Optimal Placement of Monitoring Sensors in Lakes

Abstract

Long-term surveillance of aquatic environments is a costly endeavor, thus a sound strategy is necessary to select the locations of monitoring stations to improve the quality of a monitoring system. This can be accomplished by optimizing the locations of a sensor network with respect to the hydraulic and fate and transport characteristics of the surface water system. It is expected that such an approach may improve the effectiveness of the monitoring system and also reduce the cost. To determine the optimal sensor placement pattern, the hydrodynamics of the target environment must be understood first. Based on this information, the contaminant transport behavior can be simulated. Since the hydrodynamics and the contaminant pathways in the surface water environment are complex, it is not possible to optimize a surface water network using simple optimizing methods. To solve this problem in a simplified form, a 2-dimensional hydrodynamic simulation model is developed using a finite element method. The best sensor locations are selected to minimize detection time of the contaminant presence. Genetic algorithm is used for the solution of the optimization formulation in selection of the best sensor network due to the nonlinear nature of the model. Further, these computation processes require a significant amount of computational time, thus parallel computing is adopted to reduce computation time.