EPA/600/R-14/459 | December 2014 | www2.epa.gov/water-research
United States
Environmental Protection
Agency
Intrusion of Soil Water
through Pipe Cracks
r
Office of Research and Development
Water Supply and Water Resources Division
-------
EPA/600/R-14/459
December 2014
Intrusion of Soil
Water through Pipe Cracks
by
Lewis Rossman, Jill Neal, and Michelle Simon
United States Environmental Protection Agency
Urban Watershed Management Branch
National Risk Management Research Laboratory
Water Supply and Water Resources Division
Cincinnati, OH 45268
Srinivas Panguluri, Rendahandi Silva, and Radha Krishnan
CB&I Federal Services, LLC
National Risk Management Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
26 Martin Luther King Drive
Cincinnati, OH 45268
-------
Disclaimer
The U.S. Environmental Protection Agency (EPA), through its Office of Research and
Development (ORD) National Risk Management Research Laboratory (NRMRL) funded and
managed this experimental effort through EPA Contract No. EP-C-09-041. This report has been
both peer and administratively reviewed and approved for publication as an EPA document.
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use of a specific product.
Questions concerning this document or its application should be addressed to:
Michelle Simon, Ph.D., P.E.
Water Supply and Water Resources Division
National Risk Management Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
26 West Martin Luther King Dr.
Cincinnati, OH 45268
Simon.michelle@epa.gov
-------
Abstract
This report describes a series of experiments conducted at U.S. EPA's Test and Evaluation
Facility in 2013-2014 to study the intrusion of contaminated soil water into a pipe crack during
simulated backflow events. A test rig was used consisting of a 3' x 3' x 3'l acrylic soil box
with a one-inch diameter pipe running along 2 inches1 above the bottom of the soil box. The
pipe had a 1/16-inch1 hole at its top, positioned in the center of the box. Each experiment
consisted of filling the box with soil media, saturating the media with a solution containing
both a microbial and chemical tracer, running tap water through the pipe at a specific pressure
to represent normal conditions where clean water leaks out into the soil, and then turning off
the pipe flow and sampling the water drawn back into the pipe through the crack either by
gravity or forced pumping. Ten experimental runs were performed under various conditions -
backflow method (gravity drainage or forced pumping); type of soil media (sand or gravel);
microbial tracer (B. globigii orE. coli), and leak pressures (20, 40 and 55 psi1). All of the tests
indicated that significant levels of microbial tracer re-entered the pipe during the first five
minutes of backflow while the chemical tracer remained essentially at background. This
behavior can be explained by the displacement of soil water around the hole with clean water
during normal operation which removes the dissolved chemical tracer, but allows some
microbial particles to remain due to filtration through the soil media. The sand media provided
higher filter efficiency than the gravel media resulting in lower numbers of microorganisms
entering the pipe during backflow. Lower backflow rates produced lower average
concentration of microorganisms in the intruded soil water. As the gravity backflow period
extended beyond 5 minutes, microbial levels tended to level out or be reduced while the
chemical tracer concentrations began to increase.
Note the (') = foot = 0.3048 meter; 1 inch = 2.54 cm; 14.5 psi = 760 mm Hg
3
-------
Acknowledgments
Principal authors of this report were:
United States Environmental Protection Agency
Lewis Rossman
Michelle Simon
Jill Neal
CB&I Federal Services LLC
Srinivas Panguluri
Gune Silva
Radha Krishnan
Contributions of the following individuals and organizations that assisted in performing this effort
are gratefully acknowledged.
Peer Reviewers:
Michael Royer (EPA)
Ariamalar Selvakumar (EPA)
Jeff Yang (EPA)
Test-Rig Fabrication and Experiment Setup:
Timothy Kling (CB&I)
Dave El stun (CB&I)
Analysis:
Lee Heckman - Microbiology (CB&I)
David Griffith - Bromide (EPA SEE Employee)
-------
Table of Contents
Disclaimer 2
Abstract 3
Acknowledgments 4
Executive Summary 8
1.0 Introduction 10
1.1 Background 10
1.2 Experimental Overview 10
2.0 Experimental Setup and Test Protocol 11
2.1 Experimental Design 11
2.2 Experimental Test Run Protocol 12
2.3 Contaminant Tracer Stock Preparation and Analysis 14
3.0 Experimental Results 15
3.1 Test Run 1 15
3.2 Test Run 2 16
3.3 Test Run 3 17
3.4 Test Run 4 18
3.5 Test Run 5 - Control Run 19
3.6 Test Run 6 20
3.7 Test Run 7 21
3.8 Test Run 8 22
3.9 Test Run 9 24
3.10 Test Run 10 25
4.0 Analysis of Results 27
4.1 Intrusion Flow Rates and Pressures 27
4.2 Short Term Intrusion Behavior 28
4.3 Effect of Media and Backflow Method 30
4.4 B. globigii versus E. coli 30
4.5 Effect of Leak Pressure 31
-------
4.6 Longer Term Intrusion Behavior 32
5.0 Conclusions and Recommendations 34
6.0 References 36
List of Tables
Table 2-1 Experimental Test Conditions 12
Table 3-1 Test Run 1 Results 15
Table 3-2 Test Run 2 Results 16
Table 3-3 Test Run 3 Results 17
Table 3-4 Test Run 4 Results 18
Table 3-5 Test Run 5 (Control) Results 19
Table 3-6 Test Run 6 Results 20
Table 3-7 Test Run 7 Results 21
Table 3-8 Test Run 8 Results 23
Table 3-9 Test Run 9 Results 24
Table 3-10 Test Run 10 Results 25
Table 4-1 Observed 5-Minute Backflow Rates 27
Table 4-2 Log Reductions of B. globigii 30
List of Figures
Figure 1-1 Schematic Diagram of Experimental Setup (1 - pressure sensor; 2 - shutoff valves; 3 -
peristaltic pump) 8
Figure 2-1 Schematic Diagram of Experimental Setup (1 - pressure sensor; 2 - shutoff valves; 3 -
peristaltic pump) 11
Figure 4-1 Intrusion of Tracer into Pipe for Sand Media 29
Figure 4-2 Intrusion of Tracer into Pipe for Gravel Media 29
Figure 4-3 Comparison of Microbial Tracers 31
Figure 4-4 Effect of Leak Pressure on Intrusion of Microbial Tracer 32
Figure 4-5 Longer Term Intrusion under Gravity Backflow 33
-------
Abbreviations
AL after leak
B. globigii Bacillus globigii
BL before leak
CPU colony forming units
DRN drain
E. coli Escherichia coli - K12 strain
EPA Environmental Protection Agency
gpm gallons per minute
HASP Health and Safety Plan
L liter
min minute
mg/L milligram per liter
mL milliliter
mL/min milligram per minute
NA not available
OVF overflow
ppm parts per million
PSI pounds per square inch
PVC polyvinyl chloride
QAPP Quality Assurance Proj ect Plan
rpm revolutions per minute
SMP, SMPL sample
SOP Standard Operating Procedure
T&E EPA Test & Evaluation Facility
TMTC too many to count
Note the (') = foot = 0.3048 meter; 1 inch = 2.54 cm; 14.5 psi = 760 mm Hg
-------
Executive Summary
Does water leaking out of the pipe during normal pressure operation create a "soil washing"
effect that protects against contaminant intrusion during low pressure events? Or do
contaminants backflow from the surrounding media into a hole in a pipe when water is no longer
flowing from the pipe out into the surrounding media? A series of experiments were designed
to answer these questions.
