The hydraulic integrity of a water distribution system is defined as its ability to provide a reliable water supply at an acceptable level of service—that is, meeting all demands placed upon the system with provisions for adequate pressure, fire protection, and reliability of uninterrupted supply (Cesario, 1995; AWWA, 2005). Water demand is the driving force for the operation of municipal water systems. Because water demands are stochastic in nature, water system operation requires an understanding of the amount of water being used, where it is being used, and how this usage varies with time. For most water systems the ratio of the maximum day water demand to the average day water demand ranges from 1.2 to 3.0, and the ratio of the peak hour to the average day is typically between 3.0 and 6.0. Of course, these values are system specific, and seasonal variations may make these ratios even more extreme (Walski et al., 2003). Demands may be classified as follows (Clark et al., 2004):
Baseline demands, which usually correspond to consumer demands and unaccounted-for-water associated with average day conditions.
Seasonal variations in demand because water use typically varies over the course of the year with higher demands occurring in the warmer months.
Fire demands, which may be the most important consideration for water system design.
Diurnal variations due to the continuously varying demands which are inherent in water systems.
There is a need for research that relates distribution system design to demand in a stochastic framework. Pioneering work by Buchberger and Wu (1995), Buchberger and Wells (1996), and Buchberger at al. (2003) has found that residential water use follows a Poisson arrival process with a time dependent rate parameter. Variations in demand have an important influence on water distribution system operation and in the determination of water age which in turn influences water quality, as discussed later in the chapter.
From an infrastructure perspective, a water distribution system is an elaborate conveyance structure in which pumps move water through the system, control valves allow water pressure and flow direction to be regulated, and reservoirs smooth out the effects of fluctuating demands (flow equalization) and provide reserve capacity for fire suppression and other emergencies. All these distribution system components and their operations and complex interactions can
produce significant variations in critical hydraulic parameters, such that many opportunities exist for the loss of hydraulic integrity and degradation of service. This, in turn, may lead to serious water quality problems, some of which may threaten public health.
One of the most critical components of hydraulic integrity is the maintenance of adequate pressure, defined in terms of the minimum and maximum design pressure supplied to customers under specific demand conditions. Low pressures, caused for example by failure of a pump or valve, may lead to inadequate supply and reduced fire suppression capability or, in the extreme, intrusion of potentially contaminated water. High pressures will intensify wear on valves and fittings and will increase leakage and may cause additional leaks or breaks with subsequent repercussions on water quality. High pressures will also increase external load on water heaters and other fixtures. Pipes and pumps must be sized to overcome the head loss caused by friction at the pipe walls and thus to provide acceptable pressure under specific demands, while sizing of control valves is based on the desired flow conditions, velocity, and pressure differential. A related need is to ensure that pressure fluctuations associated with surge conditions are kept below an acceptable limit. Excessive pressure surges generate high fluid velocity fluctuations and may cause resuspension of settled particles as well as biofilm detachment.
A second element of hydraulic integrity is the reliability of supply, which refers to the ability of the system to maintain the desirable flow rate even when components are out of service (e.g., facility outage, pipe break) and is normally accomplished by providing redundancy in the system. Examples include looping of the pipe network and the development of backup sources to ensure multiple delivery points to all areas.
Many water quality parameters change with length of time in the distribution system, a factor directly related to the hydraulic design of the system. For example, chlorine residuals decrease with the increasing age of water and may be completely lost, and trihalomethanes concentrations may increase with time. In addition, higher concentrations of substances may leach from pipe materials and linings if the contact time with the water is increased. Low velocities in pipes create long travel times, resulting in pipe sections where sediments can collect and accumulate and microbes can grow and be protected from disinfectants. Furthermore, sediment deposition will result in rougher pipes with reduced hydraulic capacity. If peak velocity is increased or flow reverses in these pipe sections due to any operational change or shock loading, such as tank filling or draining, valve opening or closing, pump going on- or off-line, unexpected higher system pressure, or hydrant flushing, there is a risk that deposits will be suspended and carried to consumers. Long detention times can also greatly reduce corrosion control effectiveness by effecting phosphate inhibitors and pH management. Thus, reducing residence time is an important hydraulic issue both in pipes and in storage facilities.
A final component of hydraulic integrity is maintaining sufficient mixing and turnover rates in storage facilities. Insufficient turnover rates and incom-
plete (uneven) hydraulic mixing in reservoirs can allow short-circuiting between the tank inlet and outlet and generate pockets of stagnant water with depleted disinfectant residual. This can lead to bacterial regrowth and other biological changes in the water, including nitrification and taste and odor problems.
This chapter discusses the factors that can cause the loss of hydraulic integrity, the consequences of losing hydraulic integrity, how to detect loss of hydraulic integrity, techniques for maintaining hydraulic integrity, and how to recover system hydraulic integrity once it is lost.
FACTORS CAUSING LOSS OF HYDRAULIC INTEGRITY
There are many different ways that a water distribution system can lose its hydraulic integrity, such that water quality becomes impaired. A loss of hydraulic integrity implies a loss of positive line pressures, flow reversals, rapid changes in velocity, a reduction in hydraulic capacity, a detrimental increase in water residence time, or a combination of these events. Factors causing a loss of system hydraulic integrity include (1) pipe leaks and breaks, (2) rapid changes in pressure and flow conditions, (3) planned maintenance activities and emergencies, (4) tuberculation and scale formation in pipes, and (5) improper operational control.
Pipe deterioration resulting in leaks or breaks can lead to a loss of hydraulic integrity because adequate pressures can no longer be maintained. As discussed in detail in Chapter 4, all pipe materials are vulnerable to some kind of chemical or physical deterioration, and all water mains will require rehabilitation and eventual replacement. Aging pipe infrastructure and chronic water main breaks are a common problem for many water utilities. Analysis of water industry data showed that on average, main breaks occur 700 times per day in the United States (Cromwell et al., 2001). The condition of distribution system pipes is influenced by material type and age, line pressure, type of soil, installation procedures, and many other factors, making it difficult to predict where breaks and leaks will occur. Chapter 4 discusses the roles of leak detection and condition assessment in determining the current condition of distribution system infrastructure.
Pressure Transients and Changes in Flow Regime
Rapid changes in pressure and flow caused by events such as rapid valve closures or pump stoppages and hydrant flushing can create pressure surges of excessive magnitude. These transient pressures, which are superimposed on the
normal static pressures present in the water line at the time the transient occurs, can strain the system leading to increased leakage and decreased system reliability, equipment failure, and even pipe rupture in extreme cases. High-flow velocities can remove protective scale and tubercles, which will increase the rate of corrosion. Uncontrolled pump shutdown can lead to the undesirable occurrence of water-column separation, which can result in catastrophic pipeline failures due to severe pressure rises following the collapse of the vapor cavities. Vacuum conditions can create high stresses and strains that are much greater than those occurring during normal operating regimes. They can cause the collapse of thin-walled pipes or reinforced concrete sections, particularly if these sections were not designed to withstand such strains. In less drastic cases, strong pressure surges may cause cracks in internal lining, damage connections between pipe sections, and destroy or cause deformation to equipment such as pipeline valves, air valves, or other surge protection devices. Sometimes the damage is not realized at the time, but may cause the pipeline to collapse in the future, especially if combined with repeated transients.
