TE_F-14_HAAF_Unidirectional_Distribution_Flushing_Plan.pdf

DEPARTMENT OF THE ARMY

PUBLIC HEALTH COMMAND – ATLANTIC (PROVISIONAL)

4411 LLEWELLYN AVE

FORT GEORGE G. MEADE, MARYLAND 20755-5225

MCHB-RH-EH 20 NOV 2015

MEMORANDUM FOR Directorate of Public Works (IMAP-PWE/Mr. Stanley Thomas), U.S. Army Garrison Hunter Army Airfield, 525 Leonard Neal Street, Savannah, GA 31409

SUBJECT: Water Distribution System Unidirectional Flushing Plan, Project No. S.0033519-15, Hunter Army Airfield, 4 - 8 May 2015

1. Subject document is enclosed. Please contact us if we can be of further assistance.

2. The point of contact for this evaluation is 1LT(P) Luke Grant, Environmental Health and Engineering Division (EHED), at commercial (301) 677-3698 / email luke.a.grant.mil@mail.mil

FOR THE DIRECTOR:

Encl

James.Maddox Typewritten Text

TE F-14

Distribution limited to U.S. government agencies only;

protection of privileged information evaluating another command; November 2011. Requests for this document must be referred to Directorate of Public Works (SMDC-RDTC-TEK- W/ Mr. Anthony Hoover), U.S. Army Kwajalein Atoll, P.O. Box

26, APO, AP 96555-2627.

WATER DISTRIBUTION SYSTEM

UNIDIRECTIONAL FLUSHING PLAN

PROJECT NO. S.0033519-15

HUNTER ARMY AIRFIELD, SAVANNAH GA

4-8 MAY 2015

Distribution authorized to U.S. Government Agencies only; protection of privileged information evaluating another command: February 2015.

Requests for this document must be referred to Hunter Army Airfield, Directorate of Public Works (IMAP-PWE/Mr. Stanley Thomas), 525

Leonard Neal Street, Savannah, GA 31409

ARIMS designation: 40-5f1

DEPARTMENT OF THE ARMY

PUBLIC HEALTH COMMAND – ATLANTIC (PROVISIONAL)

4411 LLEWELLYN AVE

FORT GEORGE G. MEADE, MARYLAND 20755-5225

MCHB-RN-EH

EXECUTIVE SUMMARY

UNIDIRECTIONAL FLUSHING PLAN

PROJECT NO. S.0033519-15

HUNTER ARMY AIRFIELD, SAVANNAH GA

4-8 MAY 2015

1. PURPOSE. The purpose of this project is to provide a Unidirectional Flushing (UDF) Plan for the Hunter Army Airfield water system. The plan is designed to bring water through the system in a controlled fashion at velocities sufficient to provide a scouring action within the distribution piping. This plan complies with Army Regulation (AR) 420- 1 and the Unified Facilities Criteria (UFC) 3-230-02, which require that a systemic flushing plan of water distribution lines be developed and that the systemic flushing plan of water distribution lines provide adequate scouring velocities.

2. DISCUSSION. Flushing will start at the water supply source and continue to the periphery of the distribution system by using isolation valves and exercising hydrants in an organized and sequential manner. The Hunter Army Airfield (HAAF) flushing plan was developed using a hydraulic model WaterGEMs®. Implementing the UDF program will result in water quality improvements, more efficient use of operator time, and more efficient water use compared to other flushing methods. The exact flushing sequences can be found in Appendix E - H. This flushing plan should be considered a living document and should be updated as necessary or as the water distribution system changes.

3. CONCLUSION.

a. A routine UDF program is part of a good overall water system maintenance program. The flushing approach, velocity, frequency, and duration are specific to HAAF system's water quality concerns, hydraulic conditions, source water treatment, and system design.

b. A hydraulic model was used to design the UDF plan for HAAF. The goal of hydraulic modeling is to create a useful realistic model from which to gain insight into the behavior of a given water distribution system.

WaterGEMS is a registered trademark of Bentley Systems, Inc. Watertown, CT. Use of trademarked name(s) does not imply endorsement by the U.S. Army, but is intended only to assist in identification of a specific product.

EXSUM, Unidirectional Flushing Plan, Project No. S.0033519-15, 4-8 May 15

ES-2

c. Appendices E – H outline the flushing procedures for HAAF indicating locations for both isolation valve operation and flush hydrants.

d. The flushing times shown within the flushing procedures in Appendices E – H are estimated times based on the hydraulic model to clear the entire water volume within the piping for each flushing event. The actual time required to flush each hydrant may be more depending on the hydrant water quality.

e. Data collection is imperative to understanding the benefits, costs, and secondary impacts due to flushing. If adequate data are not collected, program refinement and a cost/benefit analysis cannot be conducted. Appendix C provides supplemental information in evaluating flushing effectiveness and each flushing event comes with an operation record keeping page.

f. During the first implementation of the flushing plan, several problems not anticipated on the hydraulic model could occur. An isolation valve’s operational status may be different from what was anticipated or broken valves may be discovered.