An apparatus was constructed at EPA's Technology and Evaluation Center in 2013.
Soil Box
Pipe Outflow to
Drain (DRN)
Soil Box Overflow
(OVF)
Pipe Under-Drain Sampling Line
(SMP) Pumped or Gravity
Domestic De-
chlorinated
Water Line
Figure 1-1 Schematic Diagram of Experimental Setup (1 - pressure sensor; 2 - shutoff
valves; 3 - peristaltic pump)
The details for this apparatus are presented in Section 2.0. The soil box was filled with sand or
gravel media saturated with microbial and bromide tracers. The under-drain valve was closed
and Greater Cincinnati MSD water flowed through the pipe for one hour. The pipe's inlet and
outlet valves were closed and the under-drain valve was opened and pump operated if needed.
The water in the pipe was sampled at 30 second intervals for five minutes, then five minutes, and
then every ten minutes until one hour after the under-drain valve was opened. Based on the
results of the experiments reported here, the following conclusions can be drawn for short term
backflow events (less than 5 minutes):
1. There will be no intrusion of chemical tracer into the pipe because of tracer washout in the
vicinity of the crack during normal operation of the leaking pipe.
2. There will however, be intrusion of microbial tracer since only partial washout occurs
because of filtering by the soil media
3. Sand media provides higher filter efficiency than does gravel media in the direction of
backflow resulting in lower numbers of microorganisms entering the pipe during backflow.
-------
4. The lower the backflow flow rate the lower the average concentration of microorganisms
in the back-flowed soil water.
5. Higher leak pressures result in lower microbial concentrations in backflow from sand
media. The same could not be shown for gravel.
As the backflow period was extended up to 60 minutes, the microbial tracer level tended to
decrease for sand and level out for gravel. The chemical tracer continued to stay at background
level for sand but started to rise for gravel.
-------
1.0 Introduction
1.1 Background
The U. S. Environmental Protection Agency (EPA) has identified the possible intrusion of
pathogenic organisms from surrounding soil into pipeline cracks during low pressure events as
being a potentially serious source of drinking water contamination (EPA 2006). During normal
periods of operation, water flowing under pressure in a pipe would leak out of any cracks and
present a hydraulic barrier against the entry of any contaminants from the soil surrounding the
pipe. However, short-term transient pressure drops can occur when pumps and valves are closed
during the course of normal system operations. In certain cases, these pressure drops can be large
enough to create a negative pressure inside the pipe that would allow outside water to be drawn
into the pipe through a crack. These transients normally last only for a few seconds. A similar
condition could occur after a pipe break where cracks downstream of the break could allow
contaminated soil and water to enter during the period of time that water continues to flow in a de-
pressurized state. Although these kinds of hydraulic conditions have been known to occur
(AWWARF 2004), it has not been established whether any water drawn back into the pipe through
the crack would be contaminated or not, since the soil in the vicinity of the crack is being washed
with clean water for the majority of the time. EPA sponsored an experimental test program to
address this issue.
1.2 Experimental Overview
The experimental test runs documented in this report were designed to simulate the conditions
described in Section 1.1. Specifically, the objective was to determine the extent of contamination
due to soil water entering a pipe crack after a low pressure condition is induced. In a series of test
runs, a leaking pipe was placed in a soil box whose media was saturated with "contaminated"
tracer solution, and clean water was subsequently allowed to leak into the surrounding soil over a
period of time. Then the clean water pipe flow was shut off, allowing soil water to flow back
through the leak opening into the pipe main. Water samples were then drawn through a drain
port/line connected to the main pipe, either by gravity or by applying suction (using a peristaltic
pump). Tracer concentrations of the backflow samples collected through the drain line were
monitored over time. The tracer solution contained both an inert non-reactive inorganic salt
(sodium bromide) and non-pathogenic microbiological contaminants [either Bacillus globigii (B.
globigii) or Escherichia coli (E. coli) -K12 strain]. A detailed protocol is provided in Section 2.0
of the document. A project specific Quality Assurance Project Plan (QAPP) was developed by
CB&I Federal Services LLC, formerly Shaw Environmental & Infrastructure, Inc. (CB&I 2014),
for conducting these tests. Both of these documents were approved by EPA, prior to conducting
the experimental test runs.
10
-------
2.0 Experimental Setup and Test Protocol
2.1 Experimental Design
Figure 2-1 shows the schematic of the experimental setup at the EPA Test & Evaluation (T&E)
Facility. It consists of an open 3'x3'x3' acrylic box with sections of a 1" diameter clear Polyvinyl
Chloride (PVC) pipe running through it. During the tests, the section of pipe within the box was
replaced with one that had a crack of desired size and shape drilled into it. One corner of the box
was sectioned off with pluggable holes at different vertical depths to provide an outlet to capture
any overflow when the box was operated with the pipe leaking water into it. Preliminary tests
indicated that it was best to plug all of the holes except at the top of the soil level to allow for water
to build up and overflow.
Soil Box
2
< — i y ,
Pipe Outflow to 2 A
Drain (DRN)
3 £
j t
:
^
V
CD
Pipe Under-Drain Sampling I
•^ " rSMP^ Pnmned or Gravitv
Soil Box Overflow
* (OVF)
1
PN 1 4
Domestic De
chlorinated
,ine Water Line
Figure 2-1 Schematic Diagram of Experimental Setup (1 - pressure sensor; 2 - shutoff
valves; 3 - peristaltic pump)
Table 2-1 presents the experimental conditions for a series often experimental runs that were
performed using the soil box for this study. Commercially available all-purpose sand or pea sized
gravel (gravel size ranging between 0.5 to 1.5 centimeters) purchased from a local retailer (Home
Depot) was used as the media in the soil box. The pipe leak opening was simulated as a 0.0625
inch (1.59 mm) diameter round hole. Once flow through the pipe was stopped, samples were
withdrawn from the pipe either under gravity flow or by induced suction (i.e., pumped) flow. The
pumped-flow was performed using a peristaltic pump (GeoPump Series II, Geotech Environmental
Equipment, Inc., Denver, Colorado). The pumped flow simulates a worst-case scenario of potential
intrusion of contaminants occurring due to transient negative-pressure conditions (created by
pressure surges and water hammer) in a typical water supply system. The GeoPump Series II model
pump is designed to deliver water at a rate of 1.67 milliliters (mL) per revolution with operating
11
-------
speeds ranging between 60 to 600 revolutions per minute (rpm), effectively delivering water at a
rate between 100 and 1,000 mL per minute (mL/min). Whenever utilized for testing, the pump was
operated at the maximum allowable speed.
Table 2-1 Experimental Test Conditions
Experimental
Test Run No.
1
2
3
4
5
6
7
8
9
10
Leak
Pressure
40psi
40psi
40psi
40psi
40psi
40psi
55 psi
55 psi
20 psi
20 psi
Soil
Medium
Sand
Sand
Gravel
Gravel
Sand
Gravel
Gravel
Sand
Sand
Gravel
Contaminant Tracer
in Soil-Box Media
Bromide and B. globigii
Bromide and B. globigii
Bromide and B. globigii
Bromide and B. globigii
Clean Water (Control)
Bromide and E. coli
Bromide and E. coli
Bromide and B. globigii
Bromide and B. globigii
Bromide and E. coli
Sample Withdrawal
Method
Gravity
Pumped
Gravity
Pumped
Gravity
Pumped
Pumped
Gravity
Gravity
Pumped
The soil medium used in each experiment was initially saturated with a tracer solution of de-
chlorinated water containing approximately 106and lO8^. globigii spores or E. coli, respectively
reported as Colony Forming Units per 100 mL (CFU/100 mL) and 30 mg/L of bromide ion. The
water flowing through the leaky pipe originated from a de-chlorinated city water line at various
target pressures between 20 and 55 pounds per square inch (psi) and a pipe main flow rate between
7 and 18 gallons per minute (gpm). During the testing, the measured pressure was the key criterion
and the main pipe flow rate was simply monitored at the stable condition that was achieved during
the individual test run. Tests under the same pressure criteria were run with similar main flow
rates.