Transient pressure and flow regimes are inevitable. All systems will, at some time, be started up, switched off, or undergo rapid flow changes such as those caused by hydrant flushing, and they will likely experience the effects of human errors, equipment breakdowns, earthquakes, or other risky disturbances (Wood et al., 2005). Figure 5-1 illustrates typical hydraulic events following a pump trip.
Low pressure transients may promote the collapse of water mains, leakage into the pipes at joints and seals under sub-atmospheric pressures, and backsiphonage (see Chapter 4). There is also evidence that pressure transients can lead to the intrusion of contaminants into the distribution system. LeChevallier et al. (2003) reported the existence of low and negative pressure transients in a number of distribution systems. Gullick et al. (2004) studied intrusion occurrences in distribution systems and observed 15 surge events that resulted in a negative pressure. Most were caused by the sudden shutdown of pumps at a pump station because of either unintentional (e.g., power outages) or intentional (e.g., pump stoppage or startup tests) circumstances. Friedman et al. (2004) confirmed that negative pressure transients can occur in the distribution system and that the intruded water can travel downstream from the site of entry. Locations with the highest potential for intrusion were sites experiencing leaks and breaks, areas of high water table, and flooded air-vacuum valve vaults.
Hydraulic Changes during Maintenance and Emergencies
Water distribution systems are occasionally subject to emergencies or planned maintenance activities in which certain components become inoperable and the system can no longer provide the minimum level of service to customers (AWWA, 2005). Planned maintenance activities include supplies going off line (e.g., stopping the treatment plant or shutting down a well); reservoir shutdown for inspection, cleaning, or repairs; installation of new pipe connections; pipe rehabilitation or break repairs; and transmission main valve repairs. Examples of emergency situations include earthquakes, hurricanes, power failures, equipment failures, or transmission main failures. All these activities can result in a reduction in system capacity and supply pressure and changes to the flow paths of water within the distribution system.
Tuberculation and Scale
The hydraulic capacity of distribution systems can be compromised by deposits on the internal surface of the pipelines. The deposition of corrosion products in the form of tubercles and other types of scales on the interior of the pipes can seriously clog water lines and thus restrict the flow of water. Scales may also form because metal salts such as calcium carbonate, aluminum silicate, etc. (see Chapter 6) in treated water entering the network are supersaturated, leading to their precipitation on the pipe walls. Excessive pressure may be necessary to deliver the required flow of water in pipes with tuberculation and scales, further weakening aging pipes. The reduction in hydraulic capacity is caused by the increases in head loss due to the roughness of the deposits and to the decrease in pipe diameter that they cause.
Inadequate Operational Control
Historically, utilities have focused on the quality of water leaving the treatment plant, because of regulatory drivers, and on the quantity of water supplied by the distribution system, because of their mission to satisfy water demand and maintain system pressure. Thus, it is not surprising that distribution system operations at many utilities and their associated professionals (designers, builders, plumbers, inspectors, etc.) have been water quantity focused rather than water quality focused.
There is now greater recognition of the water quality effects of how long water is retained in the various elements of the distribution system. Retention time or water age is strongly related to the characteristics of the system and its operation. For example pipe roughness, which affects water flow and residence time, may be modified by repair or rehabilitation. Operational activities, such as
pump scheduling and planned maintenance, or unplanned effects, such as unexpected changes in demand, will all have an effect on water age. A particularly important issue that demonstrates the interaction of system operation and water quality is the ability or inability of utilities to ensure adequate mixing intensity and time in storage tanks to minimize short circuiting and to limit residence times to be within acceptable limits. Interestingly, the design of tanks to ensure adequate turnover is required in only 15 of 34 states that responded to a survey of drinking water programs conducted by the Association of State Drinking Water Administrators in March 2003 (see Table 2-4). Dealing with these issues is discussed in the context of system operation later in this chapter.
CONSEQUENCES OF A LOSS IN HYDRAULIC INTEGRITY
There are several detrimental consequences of losing system hydraulic integrity, including contamination of the distribution system via intrusion, sedimentation, a reduction in hydraulic capacity, loosening of scale, and extended water age. Each of these has attendant water quality implications, as described below.
A distribution system can become contaminated by the external environment for several reasons. The most well documented contamination events are backflow and direct contamination at breaks and repair sites, discussed in Chapter 4. A specific type of backflow event related to a loss of hydraulic integrity is called intrusion, which refers to the entrance of contamination into the water distribution system through leaks (caused by corroded areas, cracks, and loose joints) because of sustained low or negative pressures or a pressure transient. When a section of the distribution system is depressurized due to a normal shutdown, failure of a main or a pump, routine flushing, or emergency fire-fighting water drawdown, contaminated water can be pulled into the main. For example, during a large fire, a pump is connected to a hydrant. High flows pumped out of the distribution system can result in a significantly reduced water pressure around the withdrawal point. A partial vacuum is created in the system, which can cause suction of contaminated water into the potable water system through nearby leaks. During such conditions, it is possible for water to be withdrawn from nonpotable sources into the distribution system and subsequently distributed to homes and buildings located near the fire. The same conditions can be caused by a water main break.
Sustained low pressure events and transient pressure events that lead to intrusion of contaminated water have the potential for substantial water quality and health implications. The potential for intrusion of contaminated groundwater into pipes with leaky joints or cracks seems greatest in systems with pipes below the water table and where pathogens or chemicals are in close proximity to the pipe. As discussed in Chapter 4, two recent studies (Kirmeyer et al., 2001; Karim et al., 2003) have established that soil and water samples collected immediately adjacent to pipelines can contain high fecal coliform concentrations and viruses. In the event of a large intrusion of pathogens, the disinfectant residual normally sustained in drinking water distribution systems may be insufficient to neutralize contaminated water (see Chapter 6 discussion on Adequate Disinfectant Residual). Transient events can also generate high intensities of fluid shear and may cause resuspension of settled particles as well as biofilm detachment.
When water is moving slowly through a pipe, particles suspended in the water may settle out into the pipe. The accumulated sediment reduces the pipe’s hydraulic capacity. They also serve as a food source for bacteria and create a hospitable environment for microbial growth. If not removed these materials may cause water quality deterioration, taste and odor problems, or discoloration of the water. This is particularly evident if the sediments are disturbed (stirred up) by changes in the flow of water, such as when a main break occurs, a service
reservoir is filling or draining, a pump is going on or off line, or during normal hydrant flushing activities. The normal flow of water through the system will reduce some but not all sediment accumulation over time, and supplemental measures are periodically needed to clear out the system.
