Reservoir
On-line Sample
Introduction Port
Figure 3-1. Schematic Representation of Recirculation System
between these two methods is in the choice of light source. Method 180.1 requires the use of a
broadband incandescent tungsten lamp, while ISO 7027 requires the use of a narrowband IR
source. Since the TST-SC is designed to comply with ISO 7027 requirements, that reference
method is the basis for this verification. Comparisons of data with Method 180.1 are also shown,
however, because of the widespread recognition and use of that method. Method 180.1 com-
parisons are secondary to the ISO 7027 comparisons used for verification. The bench-top
turbidimeters used as the reference methods were the Hach 2100AN (Serial Number
980300001366) and the Hach 2100AN IS (Serial Number 950700000173), which, according to
the manufacturer's literature, comply with the design specifications described in ISO 7027(1) and
EPA Method 180.1(3), respectively. Throughout the test the reference turbidimeters were operated
in the non-ratio mode.
3.5 Off-Line Testing
The off-line phase of the verification test involved off-line sample introduction aimed at
assessing the linearity, accuracy, and precision of the Monitek TST-SC on-line turbidimeter
relative to the reference methods. Additionally, response to various upset conditions was
quantified. As a means of testing these parameters, the off-line test phase included the intro-
duction of standard formazin solutions or other samples, and the intentional manipulation of flow
and water quality parameters.
6

-------
Throughout the verification test, continuous turbidity measurements from the Monitek
turbidimeter were recorded at preset intervals using LabTech Notebook software. Grab samples
were collected simultaneously with some of these recorded measurements and analyzed using
bench-top reference turbidimeters to provide a basis of comparison for the performance
evaluations. The collection of grab samples was timed to coincide within 10 seconds with the
recording of real-time turbidity measurements from the Monitek TST-SC, and the grab samples
were analyzed within three minutes after collection to minimize possible temperature and settling
effects.
Additionally, off-line testing included monitoring the instrumental responses of the Monitek
turbidimeter to variations in water temperature, flow rate, and color. Each of these parameters
was varied within a range consistent with conditions encountered under typical plant operation.
The following subsections describe the procedures used for the off-line phase of the verification
test.
Table 3-2 provides a summary of the parameters tested in the off-line phase, the test solutions
used, and the number of readings recorded for each parameter.
Table 3-2. Summary of Measurements for Off-Line Testing
Parameter Tested
Test Solution
Number of Readings
Linearity
Filtered water
5

(<0.1 NTU)

Linearity (accuracy, precision)®
0.3 NTUFormazin
5
Linearity (accuracy, precision)
0.5 NTU Formazin
5
Linearity (accuracy, precision)
2 NTU Formazin
5
Linearity (accuracy, precision)
5 NTU Formazin
5
Water Temperature Effect
0.3 NTUFormazin
5 each at 16, 21, 27°C
Water Temperature Effect
5 NTU Formazin
5 each at 16, 21, 27°C
Flow Rate Effect
0.3 NTUFormazin
5
Flow Rate Effect
5 NTU Formazin
5
Color Effect
-0.1 NTU
5 each at 5, 15, 30 CU
Color Effect
5 NTU Formazin
5 each at 35, 45, 60 CU
a: () indicates additional parameters analyzed using collected data.
7

-------
3.5.1	Linearity
Linearity was measured in the range from approximately 0.05 to 5 NTU as an initial check in the
off-line phase. The recirculation system was filled with distilled, dionized water, which was then
recirculated and filtered in the test apparatus using a 0.2-|j,m pleated polypropylene filter for
24 hours. After filtering, the in-line filter was bypassed and the turbidity of the water in the
recirculation system was measured to be approximately 0.05 NTU. A series of five turbidity
measurements was taken at that turbidity level with intervals of at least five minutes between
successive measurements. A corresponding set of five measurements also was recorded at
approximately 0.3, 0.5, 2, and 5 NTU. To reach each turbidity level, a small amount of 4000
NTU StablCal formazin stock solution was diluted in the recirculation system and allowed to
flow through the recirculation system unfiltered for at least 15 minutes before turbidity readings
were recorded. At each turbidity level, a series of five turbidity readings was recorded with at
least a 5-minute interval between successive readings. These readings were compared to the
reference measurements of grab samples collected simultaneously with each reading; that is, the
turbidity of the solutions was determined by measurement with the reference turbidimeters,
rather than simply by calculations based on the dilution process. After the prescribed
measurements were recorded at each turbidity level, additional formazin stock solution was
added to the recirculation system to increase the turbidity of the solution to the next value in the
series.
Before measurements were recorded, the calibration of the reference turbidimeters was checked
using a 0.5 NTU StablCal formazin solution purchased from Hach Company, Loveland,
Colorado. Pursuant to the requirements of the test/QA plan,(1) agreement between the reference
measurement and the certified turbidity of the standard was required to be within 10% before
recording any series of measurements. After each series of measurements, the calibration of the
reference turbidimeters was again checked with the same standard, and the same acceptance
limits were applied. In addition to the 0.5 NTU calibration checks, before and after the measure-
ments on the filtered water, a < 0.1 NTU blank standard also was measured to ensure proper
calibration of the reference instruments at low levels. The < 0.1 NTU standard also was pur-
chased from Hach Company; agreement between the reference measurement and the turbidity
reported on the certificate of analysis was required to be within 0.02 NTU.
3.5.2	Accuracy and Precision
Data obtained from the linearity measurements were used to establish the accuracy and precision
of the Monitek turbidimeter measuring formazin solutions. Accuracy was assessed by comparing
continuous turbidity measurements with those from the ISO 7027 reference turbidimeter.
Precision was assessed from the five replicate results at each turbidity level.
3.5.3	Water Temperature
Variations in the temperature of the water stream were introduced to simulate a range of
conditions under which the on-line turbidimeters may typically operate. During off-line testing,
the temperature of the recirculating water equilibrated in the range from 27 to 29 °C, which was
8

-------
approximately 3 to 6°C above the room temperature at the water plant during testing. To assess
the effect of temperature on the turbidimeter performance, the temperature of the recirculating
solution was lowered using an immersion type chiller, and replicate turbidity measurements were
recorded at approximately 21 °C and again at 16°C. In these tests, the solution temperature in the
reservoir was held within 2.5 degrees of the nominal 16°C and 21 °C targets, while a series of
five measurements was recorded at each temperature. To ensure equilibration, the solution was
allowed to recirculate for one hour before the turbidity measurements were recorded. For the
temperature tests at 16°C and 21 °C, the temperature of the sample stream was recorded at the
grab sample port within 30 seconds of sample collection, and the temperature of the grab sample
was measured within 30 seconds of completion of the reference measurement. To assess
temperature effects at different turbidities, this test was conducted with both 0.3 and 5 NTU
solutions.
3.5.4	Flow Rate
The flow rate of the sample stream through the Monitek TST-SC turbidimeter was manipulated
to assess the response of the turbidimeter to various realistic operational conditions. Manual
valves, which were included upstream and downstream of the Monitek turbidimeter, were
adjusted to vary the flow rate through the turbidimeter while maintaining a back pressure on the
line. During normal testing, the flow rate was held in the range from 0.2 to 0.5 gallons per minute
(gpm). In contrast, the flow test was performed at a minimum flow rate of 0.1 gpm and at a
maximum flow rate of 1.0 gpm. To assess the effect of flow rate on performance, measurements
were made at both the minimum and maximum flow rates at turbidity levels of both 0.3 NTU and
5 NTU.
3.5.5	Color
Changes in water color were introduced by spiking the sample stream with colored solutions
prepared from commercial food coloring dye. Stock solution was added to the system reservoir to
give sample solutions approximately 5, 15, and 30 color units (CU) successively, and the
instrumental response to these color changes was monitored. Five measurements were made for
each color level at both low turbidity (~ 0.1 NTU) and higher turbidity (~ 5 NTU).
The color of the recirculated solution was determined by analyzing the grab samples instru-
mentally using the Hach 2100AN reference turbidimeter with the supplied light filter. This
reference turbidimeter was calibrated for color measurements according to the instrument
manual. Solutions used in the color calibration of the reference turbidimeter were prepared by
dilution of a commercial cobalt-platinum color standard(4) (Hach Company, Loveland, Colorado).
At ~ 0.1 NTU, the color of the solution before the addition of the dye was approximately 0 color
unit (CU). However, at the 5 NTU level, light scattering from the presence of formazin intro-
duced an apparent color to the solution of approximately 30 CU. Consequently, for the 0.5 NTU
test, dye solution was added to increase the color by 5, 15, and 30 CU; i.e., to bring the absolute
color to approximately 35, 45, and 60 CU, respectively.
9