2.2 Experimental Test Run Protocol
Overall, each test run comprised of the following three categorical steps: 1) Preparing/sampling
the stock solutions and applying it to the selected soil medium, 2) Establishing initial pressurized
pipe condition with clean tap water leaking out into the soil-box for an hour and collecting
specified timed-samples (from pipe outflow and soil-box overflow), 3) Inducing backflow by
shutting off the clean water supply and collecting specified timed samples from the sampling line
(pipe under drain). Appendix A of the QAPP contains the datasheet that describes the sample times
and sample locations. The general procedure used for conducting the intrusion tests are outlined
below:
12
-------
Preparing/sampling the stock solutions and applying it to the selected soil medium.
1. The stock tracer solutions were prepared for each test run and analyzed for bromide and B.
globigii (or E1. coif). The prepared stock was mixed with bromide and de-chlorinated water
in a 55-gallon drum immediately before application to soil (Sample IDs T-XX-Stock 1 and
T-XX-Stock 2).
2. Starting with an empty soil box with all drain valves closed, six inches of dry media were
filled and saturated with tracer solution (the media was raked as needed to achieve a
relatively even distribution of soil and tracer solution).
3. Step 2 was repeated for two more six-inch layers of media for a total of 18 inches of media.
4. Another six inches of selected media were added without any tracer solution for a total of
24 inches of media.
5. The under-drain valve was opened and allowed to flow for a few minutes before taking a
sample (Sample IDs T-XX-SMPL-BL [before the main pipe line was pressurized and a
leak was induced]). This sample was considered to be representative of the contaminated
water that could potentially infiltrate into the main pipe during a back-siphoning event.
Establishing initial pressurized pipe-condition with de-chlorinated tap-water leaking out into the
soil-box for an hour and collecting specified timed-samples (from pipe outflow to drain and soil-
box overflow shown in Figure 3-1).
6. The under-drain valve was closed and the main pipe shut-off valves were opened to allow
de-chlorinated city water to run through the pipe at the selected pressure rate. The pressure
rate was adjusted by restricting the shut-off valves until the desired pressure was reached
(which took about a minute).
7. At fifteen minute intervals, samples were collected from the main pipe outflow drain line
(Sample IDs T-XX-DRN 15 through T-XX-DRN 60) and soil box overflow lines (Sample
IDs T-XX-OVF 15 through T-XX-OVF 60). Flow rates and main line pressure were
recorded to ensure the desired pressure was maintained.
8. After 55 minutes, with water still running through the main pipe, the under-drain line was
opened to allow the water to flow freely for 1-2 minutes to remove air bubbles and any
contamination in the line from Step 5.
Inducing backflow by shutting off the clean water supply and collecting specified timed samples
from sampling line (pipe under drain shown in Figure 2-1).
9. At 60 minutes, sample was taken from the under-drain line (Sample IDs T-XX-SMPL-AL
[after the main pipe line is pressurized and leak was induced]) and then the valves were
closed on both ends of the pipe to stop all flow through it.
10. Immediately a sample was taken from the under-drain line (Sample ID T-XX-SMPL 0.1)
and subsequently at 30-second intervals (Sample IDs T-XX-SMPL 0.5 through T-XX-
SMPL 5.0) until five minutes elapsed since flow to the pipe was stopped. During gravity
13
-------
discharge sampling, the flow rate was slow and did not allow for sampling at the 30-second
rate. Therefore, samples were collected at the maximum rate possible during this initial
post-leak 5-minute period.
11. After the initial 5-minute high-rate sampling activities were completed, the flow from the
under-drain sample line (SMPL) was directed into a graduated cylinder or a pre-weighed
container and recorded the water level or weight for one-minute to compute sample flow
rates was recorded.
12. At 10-minutes post-leak time, another sample was collected (Sample ID T-XX-SMPL 10)
and sample flow rate was recorded using the methodology described in Step 11.
13. Step 12 was repeated at ten minute intervals until 60 minutes post-leak time had elapsed.
14. At this point, the test was considered complete and water was drained from the soil box.
Collected samples were transferred for analysis.
After completion of each test, the soil box media was decontaminated and disposed in accordance
to the EPA approved HASP procedures.
2.3 Contaminant Tracer Stock Preparation and Analysis
The stock of B. globigii spores or E. coll with approximate concentration of 109/mL were grown
at the EPA T&E Facility BSL-2 laboratory following CB&I T&E SOP 309 and T&E SOP 310
for B. globigii and E. coll., respectively. The stock B. globigii with the target concentrations were
prepared by mixing a culture of vegetative cells/stock with generic sporulation media and
incubating by gentle shaking (-150 rpm) at 35°C for five days. The presence of spores was
confirmed using phase-contrast microscopy (<0.1% vegetative cells). Stock E. coli was prepared
by sub-culturing a pre-existing flask in nutrient broth. Cultures of E. coli were incubated at 37°C
with gentle shaking (-150 rpm) for twenty-four hours. The injection suspension was prepared by
mixing an appropriate amount of the B. globigii or E. coli stock in 1 L of 0.01% Tween 20 (a
dispersing agent). The amount of the stock was estimated based on the target influent concentration
(~ 106/100 mL) and the pore volume of the media placed in the soil box.
All samples collected from the stock tracer solution, the pipe outflow drain line, the soil box
overflow line, and the soil box under-drain line were analyzed for bromide ion using CB&I SOP
405 Ion Chromotography (CB&I 2014 Appendix C) and B. globigii (CB&I 2014 Appendix B SOP
309) (orE. coli) (CB&I 2014 Appendix D SOP310). Additional sampling plan detail is included
in the project QAPP (CB&I 2014).
14
-------
3.0 Experimental Results
3.1 Test Run 1
The test run conditions for this test were as follows: leak pressure - 40 psi; soil media - sand;
selected contaminant tracer in soil-box media - bromide and B. globigii; and backflow sample
withdrawal method - gravity. Table 3-1 summarizes the results from Test Run 1.
Table 3-1 Test Run 1 Results
Sample ID
T-01-Stock-01
T-01-Stock-02
T-01-SMPL-BL
T-01 SMPL-AL
T-01-SMPL-0.1
T-01-SMPL-1.20
T-01-SMPL-2.20
T-01-SMPL-3.30
T-01-SMPL-4.40
T-01-SMPL-10
T-01-SMPL-20
T-01-SMPL-30
T-01-SMPL-40
T-01-SMPL-50
T-01-SMPL-60
T-01-OVF-15
T-01-OVF-30
T-01-OVF-45
T-01-OVF-60
T-01-DRN-15
T-01-DRN-30
T-01-DRN-45
T-01-DRN-60
B. globigii
(CFU/100ml_)
3.50E+07
2.90E+07
2.80E+05
5
50
38
663
2550
2450
260
225
350
ND
ND
10
TMTC*
1.90E+04
4.00E+03
1.30E+03
ND
5
ND
NDO
B. globigii
(log values)
7.54
7.46
5.45
0.70
1.70
1.58
2.82
3.41
3.39
2.41
2.35
2.54
ND
ND
1.00
NA**
4.28
3.60
3.11
ND
0.70
ND
ND
Bromide
(mg/L)
29.73
29.67
27.76
0.04
0.08
0.03
0.03
0.03
0.03
0.04
0.05
0.06
0.07
0.08
0.09
15.45
8.39
6.54
4.94
0.04
0.03
0.03
0.03
*TMTC - too many to count, **NA - not available, ND-not detected.
