Hailiang Shen

HYDRAULIC CALIBRATION FOR A SMALL WATER DISTRIBUTION NETWORK

Procedures utilized for hydraulic model calibration for the C-Town network, which is the first step prior to a model being useful for operation and maintenance, and water quality model construction, is described. For the C-Town network as provided by BWCN, seven groups of parameters are identified, namely, pipe roughness, elevation, leakage coefficient, pump curve, base demand, pattern value, and pump and valve control. The calibration is formulated as an optimization problem, aimed at minimizing the discrepancy between observed and simulated data. A toolkit is developed within VC++ 2008 Express to solve the optimization problem, by employing a flexible genetic algorithm library GAlib as the optimization engine, and EPANET as the hydraulic solver. Monte Carlo simulation is applied for sensitivity analyses to identify sensitive parameters to feed the calibration process. It is shown the seven groups of parameters have similar sensitivity and all feed to calibration process. The tank levels are relatively well calibrated, comparing with the pump station flow rates.

Application of parallel computing in data mining for contaminant source identification in water distribution systems

Contaminant source identification (CSI) procedures are drawing increasing attention due to the possibility of accidental and/or deliberate contaminant intrusion into water distribution systems. However, uncertainties that exist in the modeling have the potential to dramatically impact the capabilities of CSI procedures. Nodal demand uncertainties, as they influence false negative and false positive rates of contaminant detection, are examined. A procedure to quantify the false negative rate is provided, and the false positive issue is shown to be related to a parameter ‘m’. Addressing the false positive and negative issues is demonstrated as feasible due to the use of parallel computing in a super-computer, which reduces the elapsed time for 150 scenario simulations from 37.5 hrs to only 15 min in the case study. By increasing the number of scenarios in the database for CSI through the use of a super-computer, the opportunity exists to decrease the false negative rate and reduce the number of false possible intrusion nodes.

DATA MINING AS A TOOL TO IDENTIFY CONTAMINANT SOURCES IN WATER DISTRIBUTION SYSTEMS

With growing concerns related to potential contamination ingress via backflow and/or terrorist threats to drinking water, useful methodologies are needed to assist in identifying locations from which ingress may have occurred. An efficient data mining approach in conjunction with a maximum likelihood procedure is described, which provides a means to identify the location and timing of an intrusion event, based on limited sensor data. The effectiveness of the data mining method is demonstrated using a case study network where it takes only approximately 5 minutes to identify an injection event using 5 sensors in a 285 node water distribution network. The effectiveness of the method is demonstrated using a number of alternative applications of the data mining methodology, which ensures the robustness of the methodology in locating potential sources of the ingress event.

Sensor Placement Under Nodal Demand Uncertainty for Water Distribution Systems

The governing requirements for security for water supply systems in Canada are provincially mandated, not federally regulated. The national strategy works on the basis of collaborative efforts from federal, provincial, territorial, and critical infrastructure sectors to provide the infrastructure resiliency.

AN ARCGIS TOOL FOR RAPID CONTAMINANT SOURCE IDENTIFICATION

The contaminant source identification (CSI) problem consists of five parts: i) identification of possible intrusion nodes (PINs), ii) quantification of the probability of each PIN as the true intrusion node, iii) identification of priority nodes (PNs) from PINs which are upstream of important nodes such as schools, hospitals, and multi-story buildings, iv) quantification of the priority degree of each PN, which indicates importance for emergency response, and iv) identifying whether a PIN connects aging pipe(s) which possess high potential for contaminant intrusion. The locations of PINs, PNs, and the connection of aging pipe(s), require involvement of geographic information system (GIS). An ArcGIS tool namely GIS-CSI is developed to integrate various data sources into ArcGIS feature class, and implement the CSI algorithm developed in Shen et al. (2009b) to solve the CSI problem, and to display the CSI outputs, i.e., PINs and PNs locations, their probabilities as intrusion nodes and priority degree respectively. In the tool demonstration, immediately after each sensor alarm, GIS-CSI can run the CSI algorithm within 5 min, and export the outputs to the water distribution system map to greatly facilitate emergency response.

Utility of Supercomputers in Trace-Back Algorithms for City-Sized Distribution Systems

A contaminant source identification procedure intended to protect water distribution systems has to be both rapid and able to incorporate uncertainties, when identifying possible intrusion nodes (PINs). PIN identification has two major issues, the false-negative rate (failure to identify the true ingress location) and the false-positive issue (falsely identifying a location which is not the true ingress location). A data mining procedure is described and applied, which involves mining an off-line-built database, to select PINs that possess first detection times within ±m from the online sensor first detection time. The “m” value is a statistical characterization of the array of events of the offset values between online sensor first detection time under uncertainty and the one corresponding to the same intrusion event stored in the off-line database; with “m” selected, issues of controlling false negatives and positives are addressed. The approach described herein is made possible through the power of parallel computing in supercomputers, which demonstrates huge potential by simulating scenarios simultaneously. The online data mining procedure, i.e., the PIN identification, is integrated into a geographic information system toolkit for rapid emergency response. In the case studies, simulation of scenarios is reduced linearly to the number of processors applied. Results show that increasing the number of scenarios in the database can provide input to compute the “m” value, always reduce the false-negative rate of each sensor, and usually reduce the number of false-positive PINs.