Reduction in Hydraulic Capacity and Associated Increase in Pumping Costs
As metal pipes age their roughness tends to increase and their cross sectional area tends to decrease due to encrustation and tuberculation of corrosion products on the pipe walls. This increase in hydraulic roughness and decrease in effective diameter will increase the resistance to flow and reduce the hydraulic capacity of the affected mains. Other deposits such as microbial slimes can also result in a significant decrease in the hydraulic capacity of water mains. The reduction in the hydraulic capacity can lead to a subsequent unwanted reduction in system pressure due to the higher head loss. The loss in system pressure can result in a water system that cannot deliver the necessary fire flows and, in the extreme, it provides the potential for backflow of contaminants.
In order to meet demand in such systems, higher pumping rates are needed to overcome the higher head losses and to avoid or postpone the replacement, duplication, or rehabilitation of tuberculated mains. This can overload motors and result in a significant increase in energy consumption and operational and maintenance costs of a water utility. Furthermore, the additional pumping can over-pressurize certain portions of the distribution system, thereby increasing leaks and breaks, and it can lead to ineffective utilization of storage tanks and reservoirs because high pressure in the mains prevents outflow from the reservoirs. If these reservoirs are subsequently put back into service during peak times when consumption is high, this may result in the provision of “old” (poor quality) water.
Poor Water Quality from Sediment Suspension and Removal of Scales
Changes in flow (magnitude and direction) within the water distribution system as a result of hydrant flushing and valve and pump operation can scour sediments, tubercles, and scales from the interior pipe walls and degrade water quality. For example, when the water velocity is increased or flow direction is reversed, sediment deposited on the pipe walls during periods of low flow may be re-suspended and scales may detach. These materials may cause the water to be colored, turbid, and sometimes odorous. Also, it is possible that these particles have adsorbed contaminants such as arsenic and other metals that originated in the source water, as discussed in Chapter 6.
Hydraulic Integrity and Water Age
As distribution system water ages, its quality degrades, such that delivering “younger” water is a desirable operational goal for water utilities. However, the concept of water age is complex. Water age at a given location and time in a water distribution system is actually a mixture of water parcels that have traveled along different paths through the distribution system with correspondingly different travel times. Therefore, the age of water at a given point in the distribution network is not a single value, but rather a distribution of values, termed a residence time distribution (Levenspeil, 2002). This concept is illustrated in Figure 5-2, which shows the results of a study conducted by EPA in collaboration with the Greater Cincinnati Water Works in which a calcium chloride tracer was introduced into an isolated portion of the distribution system (Clark et al., 2004; Panguluri et al., 2005). Figure 5-2 shows the field results from 34 continuously recording specific conductivity meters that were deployed at various nodes in the system, with an EPANET modeling prediction superimposed on the data. The three concentration peaks represent the different parcels of water that have taken different routes to the monitoring point, resulting in a residence time distribution at that monitoring station at the time the data were collected.
For the purposes of this report, water age at a specific point in the distribution system is assumed to be the mean of the residence time distribution. The report uses the term “water age” as a surrogate for water quality. However, it should be noted that while water quality may depend on the age of the water, it may also depend on the specific residence time distribution at that point in the network or on one of its statistics (such as its variance). These complexities are infrequently considered in studies where water age is measured, making this an area ripe for additional research.
In addition to water age at any one point in the network being a distribution of values, the age of water delivered to all consumers is also a distribution of values, the shape of which depends on the location of the consumer, seasonality, whether the network is looped versus one way, the existence of storage facilities, etc. A typical system may deliver water to consumers that has resided in the network for a few days, but many systems have some portion of the network where residence time is much longer. For example, in Blacksburg, Virginia, 97 percent of the water in the main distribution system has a water age of less than 7 days, but 1 percent of the system has a residence time longer than 28 days. Premise plumbing adds another layer of complexity that is addressed in Chapter 8.
Hydraulically, increased water age is a consequence of many factors, including the inevitable loss of carrying capacity as pipes age. However, system design and operation have the most significant impact on water age, particularly where water storage facilities are concerned. For example, high residence time in these facilities can allow the disinfectant residual to be completely depleted, thereby preventing the protection of finished water from additional microbial contaminants that may be present in the distribution system downstream of the
facilities. A survey of water utilities found that bacterial regrowth became a problem in free chlorine residual systems when water age reached three days whereas in monochloramine residual systems regrowth was not a problem until water age reached or exceeded seven days (Baribeau et al., 2005). Other negative consequences of increased water age are discussed in Chapter 6.
DETECTING LOSS OF HYDRAULIC INTEGRITY
Ideally, the verification of hydraulic integrity should involve real-time monitoring of pressure, flow direction, and velocity based on telemetry data. This type of data can be transmitted electronically from permanently installed measurement devices in the field. Typical measurement locations should include treatment plants and wells, pump and booster stations, reservoirs, valves, and other critical points in the system such as elevated sites.
An effective system-wide monitoring program can capture local variations in hydraulic behavior (e.g., pressure, flow) at a specific point in a water distribution system but cannot provide an overall understanding of the spatial and temporal changes, complex flow pathways, and interactions among the various water system characteristics. Thus, water distribution system network models are attractive supplements to monitoring for evaluating hydraulic and water quality changes throughout the distribution system. By combining telemetry data and modeling information, water utilities can gain a more complete and accurate picture of their systems hydraulic and water quality operation and performance capabilities. For example, the North Marin Water Authority in North Marin, California, draws its water from two sources, one of very poor quality with high levels of natural organic matter and another source of very high quality. Because of demand variations there is a great deal of mixing between the water sources at various nodes in the system leading to wide variations in trihalomethane (THM) values over a given day. On the surface these variations in THM concentration were unexplainable until hydraulic modeling techniques were applied which clearly showed that these variations were the result of the mixing effect from the two sources of water (Clark and Buchberger, 2004). Hydraulic integrity is best measured by monitoring and modeling of the system hydraulic parameters, as discussed below.
Monitoring the operation of water distribution system components yields data used to detect the system hydraulic integrity. This can be accomplished in real-time by means of a Supervisory Control and Data Acquisition (SCADA) system, which provides local and remote (supervisory) real-time control and monitoring of selected process equipment and parameters at strategic locations throughout the water distribution system. Any parameter with a proper sensor
and transmitter that can produce an analog signal (e.g., 4-20 mADC) proportional to the variation of the measured parameter can be monitored in real-time or historically via the SCADA system. The acquired data can be viewed on a real-time basis and also archived in a database for historical evaluation at a later date. The data generated from the sensors and transmitters is conveyed to the central control system using various communication media such as telephone lines, fiber optic cables, or radio and cellular systems. The amount of data collected is determined by the polling frequency of the SCADA system.