g. This UDF was designed to ensure that no customer will be without water during the implementation of the flushing plan. However, throughout the implementation of the UDF, flushing crews may realize that some distribution system valves that were not assessed in the flushing plan may be in the closed position. Nationwide distribution system survey data show that 10-20% of isolation valves within water distribution systems are in the wrong position and up to 40% of isolation valves are inoperable.

h. A valve exercise program is part of a normal operation and maintenance program and will aid in determining isolation valve status and operability on HAAF.

i. Future modification to the distribution system, adding or removing water mains, can affect the flushing procedure and effectiveness. Relevant portions of the system may need to be appended to continually achieve effective flushing,

4. RECOMMENDATIONS.

a. Implement a flushing plan as laid out in Appendices E – H on an annual basis or when water quality issues arise.

b. Each hydrant should be flushed using the time specified within the flushing procedures in Appendices E – H at a minimum. If the water from a hydrant is not clear, the hydrant flush time may need to be increased.

c. Collect water quality data and take notes after each flush to gain insight and improve future water system flushes.

EXSUM, Unidirectional Flushing Plan, Project No. S.0033519-15, 4-8 May 15

ES-3

d. To minimize the risk of low pressure events (less than 20 psi) from malfunctioning valves or unintentionally closed isolation valves when performing this UDF plan, it is important to measure upstream (or nearby) hydrant pressures, especially after closing isolation valves.

e. Before implementing the UDF plan, time should be spent identifying, verifying operability, and assessing the condition of the targeted isolation valves and hydrants in each flushing section.

f. If specific isolation valves cannot be found, the flushing should occur and the flushing sequence should continue to be followed. Flushing with some of the isolation valves closed and abiding by the flushing sequence is still highly beneficial.

g. Purchase needed equipment as identified in Appendix C to monitor pressure and water quality data.

h. In a system that has not previously performed unidirectional flushing, maintenance crews should be available for repair activities.

i. An annual valve exercise program is recommended for HAAF.

j. Conduct an inventory of all isolation valves on HAAF.

k. Update GIS data. Create a single, current, accurate set of shape files in accordance with the 2013 survey of the water system.

l. Notify customers when flushing program will begin and end as described in Appendix D.

m. Conduct a criticality/reliability analysis once HAAF obtains an accurate inventory of isolation valves.

Unidirectional Flushing Plan, Project No. S.0033519-15, 4-8 May 15 i

TABLE OF CONTENTS

PARAGRAPH PAGE

1. AUTHORITY

2. REFERENCES

3. PURPOSE

4. GENERAL

a. Survey Team

b. Personnel Contacted

5. WATER SYSTEM PROFILE

a. Location

b. Water System Characteristics and Description

6. REGULATORY REQUIREMENTS AND GUIDANCE

a. AR 420-1

b. UFC 3-230-02

c. Technical Bulletin Medical (TB MED) 576

7. DISCUSSION

a. Types of Flushing

b. Flushing Plan Development

c. Hydraulic Model Development and Calibration

d. Accuracy

e. Sources of Error/Factors of Note

f. General Flushing Plan Guidelines

g. Flushing Implementation

8. CONCLUSION

9. RECOMMENDATIONS

ii

APPENDICES

A. REFERENCES ...................................................................................................... A-1 B. FLUSHING GUIDELINES ...................................................................................... B-1 C. SUPPLEMENTARY INFORMATION ..................................................................... C-1 D. FLUSHING NOTIFICATION LETTER .................................................................... D-1 E. SOUTHWEST AREA SECTION ............................................................................ E-1 F. CENTRAL AREA SECTION .................................................................................. F-1 G. EAST AREA SECTION .......................................................................................... G-1 H. NORTH AREA SECTION ...................................................................................... H-1

UNIDIRECTIONAL FLUSHING PLAN

PROJECT NO. S.0033519-15

HUNTER ARMY AIRFIELD, SAVANNAH GA

04-08 MAY 2015

1. AUTHORITY. Electronic communication between Mr. Stanley Thomas, Water Program Manager, Hunter Army Airfield and Mr. John Brokaw, Drinking Water and Sanitation Program Section Chief, U.S. Army Institute of Public Health (AIPH) providing military interdepartmental purchase request (MIPR) to conduct unidirectional flushing plan on Hunter Army Airfield (HAAF), 13 March 2015.

2. REFERENCES. A list of references is provided in Appendix A.

3. PURPOSE. The purpose of this project is to provide a Unidirectional Flushing (UDF) Plan for the HAAF water system. The plan is designed to bring water through the system in a controlled fashion at velocities sufficient to provide a scouring action within the distribution piping. This plan complies with Army Regulation (AR) 420-1 (reference

1) and the Unified Facilities Criteria (UFC) 3-230-02 (reference 2), which require that a systemic flushing plan of water distribution lines be developed and that the systemic flushing plan of water distribution lines provide adequate scouring velocities.

4. GENERAL.

a. Survey Team.

(1) 1LT(P) Luke Grant, Environmental Science and Engineering Officer, Public Health

Command Region North (USAPHCR-North), Team Leader

(2) Mr. Jerry Manint, P.E., Environmental Engineer, USAPHCR-North

(3) Mr. Matthew Waterbury, Hydrologist, USAIPH

b. Personnel Contacted.

