UNI-DIRECTIONAL FLUSHING IN REIDSVILLE, NC: A TEAM EFFORT

UNI-DIRECTIONAL FLUSHING IN REIDSVILLE, NC: A TEAM EFFORT ABSTRACT Kevin Eason, PE, City of Reidsville Shane Majetich, Mueller Service Company Jeff Cruickshank, PE, Hazen and Sawyer* 629 Green Valley Road, Suite 200 Greensboro, NC 27408 336-292-7490 Uni-directional flushing (UDF) is getting increased attention from water utilities because of the Stage 2 Disinfectants and Disinfection Byproducts (DBP) Rule. This rule focuses on water quality in distribution systems, where DBPs form when chlorine reacts with water-borne organic material and bio-films on pipe walls. UDF expels sediment from water mains and scours bio-films. In distribution systems that use chloramines as the residual disinfectant, UDF helps control nitrification in areas with excessive water age. UDF induces high velocities by closing valves so water reaches flowing hydrants from only one direction through a single pipe. UDF is labor intensive because it requires hundreds of valve operations. Careful planning is required so water is routed to flowing hydrants only through pipes that have been flushed previously. Specialized software helps plan UDF using hydraulic models. The City of Reidsville decided on UDF as part of its compliance strategy for the Stage 2 DBP Rule. However, city crews did not have sufficient manpower to conduct UDF in a timely manner. In addition, parts of Reidsville s distribution system are more than 75 years old, and the City s GIS did not include sufficient information about the old valves and hydrants to effectively plan and implement a UDF program. The City decided that outsourcing was the best method of getting the job done. This presentation describes a team effort that successfully implemented UDF in Reidsville, covering the entire distribution system comprising about 100 miles of pipe. The team consisted of utility staff, an environmental engineering consultant that planned the UDF sequencing, and a municipal infrastructure assessment company that initially collected information on the valves and hydrants and subsequently executed the UDF program. Lessons learned from this project are discussed from the perspective of the City, the consultant and the infrastructure assessment company. The emphasis is on practical knowledge concerning both planning and execution of a UDF program. The discussion also explains a number of indirect benefits to the City. This presentation will help other water utilities understand UDF, recognize potential problems and work around them. KEYWORDS Distribution Systems Flushing Stage 2 Disinfectants and Disinfection Byproducts (DBP) Rule Uni-Directional Flushing Water Distribution Water Quality

INTRODUCTION The City of Reidsville is located in Rockingham County on US 29 between Greensboro, NC, and Danville, VA. The population was 14,520 in the 2010 census. The City built its first filtration plant in 1912 along with a distribution system consisting of unlined cast iron pipes. The distribution system grew from 41 miles of pipe in 1958 to 90 miles in 1981, and currently includes about 125 miles of pipe. Pipe sizes range from ¾-inch services to 24-inch transmission mains. The average day demand is approximately 3 mgd. Demand has decreased in recent years due to the loss of several large customers, mostly textile mills. Reidsville sells water to the City of Greensboro and is part of a regional combined water system. In 2011, the regional system changed its residual disinfectant from free chlorine to chloramines in response to the Stage 2 Disinfectants and Disinfection Byproducts (DBP) Rule. Before the conversion, Reidsville had occasional compliance issues with the Stage 1 DBP Rule and has also had water quality problems caused by manganese. The City decided to conduct unidirectional flushing (UDF) to prevent potential nitrification problems associated with chloramines and remove manganese deposits. UDF is a systematic method of closing valves and opening hydrants to direct water at high velocities through targeted segments of pipe. Flushing begins near clean water supply sources such as water plants, pump stations and tanks. Closing certain valves in a prescribed sequence creates one-way flow into each segment from other pipes that have been flushed previously. Hydrants are opened to induce water velocities high enough to scour sediment and biofilm from the walls of the pipes. UDF is practical only for 12-inch and smaller pipes. The flow required to produce scour velocities in transmission mains exceeds what can be obtained from hydrants. Prior to the UDF program, the City conducted reactive flushing in response to customer complaints, and the fire department did random flushing that frequently caused subsequent water quality complaints. The City also installed 11 auto-flushers that flush about two hours per day in problem areas. METHODOLOGY This project used a specialized UDF planning module that worked with the city s hydraulic model. In a 2009 project, Hazen and Sawyer used a UDF module and modeling software called InfoWater by Innovyze to import the City s existing EPANET model, a hydrant shapefile from GIS, and a valve shapefile created from an AutoCAD drawing of the distribution system. The drawing included tick marks showing approximate valve locations. The City s EPANET model was based on the same AutoCAD drawing. The UDF software automatically associated the hydrants and valves with the nearest pipe in the model. The automated associations were reasonably good, but some manual corrections were needed because the hydrants were located from GPS coordinates but the valves and model were based on approximate locations from the AutoCAD drawing. The UDF software generated a field journal with 14 flushing zones and 325 flushing sequences, one on each page. For each flushing sequence, the field journal showed valves to close, hydrants to open, required flow for 5 fps scour velocity, minimum flushing time for one turnover, flushing duration for three turnovers and volume flushed, as well as model-predicted static and residual pressures.