-------
3.6 On-Line Testing
The on-line test phase focused on assessing the long-term performance of the Monitek
turbidimeter under realistic unattended operating conditions and assessing its accuracy in
monitoring an actual sample stream. Specifically, this phase of testing addressed the calibration
and drift characteristics of the turbidimeter over a five-week period of monitoring a sample
stream from the water plant. Routine reference measurements were used for comparison with the
on-line readings to assess accuracy, and a re-evaluation of the calibration at the end of the test
period helped establish drift characteristics. Natural meteorological and demand changes
contributed to the variability of water quality in the treatment facility and provided a natural
range of turbidity for characterizing performance.
Table 3-3 provides a summary of the parameters tested in the on-line phase, the test solutions
used, and the number of readings recorded for each parameter.
3.6.1 Accuracy
In the on-line testing, the accuracy of the Monitek TST-SC turbidimeter relative to the ISO 7027
reference method was assessed on water samples from the plant stream. A sample stream was
drawn from a flocculation settling basin at the Dublin Road Water Plant facility, containing
unfiltered water that had been treated with lime, caustic, and alum. The sample stream was
directed to the Monitek TST-SC turbidimeter through the distribution manifold. Two grab
samples of this stream were collected and analyzed by the reference turbidimeters each weekday
(Monday through Friday) for the four weeks of testing. The reference measurements of these
samples were compared with the simultaneous results from the Monitek TST-SC turbidimeter.
Table 3-3. Summary of Measurements for On-Line Testing
Parameter Tested	Test Solution	Number of Readings
Accuracy
Drift
Drift
Drift
Drift
3.6.2 Drift
Drift was determined in two ways: (1) through off-line calibration checks conducted regularly
throughout the course of the verification test using formazin solutions and (2) through a
Plant Water
0.3 NTU Standard
0.5 NTU Standard
2 NTU Standard
5 NTU Standard
2 per weekday for 4 weeks
(40 total)
5 for final linearity check
5 each for eight calibration checks
and 5 for final linearity check
5 for final linearity check
5 for final linearity check
10

-------
comparison of multi-point linearity checks performed initially during the off-line phase described
in Section 3.5.1 and subsequently after the completion of the on-line phase. The Monitek
TST-SC turbidimeter was calibrated by the vendor prior to shipment and installation at the water
plant. After that calibration, no further manual calibration or adjustment was performed
for the duration of the verification test period. However, during the course of the on-line
testing the Monitek TST-SC turbidimeter was observed to drift toward higher turbidity relative to
the reference measurements on several occasions. When this behavior was observed, the Monitek
TST-SC turbidimeter was taken offline briefly for cleaning, and the cylindrical window was
cleaned with lint-free laboratory towels.
The Monitek TST-SC turbidimeter was taken offline briefly twice each week for routine
calibration checks against a 0.5 NTU formazin solution. These intermediate calibration checks
were performed twice weekly for four consecutive weeks. Freshly prepared 0.5 NTU formazin
solutions were used for the standard solution.
Upon completion of the four-week period, calibration and linearity were checked again by
comparison with the reference measurements using standard solutions of 0.3, 0.5, 2, and 5 NTU.
A linear fit of these data was compared with the initial linearity check performed in the off-line
phase to assess the degree of calibration drift.
11

-------
Chapter 4
Quality Assurance/Quality Control
Quality control (QC) procedures were performed in accordance with the quality management
plan (QMP) for the AMS pilot(5) and the test/QA plan(2) for this verification test.
4.1	Data Review and Validation
Test data were reviewed and approved according to the AMS pilot QMP,(5) the test/QA plan,(2)
and Battelle's one-over-one policy. The Verification Test Coordinator, or the Verification Test
Leader, reviewed the raw data and the data sheets that were generated each day and approved
them by adding their signature and the date. Laboratory record notebook entries were also
reviewed, signed, and dated.
4.2	Deviations from the Test/QA Plan
During the preparation and performance of the verification test, deviations from the test/QA plan
were implemented to better accommodate differences in vendor equipment and other changes or
improvements. Any deviation required the approval signature of Battelle's Verification Testing
Leader. A planned deviation form was used for documentation and approval of the following
changes:
1.	Commercial food coloring dye was used for the color test instead of diluted color
standard owing to the strongly acidic nature of the cobalt-platinum standard solution.
2.	Calibration of the pH meter was performed only once during the test, and the meter was
not readjusted to account for variations in ambient temperature. Recalibration was to be
performed under the conditions of the test. However, the pH measurements were used
only to assess changes and not for absolute measurements.
3.	Only one in-line filter was used in the recirculation system.
4.	The schedule of tests was lengthened and the order of testing was changed to better
group series of parameter evaluations and to respond to unexpected experimental
results.
These deviations had no significant impact on the test results used to verify the performance of
the on-line turbidimeters.
12

-------
4.3 Calibration
4.3.1 Reference Turbidimeters
The reference turbidimeters were calibrated according to the procedures described in their
respective instrument manuals. The calibrations were performed on August 23, 1999. Calibration
was performed using a blank, and 20, 200, 2000, and 7500 NTU StablCal calibration standards
(Hach Company, Loveland, Colorado). After calibration and before proceeding with the
verification test, the calibration of each reference turbidimeter also was checked through a five-
point linearity test using solutions with the following turbidities: < 0.1, 0.3, 0.5, 2, and 5 NTU.
The < 0.1, 0.3, and 0.5 NTU solutions were purchased and used as is, whereas the 2 and 5 NTU
solutions were prepared by diluting a purchased 20 NTU StablCal formazin standard solution.
The results of the linearity check are summarized in Table 4-1, indicating that the two reference
turbidimeters gave essentially identical results. For each reference turbidimeter, the slope of this
linear fit was within the 0.90 and 1.10 limits prescribed in the test/QA plan,(1) and each fit had an
r2 > 0.98 as called for in the test/QA plan.(1)
Table 4-1. Results of Linearity Check of Reference Turbidimeters
Parameter
Hach 2100AN IS (ISO 7027)
Hach 2100AN (180.1)
Slope
1.086
1.086
Intercept (NTU)
0.004
0.010
r2
0.9991
0.9996
The calibration of each reference turbidimeter also was checked both before and after each series
of test measurements, using a nominal 0.5 NTU StablCal standard solution. The reference
turbidimeters were to be recalibrated if agreement between the turbidity reading and the certified
0.521 NTU turbidity value of this standard solution was not within ± 10% (i.e., 0.469 to
0.573 NTU). If this calibration check criterion was met before, but not after, a series of test
measurements were performed, those measurements were to be repeated after recalibration of the
reference turbidimeters. Throughout the course of the verification test, neither reference
turbidimeter was ever found to be out of calibration, and consequently no recalibration of the
reference turbidimeters was performed.
Before the background readings were measured for the detection limit determination, an
additional calibration check with <0.1 NTU standard was also performed on the reference
turbidimeters to ensure proper calibration at low levels. These calibration checks were performed
on September 9, 1999, for the initial linearity test and October 20, 1999 for the final linearity test.
The results showed agreement within 0.02 NTU between the turbidity reading of the <0.1 NTU
standard, and the value as reported on the certificate of analysis.
13