(1) Mr. Stanley Thomas, Water Program Manager, HAAF, Directorate of Public Works

(DPW)

(2) Mr. Nate Williams, Water Compliance Manager, HAAF, DPW

(3) Mr. Robert Stewart, Storm Water Specialist, HAAF, DPW

5. WATER SYSTEM PROFILE.

a. Location HAAF is located in Savannah, Georgia (GA). It covers 5,400 acres with approximately 5,000 soldiers on Post. It is home for units of the 3rd Infantry Division (Mechanized) headquartered at Fort Stewart, which lies approximately 25 miles to the southwest.

b. Water System Characteristics and Description.

(1) During the site visit, HAAF was using up to five ground wells to supply their distribution system. Water treatment occurred at the point of supply.

(2) Average water use based upon the previous year’s data was approximately 414 gallons per minute (gpm) or 597,000 gallons per day (gpd).

(3) HAAF has four primary water tanks on the installation. These four tanks provide the necessary water storage for the installation and are distributed throughout the distribution system to ensure adequate pressure throughout the installation during normal and peak demands.

(4) Water distribution mains were 12 inches in diameter from each of the four water tanks. Water pipes were also 10 inch, 8 inch, and 6 inch diameters throughout the distribution system. The 6 inch diameter is the smallest pipe on which a hydrant should be installed.

6. REGULATORY REQUIREMENTS AND GUIDANCE.

a. AR 420-1. Also known as the Army Facilities Management document states “Water distribution systems will be flushed as necessary to remove accumulated debris and that a systemic flushing plan of water distribution lines must provide adequate scouring velocities (reference 1).”

b. UFC 3-230-02. Formally known as Technical Manual 5-660, section 7.3.2.2 states that "Matter deposited in the distribution system over time can often be removed by periodically flushing the system. The frequency varies for different waters and pipe materials, but annual flushing is usually enough. Dead-end pipes should be flushed and disinfected at least once a year. Flushing and disinfection are required whenever mains are opened for repair (reference 2)."

c. Technical Bulletin Medical (TB MED) 576. Paragraph 4-3.b states that "Existing water mains should be flushed as a matter of policy, at least annually (reference 3)."

7. DISCUSSION.

a. Types of Flushing. Flushing is one of the most powerful tools available to water utilities for maintaining distribution system water quality. The three basic approaches to distribution system flushing are conventional, continuous blow off, and unidirectional.

Each technique can be implemented on a comprehensive system-wide basis or on a narrower “spot” basis. Whether to flush an entire system or simply one section or troubling area depends on the system configuration and water quality goals.

TB MED 576 along with other Army guidance recommends that comprehensive flushing be conducted at least annually with the option to flush areas with acute problems more frequently.

(1) Conventional Flushing. This flushing option is defined as the opening of fire hydrants in a specific area of the distribution system until water quality criteria (color, turbidity levels, disinfectant residuals, etc.) are met. Conventional flushing is the most common flushing method used by water utilities. This process may or may not be sequential (working from the water source out to the periphery of the system), but by definition it does not utilize the system’s valves to isolate the mains as part of the process. When not performed sequentially, conventional flushing can result in moving sediment and turbid water from one part of the system to another, rather than systematically removing it from the system. Also, by not utilizing valves to isolate pipe sections, water usually flows to the hydrants from several different mains. The resulting velocities in each main end up less than the suggested 2.5 feet per second (fps) required for effective flushing. Often conventional flushing results in the following water quality problems: increased customer complaints during and immediately after implementation, a considerable waste of water, short-lived water quality benefits, and the potential for increased coliform bacteria occurrences following flushing.

(2) Continuous Blow Off. Continuous blow off or bleeding is often used for oversized or dead-end mains. This strategy forces low-velocity flow through a small portion of the system to remove stale water, and raise the disinfectant residual. This flushing method typically results in velocities around 1 fps and does not provide the scouring action required to remove accumulated sediments. Continuous blow off wastes large amounts of water without completely removing sediment and does not provide a permanent solution. It is not recommended for management of a system’s dead ends or areas with acute problems.

(3) Unidirectional Flushing.

(a) Unidirectional flushing is designed to bring the water through the system in a controlled fashion at velocities sufficient to provide a scouring action within the distribution piping. The technique consists of isolating a particular pipe section or loop (typically through closing appropriate valves) and exercising the hydrants in an organized, sequential manner. This process moves fresh water from the source to the periphery of the system, and always moves from large-diameter pipes into smaller diameter pipes. Unidirectional flushing is most often associated with establishing velocities of greater than 2.5 fps (references 4-5) within each pipe segment being flushed. This velocity promotes a scouring action within the pipe that helps remove sediments, biofilms, and loose deposits. Unidirectional flushing in a sequential manner helps ensure that each pipe section is completely flushed.

(b) Unidirectional flushing programs have a variety of water quality and hydraulic benefits which are listed below (reference 4).

(1) Reduction of customer complaints regarding water quality.

(2) Reduces sediment deposits, which in turn reduces chlorine demand, reduces potential media for bacterial growth, and masking of disinfectant effectiveness.