The initial UDF plan called for 700 valve operations involving about 300 valves. The City decided they did not have sufficient staff for a field operation of this magnitude. In addition, the City had concerns about their ability to locate and operate such a large percentage of the system s valves. Therefore, the City decided to conduct a valve and hydrant assessment program before proceeding with UDF. The City issued a RFP for a project that combined assessing the valves and hydrants, revising the UDF plan based on the valve assessment, and executing the UDF. The City selected a team consisting of Mueller Service Company with Hazen and Sawyer as a subcontractor to revise the UDF plan. The valve and hydrant assessment by Mueller Service Company included: Locating each valve in the field Identifying each valve with a unique label Accessing each valve by removing the valve box cover Cleaning out and inspecting each valve box Exercising each valve through two full cycles and recording operating torques Marking each valve with paint after inspection Delivering a database with condition, torques, GPS data and other attributes for each valve The hydrant assessment by Mueller Service Company included: Inspecting, testing and performing maintenance on every hydrant and hydrant isolation valve Delivering a database with make, model, static and residual pressures and test flow for all hydrants After completion of the valve and hydrant assessment, Hazen and Sawyer attempted to revise the UDF plan using the GPS locations of the valves and their reported condition, avoiding those that were inoperable. However, the UDF software was unable to associate most of the valves with the correct pipes because the valves now had correct GPS coordinates, while the model still was based on the old AutoCAD drawing. The City therefore decided to expand the scope of the project to include moving the pipes in the model to their correct locations based on valve and hydrant locations and aerial images from Google maps. After Hazen and Sawyer made this adjustment, the UDF software easily associated the valves and hydrants with the correct pipes in the model. Adjusting the model also allowed model results to be mapped with other GIS data for other applications. The revised UDF plan used the same 14 flushing zones. The field journal included 320 flushing sequences involving 738 valve operations. The total length of the flushing sequences was 103 miles of pipe, excluding transmission mains. The City used an automatic calling system to notify customers about the flushing program. The City used GIS to define a grid for each day of flushing, and the City s Code Red system called all the customers within the designated area two days before flushing was conducted. Mueller Service Company conducted the UDF program using their crews and fully equipped vehicles. Flushed water was dechlorinated using sodium sulphite. The flushing crews followed the revised field journal and frequently checked the water being flushed using a white bucket. The City provided an inspector who accompanied the flushing crews, made sure water drained properly and explained the purpose of the work to curious customers.

