Efficient Contamination Events Sampling for Sensors Layout Design

Sensors network design models involve exhaustive simulations of pollution event scenarios. To satisfactory evaluate a sensors design it is necessary to simulate the impacts of as much as possible contamination events, a process which becomes impractical even for a small sized water distribution system. Previous studies showed that using different random event samples yields diversed sensor designs. The objective of this study is to develop and demonstrate a methodology for minimizing the events sample size for sensors layout design while keeping its `representative' quality. The suggested algorithm finds a small contamination events matrix that minimizes the difference between its normal expectation μ and variance σ², and other pollution event matrices characteristics μ and σ². The effectiveness of the suggested approach is demonstrated on an example water distribution system, using a representative sensor design objective.