(3) Reduces turbidity, color, total dissolved solids, odor, and iron.

(4) Removes mineral deposits and reduces corrosion, which can increase the service life of piping.

(5) Reduces maintenance and replacement costs of hydrants by keeping them operating more effectively through annual exercising.

(6) Consumes less water than conventional flushing. Savings are greater than 40% in some estimates.

(7) Flushing procedures are standardized providing a uniform basis of comparison between different flushing events.

(8) Isolating segments of the water system with valves creates higher flows and increases scouring velocities in water mains.

b. Flushing Plan Development.

(1) A UDF Plan involves choosing water main loops within the distribution system and determining a sequence of opening and closing fire hydrants and valves to move water from the system’s source to the periphery of the system. The overall flushing strategy of this flushing plan involves delineating the distribution system at HAAF into manageable sections. This flushing plan is separated into four flushing sections or areas which correspond to the tank locations within the distribution system. These areas are delineated in appendices E – H. Each section has a number of flushing events that should be conducted in the order presented in the flushing plan to ensure a high scouring velocity that moves water in the correct direction and prevents dirty water from entering mains that have been previously flushed.

(2) Each flushing section should take between 3-6 workdays to complete all flushing events. This is reasonable considering American Water Works Association (AWWA) have found that a flushing team can usually complete several miles of pipe per day (reference 5). Time assumptions for opening and closing isolation valves and hydrants were built into the time assumption to complete each section. Table 1 indicates the section location and time allotted to complete each flushing section. Provided the hours listed and considering the length and structure of a typical work day, a flushing team will take approximately 4 weeks to complete the flushing of all sections. The Southwest area will take approximately 3 days, the Central area will take approx. 6 days, the East area will take approx. 4 days, and the North will take approx. 5 days.

(3) Ideally, each flushing sections be conducted in the order presented in the flushing plan and in Table 1 during an annual comprehensive flushing event. However, if there are water quality issues associated with certain flushing sections, these sections can be flushed individually and more often than annually.

(4) It is possible that once the flushing plan is initiated, water system personnel may find that certain sections may take more or less time than noted in Table 1.

Because this is a living document, changes and improvements can be made to the flushing plan which may shorten or lengthen the time needed to adequately flush each section.

(5) It is recommended that time is spent locating isolation valves and hydrants and that their operability be verified prior to starting a flushing section. This familiarization will optimize the time spent executing the flushing plan.

Table 1. Delineated Flushing Sections

Order* HAAF Area Descriptor Estimated Time to

Complete (Hours)**

Locaiton on Distribution

System Map

1 Southwest 12

2 Central 32

3 East 21

4 North 27

*Sequence Order to be followed during annual comprehensive flushing event only. If there are water quality issues associated with these sections, these sections can be flushed individually and more often than annually.

** This includes total flush time with assumed hydrant and isolation valve open/close time. It does not include the time necessary to find isolation valves and fire hydrants.

c. Hydraulic Model Development and Calibration.

(1) To determine the accuracy and effectiveness of the valve and hydrant flushing sequence, a hydraulic model was developed. Hydraulic model development involves knowledge about the entire water distribution system. Components such as water isolation valves, fire hydrants, and piping must be correctly georeferenced. In addition, information must be known about the pipe material, diameter, length, and ground elevation within the water distribution system. This information is contained in GIS files and is transferred to the hydraulic model, WaterGEMs®.

(2) To effectively simulate a water distribution system, it is important to collect in-field pressure and flow data during normal operation. This information is used to gauge how well the model reflects the actual performance of the water distribution system and is also used to adjust or calibrate the model as needed to fit the real-time field data.

Pressure and flow field data were collected by placing pressure data loggers on hydrants throughout the water distribution system and taking flow measurements using a pitot diffuser when flushing hydrants.

(3) Calibration is an iterative process of adjusting parameters within the WaterGEMS software so that the model accurately reflects values found during the data collection stage. The purpose of calibration is to establish model credibility, establish a predictive tool, gain knowledge of the current system operation and performance, and discover errors or unknowns in the distribution system from inconsistencies found during simulation.

(4) Calibration is complicated by the fact that some data are known and unchanging, some are variable over time, and some are assumed. Generally, modeling parameters such as C-factors and domestic water demand allocations at demand nodes are modified to adjust the hydraulic model so that it accurately reflects the conditions when the field data was collected. While infrequently changed, components such as ground elevation or isolation valve status can be changed to accurately represent the observed conditions in the simulated model if conditions warrant. Calibration parameters are adjusted both manually and with the assistance of genetic algorithms to find the best possible realistic solution.

WaterGEMS is a registered trademark of Bentley Systems, Inc. Watertown, CT. Use of trademarked name(s) does not imply endorsement by the U.S. Army, but is intended only to assist in identification of a specific product.

d. Accuracy.

(1) The goal of hydraulic modeling is to create a useful realistic model from which to gain insight into the behavior of a given water distribution system. Currently, no specific accuracy requirements exist. However, many modelers agree that the level of accuracy should depend on the site-specific conditions as well as the model use.