To detect changes in hydraulic integrity, certain hydraulic characteristics of water system components should be monitored continually in the distribution system via SCADA. These include reservoir inflow/outflow rates, water volumes and levels (used to calculate daily volume turnover), pump station operation such as status and speed settings, pump discharge flows and pressures, valve positions, regulating valve downstream (and /or upstream) pressures, pipe flow rates, and pressures at strategic sites. In addition, disinfectant residual, temperature, conductivity, turbidity, dissolved oxygen, and pH can be continuously monitored at the treatment plant. Temperature in storage tanks and reservoirs could also be monitored via SCADA to detect thermal stratification that results from poor mixing characteristics. Temperature differential between the inflow and the bulk water in the reservoir can result in density gradients inside the storage facility and cause stratification and poor hydraulic mixing and, thus, the greatest potential for water quality deterioration (Mahmood et al., 2005).
Continuous system-wide monitoring provides insight into the patterns of operational characteristics throughout the distribution system. An analysis of these patterns can directly determine if the system hydraulic integrity is not compromised and the system is operating as designed, or detect any unanticipated operational anomalies. For example, high night-time flows in specific areas could be an indicator of high leakage. Sonic leak detection equipment (discussed in Chapter 4) can be used to pinpoint the exact location of those leaks, which can then be isolated and repaired. Similarly, unexpected low pressure readings, excessive pumping, or a drop in reservoir levels in a specific area could indicate a large main break that may increase the potential for backflow.
Another function of SCADA is the ability to monitor and remotely control local conditions of water system components based on any desired range of operating conditions or set points. For example, a pump can be set to turn on or off automatically when the pressure at a critical location or the water level in a reservoir drops to a specified lower limit or goes above a specified upper limit. Alarms can be set to alert operators when a fault within the system equipment (e.g., equipment operating out of its normal range or overheating of a pump) and any breach in the system hydraulic integrity is detected. For example, extreme fluctuations in pressure and flow readings could result from pressure surges generated from a power failure at a pump station. SCADA could then divert water to the affected region from a different pump station, thus ensuring adequate supply and fire flow protection.
SCADA systems also contain pertinent system operational information required for water distribution network modeling (Cesario, 1995), such as the boundary conditions (e.g., tank water levels, valve and pump statuses and settings) for the network model as well as local flow and pressure conditions. These data can be used for calibrating network models (the process of adjusting model parameters so that modeled values reasonably match with measured data), confirming normal system operation, verifying daily variation in total system demands (based on a mass balance of the flows from the treatment plant and wells and in and out of the reservoirs), estimating water losses during main breaks, and investigating and solving operational problems. Operating data can be time specific or represent several consecutive points in time for comprehensive dynamic (extended period simulation) network modeling (e.g., 24-hour simulation) (see Chapter 7 for details on modeling). Clark et al. (2004) list many benefits of remote monitoring and network modeling for water security protection.
Beyond remote controlled, real-time monitoring provided by SCADA, actual field measurements can be made to detect any potential loss of system hydraulic integrity. Hydrant tests are performed to determine if fire flow requirements are met as an indicator of the hydraulic strength of the water system. Head loss tests are conducted to determine the hydraulic capacity of pipes as an indication of system hydraulic performance capability. Pump efficiency tests can be used to determine whether or not pump performance (e.g., overall system efficiency, electrical motor performance, and pump hydraulics) is degrading with time and if replacement or upgrading of equipment is warranted. Hydraulic grade line tests of a pipeline profile (stretches of pipes) help locate partially closed valves and deteriorated pipes with poor hydraulic capacity (high roughness). Field measurements of pressure, flow conditions, velocity, and other water system characteristics can also be carried out using a variety of measurement devices at any facility to verify questionable SCADA readings.
Computer based mathematical models provide an effective and viable means of analyzing hydraulic and water quality conditions in distribution systems (see Clark and Grayman, 1998; Lansey and Boulos, 2005; Panguluri et al., 2005; Boulos et al., 2006). They can calculate the spatial and temporal variations of flow, pressure, velocity, reservoir level, water age, source contribution, disinfectant concentration, and other hydraulic and water quality parameters throughout the distribution system. These predictive capabilities are useful for detecting a loss of system hydraulic integrity. For example, model results can help identify areas of low pressures, excessive head losses, and high water age; compute water losses; locate partially closed valves; verify that the replacement or addition of a new supply source (e.g., emergency service connections or adding a new reservoir or well) will have little or no effect on the flow, velocity,
and pressure patterns and residence times; estimate filling and draining cycles of reservoirs; detect oversized facilities; calculate interzone water transfers; and determine the adequacy of the system to supply fire flows under a variety of demand loading and operating conditions.
A few specific models are of particular importance to maintaining hydraulic integrity. First, surge models can be used to assess the hydraulic adequacy of the system under various transient conditions, identify weak spots, and evaluate the efficacy of surge control devices. These models could be instrumental in future research to better understand the potential for intrusion to contaminate distribution systems. Second, computational fluid dynamics (CFD) modeling has potential for investigating hydraulic mixing and transport characteristics in storage facilities and pipes for a wide range of designs and system operational conditions (Panguluri et al., 2005). CFD models predict flow patterns, heat transfer, and chemical reactions via the solution of partial differential equations that describe conservation of mass, momentum, and energy in a two- or three-dimensional grid that approximates the pipe or tank geometry. CFD models are used to simulate temperature profiles, unsteady hydraulic and water quality conditions, and decay of constituents in bulk flow and in storage facilities. However CFD modeling requires experienced and skilled programmers for effective application (Panguluri et al., 2005). Such network modeling applications greatly enhance the ability of water utilities to effectively manage, operate, and maintain their water distribution systems and deliver an adequate level of service to their customers.
MAINTAINING HYDRAULIC INTEGRITY
Water utilities often find themselves choosing between two approaches to preserve system hydraulic integrity: (1) reacting only to emergencies or (2) acting to prevent problems from occurring. The desirable approach is to develop an active program to prevent future problems and service interruptions.
To maintain the hydraulic integrity of water distribution systems and ensure the highest possible water quality, travel times in the system should be kept as short as possible and large fluctuations in the hydraulic regime and low flow and pressure conditions should be avoided. This can be accomplished by implementing best design, management, operational, and maintenance practices, as discussed below. Hydraulic modeling, discussed in the previous section, is also a critical component that can be used to identify problems areas within the distribution system and to develop design and operational alternatives that address the deficiencies. Those practices necessary to maintain both physical and hydraulic integrity, such as preventing the formation of leaks and cracks in pipe mains and using backflow prevention devices, are discussed in the previous chapter.
Reliability of water distribution systems, which is necessary to minimize outages, is provided by building redundancy in the system in the form of looping and backup sources. A looped (as opposed to branched) multi-source system has the hydraulic advantage of carrying water to any location from more than one direction when a high rate is required (e.g., a fire flow demand) or when a pipe or source is out of service (see Chapter 1). Sufficient interconnections between the distribution mains are necessary to improve the ability of the system to maintain the normal supply by re-routing the water when a breakdown occurs. Dead-end distribution lines should be avoided. A fire-flow demand or large water use on a dead-end main can only draw water through a single pipe, with the maximum flow dictated by the size and length of the pipe. In addition, during scheduled maintenance or repairs on dead-end mains both the supplied customers and available fire flows will be affected. Availability of back-up power (e.g., generators in pump stations), extra pumps, additional reservoirs, standby wells, and emergency interconnections with other systems will provide the necessary redundant sources.
