Improving Contaminant Mixing Models For Water Distribution Pipe Improving Contaminant Mixing Models For Water Distribution Pipe Networks Networks Siri Sahib S. Khalsa Siri Sahib S. Khalsa University of Virginia University of Virginia Charlottesville, VA Charlottesville, VA [email protected] [email protected] Clifford K. Ho Clifford K. Ho Sandia National Sandia National Laboratories Laboratories Albuquerque, NM Albuquerque, NM [email protected] [email protected] Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy’s National Nuclear Security Administration under Contract DE-AC04-94AL85000. Problem Problem Development of New Mixing Model Development of New Mixing Model Objective: Predict concentrations resulting from two pipe flows intersecting at a junction Objective and Objective and Approach Approach Implementation and Data Testing Implementation and Data Testing Ongoing Research Ongoing Research Conclusions Conclusions Mixing in a cross junction is incomplete and physically bounded by predictions of the complete-mixing and bulk- mixing models Our new models have been implemented into EPANET, and results show significant improvement in contaminant transport predictions Employ computational fluid dynamics software to study contaminant mixing in cross junctions, develop a new mixing model, and implement it into EPANET Understanding and predicting solute transport through water distribution pipe networks are important to mitigate potential contamination events Much of the water distribution industry relies on the Environmental Protection Agency’s water quality simulation software (EPANET) However, EPANET incorrectly assumes contaminants mix instantaneously and completely in junctions, leading to potentially inaccurate transport predictions Use laboratory data to test and improve our new mixing model Flow Simulation Flow Simulation 3-D simulations with computational fluid dynamics software reveal incomplete mixing within junctions, contrary to the assumption of the current mixing model Equal Flow Rates Unequal Flow Rates • Bulk fluid momentum is retained in the mixing process • Our new mixing model honors bulk advective transport as a lower bound to the amount of mixing in a junction • A mixing parameter is used to scale the results between the complete-mixing and bulk-mixing model predictions Tracer Inlet Clean Inlet Tracer Inlet Clean Inlet The new mixing model was implemented into EPANET and used to predict contaminant concentrations at various sensor locations in laboratory pipe networks 3x3 Network: Tracer Inlet Flow Rate > Clean Inlet Flow Rate New Mixing Model Solution Original EPANET Solution Original EPANET Solution New Mixing Model Solution 3x3 Network: Clean Inlet Flow Rate > Tracer Inlet Flow Rate Predictions by EPANET implemented with the new mixing model agree with laboratory measurements when appropriate mixing parameters are used Sensor Sensor Concentrations in 3x3 Network Sensor Normalized Concentration Sensor Concentrations in 3x3 Network Clean Inlet Flow Rate > Tracer Inlet Flow Rate Source: http://toxics.usgs.gov/highlights/gw_ces sation.html Analysis of mixing behavior in other pipe-junction configurations U Junction Double-T Junction Inverse modeling to calibrate combined mixing model to pipe-network data Development of mixing parameter regressions as functions of different flow conditions Working with the Environmental Protection Agency to implement and distribute our new models Contaminated Water Clean Water Validation of combined mixing model using both large-scale and laboratory pipe-network data 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1 2 3 4 5 6 7 8 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1 2 3 4 5 6 7 8 Normalized Concentration Tracer Inlet Flow Rate > Clean Inlet Flow Rate 0 clean C 3 6 2 1 4 7 5 Sensor Locatio ns 8 1 tracer C 3 6 2 1 4 7 5 Sensor Locatio ns 8 0 clean C 1 tracer C 0 clean C 3 6 2 1 4 7 5 8 1 tracer C Sensor Locatio ns 3 6 8 2 1 4 7 5 Sensor Locatio ns 0 clean C 1 tracer C