(2) Several references do give guidance on hydraulic model accuracy. A model should reproduce actual discharge from fire hydrants to within 10-20 percent of observed flow to be considered calibrated when using it to develop a flushing plan (reference 6-8). More stringent criteria compare observed and simulated pressure as an accuracy indicator when developing models. One such criterion dictates that the model may be considered calibrated if observed and simulated hydraulic grade lines (HGLs) are within 10-20 ft of one another (references 7 and 8).

(3) For the development of the HAAF model, all fire hydrant flows and HGLs simulated were within the accuracy criteria when compared to observed conditions after calibration.

e. Sources of Error/Factors of Note. As mentioned above, the primary objective of a model simulation is to reproduce the behavior of a real system and its spatial and dynamic characteristics in a useful way. To accomplish this goal, data are supplied that depict the physical characteristics of the system, the loads placed on the systems, and the boundary conditions in effect. Even if all of the data gathered describing the model match the real system exactly, it is unlikely that the pressures and flows computed by the hydraulic or simulation model will absolutely agree with the observed pressures and flows found in the distribution system. There are significant mathematical assumptions employed by the simulation software to make the simulation computationally tractable, yet allow the simulated results to be meaningful and useful. Thus, modeling is essentially, a balance between both the physical and simulated reality and the effort necessary to make the two agree. Performing annual or routine calibrations to the hydraulic model will aid in reducing any errors or changes that may occur over time within the distribution system. This involves updating water demands, importing new pipes or elements that may have been installed into the hydraulic model, updating and including all piping material, and comparing simulated and actual pressure and flow data to achieve the desired degree of accuracy. Below are several sources of error that may need to be explored further.

(1) GIS Data. Sources of error can be present in the components found in the GIS data such as piping diameters, hydrant locations, isolation valve locations, and elevation data. There were several instances in which fire hydrants were shown to be in one location in the GIS data, however upon searching for them during the site visit it was found that they were located up to 40 meters away on the same water line. In the GIS data, there were many instances of pipes that were abandoned and were no longer functioning in the actual distribution system. Accuracy of a hydraulic model is directly representative of the accuracy of the components embedded in the GIS data.

Therefore, some error may exist in the hydraulic model due to inaccuracies within the GIS data.

(2) Isolation Valves. Every effort was made to develop a UDF plan that would ensure velocity, pressure, and flow direction constraints were met. However, velocity, pressure, and flow direction were heavily influenced by the position and number of isolation valves shown in the GIS data. If these isolation valves do not exist, it may create flushing events that do not meet their intended goals. Therefore, every effort should be made to find and ensure the isolation valves noted in this plan exist and are operational. If they do not, it should be noted and the flushing plan should be adjusted accordingly. During a return site visit, an attempt was made to locate all isolation valves to be used in the UDF plan. The majority of valves were located, but many could not be found that were located in the GIS. The most common reason was that grass or sediment was likely covering them. In the Southwest area approx. 25% of valves could not be located. In the Central area approx. 45% of valves could not be located. In the East area approx. 40% of valves could not be located. In the West area approx. 40% of valves could not be located.

(3) Isolation Valve Status. It is possible there are partially closed or closed valves that are inhibiting flow within the HAAF water system. It has been estimated in nationwide surveys that as high as 20% of the isolation valves found in a typical distribution system are in the wrong position and up to 40% are inoperable (references 9-11). Partially closed or closed valves can have localized effects on the distribution system by inhibiting flow. Unknown partially closed or closed valves could cut off water service to some customers during the implementation of this flushing plan. The best way to determine if there are partially closed or closed isolation valves in a water system and to gain insight into their functionality is to perform annual isolation valve exercises. An annual valve exercise program is recommended as part of a normal operation and maintenance program according to the Unified Facilities Criteria for Department of Defense (DoD) installations (reference 2). While, implementing this UDF plan will help exercise some isolation valves, a true valve exercise program will aid in obtaining an accurate inventory of the isolation valves located in the HAAF water system.

(4) Criticality/Reliability. Once HAAF obtains an accurate inventory through a valve exercise program, updates should be made to their GIS system. HAAF should then consider performing a criticality / reliability analysis to determine if the current operable inventory of isolation valves can adequately isolate the distribution system during an emergency event based on industry design standards.

(4) Water Demand.

(a) Monthly water usage records were provided by HAAF DPW over the calendar year 2014 and 2015. These monthly figures were used to estimate the average daily demand. The average water demand was found to be 414 gpm.

(b) Diurnal demand is the variation in water demand over a 24 hour period. Water demand diurnal curves for utilities and installations can vary considerably over the course of a day. Therefore, when conducting hydrant flow test studies for model calibration, it is important to know the total system demand when tests are conducted.

Unfortunately, during the site visit on HAAF, it was not possible to know how much total consumption the system was consuming at any specific time. According to DPW personnel, only daily usage is recorded and tracked. Therefore, for calibration purposes, it was assumed that hydrant flow tests conducted during the site visit occurred during a demand that was greater than the average daily demand. Standard peaking factors used by industry to convert average daily demand to peak hourly demand range from 3.0 – 6.0 (reference 6). However, given that higher peaking demands tend to be in areas of high residential populations, a lower peaking factor of

2.0 was used to estimate the total water usage on HAAF at the time of hydrant flow tests. The estimate of water use may be a source of error in the model calibration.