Redundancy can also be facilitated by ensuring an adequate number of operable valves and hydrants, as well as their strategic placement to allow for control of the system and for shutdown of sections for emergency repair and planned maintenance (Male and Walski, 1991).
Management of Pressure Zones
Water distribution systems work best with minimal fluctuations in pressure. The pressure differential range, which specifies the operating values for maximum and minimum pressure to be maintained, is based on local engineering standards and conditions. Many states have established requirements for the design, construction, operation, and maintenance of drinking water distribution systems that relate to hydraulic parameters. For example, 32 of 34 responding state require that distribution systems be designed for an operational pressure of at least 20 psi under all flow conditions (see Table 2-3). Further, nine of 34 require both a minimum and maximum velocity through pipes. These requirements determine the maximum and minimum ground elevations that can be supplied. The minimum pressure establishes the highest ground elevation that can be supplied, and the maximum pressure defines the lowest ground elevation. The former criterion ensures that the highest customers will be supplied with at least the minimum pressure, while the latter ensures that the lowest customers will not experience objectionably high pressures.
To supply water at acceptable pressure, the distribution system is thus divided into a number of distinct pressure zones. The maximum change in elevation across each zone is determined by the difference between the maximum and minimum design pressure values. Adding new pressure zones or adjusting exist-
ing pressure zone boundaries is needed when pressure differentials are outside their desirable values. Pressure zone boundaries are delineated through the use of closed valves. To improve reliability, pressure-regulating valves (or pumps) are normally installed between the zones (along the pressure zone boundaries), and stretches of new pipe are added to eliminate dead ends.
Pressure zoning is desirable but requires careful planning and design (for details, see Boulos et al., 2006). Proper design of pressure zones will reduce leaks (because leakage normally varies exponentially with pressure and will be reduced with a fall in system pressures), breaks, and pumping costs; improve reservoir turnover rates; and avoid over-pressurizing the system. Existing facilities (e.g., reservoirs, pumps, pressure regulating valves) and natural (e.g., rives, lakes) or political boundaries (e.g., city limits, county and state boundaries) will influence the design and modification of pressure zones (Cesario, 1995).
Pressure events or surges that can allow intrusion to occur are caused by sudden changes in water velocity due to loss of power, sudden valve or hydrant closure or opening, a main break, fire flow, or an uncontrolled change in on/off pump status (Boyd et al., 2004). Intrusion can be minimized by knowing the causes of pressure surges, defining the system’s response to surges, and estimating the system’s susceptibility to contamination when surges occur (Friedman et al., 2004). Pressure transients in distribution systems are usually most severe at pump stations and control valves, in high-elevation areas, in locations with low static pressures, and in remote locations that are distanced from overhead storage (Friedman et al., 2005).
A number of devices can be used for controlling transients in pipeline systems (Boulos et al., 2005, 2006; Wood et al., 2005). The general principles of pressure surge control devices are to store water or otherwise delay the change of flow or to discharge water from the line so that rapid or extreme fluctuations in the flow regime are minimized. Devices such as pressure-relief valves, surge anticipation valves, surge vessels, surge tanks, pump bypass lines, or any combination thereof are commonly used to control maximum pressures. Storage tanks with a free water surface can be effective in controlling surges. Minimum pressures can be controlled by increasing pump inertia or by adding surge vessels, surge tanks, air-release/vacuum valves, pump bypass lines, or any combination of these components. The overriding objective is to reduce the rate at which flow changes occur. Figure 5-3 illustrates typical locations for the various surge protection devices in a water distribution system.
Because no two distribution systems are hydraulically the same, there are no general rules or universally applicable guidelines for eliminating objectionable pressures in distribution systems. Any surge protection device must be chosen accordingly. The final choice will be based on the initial cause and location of the transient disturbance(s), the system itself, the consequences if
remedial action is not taken, and the cost of the protection measures. A combination of devices may prove to be the most effective and economical. Final determination of the adequacy and efficacy of the proposed measure should be checked and validated using detailed surge modeling. Boulos et al. (2005, 2006) provide a detailed transient flow chart that offers a comprehensive guide to the selection of components for surge control and suppression in distribution systems. Good maintenance, pressure management, an adequate disinfectant residual, and routine monitoring programs are also essential components of transient protection.
Flushing Water Mains
Flushing is one of the most ubiquitous activities of water utilities for both maintaining and recovering the integrity of distribution systems because it is the primary means by which to remove contaminated water from the system. It was discussed briefly in Chapter 4 in association with the cleaning and disinfection of water mains following pipe installation, repair, and replacement. It is a topic of Chapter 6, which focuses on water quality integrity, because flushing is routine in areas with repeat customer complaints about color, taste, or odor; in dead ends mains; and in storage facilities. Its importance with respect to maintaining hydraulic integrity is that flushing removes accumulated sediment and corrosion
products that reduce the hydraulic capacity of the pipe, improving the flow of water through the distribution system.
Flushing (discussed in greater detail in a subsequent section) is performed by isolating sections of the distribution system and opening fire hydrants (or flushing valves) to cause a large volume of flow to pass through the isolated pipelines so that a scouring action is created. Water is then discharged through a hydrant, which in turn removes any material buildup from the pipe. When flushing pipes, it is important to ensure that the flushing velocity is sufficient to suspend loose sediments. Flushing should continue until the water has cleared and disinfectant residual has reached normal expected levels. To minimize any negative environmental impacts (as flushed water may be high in suspended solids and other contaminants that can harm waterbodies), flushed water is normally discharged into sanitary or combined sewers or storm water management facilities. It is important to optimize flushing programs, as excessive flushing can waste significant volumes of water.
Operation and Design for Water Age Minimization
As discussed in Chapter 1, a primary reason for water quality problems within distribution systems is the advanced water age necessitated by the provision of adequate standby fire flow and redundant capacity. This requires that utilities use standpipes, elevated tanks, and large storage reservoirs, as well as
larger-sized pipes than would otherwise be necessary. The effect of designing and operating a system to maintain adequate fire flow and redundant capacity can result in long travel times and low velocities between the treatment plant and the consumer, which can be detrimental to water quality.
Brandt et al. (2004, 2006) have recently completed a two-volume study sponsored by the American Water Works Association Research Foundation to suggest ways to minimize water age (retention times) while at the same time controlling water quality degradation and providing the pressure and quantity constraints that are required to maintain water service. In particular, Brandt et al. (2006) have developed a diagnostic methodology by which a water utility can assess and then minimize water quality problems associated with excessive retention times. Best management practices for controlling retention time can generally be categorized into storage and network methods. Storage methods include adjusting pump schedules, reducing the operational top water level, removing storage tanks from service, and reconfiguring reservoir and storage tanks to avoid dead zones. Network methods include altering network valving patterns, installing time actuated valves, flushing (manual and automated), and abandoning and downsizing mains (Brandt, 2006).