As mentioned previously (in Section 2.2, Step 10), the gravity-based sample flow rate encountered
during the test run was too slow to allow for sampling every 30 seconds (first 5-minutes post-leak
sampling requirement). Field measurements indicated that the gravity-based sample flow rates
ranged from 215 ml/min (at five minutes) to 160 ml/min (towards the end of the 60-minute
sampling period). Therefore, samples were collected as often as the flow rate allowed and the time
stamps were noted in the Sample IDs. For example, in Table 3-1, the Sample ID T-01-SMPL-1.20
15
-------
represents a sample collected at the 1.2 minute interval post leak. In general, for all test runs, if the
B. globigii values in the first pass were below detection (i.e., lower than 3 orders of magnitude
based on expected value from preliminary and/or previous test runs), the samples were reanalyzed
if sufficient sample volume was available. In this test, the sample ID T-01-OVF-15 yielded counts
at a much higher than expected three orders of magnitude range. Prior to analysis, each B. globigii
sample was diluted and analyzed to fit within an expected 3 orders of magnitude range. However,
this was the first test and there was no clear expected range, and a reanalysis for this sample was
not possible because the sample volume was spent during the initial analysis.
3.2 Test Run 2
The test run conditions for this test were as follows: leak pressure - 40 psi; soil media - sand;
selected contaminant tracer in soil-box media - bromide and B. globigii; and backflow sample
withdrawal method - pumped. Table 3-2 summarizes the results from Test Run 2.
Table 3-2 Test Run 2 Results
Sample ID
T-02-Stock-01
T-02-Stock-02
T-02-SMPL-BL
T-02-SMPL-AL
T-02-SMPL-0.1
T-02-SMPL-0.5
T-02-SMPL-1 .0
T-02-SMPL-1 .5
T-02-SMPL-2.0
T-02-SMPL-2.5
T-02-SMPL-3.0
T-02-SMPL-3.5
T-02-SMPL-4.0
T-02-SMPL-4.5
T-02-SMPL-5.0
T-02-SMPL-10
T-02-SMPL-20
T-02-SMPL-30
T-02-SMPL-40
T-02-SMPL-50
T-02-SMPL-60
B. globigii
(CFU/100ml_)
3.65E+07
6.60E+07
1.50E+06
*
5.20E+04
*
2.05E+04
6.07E+04
8.00E+03
7.37E+03
5.80E+03
5.65E+03
2.95E+03
2.10E+03
1.90E+03
2.70E+03
9.50E+02
1.00E+02
1.05E+02
1.05E+02
4.50E+01
B. globigii
(log values)
7.56
7.82
6.18
4.72
4.31
4.78
3.90
3.87
3.76
3.75
3.47
3.32
3.28
3.43
2.98
2.00
2.02
2.02
1.65
Bromide
(mg/L)
30.08
29.96
30.28
0.02
0.02
0.02
0.02
0.02
0.02
0.03
0.02
0.02
0.03
0.03
0.03
0.03
0.04
0.05
0.06
0.06
0.07
16
-------
Sample ID
T-02-OVF-15
T-02-OVF-30
T-02-OVF-45
T-02-OVF-60
T-02-DRN-15
T-02-DRN-30
T-02-DRN-45
T-02-DRN-60
B. globigii
(CFU/100ml_)
2.50E+04
8.00E+02
5.00E+02
7.38E+02
ND
ND
ND
40
B. globigii
(log values)
4.40
2.90
2.70
2.87
ND
ND
ND
1.60
Bromide
(mg/L)
9.026
19.85
12.50
8.05
0.05
0.03
0.02
0.02
* Samples lost during heat-shock treatment (lids popped), ND-not detected.
Field measurements indicated that the pumped sample flow rate was approximately 740 ml/min
and relatively constant over time.
3.3 Test Run 3
The test run conditions for this test were as follows: leak pressure - 40 psi; soil media - gravel;
selected contaminant tracer in soil-box media - bromide and B. globigii; and backflow sample
withdrawal method - gravity. Table 3-3 summarizes the results from Test Run 3.
Table 3-3 Test Run 3 Results
Sample ID
T-03-Stock-01
T-03-Stock-02
T-03-SMPL-BL
T-03-SMPL-AL
T-03-SMPL-0.1
T-03-SMPL-0.5
T-03-SMPL-1 .0
T-03-SMPL-1 .5
T-03-SMPL-2.0
T-03-SMPL-2.5
T-03-SMPL-3.0
T-03-SMPL-3.5
T-03-SMPL-4.0
T-03-SMPL-4.5
T-03-SMPL-5.0
T-03-SMPL-10
T-03-SMPL-20
T-03-SMPL-30
B. globigii
(CFU/100ml_)
8.40E+07
7.43E+07
2.60E+05
5.00E+00
ND
7.50E+03
1.75E+04
2.60E+04
6.00E+03
1.05E+04
1.68E+04
7.10E+03
9.23E+03
8.50E+03
5.00E+03
1.24E+03
1.43E+03
2.20E+03
B. globigii
(log values)
7.92
7.87
5.41
0.70
ND
3.88
4.24
4.41
3.78
4.02
4.23
3.85
3.97
3.93
3.70
3.09
3.15
3.34
Bromide
(mg/L)
30.53
30.18
26.48
0.02
0.02
0.02
0.02
0.03
0.04
0.04
0.04
0.05
0.05
0.06
0.06
0.12
0.47
0.75
17
-------
Sample ID
T-03-SMPL-40
T-03-SMPL-50
T-03-SMPL-60
T-03-OVF-15
T-03-OVF-30
T-03-OVF-45
T-03-OVF-60
T-03-DRN-15
T-03-DRN-30
T-03-DRN-45
T-03-DRN-60
B. globigii
(CFU/100ml_)
1.80E+03
1.80E+03
1.70E+03
8.55E+05
5.75E+05
1.80E+05
8.95E+04
5
ND
ND
ND
B. globigii
(log values)
3.26
3.26
3.23
5.93
5.76
5.26
4.95
0.70
ND
ND
ND
Bromide
(mg/L)
0.87
1.02
0.94
25.32
24.58
22.11
13.59
0.03
0.03
0.02
0.02
ND-not detected
Field measurements indicated that the gravity sample flow rate ranged from 646 ml/min (at five
minutes) to 583 ml/min (towards the end of the 60-minute sampling period).
3.4 Test Run 4
The test run conditions for this test were as follows: leak pressure - 40 psi; soil media - gravel;
selected contaminant tracer in soil-box media - bromide and B. globigii; and backflow sample
withdrawal method - pumped. Table 3-4 summarizes the results from Test Run 4.