(c) There were eight water meters on HAAF that recorded a statistically significant amount of water use. These eight meters accounted for approximately 11% of all water use. The remaining HAAF water demands were allocated to nodes based on the building density in the area of nodes. The demand allocation of water use may be a source of error in the model calibration although it has been found that modest errors in demand loading within the hydraulic model have very little impact on hydraulic modeling results (references 6 and 7).

f. General Flushing Plan Guidelines. Appendix B provides additional flushing guidelines.

(1) Flush water system at least annually using the flushing sections and events located in Appendix E - H.

(2) Annual flushing should take place during the spring or fall (March through May, or September through November) when water consumption is likely to be at a minimum. If flushing during the summer months, ensure the irrigation systems are temporally shut off.

(3) Consider flushing during the late night hours or weekends to avoid customer complaints and to utilize extra water volume. Appendix D provides an example notification letter to distribute to customers before flushing.

(4) Even though flushing times are provided in the flushing events, flushing should continue until the water flows clear, or until certain criteria such as turbidity or chlorine residual goals are met.

(5) It is recommended that flushing records be maintained. Items to document include: times, dates, duration, flow rate (to ensure 2.5 fps velocity), pressure, and water quality monitoring such as chlorine residual, pH, and turbidity. Each flushing event is accompanied by a log sheet that should be filled out for each flushing event in order to track the effectiveness of the plan over time.

(6) During implementation of the flushing plan, several problems not anticipated on the hydraulic model could occur. In most cases, a small change to the sequence will rectify the problem and result in effective flushing. Future modification to the distribution system, adding or removing water mains, can also affect the flushing procedure. The relevant portions of the system may need to be appended to continually achieve effective flushing.

(6) This UDF was designed to ensure that no customer will be without water during the implementation of the flushing plan. However, throughout the implementation of the UDF, flushing crews may realize that some customers may be without water. Previously closed isolation may be valves that were not assessed in the flushing plan and which may be in the closed position. Often this occurs due to past repairs where the isolation valve is left in the closed position. Additionally, many isolation valves and pipes are incorrectly mapped within GIS data and may not accurately field conditions. Therefore, it is important to take upstream hydrant pressure readings when flushing hydrants especially after closing isolation valves and be attuned and ready to quickly return an isolated water line to service until more investigation can be done to determine the cause of the interrupted water supply.

(7) Before implementing a particular UDF sequence, time should be spent identifying, verifying operability, and assessing the condition of the targeted isolation valves and hydrants in each flushing section.

(8) In a system that has not previously performed unidirectional flushing, maintenance crews should be available for repair activities as valve status may be different from what was anticipated, broken valves may be discovered, hydrants may have marginal discharge or fail during flushing, and weak pipes can fail. Therefore, maintenance crews should be available for repair activities.

(10) Documenting the success or failure of a flushing event is an important way to determine if adjustment of the flushing plan is required. It is possible that all anticipated benefits from the procedure will not be realized immediately and experimentation with velocities, duration, and frequency will be required. To conduct an evaluation, it will be necessary to carefully document the costs (time, labor, equipment, etc.), potential savings, benefits, and problems, as well as the secondary effects that can result from the flushing program. The ability to evaluate flushing effectiveness depends on proper documentation while performing the flushing program. The last page of each flushing report (See Figure 3) contains an example data collection spreadsheet for use during a flushing event. This sheet or a similar sheet should be kept for record keeping purposes and should serve as a baseline to improve future flushing attempts. Appendix C provides supplemental information of the evaluation of flushing effectiveness.

g. Flushing Implementation. Appendix E – H contains the flushing procedures for HAAF, including isolation valve and hydrant locations. Below is a brief introduction and outline on how to use the flushing plan.

(1) Each flushing event has an associated report that should be used by operations when implementing this flushing plan. An example report is shown in Figures 1-3. The report for each flushing event contains three parts: summary page (see Figure 1);

system diagram page(s) (see Figure 2); operator report page (see Figure 3)

(2) Figure 1 details the distribution system section where the flushing will be conducted, the sequence order of the flush, the fire hydrant identifier that will be flushed, the minimum time necessary to flush the hydrant, and the isolation valves to be closed. The element labels for both the isolation valves and flush hydrants are those that were given in the GIS data. For the fire hydrants evaluated during the site visit, the GIS element labels matched the hydrant labels found in the HAAF distribution system.

The flush times on each hydrant is calculated to clear the piping volume one time and should be followed at a minimum. Ideally, each hydrant should be flushed long enough to clear twice the volume. However, this is a general guideline that depends on time and operation staff constraints.

(3) Figure 2 gives a diagram of the location of the closed isolation valves, the flush hydrant, and the intended flushing path. Sometimes, the report will contain a second diagram page for further clarity. Do not depend on memory when opening and closing isolation valves. Use these pages similar to what is shown in Figure 2 to mark closed valves when they are closed and erase marks after the valves are reopened so that the operational status of valves are not in question.

