Author(s)

J. Guan, M. M. Aral & M. L. Maslia

Abstract

Water quality in a water-distribution system is a health concern. When a contaminant is released at an entry point, it rapidly spreads within the waterdistribution system. The identification of contaminant source locations and their release histories are data necessary for exposure analyses and security assessment studies. This data can be characterized based on the measured contaminant distribution data at selected monitoring locations. Topologic-based methods are generally inefficient in assessing this problem. Here a new approach is proposed to solve this problem. This approach is based on a water-distribution system simulator and optimization analysis. EPANET is used to generate simulated concentrations at selected monitoring locations by specifying release histories of potential contaminant sources under known hydraulic operational patterns. The optimization model is then used as a predictor to estimate the release histories of these contaminant sources based on the similarity of response between simulation and measured data at monitoring locations. This information exchange can be developed as a closed-loop convergent algorithm. To test the proposed algorithm, water-distribution systems serving the Dover Township, NJ, area in 1962 and 1996 are used and findings are discussed. Keywords: contaminant source, EPANET, monitoring locations, optimization, reduced gradient method, release history, water-distribution system.

Keywords

contaminant source, EPANET, monitoring locations, optimization, reduced gradient method, release history, water-distribution system.