Table 3-4 Test Run 4 Results
Sample ID
T-04-Stock-01
T-04-Stock-02
T-04-SMPL-BL
T-04-SMPL-AL
T-04-SMPL-0.1
T-04-SMPL-0.5
T-04-SMPL-1 .0
T-04-SMPL-1 .5
T-04-SMPL-2.0
T-04-SMPL-2.5
T-04-SMPL-3.0
T-04-SMPL-3.5
T-04-SMPL-4.0
T-04-SMPL-4.5
T-04-SMPL-5.0
T-04-SMPL-10
B. globigii
(CFU/100ml_)
7.37E+07
8.03E+07
5.20E+06
1.00E+01
TMTC*
2.23E+04
4.60E+04
3.95E+04
4.05E+04
4.35E+04
2.03E+04
2.47E+04
1.53E+04
1.35E+04
1.15E+04
8.05E+03
B. globigii
(log values)
7.87
7.90
6.72
1.00
NA**
4.35
4.66
4.60
4.61
4.64
4.31
4.39
4.19
4.13
4.06
3.91
Bromide
(mg/L)
30.04
29.96
27.65
0.06
0.02
0.03
0.03
0.03
0.03
0.04
0.04
0.04
0.05
0.05
0.05
0.11
18
-------
Sample ID
T-04-SMPL-20
T-04-SMPL-30
T-04-SMPL-40
T-04-SMPL-50
T-04-SMPL-60
T-04-OVF-15
T-04-OVF-30
T-04-OVF-45
T-04-OVF-60
T-04-DRN-15
T-04-DRN-30
T-04-DRN-45
T-04-DRN-60
B. globigii
(CFU/100ml_)
3.10E+04
3.90E+04
3.20E+04
3.70E+04
4.40E+04
3.10E+06
3.00E+06
1.30E+06
4.35E+05
5
ND
5
ND
B. globigii
(log values)
4.49
4.59
4.51
4.57
4.64
6.49
6.48
6.11
5.64
0.70
ND
0.70
ND
Bromide
(mg/L)
0.40
0.64
0.75
0.84
0.98
24.66
24.68
18.97
9.39
0.03
0.03
0.02
0.02
*TMTC - too many to count, values outside of expected range. **NA - not available,
ND-not detected
Field measurements indicated that the pumped sample flow rate ranged from 800 ml/rnin (at five
minutes) to 735 ml/min (towards the end of the 60-minute sampling period).
3.5 Test Run 5 - Control Run
The test run conditions for this test were as follows: leak pressure - 40 psi; soil media - sand;
selected contaminant tracer in soil-box media - de-chlorinated water (control run); and backflow
sample withdrawal method - gravity. Table 3-5 summarizes the results from Test Run 5.
Table 3-5 Test Run 5 (Control) Results
Sample ID
T-05-Stock-01
T-05-Stock-02
T-05-SMPL-BL
T-05-SMPL-AL
T-05-SMPL-0.1
T-05-SMPL-0.5
T-05-SMPL-1.5
T-05-SMPL-2.5
T-05-SMPL-3.5
T-05-SMPL-4.5
T-05-SMPL-5.0
T-05-SMPL-10
B. globigii
(CFU/100ml_)
NA
NA
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
B. globigii
(log values)
NA
NA
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
Bromide
(mg/L)
NA
NA
0.66
0.01
0.014
0.02
0.02
0.02
0.02
0.02
0.02
0.02
19
-------
Sample ID
T-05-SMPL-20
T-05-SMPL-30
T-05-SMPL-40
T-05-SMPL-50
T-05-SMPL-60
T-05-OVF-15
T-05-OVF-30
T-05-OVF-45
T-05-OVF-60
T-05-DRN-15
T-05-DRN-30
T-05-DRN-45
T-05-DRN-60
B. globigii
(CFU/100ml_)
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
B. globigii
(log values)
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
Bromide
(mg/L)
0.03
0.03
0.03
0.04
0.03
1.16
0.71
0.32
NA
0.03
0.02
0.01
0.02
ND-not detected
As mentioned previously, the gravity-based sample flow rate encountered during the test run was
too slow to allow for sampling every 30 seconds (first 5-minutes post-leak sampling requirement).
Therefore, samples were collected as often as the flow rate allowed and the time stamps were noted
in the Sample IDs. For example, in Table 3-5, the Sample ID T-05-SMPL-1.5 represents a sample
collected at the 1.5 minute interval post leak. Field measurements indicated that the gravity sample
flow rate ranged from 215 ml/min (at five minutes) to 160 ml/min (towards the end of the 60-
minute sampling period).
3.6 Test Run 6
The test run conditions for this test were as follows: leak pressure - 40 psi; soil media - gravel;
selected contaminant tracer in soil-box media - bromide and E. coli; and backflow sample
withdrawal method - pumped. Table 3-6 summarizes the results from Test Run 6.
Table 3-6 Test Run 6 Results
Sample ID
T-06-Stock-01
T-06-Stock-02
T-06-SMPL-BL
T-06-SMPL-AL
T-06-SMPL-0.1
T-06-SMPL-0.5
T-06-SMPL-1 .0
E. co//
(CFU/100ml_)
4.10E+05
5.20E+05
3.50E+05
ND
1.00E+02
9.70E+01
1.48E+02
E. co// (log
values)
5.61
5.72
5.54
ND
2.00
1.99
2.17
Bromide
(mg/L)
30.22
30.31
26.29
0.03
0.03
0.03
0.03
20
-------
Sample ID
T-06-SMPL-1 .5
T-06-SMPL-2.0
T-06-SMPL-2.5
T-06-SMPL-3.0
T-06-SMPL-3.5
T-06-SMPL-4.0
T-06-SMPL-4.5
T-06-SMPL-5.0
T-06-SMPL-10
T-06-SMPL-20
T-06-SMPL-30
T-06-SMPL-40
T-06-SMPL-50
T-06-SMPL-60
T-06-OVF-15
T-06-OVF-30
T-06-OVF-45
T-06-OVF-60
T-06-DRN-15
T-06-DRN-30
T-06-DRN-45
T-06-DRN-60
E. co//
(CFU/100ml_)
2.41 E+02
1.75E+02
2.18E+02
7.17E+02
1.45E+03
4.35E+03
6.13E+03
8.16E+03
3.00E+04
6.20E+04
5.00E+04
2.42E+03
2.42E+03
2.42E+03
2.00E+05
2.00E+05
1.10E+05
1.60E+05
ND
ND
ND
ND
E. co// (log
values)
2.38
2.24
2.34
2.86
3.16
3.64
3.79
3.91
4.48
4.79
4.70
3.38
3.38
3.38
5.30
5.30
5.04
5.20
ND
ND
ND
ND
Bromide
(mg/L)
0.04
0.04
0.06
0.13
0.30
0.54
0.86
1.21
4.65
8.49
11.29
13.17
14.20
14.84
26.97
24.39
21.67
19.30
0.05
0.04
0.03
0.03
ND-not detected
Field measurements indicated that the pumped sample flow rate ranged from 725 ml/rnin (at five
minutes) to 670 ml/min (towards the end of the 60-minute sampling period).
3.7 Test Run 7
The test run conditions for this test were as follows: leak pressure - 55 psi; soil media - gravel;
selected contaminant tracer in soil-box media - bromide and E. coli; and backflow sample
withdrawal method - pumped. Table 3-7 summarizes the results from Test Run 7.