An important aspect of hydraulic integrity maintenance is to ensure sufficient mixing and to minimize water age in storage facilities—issues which if not addressed can generate pockets of stagnant water with depleted disinfectant residual and associated water quality problems. Mixing will eliminate internal dead zones within a storage facility and prevent short-circuiting between the tank inlet and outlet. Completely mixed flow can be achieved by using a turbulent (high velocity) inlet jet, mechanical mixers, or hydraulic circulation systems. Controlling pumping rates and fill and discharge rates can also provide adequate intensities to achieve complete mixing. For example, Grayman et al. (2000) recommend that to avoid stratification in distribution storage facilities, the fill time should exceed the mixing time. A utility’s SCADA system can be used to monitor the real-time mixing intensity within a storage facility, and as such is useful for process control. It should be noted that utilities may be constrained in their ability to provide complete mixing due to the increased energy requirements.
Both poor mixing and improper tank discharge management can increase the residence time of water in a service reservoir. To combat this potential problem, frequent exercising of reservoirs (i.e., continuously mixing the water and making sure that fresh water replaces stagnant water) is required. Grayman et al. (2000) used various modeling techniques to develop a set of general guidelines for reducing water quality deterioration associated with inadequate mixing and excessive water age in distribution storage facilities. They reviewed the application of CFD, compartment, and physical scale models. A stand-alone model called CompTank is presented which provides a wide range of alternatives and allows the user to model water age and the concentrations of reactive or conservative substances over long time periods.
There is limited information about how to operationally reduce water age in an existing system while taking into account larger issues such as minimizing operational costs and maintaining the other aspects of hydraulic integrity, such as reliability of supply and adequate pressure for all water uses. At the present time, there is so much variability in the system parameters affecting distribution system operation that it is not possible, for example, to quantify the tradeoff between the risk of running out of water and the risk of delivering water of poor quality. This quandary is manifested in our inability to optimally maintain and operate storage facilities. The benefits of large storage tanks are not clear, nor is it easy to determine whether to remove a tank from service or reduce its volume. Answering such questions will require research that quantifies how various actions (such as removing a storage tank from service) will affect other aspects of hydraulic integrity (such providing fire flow and minimizing water age) within a given distribution system.
RECOVERING HYDRAULIC INTEGRITY
When a distribution system experiences high head losses, inadequate pressures or flows, high turbidity from scale loosening or resuspension of sediment, or low disinfectant residual and high bacterial counts from advanced water, there are several steps the utility should take. One of the first steps is to consider one or more of the standard techniques available to remove any loose sediment, biofilm, and tubercles that may be the cause of the problem. These procedures can restore most of the pipes’ original hydraulic capacity, and include conventional and unidirectional flushing, air scouring, swabbing, abrasive pigging, chemical cleaning, mechanical cleaning and lining (nonstructural, cement or epoxy applied linings), and structural lining. If the problems persist even after the application of these techniques, replacement of the pipes should be considered (see Chapter 4). A brief discussion of each technique follows.
It should be noted that to overcome increasing head losses and local deficiencies in system pressure and to increase the carrying capacity of water mains, increased levels of pumping are usually needed. This will result in increases in energy consumption and increased operational costs for a water utility.
Conventional flushing generally involves opening hydrants in a specific area of the distribution system until the water visually runs clear. While effective in quickly removing loose particles, this type of flushing is usually not effective in dislodging well-attached deposits and cannot remove scales and tuberculation. Because in a looped system the water will flow to the hydrant from
multiple mains and directions, it becomes very difficult to achieve the high-velocity flushing required to scour and remove deposits (as shown in Figure 5-4). As a result, some sediment and biofilm may not be removed, and the cleanup method requires a substantial quantity of water. In addition, because the dynamics of the entire distribution system are not considered, it is possible that the water used to the flush the system may come from a component that has not been previously cleaned. Therefore, sediments, detached biofilm, etc., may simply be transported from one part of the distribution system to another.
Unidirectional flushing involves the closure of valves and opening of hydrants to create a one-way flow in the water mains (see Figure 5-5). This increases the speed of the water flow so that the shear velocity near the pipe wall is maximized, producing a scouring action in the mains, effectively removing sediment deposits and biofilm. Flushing should start at a clean water source (e.g., pump station) and proceed outward in the system so that flushing water is drawn from previously flushed reaches. This ensures that clean water is always used to flush the mains. No special equipment is needed; however, substantial planning time is required to define the flushing zones, the valves and hydrants to be operated, the duration of the flush for each zone, the required velocities, and the sequence of operation. A hydraulic model of the distribution system can
greatly simplify and expedite the planning process, especially for estimating pipe flow rates, velocities, and flushing times. While more costly and time consuming than conventional flushing, unidirectional flushing is more effective and uses less water (Hasit et al., 2004). There are often long-term water quality benefits because deposits and water of questionable quality are actually removed rather than being re-routed to other parts of the distribution system.
Work done by Slaats (2001) demonstrated the velocities needed to entrain sediments, and these were within the range of velocities used for flushing. Carriere et al. (2002) showed that loose deposits could be removed by unidirectional flushing as a function of time, pipe material, and water characteristics. Gauthier et al. (1997) showed that loose deposits in a French system removed by flushing contained organisms including invertebrates, protozoa, and bacteria (although it should be noted that French distribution systems maintain no disinfectant residual such that their ecology is not representative of U.S. distribution systems). The abiotic constituents were primarily iron, volatile solids, calcium, aluminum, and other insoluble materials. Deposits flushed from four systems in the United Kingdom were all high in iron and manganese (Marshall, 2000).
Not all systems can or will routinely flush. There may be water restrictions that preclude flushing, and customers may be upset if they see water being
“wasted” while they are being told to conserve. Additionally, there is often a requirement that disinfectant residuals in the flushed water be neutralized, and this may be more complicated if chloramines are present compared to chlorine.
Air Scouring, Swabbing, and Abrasive Pigging
There is a long history of cleaning pipelines in order to remove accumulated material resulting from corrosion, improper pH adjustment, post precipitation of water treatment chemicals, and biofilm growth. Cleaning usually is a precursor to another process like lining or insertion rather than a process onto itself. This is due to the fact that cleaning potentially exposes unprotected metal pipe which would result in additional water quality problems.
Scouring, swabbing, and abrasive pigging are progressively more aggressive cleanup techniques that involve more specialized equipment and specialized skills. Although a few water utilities have implemented these methods using their own staff, typically these methods are contracted to specialty firms. Air scouring involves the continuous injection of filtered, compressed air into the pipe, along with a continuous but smaller flow of water. Given a continuous supply of water and air in the right proportions, discrete “slugs” of water are formed in the pipe and driven along by the compressed air at high velocity. The high velocity slugs tend to remove silt, sediment, loose matter, and debris from the base of the pipe. No disassembly of the pipe is necessary. Water scouring involves the insertion of a high-pressure water jet into the pipe to remove deposited materials. The water jet pressure can be adjusted to remove the deposits without damaging the piping material. The jet will back flush the deposited material to the insertion point in the pipeline. While jetting is very effective, it is limited to the length of the jetting equipment, which will result in frequent insertion points, and to small diameter pipes.