-------
4.3.2	Temperature Sensors
A Fluke 52 thermocouple (Battelle Asset Number 570080) was used throughout the verification
test to determine water temperature and the ambient room temperature. This thermocouple was
calibrated on June 30, 1999, against a calibrated temperature standard (Fluke 5500A, Battelle
Asset Number SN-714755).
4.3.3	Flow Meters
The flow meter used in the verification test to measure the water flow through the Monitek TST-
SC turbidimeter was a panel-mounted, direct-reading meter purchased from Cole-Parmer
(Catalog Number P-03248-56), capable of measuring up to 1 gpm. The flow meter was factory
calibrated, and was checked once during the verification test by measuring the time required to
fill a container of known volume through the meter at settings of 0.2 gpm and 1.0 gpm. Table 4-2
summarizes the results of the flow rate checks.
Table 4-2. Summary of Flow Meter Calibration Check
Flow Meter Setting
Volume
Time
Calculated Rate
(gpm)
(gallon)
(seconds)
(gpm)
0.2
2
590
0.20
1.0
1
55
1.09
1.0
2
115
1.04
1.0
3
178
1.01
The calibration check was performed on August 26, 1999, and indicated agreement within the
10% criterion established in the test/QA plan(2) at both the minimum and maximum flow rates.
4.3.4 pHMeter
The pH meter was calibrated once during the verification test, with no further adjustment of the
meter. Calibration included standardization at a pH of 7 and a pH of 10 using buffer solutions.
Calibration checks performed during the color test indicated a bias of 0.1 to 0.3 pH units. Biases
above 0.2 pH units fall outside of the acceptance criterion for the verification test and introduce
an uncertainty to the absolute magnitude of the pH readings. However, the pH readings were
recorded as a means of assessing if changes in the acidity of the solution occurred as a result of
adding the color solution, rather than as an absolute measure of the pH itself. The pH readings
recorded during the test indicated no evidence of pH change in the test solution as the result of
adding dye to the test solution.
14

-------
4.4 Data Collection
Electronic data were collected and stored by a PC-based data acquisition system using LabTech
Notebook software (Version 8.0.1). Data were collected from the Monitek TST-SC turbidimeter
every 10 seconds over much of the course of verification testing. These data were saved in ASCII
files along with the time of collection. Data files were stored electronically both on the hard drive
of the data collection system and on floppy discs for backup purposes. Data collected manually
included turbidity readings of the reference turbidimeters, flow rates, and water and ambient air
temperature measurements. An example of the data recording sheet used to record these data is
shown in Appendix A.
4.5 Assessments and Audits
4.5.1	Technical Systems Audit
Battelle's Quality Manager performed a technical systems audit once during the verification test.
The purpose of this audit was to ensure that the verification test was performed in accordance
with the test/QA plan(2) and that all QA/QC procedures were implemented. In this audit, the
Quality Manager reviewed the calibration standards and reference methods used, compared
actual test procedures with those specified in the test/QA plan, and reviewed data acquisition and
handling procedures. A report on this audit is provided in Appendix B.
4.5.2	Performance Evaluation Audit
Performance evaluation audits were conducted to assess the quality of the measurements made in
the verification test. These audits addressed only those measurements made by Battelle staff in
conducting the verification test, i.e., the reference turbidimeter readings and temperature
measurements. The audits were conducted by analyzing the standards or comparing them with
references that were independent of those used in the verification test. Each audit was made at
least once during the verification test.
The audit of the reference turbidimeters was performed by analyzing a reference solution that
was independent of the formazin standards used for calibration of the reference turbidimeters
during the verification test. The independent reference solution was an AMCO AEPA-1
0.500 NTU standard solution obtained from APS Analytical Standards, Redwood City,
California. This audit was conducted once daily throughout the verification test and served as an
independent verification of the calibration of the reference turbidimeters. Agreement between the
National Institute of Standards and Technology (NIST) traceable turbidity value of the AMCO
AEPA-1 solution and the turbidity readings from each reference turbidimeter was recorded and
tracked graphically using a control chart. Furthermore, similar calibration assessments were
performed daily using a purchased 0.521 NTU StablCal formazin standard (Hach Company,
Loveland, Colorado), as described in Section 4.3.1. The results of these daily calibration
assessments always showed agreement between the turbidity reading from each reference
turbidimeter and the true turbidity within ± 10% as required in the test/QA plan.(2) The results of
15

-------
the daily calibration assessments are shown in Figures 4-la and 4-lb, for both the AMCO
AEPA-1 standard and the formazin standard on the ISO 7027 and Method 180.1 reference
turbidimeters, respectively. (The dashed lines in the upper parts of Figures 4-la and 4-lb are at
intervals of 0.05 NTU, but are for visual reference only and are not exactly the ± 10% control
limits of the calibration checks. The bottom portion of each figure shows the ± 10% control
limits.) Throughout the course of the verification test, readings of the AMCO AEPA-1 standard,
as measured by the ISO 7027 turbidimeter, ranged from -1.2% to +6.2% relative to the certified
turbidity value for that standard, and were on average ~ 2.7% higher. For the Method 180.1
turbidimeter, the range was -1.2% to +5.4% with an average reading which was 1.7% higher than
the certified turbidity value. The daily fluctuations in these measurements resulted in standard
deviations of - 1.7% for each reference turbidimeter. Similarly, readings of the formazin
standard ranged from -5.2% to + 8.4% for the ISO 7027 and from -6.9% to +8.3% for the
Method 180.1. The average readings were higher than the certified turbidity value by ~ 2.2%
when measured by the ISO 7027 turbidimeter and by ~ 1.6% for the Method 180.1 turbidimeter,
with standard deviations of 3.5% and 3.7% respectively. Although the average deviations from
the true turbidity values for these standards were approximately the same, the scatter in the
readings was greater in the formazin readings.
The audit of the thermocouple used during the verification test consisted of a comparison of the
temperature readings from the thermocouple with those of an independent temperature sensor.
The thermocouple was checked for accuracy by comparison with an American Society for
Testing and Materials mercury-in-glass thermometer in the Battelle Instrument Laboratory on
October 13, 1999, and again on November 1, 1999. Those comparisons were done at ambient
temperature, and the results are shown in Table 4-3.
Agreement between the thermocouple used in the verification test and the mercury-in-glass
thermometer was well within the two-degree specification established in the test/QA plan.(2)
4.5.3 Verification Test Data Audit
Battelle's Quality Manager audited at least 10% of the verification data acquired during the
verification test. The Quality Manager traced the data from initial acquisition, through reduction
and statistical comparisons, and to final reporting. All calculations performed on the data
undergoing the audit were checked.
16

-------
0.575
0.525
-d
C3
1)
Pi
0.475
AMCO AEPA-1 (0.500 NTU)
StableCal Formazin (0.521 NTU)
0.425
09/09/1999
09/23/1999
10/07/1999
10/21/1999
SB
Q
-10
09/09/1999
09/23/1999
10/07/1999
10/21/1999
Figure 4-la. Control Chart for Performance Evaluation Calibration
Checks of the ISO 7027 Reference Turbidimeter
0.575
0.525
0.475
AMCO AEPA-1 (0.500 NTU)
StableCal Formazin (0.521 NTU)
0.425
09/09/1999
09/23/1999
10/07/1999
10/21/1999