RESULTS The valve and hydrant assessment generated a wealth of information. The geo-database included GPS coordinates for 784 valves. The database included the valve ID, size, depth of operating nut, number of turns, close direction and torque data, and GPS coordinates. Of the valves shown on the City s AutoCAD drawing, 169 valves or 1 percent could not be found in the field. Only 42 valves were inoperable, not nearly as many as the City expected. The number of inoperable valves was minimized by using a valve turning machine to exercise the valves through multiple cycles so they turned easier. The valve machine measured initial and final operating torques, and the average improvement was 76 foot-pounds. City staff cited this improvement in valve operability as one of the major benefits of the project. The hydrant assessment database included the location, make, model year, valve size, pumper height, static pressure, residual pressure and test flow for 700 hydrants. The assessment showed all the City s hydrants were operable. Isolation valves were not found on 20 percent of the hydrants, and 4 percent of the hydrant isolation valves were inoperable. The UDF program started April 23 and ended June 7, 2013. A two-man crew flushed five days per week but was off every third Friday, working a total of 31 days. The crew averaged 10 flush sequences per day. DISCUSSION The City has a much better understanding of the condition of its valves as a result of this project. When the valve assessment crew discovered a valve that needed repairs, a text message was sent to the City to generate a work order. The City is now confident that valves will hold when repairing leaks. The project showed the value of conducting a valve and hydrant assessment before planning UDF. This avoids planning UDF sequences that rely on inoperable valves or those that cannot be located. The City received low pressure complaints when closing valves to prepare for one of the UDF sequences. Subsequent investigations revealed a closed 12-inch valve that was not shown on any City maps, so it was not included in the valve assessment. Opening the closed valve improved reliability and increased fire protection at a nearby hospital, an unexpected benefit of conducting UDF. The project showed that the best results are obtained when a City inspector accompanies the contractor crews for both the valve and hydrant assessments. The project also showed that problems arise when mixing data that includes GPS coordinates and approximate locations from paper maps. UDF planning software works best when the hydraulic model, valve locations and hydrants locations are all based on GPS coordinates. Unlined cast iron pipes did not create unusual problems, and the flushed water generally cleared up within the calculated duration for three turnovers. In some areas, flushed water cleared up initially but then suddenly clouded up after several minutes. In other areas, flushed water cleared up and remained clear well before three turnovers. The Mueller Services crew and the City inspector checked the flushed water frequently using a white bucket, and after at least one turnover, hydrants were turned off when the water was clear.

The City s water quality sampling showed low total chlorine residuals in the three months after flushing was completed. Several factors may be contributing to this problem, but City staff believes some of the residual drop may be attributable to oxidation at pipe walls after scouring away the protective layer of corrosion inhibitor. This problem should improve over time as the corrosion inhibitor re-pacifies pipe walls. Plant operators requested coordinating UDF operations with DBP sampling so that flushing is completed well before quarterly samples are taken. Flushing crews preferred longer sequences with fewer valves and longer flush durations. Crews flushed about six hours per day, with two hours for setting up. An improvement for future UDF planning would be creating smaller flushing zones with a total flush duration just under six hours, so that each flushing zone could be completed in one day. There were few customer complaints during the flushing program. The City inspector talked to many of the customers as flushing progressed, and the advance notification minimized complaints. The City has experienced fewer water quality complaints since the UDF project was completed. Overall, the City was pleased with the results of the project and is considering repeating the UDF program periodically. CONCLUSIONS Conducting a valve assessment before planning a UDF program has many benefits. Exercising the valves during the assessment greatly improves their operability for UDF. Cooperation between UDF planners and flushing crews ensures an efficient and successful project. Cooperation between the utility and the flushing crews minimizes customer complaints. UDF is an effective method of scouring sediment, removing biofilms, and reducing water quality complaints. ACKNOWLEDGEMENTS The authors wish to thank Kevin Widderich with Hazen and Sawyer for his modeling and UDF planning efforts; Richard Vaughn with the City of Reidsville for sharing lessons learned as the field inspector and Steve Coleman, formerly with Mueller Service Company, for sharing his perspective in managing field operations.