More aggressive techniques, such as swabbing and abrasive pigging, work to varying degrees in removing heavy sediment, biofilm, adherent material, tuberculation, and even very hard scale (Ellison et al., 2003). Swabbing involves driving cylindrical foam sponges (known as swabs) through pipes using water pressure. The swabs travel along the water main and scrub the scale encrustations and slime build-up from the inner pipe walls. Loosened debris and swabs are eventually flushed out at an exit point. Currently, pigging is used primarily if there are hydraulic problems in the water mains, i.e., to improve the “C” factor (roughness coefficient) of the pipes. It involves isolating a segment of the distribution system and passing a fluid-propelled object through the pipe. A styrofoam plug is often used as the “pig,” which is normally the same or slightly bigger in diameter than the water main and is shaped like a torpedo (Deb, 1990). Increasing sizes of pigs are passed through the pipeline to gradually remove deposits within the pipeline. The abrasiveness of the pig results in varying quality of water being discharged during the cleaning process. For instance, higher
concentrations of suspended solids normally follow the more abrasive pigs, as these scour the inner lining of the pipe.
Both pigging and swabbing can be difficult to implement because they require the removal of hydrants or the installation of new pipeline appurtenances (e.g., pig launching and receiving stations). Few water utilities have implemented these methods using their own staff, such that these methods are usually contracted to specialty firms.
Chemical Cleaning, Mechanical Cleaning, and Lining
Chemical cleaning to restore old pipes involves the recirculation in an isolated pipe section of proprietary acids and surfactants to remove scale and deposit, while mechanical cleaning is accomplished by dragged scrapers. Scrapers are devices that use springs to force blades against the wall of the pipe. As the device moves through the pipe, the blades scrape the material off the walls which can then be flushed from the pipe. These techniques are typically applied in the rehabilitation of older unlined cast iron pipes that have become scaled and tuberculated. In another example, a process using a cleansing solution of an organic oxide scavenger and muriatic acid circulated through an isolated section of distribution main worked effectively for small diameter pipelines (Estrand, 1995). Compared to air scouring and pigging, chemical cleaning is infrequently used due to the cost of chemicals and their proper disposal after cleaning.
It is common practice to reline a cleaned pipeline to protect the newly exposed metallic pipeline material. The most common technique is to use concrete mortar applied to the internal surface, a technology that has been used for over 50 years. Spray-on epoxy lining is a newer method that is especially useful when the water is low in hardness, which can cause a cement lining to deteriorate. Most recently, polyurethane lining is becoming a competitive alternative to concrete mortar lining especially in long pipelines with few service connections. This type of chemical lining on the inner surface of the pipe is referred to as nonstructural lining and does not increase the pipe’s structural integrity.
CONCLUSIONS AND RECOMMENDATIONS
Maintaining the hydraulic integrity of distribution systems is vital to ensuring that water of acceptable quality is delivered in acceptable amounts. The most critical element of hydraulic integrity is adequate water pressure inside the pipes. The loss of water pressure resulting from pipe breaks, significant leakage, excessive head loss at the pipe walls, pump or valve failures, or pressure surges can impair water delivery and increase the risk of contamination of the water supply via intrusion. In addition, slow moving water or changes in the flow regime (including flow reversals) and advanced water age can negatively impact finished water quality. Proper system design, operation, and mainte-
nance, along with monitoring and modeling, can help water utilities achieve a high degree of hydraulic integrity and reliability and extend the life of their distribution systems. The following conclusions and recommendations focus on the highest priority issues.
Water residence times in pipes, storage facilities, and premise plumbing should be minimized. Excessive residence times can lead to low disinfectant residuals and leave certain service areas with a less protected drinking water supply. In addition, long residence times can promote microbial regrowth and the formation of disinfection byproducts. From an operational viewpoint it may be challenging to reduce residence time where the existing physical infrastructure and energy considerations constrain a utility’s options. Furthermore, limited understanding of the stochastic nature of water demand and water age makes it difficult to assess the water quality benefits of reduced residence time. Research is needed to investigate such questions, as well as how to achieve minimization of water residence time while maintaining other facets of hydraulic integrity (such as adequate pressure and reliability of supply).
Positive water pressure should be maintained. Low pressures in the distribution system can result not only in insufficient fire fighting capacity but can also constitute a major health concern resulting from potential intrusion of contaminants from the surrounding external environment. A minimum residual pressure of 20 psi under all operating conditions and at all locations (including at the system extremities) should be maintained. The minimum value could be adjusted based on site specific conditions.
Where feasible, surge protection devices should be installed. Because these devices provide the only practical opportunity to prevent intrusion of contaminants due to low or negative pressure events, surge tanks should be considered at all pump stations (to dampen negative pressure waves) and other surge control devices at vulnerable locations in the system such as high points. This can be aided by a comprehensive surge analysis on a representative network model of the distribution system to select, locate, and size the most effective combination of surge protection devices. Although looped networks are generally less susceptible to objectionable pressure transients than single long transmission main systems, they must still be protected against low or negative pressure transients.
Distribution system monitoring and modeling are critical to maintaining hydraulic integrity. Hydraulic parameters to be monitored should include inflows/outflows and water levels for all storage tanks, discharge flows and pressures for all pumps, flows and/or pressure for all regulating valves, and pressures at critical points. An analysis of these patterns can directly determine if the system hydraulic integrity is compromised or if the system is operating as designed, or detect any unexpected operational anomalies. Calibrated distribu-
tion system models can calculate the spatial and temporal variations of flow, pressure, velocity, reservoir level, water age, and other hydraulic and water quality parameters throughout the distribution system. Such results can, for example, help identify areas of low or negative pressure and high water age, estimate filling and draining cycles of storage facilities, and determine the adequacy of the system to supply fire flows under a variety of demand loading and operating conditions.
American Water Works Association (AWWA). 1986. Introduction to Water Distribution Principles and Practices of Water Supply Operations. Denver, CO: AWWA.
AWWA. 2005. AWWA Manual M32: Computer modeling of water distribution systems. Denver, CO: AWWA.
Baribeau, H., N. L. Pozos, L. Boulos, G. F. Crozes, G. A. Gagnon, S. Rutledge, D. Skinner, Z. Hu, R. Hofmann, R. C. Andrews, L. Wojcicka, Z. Alam, C. Chauret, S. A. Andrews, R. Dumancis, and E. Warn. 2005. Impact of Distribution System Water Quality on Disinfection Efficacy. Denver, CO: AwwaRF.
Boulos, P. F., B. W. Karney, D. J. Wood, and S. Lingireddy. 2005. Hydraulic transient guidelines for protecting water distribution systems. J. Amer. Water Works Assoc. 97(5):111–124.