PL, 09/09/1999
09/23/1999
10/07/1999
10/21/1999
Figure 4-lb. Control Chart for Performance Evaluation Calibration
Checks of the Method 180.1 Reference Turbidimeter
17

-------
Table 4-3. Results of Calibration Checks of Thermocouple Used in Verification Test

October 13,1999
November 1,1999

27.2°C
29.5°C
Fluke 52 Thermocouple


ASTM Mercury-in-Glass Thermometer
27.2°C
29.7°C
4.6 Audit Reporting
Each assessment and audit was documented in accordance with Section 2.9.7 of the Quality
Management Plan for the AMS pilot.(5) The assessment report included the following:
¦	Identification of any adverse findings or potential problems
¦	Response to adverse findings or potential problems
¦	Possible recommendations for resolving problems
¦	Citation of any noteworthy practices that may be of use to others
¦	Confirmation that corrective actions have been implemented and are effective.
A copy of the Technical Systems Audit Report is included in Appendix B of this report.
18

-------
Chapter 5
Statistical Methods
5.1 Off-Line Testing
The turbidimeter performance characteristics were quantified on the basis of statistical
comparisons of the test data. This process began by converting the files that resulted from the
data acquisition process into spreadsheet data files suitable for data analysis. The following
statistical procedures were used to make the comparisons.
5.1.1 Linearity
Linearity was assessed by linear regression, with the reference turbidity reading (R) as an
independent variable and the turbidimeter response (7) as a dependent variable. The regression
model was
T- fll x R+ fl
where |_i, and (3 are the slope and intercept of the response curve, respectively. The turbidimeter
performance was assessed in terms of the slope, intercept, and the square of the correlation
coefficient of the regression analysis.
5.1.2 Accuracy
The accuracy of the turbidimeter with respect to the reference method was assessed in terms of
the average relative bias (B), as follows:
B
{R- ry
R
x 100
where R is the turbidity reading of the reference turbidimeter, and T is the corresponding
turbidity reading of the Monitek TST-SC turbidimeter.
Accuracy relative to the reference turbidimeter was assessed both for the prepared solutions and
the samples from the plant water stream. The accuracy of the Monitek TST-SC turbidimeter was
assessed relative to the ISO 7027 reference method for verification purposes and relative to the
180.1 reference method as an illustration of performance.
19

-------
5.1.3 Precision
Precision was reported in terms of the percent relative standard deviation (RSD) of a group of
similar measurements. For a set of turbidity measurements given by Tx, T2, Tn, the standard
deviation (S) of these measurements is
5 =
1 11	— 9
	-ZfTk-T/
n -1 k=i
1/2
where T is the average of the turbidity readings. The RSD is calculated as follows:
RSD = 1x100
T
and is a measure of the dispersion of the measurement relative to the average value of the
measurements. This approach was applied to the groups of replicate measurements on each test
solution. In some cases, the turbidity of the prepared solution changed approximately linearly
with time, due to loss of particles in the recirculation system. In those cases, a linear regression of
the data was performed to assess the slope of the turbidity change as a function of time. This
slope was used to adjust the individual turbidity readings to approximately the initial concen-
tration. The precision was then calculated on the adjusted values as described above.
5.1.4 Water Temperature Effects
The effect of water temperature on the response of the Monitek TST-SC at 0.3 NTU and 5 NTU
was assessed by trend analysis. The turbidity readings relative to the ISO 7027 reference turbidi-
meter were analyzed as a function of water temperature to identify trends in the relative turbidity
at each of the two levels of turbidity. The calculations were performed using separate linear
regression analyses for the data at each turbidity level. A similar calculation was done for
illustrative purposes using the 180.1 reference data.
5.1.5	Flow Rate Sensitivity
Analysis of flow rate influence on turbidity readings was similar to that for water temperature
effects. The turbidimeter response relative to the ISO 7027 reference turbidimeter was analyzed
as a function of flow rate to assess trends in the response of the turbidimeter with changes in
sample flow rate. The analyses were performed separately for the 0.3 NTU and 5 NTU data. A
similar calculation was done for illustrative purposes using the 180.1 reference data.
5.1.6	Color Effects
The influence of color on turbidity was assessed through a linear regression analysis of the
turbidity measured for each color relative to the ISO 7027 reference turbidimeter. Separate
20

-------
analyses were performed for the measurements recorded at 0.1 NTU and those recorded at
5 NTU. A similar calculation was done for illustrative purposes using the 180.1 reference data.
5.2 On-Line Testing
5.2.1	Accuracy
As described in Section 5.1.2, accuracy in the on-line measurements was determined as a bias
relative to the ISO 7027 reference turbidimeter. Daily reference measurements of the sample
stream from the water plant were used to assess for accuracy. A similar calculation was done for
illustrative purposes using the 180.1 reference data.
5.2.2	Drift
Drift was assessed in two ways. The drift in the calibration of the Monitek TST-SC turbidimeter
was assessed by comparison of the regression analyses of the multi-point linearity tests per-
formed at the beginning and end of the verification test. This comparison was used to establish
any long-term drift in instrumental calibration during the verification test. Also, the reference and
on-line turbidity results in monitoring the plant water stream were used to assess drift associated
with the operation of the instrument (e.g., fouling of the optics). Trends in the intermediate
calibration data toward a positive bias were used to identify when the turbidimeter needed
cleaning.
21

-------
Chapter 6
Test Results
The results of the verification test are presented in this section, based upon the statistical methods
of comparison shown in Chapter 5. For all performance characteristics verified, two sets of
results are shown. The primary verification results are based on comparisons with the ISO 7027
reference method; a secondary illustration of performance is based on comparisons with the EPA
180.1 reference method.
6.1 Off-Line Testing
Off-line testing was performed to assess the performance of the Monitek TST-SC turbidimeter
when measuring known solutions under controlled conditions. The first of the off-line tests was
performed to establish the linearity of the turbidimeter response in the range from < 0.1 to
5 NTU. Data from the linearity test also were used to assess the accuracy and precision of the
TST-SC in this turbidity range. After the linearity test, the effects of sample temperature, sample
flow rate, and sample color were evaluated. The results of each of these tests are described in this
section.
6.1.1 Linearity
The verification data from the initial linearity test are shown in Figure 6-la, relative to the
ISO 7027 reference turbidimeter. A series of at least five data points was recorded at each of the
five nominal turbidity levels (approximately < 0.05, 0.3, 0.5, 2, and 5 NTU). At the two highest
NTU, a decrease in turbidity was observed in the readings of both the Monitek TST-SC
turbidimeter and the reference turbidimeter. This decrease can be seen graphically as a spread in
the data along the slope of the linearity plots. Between the first and fifth readings at 2 NTU, the
decrease in turbidity represented approximately 4 to 5% of the initial turbidity as measured by the
reference turbidimeter. This decrease in turbidity was likely the result of formazin being lost
from the solution in the recirculation system. In an attempt to prevent the formazin loss, the
solution was stirred magnetically. A second series of five measurements was recorded at the
2 NTU level after magnetic stirring of the formazin solution was introduced. After magnetic
stirring was introduced, the decrease in turbidity was still observed, however, to a slightly lesser
extent (approximately 2 to 4%).
The Monitek TST-SC does not respond to turbidity levels below approximately 0.1 NTU.
Consequently, the linearity of the Monitek TST-SC was established between 0.3 and 5 NTU.
22