Table 3-7 Test Run 7 Results
Sample ID
T-07-Stock-01
T-07-Stock-02
T-07-SMPL-BL
T-07-SMPL-AL
E. co//
(CFU/100ml_)
6.10E+06
5.30E+06
ND
ND
E. co// (log
values)
6.79
6.72
ND
ND
Bromide
(mg/L)
29.94
29.83
25.06
0.039
21
-------
Sample ID
T-07-SMPL-0.1
T-07-SMPL-0.5
T-07-SMPL-1 .0
T-07-SMPL-1 .5
T-07-SMPL-2.0
T-07-SMPL-2.5
T-07-SMPL-3.0
T-07-SMPL-3.5
T-07-SMPL-4.0
T-07-SMPL-4.5
T-07-SMPL-5.0
T-07-SMPL-10
T-07-SMPL-20
T-07-SMPL-30
T-07-SMPL-40
T-07-SMPL-50
T-07-SMPL-60
T-07-OVF-15
T-07-OVF-30
T-07-OVF-45
T-07-OVF-60
T-07-DRN-15
T-07-DRN-30
T-07-DRN-45
T-07-DRN-60
E. coli
(CFU/100ml_)
ND
1.00E+03
1.70E+03
2.00E+03
3.10E+03
2.00E+03
2.00E+03
1.00E+03
1.10E+03
3.10E+03
3.10E+03
1.20E+03
2.70E+03
1.00E+04
3.70E+03
4.30E+03
3.60E+03
6.40E+05
6.20E+05
1.60E+05
1.90E+05
ND
ND
ND
ND
E. coli (log
values)
ND
3.00
3.23
3.30
3.49
3.30
3.30
3.00
3.04
3.49
3.49
3.08
3.43
4.00
3.57
3.63
3.56
5.81
5.79
5.20
5.28
ND
ND
ND
ND
Bromide
(mg/L)
0.038
0.040
0.043
0.042
0.047
0.047
0.049
0.052
0.054
0.058
0.059
0.089
0.171
0.238
0.296
0.356
0.421
27.41
22.23
15.42
9.209
0.074
0.058
0.046
0.040
ND-not detected
Field measurements indicated that the pumped sample flow rate ranged from 840 mL/rnin (at five
minutes) to 790 mL/min (towards the end of the 60-minute sampling period). It should be noted
that the E. coli concentration in sample T-07-SMPL-BL was not detected, but bromide was
recovered as expected. In this case, a non-detect value simply indicates that E. coli was either
below the detection limit or the sample was potentially over diluted based on expected value and
did not fit in the expected 3 orders of magnitude values. Also, the E coli test kit covers only 3
orders of magnitude during analysis and with a 24-hour hold time, there is only one pass at the
laboratory analysis and re-analysis was not possible.
3.8 Test Run 8
The test run conditions for this test were as follows: leak pressure - 55 psi; soil media - sand;
selected contaminant tracer in soil-box media - bromide and B. globigii; and backflow sample
withdrawal method - gravity. Table 3-8 summarizes the results from Test Run 8.
22
-------
Field measurements indicated that the gravity sample flow rate ranged from 130 rnL/min (at five
minutes) to 91 mL/min (towards the end of the 60-minute sampling period). It is interesting to note
that the B. globigii concentration in timed samples T-08-SMPL-40 through T-08-SMPL-60 was
found to be non-detect.
Table 3-8 Test Run 8 Results
Sample ID
T-08-Stock-01
T-08-Stock-02
T-08-SMPL-BL
T-08-SMPL-AL
T-08-SMPL-0.1
T-08-SMPL-1 .0
T-08-SMPL-2.0
T-08-SMPL-3.0
T-08-SMPL-4.0
T-08-SMPL-5.0
T-08-SMPL-10
T-08-SMPL-20
T-08-SMPL-30
T-08-SMPL-40
T-08-SMPL-50
T-08-SMPL-60
T-08-OVF-15
T-08-OVF-30
T-08-OVF-45
T-08-OVF-60
T-08-DRN-15
T-08-DRN-30
T-08-DRN-45
T-08-DRN-60
B. globigii
(CFU/100ml_)
3.27E+08
7.23E+07
ND
ND
ND
ND
5.00E+02
1.00E+03
3.00E+03
3.25E+03
5.00E+02
1.00E+02
1.00E+02
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
B. globigii
(log values)
8.51
7.86
ND
ND
ND
ND
2.70
3.00
3.48
3.51
2.70
2.00
2.00
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
Bromide
(mg/L)
29.97
29.80
5.05
0.04
0.04
0.04
0.04
0.04
0.05
0.05
0.05
0.06
0.04
0.04
0.04
0.05
29.83
16.57
8.87
5.18
0.07
0.06
0.05
0.04
ND-not detected
Although bromide was recovered in the overflow samples as expected, no B. globigii were detected
in these samples (T-08-OVF-15 through T-08-OVF-60). A post-test of water-rise from the gravel
pack on the overflow drain side (which is decontaminated with bleach and rinsed thoroughly with
de-chlorinated water after each test) indicated -300 ppm of residual chlorine. This indicates
chlorine build-up over time on that line at a high level, which resulted in these anomalies.
However, these values do not impact the main intrusion results or focus of the testing. No further
23
-------
corrective action was considered necessary or implemented as the testing was considered to be
complete after the 10th test run.
3.9 Test Run 9
The test run conditions for this test were as follows: leak pressure - 20 psi; soil media - sand;
selected contaminant tracer in soil-box media - bromide and B. globigii; and backflow sample
withdrawal method - gravity. Table 3-9 summarizes the results from Test Run 9.
Table 3-9 Test Run 9 Results
Sample ID
T-09-Stock-01
T-09-Stock-02
T-09-SMPL-BL
T-09-SMPL-AL
T-09-SMPL-0.1
T-09-SMPL-0.5
T-09-SMPL-1 .5
T-09-SMPL-2.0
T-09-SMPL-3.0
T-09-SMPL-4.0
T-09-SMPL-5.0
T-09-SMPL-10
T-09-SMPL-20
T-09-SMPL-30
T-09-SMPL-40
T-09-SMPL-50
T-09-SMPL-60
T-09-OVF-15
T-09-OVF-30
T-09-OVF-45
T-09-OVF-60
T-09-DRN-15
T-09-DRN-30
T-09-DRN-45
T-09-DRN-60
B. globigii
(CFU/100ml_)
6.27E+07
5.33E+07
ND
ND
ND
3.05E+03
1.07E+04
7.50E+03
6.87E+03
2.50E+03
1.33E+03
1.70E+02
55
15
40
20
10
5.00E+03
ND
ND
ND
ND
ND
ND
ND
B. globigii
(log values)
7.80
7.73
ND
ND
ND
3.48
4.03
3.88
3.84
3.40
3.12
2.23
1.74
1.18
1.60
1.30
1.00
3.70
ND
ND
ND
ND
ND
ND
ND
Bromide
(mg/L)
29.91
29.77
1.03
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.05
0.06
0.07
0.08
0.08
0.09
8.56
27.96
18.76
10.82
0.06
0.05
0.04
0.04
ND-not detected
Field measurements indicated that the gravity sample flow rate ranged from 188 mL/rnin (at five
minutes) to 146 mL/min (towards the end of the 60-minute sampling period). Similar to Test Run
8, although bromide was recovered in the overflow samples as expected, the non-detectable
concentrations for B. globigii (T-08-OVF-30 through T-08-OVF-60) were once again likely due
to chlorine build-up on the gravel-packed line (as reported under Test Run 8).
24
-------
3.10 Test Run 10
The test run conditions for this test were as follows: leak pressure - 20 psi; soil media - gravel;
selected contaminant tracer in soil-box media - bromide and E. coli; and backflow sample
withdrawal method - pumped. Table 3-10 summarizes the results from Test Run 10.
