Susceptibility of PWS to Negative Pressure Transients Kala Fleming, PhD VA AWWA Research Committee Seminar Monday, October 22, 2007
Pump Station 5-min Pressure Recording 50 40 Pressure (psi) 30 20 10 0 6/22/2004 6/23/2004 6/24/2004 6/25/2004 6/26/2004 6/27/2004 6/28/2004 Time 2
DS Monitoring at 1 Reading per sec pump start-up pump shutdown 3
Closer Look at Negative Pressure Profile Negative for > 16 sec; as low as 10.1 psi (-69 kpa) Gullick et al. 2005. J. Water Supply & Technol. AQUA 54(2): 65-81. 4
Why Do Pressure Transients Matter? 1 2 3 sewer main leaking pipe External Pathogens Leaking Pipes Microbial Risk? Transient Low Pressure 5
Presentation Overview Overview of Transient Pressure: How do negative transients occur? Evolution of a transient pressure wave Findings of AWWARF Project #3008 IL Case Study: What locations are impacted when the largest pump station looses power? Which mitigation approach works best? Microbial Risk Assessment 6
Sources of Transient Pressures Service interruptions Power failure Main breaks Sudden change in demand Flushing operations Opening and closing a fire hydrant service interruptions routine operations demand change Routine distribution system operation Pump startup and shut down Valve operation: open/close Any sudden changes in flow 7
Transients influenced by fluid properties Fluid Density water is heavy, large forces required to change flow Fluid Compressibility water not easily compressed, small mass imbalances cause large forces 8
Transient Pressures from Unsteady Flow power loss at pump velocity change pressure wave ΔH = (c / g) ΔV ** only applicable for simple pipeline ** ΔH = instantaneous pressure head change downstream of pump c = wave speed g = acceleration ΔV = change in velocity Note: a wave is a disturbance that transmits energy and momentum from one point to another through a medium without significant displacement of matter between the two points http://www.kettering.edu/~drussell/demos/waves/ wavemotion.html 9
450 ft Pressure Wave in Single Pipeline HGL-pumping HGL-steady state Wave front 7 sec after power failure 9 sec after failure 12 sec after failure Reflection increases pressure after ~ 14 sec ΔH HGL 7 sec after power failure 232 ft h L minimum head envelope during power failure ΔH = (c / g) ΔV pump runs down in 7 seconds Wave speed is 3,500 ft/s or ~0.66 mile/sec 4.6 miles in 7 seconds g = 32.174 ft/s 2 & V = 2 ft/s 0 5 10 DISTANCE (mile) Adapted from Thorley 2006. Fluid Transients in Pipelines. 10
Pressure Wave in Single Pipeline 600 PRESSURE HEAD (feet) 500 400 300 pump run down in 7 seconds 200 additional headloss until reflected wave approaches 100 0 20 40 60 80 TIME (seconds) 11
Negative Pressure Profile In more complex systems, reflections occur with changes in diameter, changes in pipe material and at dead ends or other discontinuities Negative for > 16 sec; as low as 10.1 psi (-69 kpa) Gullick et al. 2005. J. Water Supply & Technol. AQUA 54(2): 65-81. 12
Transient Analysis Pressure pulses are generated when flow conditions change from one steady state to another Pipeline plays a relative passive role, primarily transmitting disturbances from point to point Boundary conditions (devices and connections at the end of each line) play the crucial role in determining the character and nature of system response and propagation 13
Hydraulic Modeling Used to track pressure wave initiation, propagation, reflection
Modeling is Important! If you can model a system, i.e. describe its behavior using mathematical equations, then you can predict future behavior. Key Benefits: Identify Problems Optimize System Operation Make Informed Decisions 15
Model power loss at a pump station model is desktop representation of real system use model to understand how pressure and flow vary in the system Flow Key Flow < 100 gpm Flow > 100 gpm Pressure Key negative pressure start with steady state or EPS model The term transient describes unsteady flow. continuity and momentum equations used to solve unsteady flow problems 0 to 20 psi pressure > 20 psi 16
Findings from AwwaRF Project # 3008
Project # 3008 Overview 16 participating systems Variables: system size: 0.1 39 mgd number of pumped sources ( 1 to 29) pressure zones (1 to 24) topography/elevation (flat, moderate, hilly) distribution storage facilities (0 18 floating tanks) Surge relief features 18
Project # 3008 Significant Findings Systems with steady state or EPS models already have the basics to assess potential for transient pressures! In the absence of surge mitigation at pump stations, all distribution systems were susceptible to low/negative pressure fluctuations System susceptibilities ranged from 1% to 98% water velocity, number of floating storage facilities, number of source inputs and system configuration influence system vulnerability Velocities greater that 3 ft/s downstream of pump stations increase the risk of low/negative transient pressures 19