(4) Figure 3 contains an input page that operations can use to keep a record of their flushes. Ideally, this page will be generated and filled out for every flush for future reference to help optimize this flushing plan. The notes section is a good place to record water quality parameters such as disinfectant residuals and turbidity.

(5) It is likely there will be some problems faced when implementing the UDF plan.

Isolation valves may not exist or may have been paved over. Some fire hydrants may not be operational. New isolation valves may need to be installed to enable improved directional flow and system isolation.

Figure 1. Summary Page - Flushing Event Report

Figure 2. Diagram Page - Flushing Event Report

Figure 3. Operation Record Keeping Page - Flushing Event Report

8. CONCLUSION.

a. A routine UDF program is part of a good overall water system maintenance program. The flushing approach, velocity, frequency, and duration are specific to HAAF system's water quality concerns, hydraulic conditions, source water treatment, and system design.

b. A hydraulic model was used to design the UDF plan for HAAF. The goal of hydraulic modeling is to create a useful realistic model from which to gain insight into the behavior of a given water distribution system.

c. Appendices E – H outlines the flushing procedures for HAAF indicating locations for both isolation valve operation and flush hydrants.

d. The flushing times shown within the flushing procedures in Appendices E – H are estimated times based on the hydraulic model to clear the entire water volume within the piping for each flushing event. The actual time required to flush each hydrant may be more depending on the hydrant water quality.

e. Data collection is imperative to understanding the benefits, costs, and secondary impacts due to flushing. If adequate data are not collected, program refinement and a cost/benefit analysis cannot be conducted. Appendix C provides supplemental information in evaluating flushing effectiveness and each flushing event comes with an operation record keeping page.

f. During the first implementation of the flushing plan, several problems not anticipated on the hydraulic model could occur. Isolation valves valve status may be different from what was anticipated or broken valves may be discovered.

g. This UDF was designed to ensure that no customer will be without water during the implementation of the flushing plan. However, throughout the implementation of the UDF, flushing crews may realize that some distribution system valves that were not assessed in the flushing plan may be in the closed position. Nationwide distribution system survey data show that 10-20% of isolation valves within water distribution systems are in the wrong position and up to 40% of isolation valves are inoperable.

h. A valve exercise program is part of a normal operation and maintenance program and will aid in determining isolation valve status and operability on HAAF.

i. Future modification to the distribution system, adding or removing water mains, can affect the flushing procedure and effectiveness. Relevant portions of the system may need to be appended to continually achieve effective flushing,

9. RECOMMENDATIONS.

a. Implement a flushing plan as laid out in Appendices E – H on an annual basis or when water quality issues arise.

b. Each hydrant should be flushed using the time specified within the flushing procedures in Appendices E – H at a minimum. If the water from a hydrant is not clear, the hydrant flush time may need to be increased.

c. Collect water quality data and take notes after each flush to gain insight and improve future water system flushes.

d. To minimize the risk of low pressure events (less than 20 psi) from malfunctioning valves or unintentionally closed isolation valves when performing this UDF plan, it is important to measure upstream (or nearby) hydrant pressures, especially after closing isolation valves.

e. Before implementing the UDF plan, time should be spent identifying, verifying operability, and assessing the condition of the targeted isolation valves and hydrants in each flushing section.

f. If specific isolation valves cannot be found, the flushing should occur and the flushing sequence should continue to be followed. Flushing with some of the isolation valves closed and abiding by the flushing sequence is still highly beneficial.

g. Purchase needed equipment as identified in Appendix C to monitor pressure and water quality data.

h. In a system that has not previously performed unidirectional flushing, maintenance crews should be available for repair activities.

i. An annual valve exercise program is recommended for HAAF.

j. Conduct an inventory of all isolation valves on HAAF.

k. Update GIS data. Create a single, current, accurate set of shape files in accordance with the 2013 survey of the water system.

l. Notify customers when flushing program will begin and end as described in Appendix D.

m. Conduct a criticality/reliability analysis once HAAF obtains an accurate inventory of isolation valves.

10. POINT OF CONTACT. The point of contact for this evaluation is 1LT(P) Luke Grant, Environmental Health and Engineering Division (EHED), at commercial (301) 677-3698 / email luke.a.grant.mil@mail.mil.

Encls

A-1

APPENDIX A

REFERENCES

1. Army Regulation (AR) 420-1, Army Facilities Management, 12 February 2008.

2. Unified Facilities Criteria (UFC) 3-230-02, Operations and Maintenance: Water Supply Systems, 10 July 2001.

3. Technical Bulletin, Medical 576, Sanitary Control and Surveillance of Water Supplies at Fixed Installations, March 1982.

4. Guidance Manual for Maintaining Distribution System Water Quality, American Water Works Association Research Foundation, 2000.