-------
The data from the linearity test, excluding the low turbidity readings, were fit using a linear
regression as described in Section 5.1.1, and the results of these fits are shown in Table 6-1. The
secondary comparison with the Method 180.1 data is shown in Figure 6-lb, with the regression
results also shown in Table 6-1.
5.0
4.0
$
H
£
Sd) 3.0
Ch
CD
&
.1? 2.0
£
H
1.0
0.0
0	1	2	3	4	5
Reference Turbidity (NTU)
Figure 6-la. Linearity Plot for Monitek TST-SC
Turbidimeter vs. ISO 7027 Reference
Turbidimeter
Table 6-1. Statistical Results of Initial Linearity Test on Monitek TST-SC Turbidimeter
Linear Regression Parameter
Verification Results3
Secondary Comparisonb
Slope (std. error)
0.935 (0.004)
0.946 (0.005)
Intercept (std. error) NTU
-0.107 (0.010)
-0.089 (0.011)
r2
0.9996 (0.0331)
0.9995 (0.0373)
a Comparison with ISO 7027 reference method (2100AN IS reference turbidimeter).
b Comparison with EPA Method 180.1 (2100AN reference turbidimeter).
23

-------
Z
&Q
11
ai
12 2.0
•e
£
0.0
0.0
1.0
2.0
3.0
4.0
5.0
Reference Turbidity (NTU)
Figure 6-lb. Linearity Plot for Monitek TST-SC
Turbidimeter vs. Method 180.1 Reference
Turbidimeter
The verification results of the linear regression indicate that the Monitek TST-SC turbidimeter
responded linearly to turbidity between 0.3 NTU and 5 NTU. The slope of the response curve
was -6.5% lower than unity with respect to the ISO 7027 reference turbidimeter, representing a
negative bias in turbidity readings. A negative intercept (~ 0.1 NTU) was determined for the
linearity plot, indicating an offset in the calibration curve of the Monitek TST-SC turbidimeter.
The secondary comparison in Table 6-1 shows that the TST-SC also exhibited good linearity
relative to Method 180.1, with a similar slope and negative intercept of the response curve.
6.1.2 Accuracy
Data obtained from the initial linearity test were used to assess accuracy for the off-line tests. The
results of the accuracy verification are given in Table 6-2 and are presented as the average differ-
ence between the Monitek TST-SC turbidimeter and each reference turbidimeter, as well as the
relative bias of the Monitek TST-SC turbidimeter with respect to the reference measurements.
Negative values indicate a negative bias in the Monitek TST-SC turbidimeter readings when
compared with the reference turbidimeter.
24

-------
Table 6-2. Bias of Monitek TST-SC Turbidimeter Relative to the Reference Measurements
on Prepared Test Solutions

Verification Results3
Secondary Comparisonb
Nominal Turbidity
of Test Solution
(NTU)
Average
Difference
(NTU)
Relative
Bias
(%)
Average
Difference
(NTU)
Relative
Bias
(%)
0.3
-0.114
-28.9
-0.089
-24.0
0.5
-0.109
-19.7
-0.078
-14.9
2
-0.249
-15.0
-0.217
-13.5
5
-0.407
-8.2
-0.329
-6.8
a Comparison with ISO 7027 reference method (2100AN IS reference turbidimeter).
b Comparison with EPA Method 180.1 (2100AN reference turbidimeter).
The verification results in Table 6-2 show a bias of approximately 30% at 0.3 NTU and approx-
imately 10% at 5 NTU, resulting from average measured differences of ~ 0.1 to ~ 0.4 NTU. It is
worth noting that although there is a considerable relative bias in the Monitek readings, a large
portion of this bias can be attributed to the intercept calculated from the linearity plots (Table 6-
1). At the 0.3 and the 0.5 NTU turbidity levels, the calculated intercepts account for essentially
100%) of the relative bias. Likewise, the calculated intercepts account for approximately 25 to
50%o of the relative bias at each of the two higher turbidity levels. A shift in the calibration curve
to eliminate the offset would result in a dramatic reduction in the observed bias.
The secondary comparison in Table 6-2 shows similar performance relative to Method 180.1,
with average bias results of 24% at 0.3 NTU and 6.8%> at 5 NTU.
6.1.3 Precision
Data from the linearity test were used to calculate precision at 0.3, 0.5, 2, and 5 NTU. At both the
2 NTU and 5 NTU levels, a decrease in turbidity was observed as a function of time during the
test procedure. To account for this variability in turbidity, the readings at these two levels were
analyzed by linear regression against time and adjusted to approximately the initial turbidity
value using correction factors based on the regression results. The adjusted values (T-) were
calculated using the following equation:
Ti = Tt + c(tt - t0)
where Tt is the ith turbidity reading, /, is the time at which the ith sample was collected, l0 is the
time of collection for the initial sample in the series, and c is the slope of the line determined
from the linear regression results of turbidity versus time at 2 NTU or 5 NTU. The results of the
adjustment calculations are given in Table 6-3 for the Monitek TST-SC turbidimeter. Similar
25

-------
corrections were applied to the reference readings since the reference readings showed the same
trend of decreasing turbidity with time.
Table 6-3. Adjusted Turbidity Readings for Precision Calculations of the Monitek TST-SC
Turbidimeter
Nominal
Actual
Corrected
Nominal
Actual
Corrected
Value
Reading
Reading
Value
Reading
Reading
(NTU)
(NTU)
(NTU)
(NTU)
(NTU)
(NTU)
2
1.4917
1.4917
5
4.5374
4.5374
2
1.4673
1.4841
5
4.5435
4.5535
2
1.4429
1.4766
5
4.5007
4.5207
2
1.4246
1.4751
5
4.5312
4.5612
2
1.4124
1.4797
5
4.4946
4.5346
2
1.3940
1.4782



2
1.3757
1.4767



2
1.3757
1.4935



2
1.3452
1.4798



2
1.3330
1.4845



The precision was calculated from the raw data at the 0.3 and 0.5 NTU levels, and from the
corrected data at the 2 NTU and the 5 NTU levels. The results of these calculations are shown in
Table 6-4. For comparison, the calculated precision values for the two reference turbidimeters are
also included in that table.
The results of these calculations indicate that the Monitek TST-SC turbidimeter has approxi-
mately the same precision as the bench-top reference turbidimeter through the range of turbidity
measured in this verification test. The precision of the Monitek measurements ranged from 3.3%
RSD at 0.3 NTU to 0.35% RSD at 5 NTU.
Table 6-4. Precision of Monitek TST-SC Turbidimeter and of the Reference Turbidimeters
Nominal
Monitek TST-SC
ISO 7027
Method 180.1
Turbidity
SD
RSD (%)
SD
RSD (%)
SD
RSD (%)
0.3 NTU
0.0093
3.3
0.0161
4.1
0.0051
1.4
0.5 NTU
0.0140
3.2
0.0118
2.1
0.0095
1.8
2 NTU
0.0064
0.43
0.0123
0.71
0.0108
0.64
5 NTU
0.0160
0.35
0.0126
0.26
0.0088
0.18
26