Storage Reduces Susceptibility Percent Nodes with Negative Pressure 60% 50% 40% 30% 20% 10% 0% at time of max flow to storage at time of max flow from storage R 2 = 0.9 0 20 40 60 80 100 120 Miles of Main per Floating Storage 20
Other factors System Size Surface vs Ground System Config. Smaller systems showed increased susceptibility 5 of 6 systems with < 10 mgd system delivery drew negative pressure in greater than 35% of the system with complete loss of pumping power Groundwater systems may have an increased susceptibility to low/negative pressure transients Hilly distribution systems (> 150 ft elevation difference) were less susceptible Systems with more floating storage facilities were less susceptible to negative pressures Locations at or near dead ends were more susceptible to negative pressures A few systems showed 21
IL Case Study Using modeling to prevent low/negative pressures after a power outage
IL Pressure Regulations State is currently enforcing the maintenance of pressure greater than 20 psi under all flow conditions. If pressure is less than 20 psi for even one second, a Boil Water Notice must be issued. 23
IL Water System Impacted By Regs System fed by surface water and has a relatively flat topology Primary pump station has a capacity of 30 MGD Under 2006 max day conditions, HGL at plant varied between 899 ft and 906 ft (corresponding to 64 to 72 psi) Primary pump station has unstable power supply. Currently no floating storage in system 24
More Storage = Less Transient Pressure Percent Nodes with Negative Pressure 60% 50% 40% 30% 20% 10% 0% at time of max flow to storage at time of max flow from storage R 2 = 0.9 0 20 40 60 80 100 120 Miles of Main per Floating Storage 25
Pressure Monitoring Required 26
Rationale for Selecting Monitoring Locations? 27
Low Pressure Measured 28
How to Proceed? Quick Fix Lease a generator that operates 24/7 Long Term Use model to assess extent of transient low pressures and examine range of solutions 29
Determine Susceptible Locations 30
What type of surge mitigation? Option 1 Two 20,000 gal hydropneumatic tanks $ 0.6 million Option 2 UPS sized to support 9 pumps $ 1.8 million Option 3 One 30,000 gal hydro tank & one 1MG elevated tank at Location A $ 1.5 million Option 4 One 30,000 gal hydro tank & one 1MG elevated tank at Location B $ 1.5 million Do nothing 24/7 generator $ 30,000 per month 31
Hydropneumatic Tanks compressor compressor air air water water water leaves tank to maintain pipeline pressure pipeline under steady-state conditions pipeline experiencing downsurge 32
Hydropneumatic tanks as a surge mitigation option hydropneumatic tanks installed on 4/5/05 33
Transients in Distribution Systems study of unsteady flow of liquids begins in mid-19 th century 1 Focus on transients in transmission mains 2 Is occurrence in smaller piping significant? 3 What is impact of transients on water quality? 1 Pre-2000 2 2000-2006 3 2006 future AW research demonstrated brief periods (20-50 sec) of low & negative pressure in several systems Characteristics that increase vulnerability to negative pressures investigated What are typical intrusion volumes? What is final concentration near nodes? Assess Microbial Risk 34
Microbial Risk Assessment Necessary. Provides logical approach to determine if transients can cause sufficient intrusion to impact the health of water consumers.
Microbial Risk Assessment 1 2 3 sewer main leaking pipe External Pathogens Leaking Pipes Microbial Risk? Transient Low Pressure 36
We have Leaks distribution water loss (%) = volume distribute d - (volume billed + volume 100 volume distribute d unbilled but authorized ) West South Midwest Northeast > 500,000 100,001 500,000 50,001 100,000 Distribution System Water Loss (Median Range, 25 th 75 th Percentile) Source: AWWA 2005 Benchmarking Performance Indicators for Water and WasteWater Utilities *121 Participants* 10,000 50,000 < 10,000 0 2 4 6 8 10 12 14 16 Percent 37
We have pathogens near pipe Overall 63% (20/32) of samples were positive for viruses: enteroviruses (Sabin strain), Norwalk, and Hepatitis A virus 100 80 Water Soil % Occurrence 60 40 20 0 Total Coliform Fecal Coliform Clostridium Bacillus Virus RT-PCR Phage 38
Microbial Risk Assessment frequency of power outages There is quite a bit of uncertainty in determining risk posed by intrusion microbes outside distribution system piping intrusion volumes pathogen concentrations at intrusion nodes after a power outage microbial risk for customer pipe flows; consumption patterns theoretical probability distribution U.S.EPA annual acceptable microbial risk level is 10-4 39