5. Mercer, Ken, How Important is a System-wide Flushing Program?, American Water Works Association, 2010

6. Walski, Thomas, et. al., Advanced Water Distribution Modeling and Management, Bentley Institute Press, 2007.

7. Walski, Thomas, et. Al, Water Distribution System Modeling, 1st Edition, 2001

8. Cesario, Lee, Modeling, Analysis, and Design of Water Distribution Systems, American Water Works Association, 1995

9. Maximize Your Hydraulic Model, American Water Works Association Webinar, February 2013

10. Distribution Systems Consider Water Quality and Hydraulic Quality, American Water Works Association, Opflow, December 2002

11. Criticality of Valve Operability and Technology, Ohio AWW Distribution Seminar,

12. Computer Modeling of Water Distribution Systems M32; American Water Works Association; 2005

B-1

APPENDIX B

FLUSHING GUIDELINES

1. Water mains should be flushed before consumers start complaining about poor water quality. Flushing should be conducted in the spring or fall during periods of lower water demand and when freezing or other unsuitable weather conditions can be avoided.

2. Prior to flushing mains, notify all affected customers of the dates and times the system will be flushed. Customers can be notified through door hangers, local newspapers, local radio, or local TV announcements. A sample notification flyer is included in Appendix D. Explain the intent and objective of the flushing program. Notify individuals who might be on dialysis machines, hospitals, fire department, restaurants, laundromats, and others who might be affected while the mains are being flushed.

3. The unidirectional flushing plan is a living document so flushing sequences and events can be changed if necessary. Consider flushing at night between 9:00 PM and 5:00 AM to minimize any inconvenience to customers. Night operation encounters little traffic, but traffic must be made aware of the operation if it will be affected by flushing.

Develop a traffic control plan to include signs and warning devices in accordance with local regulations. Notify the police department of the schedule of activities and roads affected.

4. Open and close the isolation valves and hydrants slowly to help prevent water hammer.

5. Direct flushing water away from traffic, pedestrians, and private lots. Avoid erosion damage to streets, lawns and yards by the use of tarpaulins and lead-off discharge devices such as a hydrant diffuser.

6. Record recommended data regarding the flushing operation as well as a description of the appearance and odor of the water flushed (see Appendix C).

7. It is strongly recommended that water personnel monitor upstream hydrants for pressure during flushing especially after making any system changes such as closing isolation valves. It is also recommended that they collect three water samples from each flowing hydrant, one in the beginning (about 1 minute after the hydrant was opened), the second sample should be taken in the middle of the flush (see Appendices E – H for flushing times for each hydrant), and the last sample should be taken just before closing the hydrant. These samples allow a check on water quality for certain

B-2 basic water quality indicators (iron, pH, chlorine residual, turbidity) and the development of water quality trends for comparison purposes. Appendix C goes into more detail about data collection.

8. After the flushing water becomes clear, slowly close the hydrant or blow-off valves.

9. In areas where the water does not become completely clear, the operator should use judgment as to the relative color and turbidity and decide when to shut down. A water sample in a clear glass or white cup will allow the operator to visually observe the color from time to time.

10. Mark closed valves on a map when they are closed and erase marks after the valves are reopened. Do this promptly and do not depend on memory.

11. After one section of the pipe has been flushed, move on to the next section to be flushed and follow the same procedure.

12. Follow the step-by-step flushing procedures laid out in Appendices E – H.

C-1

APPENDIX C

SUPPLEMENTARY INFORMATION

1. Evaluation of Flushing Effectiveness: Documenting the success or failure of the procedure is an important way to determine if adjustment is required. It is possible that all anticipated benefits from the procedure will not be realized immediately and experimentation with velocities, duration, and frequency will be required. To conduct an evaluation, it will be necessary to carefully document the costs, potential savings, benefits, and problems, as well as the secondary effects that can result from the flushing program. The ability to evaluate flushing effectiveness depends on proper documentation while performing the flushing program. Below is a list of questions to be used to determine the effectiveness of the program and the need for modifications.

Were water quality objectives met? Document improvements in water quality (e.g., improved chlorine residuals, decreased turbidity, less customer complaints, less time to reach water quality objectives from year to year).

What were the estimated costs/savings associated with the program? Assign actual costs whenever possible (e.g., labor, power, equipment, disposal requirements, planning/notification efforts). Estimate savings from conducting maintenance activities simultaneously.

Were there positive secondary effects because of the flushing program? Were operating costs reduced (e.g., lower chemical dosage requirements, reduced power costs from decreasing pipe friction, improved valve/meter/hydrant life)?

Was customer perception improved? Is there potential for increased industrial use as a result of water quality improvements?

Were there negative secondary effects from the flushing program? These might include stirred-up portions of the system in an uncontrolled fashion, increased chlorine residual that resulted in customer complaints, release of bacteria into the water column, exposed new surface of tubercles, and negative public perception/waste of water.

2. Data Collection. Below is an example of data to collect during implementation of the UDF plan. Figure C-1 which is included in every flushing event report should be completed and kept in files for future reference is shown below.

a. Pressure Data Collection. It is important that a hydrant upstream of the flushing hydrant be monitored for pressure to ensure that pressures do not drop below 20 psi especially after making any change to the system such as closing an isolation valve.

C-2

Equipment needed include hydrant adapters, a pressure gauge, and a pitot tube with equipment costs estimated at approximately $500.

b. Other Recommended Data Collection. It is recommended that utilities conduct rigorous monitoring and record keeping throughout their distribution system, from source to periphery, including storage facilities, sample stands, inlets, and dead ends, to assess water quality conditions prior to implementing flushing.