Modeling Water Quality In Drinking Water Distribution Systems

Modeling Water Quality In Drinking Water Distribution Systems: Its Potential for Enhancing Water Security

Will Discuss Concern over the vulnerability of water distribution systems to security threats 2. Basics of water quality modeling and its application in drinking water networks 3. Two examples of water quality modeling for water security 4. The future use of water quality modeling as part of a decision support framework 1.

9/11 Raised Concerns About Critical Infrastructure in the US § § § Water supply was identified as critical infrastructure It is now general consensus that the vulnerability of drinking water networks systems to security threats is a major concern Utility industry has also recognized the importance of environmental monitoring in maintaining water security

Drinking Water Systems In The U. S. § § § There are 54, 000 community water systems in US serving 264 million people 79% of the population receives drinking water from large utilities (serving 10, 000 or more), representing 14% of the systems 21% of the population receives water from small utilities (serving less than 10, 000 people) representing 86% of the systems

U. S. Water Supplies Have Common Characteristics Water source § A lake, reservoir, river or ground water from an aquifer § Surface supplies generally have conventional treatment facilities and disinfection Ground water systems § May have full range of treatment technology but some practice chlorination only or do not disinfect at all Transmission systems § Tunnels; reservoirs and/or pumping facilities; and storage facilities Distribution system § Carrying finished water pipes to consumer

Distribution System is Most Vulnerable Part of Water System § Community water supplies designed to deliver water under pressure and most of the system capacity is reserved for fire fighting purposes § Could damage or destroy a tank or reservoir § Potential for the deliberate introduction of contaminants into a distribution system (back flow, cross connections) § Need to be able to predict contaminant transport pathways and to measure concentration of contaminants in networks § Cyber attack could also provide a serious threat to an utilities operations. However many SCADA systems are not connected to the Internet

Predicting Contaminant Movement In Drinking Water Distribution Systems § § Movement of water in distribution systems is complex The ability to predict movement is still relatively crude Also need to be able to predict changes in concentration of contaminants Few attempts to integrate monitoring and modeling

Contaminants May Be Conservative, or May Experience Decay or Growth § § Changes may take place in the bulk phase or at the pipe wall Quality may be influenced by: § Cross Connections § Failures at the Treatment Barrier § Transformations in the bulk phase

Water Quality Modeling Principles § Conservation of mass within differential lengths of pipe § Complete and instantaneous mixing of the water entering pipe junctions § Appropriate kinetic expressions for the growth or decay of the substance as it flows through pipes and storage facilities This change in concentration can be expressed by a differential equation of the form:

Where: – Cji is the substance concentration mass/ft 3) at position x and time t in the link between nodes i and j – vij is the flow velocity in the link (equal to the link’s flow divided by its cross-sectional area in ft/sec – kij is the rate at which the substance reacts within the link (mass/ft 3/sec)

Storage tanks can be modeled as completely mixed, variable volume reactors where the change in volume and concentration over time are: Where: - Vs is the volume (ft 3) of the tank - Cs is the concentration in tank s

The following equation represents the concentratio leaving the junction and entering a pipe:

Where:

Model Interaction Water quality models are generally piggy backed on hydraulic models. Hydraulic Model Flows and velocities Water Quality Model Water quality results

Will Use EPANET To Illustrate the Need For Integrating Modeling and Monitoring § First example will be the application of EPANET to North Marin Water District in California • Illustrates the linkage between monitoring and modeling § Second example is the waterborne outbreak in Cabool Missouri in 1990 • Forensic application of modeling

Modeling of Contaminants § § § First field study using EPANET in North Marin California Modeled chlorine residual propagation and THM formation Applied to two source system

North Marin Water System § § Located near Novato , California Serves over 50, 000 people Virtually no rainfall during warm summer months Uses two sources of dramatically different quality

EPANET Applied to Waterborne Outbreak in Gideon Missouri in 1993 Salmonella contamination occurred in municipal tank due to failure of hatches and vents 2. Taste and odor complaints caused water officials to start flushing program 3. Out of population of approximately 1000 people, 440 became ill and 7 people died 4. Used model to track outbreak and identify source 1.

Municipal Water System in Gideon Was Old and in Disrepair § § § Tuburculation and corrosion in the distribution pipes was a major problem Two municipal tanks Another tank was located on the property of the Cotton Compress which was the major employer in the area

Waterborne Outbreak 1. 2. 3. 4. 5. On November 29, 1993 Communicable Disease Coordinator for the Missouri DOH became aware of two high school students with culture confirmed Salmonellosis Within two days five additional patients were hospitalized with confirmed salmonellosis Missouri Department of National Resources was informed that DOH suspected a water supply link to outbreak DNR samples were positive for fecal coliform City of Gideon was required to issue a boil water order

Homes with Cases Between 11/23 – 11/28 and 11/29 – 12/10 1994 in Gideon, Missouri

Comparison of Early Confirmed Cases and Salmonella Positive Sample Versus Penetration of Tank Water During First Six Hours of Flushing Program

Current Status of Water Quality/Hydraulic Models § Increasingly sophisticated § Applied to exposure studies • ATSDR study on contaminated ground water § Much research into modeling changes in water quality • Formation of DBPs and Chlorine Residuals § Tank Mixing Models

EPA Research in Real Time Monitoring Systems § § § First EPA effort was development of sensors for temperature, chlorine residual, fluoride and nitrate data with Battelle Asked to assist during MCL violation in Washington DC Initiated research on development of sensors and probes for chlorine residual, p. H and temperature using pipe loops Applied to DC water system Future efforts should focus on integrating modeling and monitoring

Summary and Conclusions § § Water systems have been classified as critical infrastructure Identified as potentially vulnerable Contaminant Propagation Can be Modeled and there are various models available EPANET is a public sector model that has become widely used

§Summary and Conclusions § § EPA has been conducting research into sensor development Applied to operation of small package plants Extended to chlorine residual monitoring in Washington DC system Future research will focus on integrating remote sensing and water quality modeling