-------
6.1.4	Water Temperature Effects
The verification data obtained for the temperature test are shown in Figure 6-2a. As a result of
gradual loss of formazin in the recirculation system during the temperature test, additional
formazin solution was added between each set of temperature measurements to maintain
turbidity levels at approximately 0.3 NTU and 5 NTU. Consequently, the absolute turbidity
readings alone cannot be used as an indication of temperature effects. Therefore, the readings
recorded for the Monitek TST-SC turbidimeter were normalized to the corresponding reference
readings to get a relative measure of turbidity. These relative values (i.e., ratios of TST-SC to
ISO 7027 data) are shown in Figure 6-2a and were analyzed by linear regression to assess the
effect of water temperature on turbidity reading. The results of the regression analysis are given
in Table 6-5.
These verification results indicate a small temperature effect at the lower turbidity (0.3 NTU),
suggesting a slight decrease in relative turbidity readings with increased temperature, i.e., the
results indicate a statistically significant change in observed turbidity relative to the reference
turbidimeters as a function of temperature. This change represents a relative difference of
approximately 0.4% per degree C relative to the ISO 7027 reference turbidimeter. At the higher
turbidity (5 NTU), the 95% confidence interval of the regression slope includes zero, and thus the
data indicate no statistically significant relation between turbidity readings and water
temperature.
The secondary results in Figure 6-2b and Table 6-5 indicate a temperature effect resulting in a
relative decrease in turbidity of 1.1 % per degree C at 0.3 NTU. At the 95% confidence level, no
statistically significant effect is observed at 5 NTU.
6.1.5	Flow Rate
The results of the flow rate test are summarized in Figure 6-3. The data are again presented and
analyzed as relative turbidity readings, rather than absolute turbidity readings to correct for a
gradual, slight loss of the formazin during the testing. The results of the statistical analysis of the
flow data are presented in Table 6-6.
27

-------
1.2
~ 0.3NTU
A 5 NTU
1
0.8
0.6
0.4
15
Temperature (C)
Figure 6-2a. Effect of Temperature on Monitek TST-SC Turbidity Readings
vs. ISO 7027 at Both 0.3 and 5 NTU
1.2
~ 0.3 NTU
A 5 NTU
1
0.8
0.6
0.4
15
Temperature (C)
Figure 6-2b. Effect of Temperature on Monitek TST-SC Turbidity Readings
vs. Method 180.1 at Both 0.3 and 5 NTU
28

-------
Table 6-5. Statistical Results of Temperature Test on the Monitek TST-SC Turbidimeter
Linear Regression
Parameter
Verification Results3
0.3 NTU 5 NTU
Secondary Comparisonb
0.3 NTU 5 NTU
Slope
-0.0042
0.0022
-0.0108
0.0026
(std. error)
(0.0013)
(0.0014)
(0.0016)
(0.0012)
Intercept (std. error)
0.784
0.869
0.844
0.869

(0.030)
(0.031)
(0.036)
(0.026)
r2
0.427
0.150
0.775
0.257
(std. error)
(0.030)
(0.026)
(0.036)
(0.021)
a Comparison with ISO 7027 reference method (2100AN IS reference turbidimeter).
b Comparison with EPA Method 180.1 (2100AN reference turbidimeter).
0.8
0.7
I
H

.>
0.6
~ ISO 7027
A Method 180.1
t
tr
l
1
~
~
~
t
i
A
0.5
0	0.4	0.8
Flow Rate (gpm)
Figure 6-3. Effect of Sample Flow Rate on Monitek TST-SC Turbidimeter Response
1.2
29

-------
Table 6-6. Statistical Results of Flow Rate Test for the Monitek TST-SC Turbidimeter
Parameter
Verification Results®
Secondary Comparisonb
Slope (std. error)
-0.010 (0.010)
-0.020 (0.011)
Intercept (std. error)
0.670 (0.007)
0.643 (0.008)
r2 (std. error)
0.0107 (0.015)
0.303 (0.015)
a Comparison with ISO 7027 reference method (2100AN IS reference turbidimeter).
b Comparison with EPA Method 180.1 (2100AN reference turbidimeter).
The 95% confidence intervals for the slopes both include zero, indicating that sample flow rate
has no significant effect on the response of the Monitek TST-SC turbidimeter in the range of 0.1
to 1.0 gpm.
6.1.6 Color Effects
The verification data obtained from the color tests at 0.1 NTU and 5 NTU are shown in
Figure 6-4a. In this figure, the data at each color level are plotted as relative values with respect
to the reference turbidimeter readings, and the statistical analysis of these data involved a linear
regression analysis of the relative data as a function of solution color. At the 5 NTU level, the
background color reading of approximately 30 CU was subtracted, and only the effect of color
added during the test is shown. The results of the statistical calculations are summarized in
Table 6-7.
The verification results in Table 6-7 show that at the 5 NTU level, color has no significant effect
(95% confidence) on the response of the Monitek TST-SC turbidimeter. At the low turbidity
level, color has a small but statistically significant effect on the response of the Monitek TST-SC
turbidimeter. However, that test was performed at a turbidity level of approximately 0.1 to
0.2 NTU, which is near the instrumental detection limit of the Monitek TST-SC turbidimeter.
Consequently, it is unclear whether the observed effect is the result of an actual response to color,
or if it is the result of measurements taken at the low end of the measuring range for the turbidi-
meter. Similar results are shown for the secondary comparison in Figure 6-4b and Table 6-7.
30

-------
1.0
~ . 1 NTU
¦ 5 NTU
0.4
10
15	20
Color (CU)
25
30
35
Figure 6-4a. Effect of Color on Relative Turbidity with the Monitek TST-SC
Turbidimeter vs. the ISO 7027 at Both 0.1 and 5 NTU
0.8
£
IS
Qi
0.6
0.4
I .1 NTU
¦ 5 NTU
10
15	20
Color (CU)
25
30
35
Figure 6-4b. Effect of Color on Relative Turbidity with the Monitek TST-SC
Turbidimeter vs. Method 180.1 at Both 0.1 and 5 NTU
31

-------
Table 6-7. Statistical Results of the Color Test with the Monitek TST-SC Turbidimeter
Reference:
Verification Results3
Secondary Comparisonb
Parameter
0.1 NTU
5 NTU
0.1 NTU
5 NTU
Slope
-0.0036
-0.0003
-0.0033
-0.0002
(std. error)
(0.0010)
(0.0001)
(0.0008)
(0.0003)
Intercept
0.6894
0.9192
0.7106
0.9521
(std. error)
(0.0201)
(0.0029)
(0.0166)
(0.0051)
r2
0.4871
0.2224
0.5340
0.0564
(std. error)
(0.0408)
(0.0059)
(0.0337)
(0.0104)
a Comparison with ISO 7027 reference method (2100AN IS reference turbidimeter).
b Comparison with EPA Method 180.1 (2100AN reference turbidimeter).
6.2 On-Line Testing
Figure 6.5 shows the results from the four weeks of on-line testing. In this figure, data from the
Monitek TST-SC and the reference turbidimeters are shown, along with additional data supplied
by the Dublin Road Water Plant (DRWP). Data from the DRWP are from a turbidimeter in the
plant sampling the same water stream at a different location, for plant operational purposes.
These DRWP data are shown to illustrate the trends in turbidity of the water stream sampled for
this test. No quantitative comparisons with the DRWP data should be made, since these data
were not collected at the same location as samples for this verification test. For convenience,
only one data point per hour is shown for the Monitek TST-SC although data were recorded at
intervals of 10 seconds throughout the on-line testing. Breaks in the data from the Monitek
turbidimeter indicate periods during which the turbidimeter was taken off-line for calibration
checks, or cleaning.
In general, this figure illustrates correlation and sometimes close quantitative agreement between
the Monitek and the reference measurements. Also, the varying turbidity levels shown by the
TST-SC indicate a temporal pattern similar to that of the DRWP data. Two episodes near the end
of the four week period show large deviations from the benchtop measurements by the Monitek
turbidimeter. These episodes occurred after adjustments to the recirculation system and are likely
to be associated with these modifications. Consequently, the data from October 4 through 5 and
the data from October 13 through 14 will not be included in the following discussions of
accuracy.
32