Table 3-10 Test Run 10 Results
Sample ID
T-10-Stock-01
T-10-Stock-02
T-10-SMPL-BL
T-10-SMPL-AL
T-10-SMPL-0.1
T-10-SMPL-0.5
T-10-SMPL-1.0
T-10-SMPL-1.5
T-10-SMPL-2.0
T-10-SMPL-2.5
T-10-SMPL-3.0
T-10-SMPL-3.5
T-10-SMPL-4.0
T-10-SMPL-4.5
T-10-SMPL-5.0
T-10-SMPL-10
T-10-SMPL-20
T-10-SMPL-30
T-10-SMPL-40
T-10-SMPL-50
T-10-SMPL-60
T-10-OVF-15
T-10-OVF-30
T-10-OVF-45
T-10-OVF-60
T-10-DRN-15
T-10-DRN-30
T-10-DRN-45
T-10-DRN-60
E. co//
(CFU/100ml_)
1.80E+06
2.40E+06
6.50E+05
1
ND
1.80E+02
41
86
1.10E+02
84
74
61
41
85
63
5.20E+02
1.00E+04
2.90E+04
4.60E+02
7.70E+04
8.10E+04
ND
ND
ND
ND
ND
ND
ND
ND
E. co// (log
values)
6.26
6.38
5.81
0.00
ND
2.26
1.61
1.93
2.04
1.92
1.87
1.79
1.61
1.93
1.80
2.72
4.00
4.46
2.66
4.89
4.91
ND
ND
ND
ND
ND
ND
ND
ND
Bromide
(mg/L)
29.78
29.79
27.51
0.04
0.04
0.04
0.04
0.04
0.04
0.05
0.05
0.05
0.06
0.06
0.06
0.13
0.79
1.85
3.10
4.62
0.04
23.65
21.64
13.11
13.46
0.04
0.04
0.04
0.04
ND-not detected
25
-------
Field measurements indicated that the pumped sample flow rate ranged from 822 mL/rnin (at five
minutes) to 795 mL/min (towards the end of the 60-minute sampling period). Similar to Test Runs
8 and 9, although bromide was recovered in the overflow samples as expected, the non-detectable
concentrations for E. coll (T-10-OVF-15 through T-08-OVF-60) were once again likely due to
chlorine build-up on the gravel-packed line (as reported under Test Runs 8 and 9).
26
-------
4.0 Analysis of Results
4.1 Intrusion Flow Rates and Pressures
Table 4-1 summarizes the flow rates obtained through both the sand and gravel media over the
first 5 minutes of backflow under both gravity and pumping. The test results indicate that the
median pumped flow rate was 3.6 times higher than the gravity rate for sand but only 1.2 times
higher for gravel.
Table 4-1 Observed 5-Minute Backflow Rates
Media
Sand
Sand
Gravel
Gravel
Backflow
Method
Gravity
Pumped
Gravity
Pumped
Min. Flow
Rate (ml/min)
130
740
646
725
Max. Flow
Rate (ml/min)
215
740
646
840
Median Flow
Rate (ml/min)
202
740
646
811
Under gravity backflow, the external head applied to the soil water equals the depth of the saturated
soil above the hole which was 24 inches (2 feet) (see footnote, page 2) in all experiments. Under
pumped flow, the additional suction pressure produced by the pump can be estimated by assuming
that the flow rate is limited by the orifice formed by the pipe crack. Thus, knowing the flow rates
for both gravity and pumped intrusion flow, and the external head for both conditions (2 feet), the
suction pressure produced under pumped fiow Pp can be expressed as follows:
where Qp is the flow rate under pumping and Qg is the flow rate under gravity. Using this equation
and the 5-minute flow rates from Table 4-1, the estimated median intrusion pressure for the
experiments using sand was 20 psi while for gravel it ranged from 1.3 to 2.1 psi.
The 20 psi suction pressure for the pumped sand experiment (there was only such experiment, Run
2) would never be observed in a real distribution system. In retrospect, it would have been better
to have adjusted the pump speed to draw around 1.3 times the gravity flow rate (about 260 ml/min)
to keep the suction pressure at a more reasonable 2 psi.
27
-------
4.2 Short Term Intrusion Behavior
Even though backflow sample through the pipe crack was collected and analyzed over a period of
60 minutes, only the first 5 minutes of data are presented here since low pressure transients rarely
(if ever) last beyond a few minutes (AWWARF 2004). Figure 4.1 presents results for sand media
while Figure 4.2 illustrates the same for the gravel media for the B. globigii runs at 40 psi leak
pressure (Runs 1- 4).
The spike concentration of B. globigii in both sand and gravel runs varied from 105 to 106 CFU/100
ml and bromide concentrations were between 27 and 30 mg/L. Also, because there are some
concerns about the reliability of the first microbial samples taken at 0.1 minutes for the gravel
experiments (Runs 3 and 4), they are not presented in Figure 4-2.
The data clearly show that the microbial and chemical tracers behave significantly different under
backflow conditions. While B. globigii enters the pipe in significant numbers, the bromide tracer
remains at essentially background level over the entire short-term duration of backflow. This
behavior is repeated in all of the other test runs, including those using E. coli, as well. One possible
explanation for this behavior is presented below.
During the leaking phase of the experiment, "clean" water enters the soil through the pipe crack
and completely displaces contaminated water within the cone of the vertical jet that forms above
the crack. In addition, there is likely to be some expansion of the media in this zone as it partially
fluidizes, which would apply more to the smaller sand particles than to the larger gravel. A
dissolved contaminant like bromide would be completely washed away as the original
contaminated pore water is replaced with clean water. However, for the suspended microbial
organisms, the sand or gravel acts as a filter media that prevents some fraction of the organisms
from being carried away from the soil above the pipe crack, most likely by interception. It is this
fraction that provides a reservoir of microbial particles in the vicinity above the crack that can then
be transported into the pipe during the backflow event. In addition, additional filtering can occur
in the opposite direction as backflow carries soil water back into the pipe.
28
-------
Intrusion Test, 40 PSI, Sand
umped B. globigii (log values)
cu
_2
as
_o
CuO
CD 2
Figure 4-1 Intrusion of Tracer into Pipe for Sand Media
Intrusion Test, 40 PSI, Gravel
•Gra
Pumped B. globigii (log values)
vity Gravel B. globigii (log values)
Pumped Bromide (mg/L
vity travel bromide
mg/L)
CO
0.4
- 0.2
2 3
Time (min)
Figure 4-2 Intrusion of Tracer into Pipe for Gravel Media
29
-------
4.3 Effect of Media and Backflow Method
Table 4-2 compares the log reduction in B. globigii from the concentration in the stock solution to
the average backflow concentration over the first five minutes for the 40 psi runs 1-4.
Table 4-2 Log Reductions of B. globigii
Media
Sand
Sand
Gravel
Gravel
Backflow
Method
Gravity
Pumped
Gravity
Pumped
Initial Tracer Concen.
Log (CFU/100 ml)
7.50
7.69
7.90
7.89
Avg. Backflow Concen.
Log(CFU/100 ml)
2.76
3.92
4.00
4.39
Avg. Log
Reduction
4.74
3.77
3.89
3.49
These results show that fewer organisms (i.e. higher log reductions) relative to the initial level of
soil contamination enter the pipe for sand than for gravel, and that the same is true for gravity as
compared to pumped backflow. This behavior is consistent with the higher filtration efficiency of
sand versus gravel and the lower pore water velocity experienced under gravity flow as compared
to pumped flow.