Boulos, P. F., K. E. Lansey, and B. W. Karney. 2006. Comprehensive Water Distribution Systems Analysis Handbook for Engineers and Planners. Second edition. Pasadena, CA: MWH Soft Pub.
Boyd, G. R., H. Wang, M. D. Britton, D. C. Howie, D. J. Wood, J. E. Funk, and M. J. Friedman. 2004. Intrusion within a simulated water distribution system due to hydraulic transients. 1: Description of test rig and chemical tracer method. J. Environ. Eng. 130(7):774–783.
Brandt, M., J. Clement, J. Powell, R. Casey, D. Holt, N. Harris, and C. T. Ta. 2004. Managing Distribution Retention Time to Improve Water Quality—Phase I. Denver CO: AwwaRF.
Brandt, M., J. Clement, J. Powell, R. Casey, D. Holt, N. Harris, and C. T. Ta. 2004. Managing Distribution Retention Time to Improve Water Quality—Phase II. Denver CO: AwwaRF.
Buchberger, S. G., J. T. Carter, Y. H. Lee, and T. G. Schade. 2003. Random Demands, Travel Times and Water Quality in Deadends. Denver, CO:AWWA.
Buchberger, S. G., and G. J. Wells. 1996. Intensity, duration and frequency of residential water demand. J. Water Resources Planning and Management 122(1):11–19.
Buchberger, S. G., and L. Wu. 1995. A model for instantaneous water demand. J. Hydraulic Eng. 121(3):232–246.
Carriere, A., B. Barbeau, V. Gauthier, C. Morissette, R. Millette and A.Lalumiere. 2002. Unidirectional flushing: loose deposits characterization in the test zones of four Canadian distribution systems. In: Proceedings of the AWWA Water Quality Technology Conference. Denver, CO: AWWA.
Cesario, L. 1995. Modeling, analysis and design of water distribution systems. Denver, CO: AWWA.
Clark, R. M., and S. G. Buchberger. 2004 Responding to a contamination threat in a drinking water network: the potential for modeling and monitoring. Pp 9.1-9.26 In: Water Supply Systems Security. L. W. Mays (ed.). New York: McGraw-Hill.
Clark, R. M., W. M. Grayman, S. G. Bucberger, Y. Lee, and D. J. Hartman. 2004. Drinking water distribution systems: an overview. Pp 4.1-4.2 In: Water Supply Systems Security. L. W. Mays (ed.). New York: McGraw-Hill.
Clark, R. M., and W. M. Grayman. 1998. Modeling water quality in drinking water distribution systems. Denver, CO: AWWA.
Clark, R. M., S. Panguluri, and R. C. Haught. 2004. Remote monitoring and network models: their potential for protecting U.S. water supplies. Pp. 14.1–14.22 In: Water Supply Systems Security. Mays, L. W. (ed). New York: Mc Graw-Hill.
Cromwell, J., G. Nestel, and R. Albani. 2001. Financial and economic optimization of water main replacement programs. Denver, CO: AwwaRF.
Deb, A. K, J. K. Snyder, J. J. Chelius, and D. K. O’Day. 1990. Assessment of Existing and Developing Water Main Rehabilitation Practices. Denver, CO: AwwaRF.
Ellison, D., S. J. Duranceau, S. Ancel, G. Deagle, and R. McCoy. 2003. Investigation of Pipe Cleaning Methods. Denver, CO: AwwaRF.
Estrand, C., A. Hicatt, and J. Ludwidg. 1995. Chemical cleaning process for water pipe systems. In: Proceedings of the Hydraulics of Pipelines Conference, ASCE, Phoenix, AZ.
Friedman, M., L. Radder, S. Harrison, D. Howie, M. Britton, G. Boyd, H. Wang, R. Gullick, D. Wood and J. Funk. 2004. Verification and Control of Pressure Transients and Intrusion in Distribution Systems. Denver, CO: AwwaRF.
Gauthier, V., C. Rosin, L. Mathieu, J. M. Portal, J. C. Block, P. Chaix, and D. Gatel. 1997. Characterization of the loose deposits in drinking water distribution systems. In: Proceedings of the AWWA Water Quality Technology Conference. Denver, CO: AWWA.
Grayman, W. M., L. A. Rossman, C. Arnold, R. A. Deininger, C. Smith, J. F. Smith, and R. Schnipke. 2000. Water quality modeling of distribution system storage facilities. Denver, CO: AwwaRF.
Gullick, R. W., M. W. LeChevallier, R. C. Svindland, and M. J. Friedman. 2004. Occurrence of transient low and negative pressures in distribution systems. J. Amer. Water Works Assoc. 96(11):52–66.
Hasit, Y. J., A. J. DeNadai, H. M. Gorill, S. B. McCammon, R. S. Raucher, and J. Whitcomb. 2004. Cost and Benefit Analysis of Flushing. Denver, CO: AwwaRF.
Karim, M., Abbaszadegan, M. and M. W. LeChevallier. 2003. Potential for pathogen intrusion during pressure transients. J. Amer. Water Works Assoc. 95(5):134–146.
Kirmeyer, G. J., M. Friedman, K. Martel, D. Howie, M. LeChevallier, M. Abbaszadegan, M. Karim, J. Funk, and J. Harbour. 2001. Pathogen intrusion into the distribution system. Report No. 90835. Denver, CO: AwwaRF and AWWA.
Lansey, K. E., and P. F. Boulos. 2005. Comprehensive Handbook on Water Quality Analysis for Distribution Systems. Pasadena, CA: MWH Soft Pub.
LeChevallier, M. W., R. W. Gullick, M. R. Karim, M. Friedman, and J. E. Funk. 2003. The potential for health risks from intrusion of contaminants into distribution systems from pressure transients. Jour. Water Health 1(1):3–14.
Levenspeil, O. 2002. Modeling in chemical engineering. Chemical Engineering Science 57: 4691–4696.
Mahmood, F., J. G. Pimblett, N. O. Grace, and W. M. Grayman. 2005. Evaluation of water mixing characteristics in distribution system storage tanks. J. Amer. Water Works Assoc. 97(3):74–88.
Marshall, G. P. 2000. Understanding and Preventing Discolored Water. UKWIR Report #01/DW/03/17. London: UKWIR Ltd.
Panguluri, S., W. M. Grayman, and R. M. Clark. 2005. Distribution system water quality report: a guide to the assessment and management of drinking water quality in distribution systems. Cincinnati, OH: EPA Office of Research and Development.
Slaats, N. (ed.). 2001. Processes Involved in the Generation of Discolored Water. Nieuwegein, the Netherlands: KIWA.
Walski, T. M., D. V. Chase, D. A. Savic, W. M. Grayman, S. Beckwith and E. Koelle. 2003. Advanced Water Distribution Modeling and Management. Waterbury, CT: Heastad Press.
Wood, D. J., S. Lingireddy, and P. F. Boulos. 2005. Pressure Wave Analysis of Transient Flow in Pipe Distribution Systems. Pasadena, CA: MWH Soft Pub.