-------
— MomtekTST-SC
i DRWP
a EPA 180.1
o ISO 7027
Data excluded from verification
1.5
.£• 1
"d
0.5
a
0
09/17/99
09/24/99
10/01/99
Date
10/08/99
° n ~ 1
^~d'lj «,«sJ
10/15/99
Figure 6-5. Summary of Stream Turbidity Data from On-Line Testing of Monitek
TST-SC
6.2.1 Accuracy
The results from the four weeks of on-line accuracy testing are given in Table 6-8 and plotted in
Figure 6-5. The results shown in the table are given as the average of the two simultaneous
readings taken each day on samples from the plant water stream with the Monitek TST-SC
turbidimeter and the reference turbidimeters. In cases where more than the prescribed two
readings were recorded, all the values are included in the reported average. Additionally, the bias
in the Monitek readings relative to the reference turbidimeters is reported.
The verification results in Table 6-8 show that the Monitek TST-SC generally agreed with the
ISO 7027 reference turbidimeter within approximately 0.1 NTU. Positive and negative biases of
up to about 50% characterize most of this data range. The TST-SC and reference data exhibited a
linear regression of the form TST-SC = 1.113 (ISO 7027) + 0.050 NTU with r2 = 0.534.
The secondary comparison in Table 6-8 shows somewhat lower accuracy of the TST-SC relative
to Method 180.1, as expected. The linear regression has the form TST-SC = 1.306 (Method
180.1) + 0.057 NTU, with r2 = 0.463.
33

-------
Table 6-8. On-Line Daily Accuracy Check Results
Monitek TST-SC Verification Results® Secondary Comparisonb
Date	NTU	(Relative bias %)	(Relative bias %)
9/17/99
0.5854
0.5320 (10.0)
0.4315 (35.7)
9/20/99
0.6860
0.5090 (34.8)
0.4165 (64.7)
9/21/99
0.7441
0.5115 (45.5)
0.4215 (76.5)
9/22/99
0.8081
0.5855 (38.0)
0.4338 (86.3)
9/23/99
0.3534
0.3725 (-5.1)
0.2990 (18.2)
9/24/99
0.4145
0.3790 (9.4)
0.3220 (28.7)
9/27/99
0.8844
0.4040 (118.9)
0.3245 (172.5)
9/28/99
0.5029
0.4885 (2.9)
0.4230 (18.9)
9/29/99
0.4175
0.4155 (0.5)
0.3560 (17.3)
9/30/99
0.6433
0.5310 (21.1)
0.4690 (37.2)
10/1/99
0.4633
0.3140 (47.5)
0.2710 (70.9)
10/6/99
0.3809
0.4240 (-10.2)
0.3605 (5.6)
10/7/99
0.2588
0.3065 (-15.6)
0.2755 (-6.1)
10/8/99
0.2934
0.4173 (-29.7)
0.3627 (-19.1)
10/11/99
0.1978
0.1660 (19.1)
0.1320 (49.8)
10/12/99
0.2283
0.1570 (45.4)
0.1340 (70.4)
10/15/99
0.3779
0.2280 (65.7)
0.1985 (90.4)
10/18/99
0.3809
0.1900 (100.5)
0.1830 (108.1)
a Comparison with ISO 7027 reference method (2100AN IS reference turbidimeter).
b Comparison with EPA Method 180.1 (2100AN reference turbidimeter).
A similar positive bias was observed for all of the on-line turbidimeters tested in this verification
test, suggesting a systematic bias in the reference data. It should be noted that, since visible
granular deposits accumulated in the test apparatus during the on-line testing, it is possible that a
systematic negative bias may have existed as a result of large particles settling in the grab sample
vial between sample collection and reference analysis.
34

-------
6.2.2 Drift
6.2.2.1 Calibration Checks
The results from the twice-weekly calibration checks at 0.5 NTU are shown in Figure 6-6 and
summarized in Table 6-9. The results from these calibration checks indicate a negative bias of
the Monitek TST-SC turbidimeter relative to the ISO 7027 reference turbidimeter, which is
consistent with the results of the linearity test and accuracy calculations discussed in Sections
6.1.1 and 6.1.2, respectively.
However, on several occasions the results from the Monitek TST-SC turbidimeter indicated a
positive bias with respect to the reference turbidimeter. A review of the data suggested that
perhaps fouling of the optics had caused a positive shift in the readings. When shifts of this
nature were observed, either from the calibration checks, or from the daily accuracy readings
discussed in Section 6.2.1, the Monitek TST-SC turbidimeter was taken off-line for cleaning. For
the final calibration check on October 18, two sets of measurements were recorded. One set was
recorded before cleaning of the turbidimeter and showed no significant bias relative to the
reference turbidimeter. The second set was recorded after cleaning and showed a negative bias of
approximately 20 to 25% relative to the reference reading. These results strongly suggest that the
observed positive shifts in the Monitek readings were caused by optics fouling. After cleaning
the optics, positive shifts of this nature were eliminated and agreement with other calibration
checks was observed.
The secondary comparison in Table 6-9 shows essentially the same behavior relative to the
Method 180.1 data as was observed relative to the ISO 7027 reference data.
6.2.2.2 Final Linearity Check
Data from the final linearity check are shown in Figure 6-7a. These data were recorded after
completion of the four weeks of on-line testing and after the Monitek TST-SC turbidimeter had
been cleaned. As with the data from the initial linearity test, these data, excluding readings for
the <0.1 NTU water, were analyzed by linear regression. The results are summarized in
Table 6-10. In Table 6-11, the results of the final linearity test are compared with those from the
initial linearity check conducted at the start of the verification as part of the off-line phase.
The verification results of the regression analysis (Table 6-10) show a high degree of linearity,
with a slight negative bias in the slope with respect to the reference turbidimeter, and a small
negative intercept of approximately 0.1 NTU.
The verification results in Table 6-11 show a change of only 0.3% in the slope of the TST-SC
response relative to the ISO 7027 reference method between the initial and final linearity tests.
Similarly, the change in intercept shown in Table 6-11 (0.005 NTU) is very small. Based on the
results of the daily calibration checks, the average difference between the reference turbidimeter
readings and the stated turbidity value of the formazin standard used for the checks was 3.2%,
with a standard deviation of 2.4%. With these uncertainties, at the 95% confidence level, the
35

-------
0.8
0.6
_3
*c3
>
& 0.4
IS
s-
3
H
0.2
0
9/18/99
I'
i:
Cleaned Turbidimeter
9/27/99
9/25/99
Cleaned Turbidimeter
10/6/99
I Monitek
¦ ISO 7027
Method 180.1
Cleaned Turbidimeter
Between Tests
10/18/99
10/2/99	10/9/99	10/16/99	10/23/99
Date
Figure 6-6. Twice-Weekly Calibration Checks During On-Line Testing of the
MonitekTST-SC Turbidimeter
Table 6-9. Results of Calibration Checks Performed During On-Line Testing
ate
Monitek TST-SC
(NTU)
Verification Results3
(Relative Bias %)
Secondary Comparisonb
(Relative Bias %)
09/23/99
0.394
0.473 (-16.7)
0.461 (-14.5)
09/24/99
0.392
0.512 (-23.5)
0.492 (-20.4)
09/27/99
0.690
0.526 (31.1)
0.518 (33.2)
09/30/99
0.482
0.542 (-11.1)
0.524 (-8.1)
10/06/99
0.303
0.458 (-33.9)
0.432 (-29.9)
10/08/99
0.393
0.530 (-25.8)
0.514 (-23.6)
10/12/99
0.401
0.505 (-20.5)
0.498 (-19.4)
10/18/99
0.560
0.556 (0.8)
0.544 (3.0)
10/18/99
0.428
0.569 (-24.7)
0.544 (-21.3)
a Comparison with ISO 7027 reference method (2100AN IS reference turbidimeter).
b Comparison with EPA Method 180.1 (2100AN reference turbidimeter).
36