One can also examine the contribution that the leak period makes in dispersing and diluting
organisms in the vicinity of the pipe crack resulting in lower microbial concentrations in this zone
as compared to the initial tracer concentration. The "BL" sample is taken in the same manner as
the gravity backflow samples are, except before the pipe is operated in leak mode. For the sand
media of Run 1, the initial tracer concentration was 7.5 logs while the BL sample was only 5.54
logs. This 2.8 log reduction of organisms was due just to the adherence of organisms on the sand
particles plus any filtering provided by the sand in the vicinity of the pipe crack opening. When
this process was repeated after an hour of allowing 40 psi clean water to leak out into the sand, the
total reduction in organism concentrations entering the pipe rose to 4.74. Thus an additional 2 logs
reduction can be attributed to the dispersal/dilution effect of pressurized water leaking out of the
pipe. The gravel media used in Run 3 saw an initial log removal of 1.4 before the leak was started
and an additional 2.5 log reduction resulting from the leak phase of the experiment.
4.4 B. globigii versus E. coli
Figure 4-3 compares the intrusion behavior of the two different microbial contaminants used in
this study, B. globigii and E. coli, under similar experimental conditions of gravel media, 40 psi
leak pressure, and pumped backflow (Runs 4 and 6). Because the stock solutions used to initially
saturate the gravel contained different concentrations of B. globigii (7.7e+07 CFU/100 ml) andE1.
coli (4.7e+05 CFU/100 ml), the data plotted in the figure are the ratios of the backflow sample
concentration to the stock solution concentration expressed as a log reduction. When plotted in
30
-------
this fashion the behavior of the two microbial tracers appears to be quite similar through the first
3 minutes of backflow. The average log reduction over the full 5 minute period was 3.5 for the B.
globigii and 3.2 for E. coli. Claghorn and Lange (2000) had previously reported that the two
organisms yielded statistically similar tracer curves when injected into porous media.
Intrusion Test, 40 PSI, Gravel, E. coli and B. globigii
o
•Pumped B. glob
•Pumped E. coli (
ii (log values)
og values)
2 3
Time (min)
Figure 4-3 Comparison of Microbial Tracers
4.5 Effect of Leak Pressure
Figure 4-4 plots the average log reduction of the microbial concentration (relative to the stock
solution) over the initial five minutes of backflow against the line pressure maintained during the
leak phase of the experiment for two sets of runs. One set consists of runs made with sand media,
gravity backflow and B. globigii (Runs 1, 8, and 9) and the second is with gravel using pumped
backflow and E. coli (Runs 6, 7, and 10). The results for sand show a clear and expected trend of
higher reductions (lower relative concentrations) of organisms with higher leak pressures due to
greater leakage volumes and expansion of the sand around the pipe crack. The results for gravel
do not follow this trend. The effect of pumping may have obscured any differences caused by
pressure-related leak flow rates in the dispersal of E. coli during the leak phase of the experiment.
31
-------
5 -
4 H
10
20
30 40
Leak Pressure, psi
50
60
•Sand, Gravity
•Gravel, Pumped
Figure 4-4 Effect of Leak Pressure on Intrusion of Microbial Tracer
4.6 Longer Term Intrusion Behavior
Low pressure intrusion conditions are usually caused by short term events such as pump shut
downs, valve closures, or temporary power outages. Longer periods of low pressure might occur
in systems that operate intermittently or when pipes are taken off-line to repair leaks. Figure 4-5
shows the intrusion of soil tracer over the full 60-minute sampling period for Runs 1 and 3. Run
1 was for sand and Run 3 for gravel, with both runs using a leak pressure of 40 psi, gravity
backflow, and B. globigii as the microbial tracer. The tracer stock solution had similar levels of
both B. globigii and bromide. The backflow rates ranged from 215 to 160 ml/min for the sand run
and from 646 to 583 ml/min for the gravel run.
As expected, the looser gravel media with the higher flow rate provided less filtration efficiency
allowing higher numbers of organisms to enter the pipe over time. The filtration efficiency of the
sand media appears to improve over time, as the flow rate decreases due to less available head,
reaching a point where only small numbers of organisms break through into the pipe. For the
dissolved bromide tracer, at some point in time the higher backflow rate and intrusion volume for
gravel begins to draw water from beyond the immediate area of the crack that had tracer completely
displaced with clean water during the leak phase, resulting in increasing bromide concentrations
over time (but still well below the initial 30 mg/L in the tracer solution). This behavior is not
32
-------
evident for the sand media, since at the smaller back flow rate it continues to draw from the clean
leaked water over the duration of the run.
0.5
Intrusion Test, 40 PSI, Sand
20
30
Time (min)
40
50
Figure 4-5 Longer Term Intrusion under Gravity Backflow
60
33
-------
5.0 Conclusions and Recommendations
This project measured the concentrations of a biological tracer (B. globigii and E. coli) and an inert
chemical tracer (bromide) added to soil media placed above a pipe with a small opening in it under
a period of pressurized flow through the pipe followed by induced backflow from the soil into the
pipe. Various experimental conditions were tested - sand versus gravel media; pumped versus
gravity backflow; E. coli versus B. globigii microorganisms; and several different leak pressures.
These conditions represent a worst case scenario since the experiments begin with fully
contaminated soil media right next to the pipe crack, only allow leakage with de-chlorinated water
for a period of one hour, and, since pipe flow is shut off during backflow, ignore any dilution of
the intruded soil water with clean water flowing through the pipe. Nevertheless, they help reveal
the potential behavior of how pipe leaks influence possible contamination during an intrusion
event.
Based on the results of the experiments reported here, the following conclusions can be drawn for
short term backflow events (less than 5 minutes):
6. There will be no intrusion of chemical tracer into the pipe because of tracer washout in the
vicinity of the crack during normal operation of the leaking pipe.
7. There will however, be intrusion of microbial tracer since only partial washout occurs
because of filtering by the soil media
8. Sand media provides higher filter efficiency than does gravel media in the direction of
backflow resulting in lower numbers of microorganisms entering the pipe during backflow.
9. The lower the backflow flow rate the lower the average concentration of microorganisms
in the back flowed soil water.
10. Higher leak pressures result in lower microbial concentrations in backflow from sand
media. The same could not be shown for gravel.
As the backflow period was extended up to 60 minutes, the microbial tracer level tended to
decrease for sand and level out for gravel. The chemical tracer continued to stay at background
level for sand but started to rise for gravel.
The following recommendations are made for future intrusion studies:
1. Ways should be found to make the experiments easier to set up and be more reproducible,
perhaps by using a smaller soil box.
2. Experiments should be performed using chlorinated tap water during the leak phase to see
the effect that disinfection would have on microbial concentrations in the backflow into the
pipe.
3. Experiments should be run under several leak/backflow cycles to be more representative
of actual field conditions.
34
-------
4. The backflow pumping rate for sand media should be adjusted so that more realistic suction
pressures are obtained.
35
-------
6.0 References
American Water Works Association Research Foundation 2004. Verification and Control of
Pressure Transients and Intrusion in Distribution Systems, Report 9100 IF, AWWA Research
Foundation, Denver, CO.
Claghorn, J. and Lange, C.R. 2000. "The Efficacy of Employing Bacillus Globigii as a Particulate
Tracer in Aquatic Systems", Proceedings of the Water Environment Federation, WEFTEC 2000,
Water Environment Federation, Alexandria, VA. pp. 468-476(9).
CB&I2014. Quality Assurance Project Plan. Category IVMeasurement Project, Back-Siphoning
of Soil Water Through Pipe Cracks. Revision 1, February 2014.
U.S. Environmental Protection Agency 2006. The Potential for Health Risks from Intrusion of
Contaminants into the Distribution System from Pressure Transients, Water Distribution System
Issue Paper, Office of Water (4601M), Office of Ground Water and Drinking, Washington DC.
36
------- |