Sustained Power Loss (>3 min) J-45 0.7 gal intrusion Total Intrusion Volume = 7.4 gal (28.0 L) Intrusion occurs at 54 demand nodes 22 nodes (41%) have intrusion volumes of 0.1 gal (0.4 L) or greater Highest intrusion volume near customers was 0.74 gal (2.8 L) 40
Power restored 2 seconds after power loss Intrusion Volumes J-45 0.26 gal intrusion Total Intrusion Volume = 3.8 gal (14.4 L) Intrusion occurs at 53 demand nodes 15 junctions (28%) have intrusion volumes of 0.1 gal (0.4 L) or greater Highest intrusion volume was 0.26 gal (1.0 L) 41
Estimate Dilution Factors 1L Intrusion @ Node J-45 Duration = 16s 38 L avg flow before transient period = 36 gpm V o C o = V f C f V o = 1L V f = 38 L Dilution factor @ J-45 = 2.6 X 10-2 0.26 gal intrusion 42
Estimate Dilution Factors 0.04 L Intrusion @ Node J-181 Duration = 2 s 10 gal avg flow before transient period = 79 gpm V o C o = V f C f V o = 0.04 L V f = 10 L Dilution factor @ J-181 = 4.0 X 10-3 0.01 gal intrusion 43
Intrusion could be responsible for coliform positive samples J-45 0.26 gal intrusion 30 MPN/100 ml J-613, J-682 & J-683 > 1000 MPN/100 ml ~.04 gal (1.5 L) intrusion 106 MPN/100 ml For intruded volume, assume total coliforms = 1.6 x 10 3 MPN/100mL & fecal coliforms = 5 x 10 2 MPN/100mL 44
New AwwaRF Project Managing Distribution System Pressures to Protect Water Quality 1. Assess microbial intrusion risk Determine daily exposure Use dose response model to determine risk of infection 2. Conduct utility survey to gauge how utility managers use pressure management to protect water quality 3. Monitor pressure & water quality continuously for two months, in six locations, in four different water systems 4. Develop Best Practices for Managing Distribution System Pressures Pressure monitoring Disinfectant residual maintenance Hydraulic modeling 45
Risk Management Through Pressure Management
Pressure Management Approach Distribution System Evaluation customer demands peak hour of max day hydraulic modeling assessments fire flows power loss at all pump stations **surge model required** Are pressures less than 20 psi at any customer locations? Can system produce required fire flows while maintaining pressure greater than 20 psi? Do any locations temporarily draw pressure less than 0 psi? measure pressure and disinfectant residual levels at low pressure locations measure pressure and disinfectant residual levels at low pressure locations Review Surge Mitigation Options: system: hydropneumatic tank elevated storage uninterruptible power pump operation 47
Recap of Key Ideas Transient pressures occur Important research in water systems questions need to be addressed: Hydraulic modeling and pressure monitoring are What are the health risks posed by intruded water? important assessments that should be conducted to determine if low/negative transient pressures occur in your system How effective are disinfectant residuals? Do chlorine and chloramine provide the same level of protection from transitory contamination? 48
Intrusion References Fleming K.K. and M.W. LeChevallier. 2007. Susceptibility of Distribution Systems to Transitory Contamination. Drinking Water Research. Vol 17, No 2. AwwaRF, Denver, CO. Fleming K.K., R.W. Gullick, J. P. Dugandzic and M.W. LeChevallier. 2006. Susceptibility of Distribution Systems to Negative Pressure Transients. American Water Works Association Research Foundation, Denver, CO. Friedman, M., L. Radder, S. Harrison, D. Howie, M. Britton, G. Boyd, H. Wang, R. Gullick, M. LeChevallier, D. Wood. And J. Funk. 2004. Verification and Control of Low Pressure Transients in Distribution Systems. AWWA Research Foundation. Denver, CO. Gullick, R.W., M.W. LeChevallier, J. Case, D.J. Wood, J.E. Funk, and M.J. Friedman. 2005. Application of pressure monitoring and modeling to detect and minimize low pressure events in distribution systems. J. Water Supply & Technol. AQUA 54(2): 65-81. Gullick, R. W., M. W. LeChevallier, R.S. Svinland, and M. J. Friedman. 2004. Occurrence of Transient Low and Negative Pressures in Distribution Systems. J. Amer. Water Works Assoc. 96(11):52 66 Karim, M, M. Abbaszadegan, and M.W. LeChevallier. 2003. Potential for Pathogen Intrusion During Pressure Transients. Journal AWWA, Vol. 95, No. 5, pp. 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. AWWA Research Foundation and American Water Works Association. Denver, CO. Walski, T.M. and T.L. Lutes. 1994. Hydraulic Transients Cause Low-Pressure Problems. Journal AWWA, 86(12):24-32. 49
Acknowledgements Project funding provided by AwwaRF and by the utility subsidiaries of American Water. Contact Information Kala K. Fleming, PhD Environmental Engineer American Water 1025 Laurel Oak Road Voorhees, NJ 08043 USA phone: (856) 309-4556 fax: (856) 782-3603 e-mail: kala.fleming@amwater.com 50