-------
5
4
$
H
£
Sd) 3
Ch
CD
Pi
.1? 2
T3
£
H
1
0
0	1	2	3	4	5
Reference Turbidity (NTU)
Figure 6-7a. Final Linearity Plot for Monitek TST-SC vs.
ISO 7027 Reference Turbidimeter
5
4
£ 2
T3 z
1
0
0
1
2
3
4
5
Reference Turbidity (NTU)
Figure 6-7b. Final Linearity Plot for Monitek
TST-SC vs. Method 180.1 Reference
Turbidimeter
37

-------
Table 6-10. Statistical Results of Final Linearity Test

Reference Turbidimeter
Linear Regression
Verification Results3
Secondary Comparisonb
Slope (std. error)
0.932 (0.003)
0.967 (0.002)
Intercept (std. error) NTU
-0.102 (0.007)
-0.104 (0.004)
r2
0.9998 (0.0289)
0.9999 (0.0176)
a Comparison with ISO 7027 reference method (2100AN IS reference turbidimeter).
b Comparison with EPA Method 180.1 (2100AN reference turbidimeter).
Table 6-11. Comparison of Results from Linearity Tests at Beginning and End of the
Verification Test
Verification Results3	Secondary Comparisonb
Slope Intercept (NTU) Slope Intercept (NTU)
Initial Linearity Test
0.935
-0.107
0.946
-0.089
Final Linearity Test
0.932
-0.102
0.967
-0.104
Difference
-0.003
0.005
0.021
-0.015
% Difference
-0.3
-
2.2
-
a Comparison with ISO 7027 reference method (2100AN IS reference turbidimeter).
b Comparison with EPA Method 180.1 (2100AN reference turbidimeter).
initial and final slopes are not significantly different from unity, and no drift can be inferred from
the difference between the slopes. The small difference in intercept is also insignificant.
The secondary comparison shown in Table 6-11 leads to a similar conclusion, as a change in the
regression slope of only 2.2% was observed, relative to Method 180.1
6.3 Other Performance Parameters
6.3.1 Cost
As tested, the cost of the Monitek TST-SC was approximately $1,000.
38

-------
6.3.2 Maintenance/Operational Factors
The Monitek TST-SC is designed for in-line or submersible operation. As tested the Monitek
TST-SC turbidimeter was configured for direct in-line pipe mounting. Time requirements for
installation were not assessed in this test, as temporary installation in the test apparatus did not
reflect the permanent installation in a water treatment facility. After installation, the Monitek
TST-SC required no operator input and provided data continuously throughout the verification
test.
The main limitation of the Monitek TST-SC is the necessity for off-line cleaning. During the off-
line testing no cleaning was required, presumably because of the relative cleanliness of the
samples being tested. During the on-line testing, the Monitek TST-SC turbidimeter required
cleaning after approximately a week of operation.
39

-------
Chapter 7
Performance Summary
The Monitek TST-SC is an on-line turbidimeter designed to provide continuous, real-time
measurement of the turbidity of aqueous solutions. The Monitek TST-SC turbidimeter provided
linear response over the tested range from 0.3 to 5 NTU. The TST-SC does not respond to
turbidity levels below -0.1 NTU, and the response of the TST-SC turbidimeter was not
measured above 5 NTU in this test. The slope of the response curve for the Monitek TST-SC
turbidimeter relative to the ISO 7027 reference turbidimeter was 0.935 at the beginning of the
test, with an intercept of - 0.107 and r2 > 0.999.
The TST-SC turbidimeter showed a negative bias relative to the reference turbidimeter in
measuring standard solutions. The observed negative bias ranged from ~ 30% at 0.3 NTU to
~ 8% at 5 NTU. This bias is consistent with the observed response curve slope of less than unity.
Much of the observed bias can be accounted for by the small negative intercept (~ -0.1 NTU). A
shift in the calibration curve to eliminate the intercept would dramatically improve the apparent
relative accuracy of the Monitek TST-SC turbidimeter. Measurement precision of the Monitek
TST-SC was approximately the same as for the ISO 7027 reference turbidimeter throughout the
range of turbidity measured in this test. The precision ranged from 3.3% RSD at 0.3 NTU to
0.35% RSD at 5 NTU.
Water temperature has a small effect on the response of the Monitek TST-SC turbidimeter at low
turbidity (0.3 NTU), but no effect at higher turbidity (5 NTU). The effect at 0.3 NTU amounts to
a decrease in observed turbidity of 0.4% per degree C relative to the ISO 7027 reference turbidi-
meter. A small effect of color on readings was found at low turbidity (~ 0.1 NTU), but none at
higher turbidity (5 NTU). However, the low turbidities used in this test approach the 0.1 NTU
lower response limit of the TST-SC, and consequently, it is unclear whether the observed effect
is in fact a response to color. In the 0.1 to 1.0 gpm range of flow rates tested for the Monitek
TST-SC turbidimeter, no statistically significant effect was found on the turbidity readings as a
function of sample flow rate.
In reading the turbidity of treated, unfiltered water from a municipal drinking water plant, the
Monitek TST-SC turbidimeter generally showed a positive bias of 0.1 to 0.2 NTU relative to the
reference turbidimeter readings of 0.1 to 0.6 NTU. This bias was observed over much of the four
weeks of measurement of the water plant sample stream. However, calibration checks of the
Monitek TST-SC turbidimeter, performed using 0.5 NTU formazin standards throughout the
same period, indicated a negative bias in the response of the Monitek TST-SC. A systematic bias
in the reference readings may have been present in the on-line test phase, and may have
40

-------
contributed to the observed differences between the Monitek TST-SC and reference readings on
stream samples. In addition, fouling of the TST-SC optics during stream sampling may have been
a factor; substantial decreases in TST-SC readings were observed when the turbidimeter optics
were cleaned.
A change of approximately 0.3% in the slopes of the response curves between the beginning and
end of the verification test was observed. This change is well within the experimental uncertainty
of the reference measurements and does not indicate a drift in the calibration of the Monitek
TST-SC. Similarly, no drift in the intercept of the response curve was observed. As noted above,
fouling of the optics during on-line testing caused intermediate drift, which was reversed upon
cleaning.
The Monitek TST-SC turbidimeter is easy to use and provides continuous on-line turbidity
readings. The turbidimeter was cleaned every 10 to 12 days during the test to remove residues
and material deposits from the optics on the inside of the turbidimeter. This was the only
maintenance required.
41

-------
Chapter 8
References
1.	"Water Quality—Determination of Turbidity," International Standard ISO 7027, Second
Edition, International Organization for Standardization, Geneva, 1990.
2.	Test/QA Plan for Verification of On-Line Turbidimeters, Battelle, Columbus, Ohio, June 3,
1999.
3.	"Determination of Turbidity by Nephelometry," Methods for the Determination of Inorganic
Substances in Environmental Samples, Method 180.1, EPA/600/R-93/100, U. S.
Environmental Protection Agency, Cincinnati, Ohio, August 1993.
4.	"Color in Water by Visual Comparison to Standards," Standard Methods for the
Examination of Water and Wastewater, 18th Edition, Method 2120-B, American Public
Health Association, 1992.
5.	Quality Management Plan (QMP) for the ETV Advanced Monitoring Systems Pilot, U.S.
EPA Environmental Technology Verification Program, Battelle, Columbus, Ohio,
September 1998.
42

-------
Appendix A
Data Recording Sheet
A-l

-------
Appendix B
Technical Systems Audit Report
B-l

-------