| Resumo: | Groundwater contamination plumes characterization is a very hard task to perform, requiring usually a large number of sampling sites. In this article a method to optimize a monitoring network for plume detection and delimitation is proposed. It is assumed that a prior extensive sampling campaign was made, and only a few sampling sites must be included in the optimal monitoring network. The objective function incorporates the prior knowledge about concentration variability, in the form of its density function, and also a measure of spatial coverage (space-filling method), in order to best distribute the stations over the field. The method was applied to a synthetic case-study with 160 sampling locations, and a final optimal monitoring network with 40 stations was obtained. Simulated annealing optimization algorithm was used to solve this very difficult combinatorial problem, which has more than 8,6x1037 possible solutions).
|
|---|