基于复合型进化算法的地下水污染反演模型

Inverse Model Based on SCE-UA Optimization Algorithm for Unknown Groundwater Pollution Identification

污染源位置和污染物排放浓度的快速确定直接决定着地下水污染的有效治理及修复,这属于地下水反演问题。通过充分分析地下水污染反演问题,耦合地下水流模拟程序MODFLOW、溶质运移模拟程序MT3DMS和优化算法SCE-UA,设计了一种模拟-优化(S/O)反演模型。通过实例验证,反演结果表明:提出的网格遍历GT算法可以自动验证潜在污染区内所有可能的污染源位置组合方式;与传统地下水污染反演模型相比,S/O模型不但能够适用于稳定流条件,而且适用于非稳定流条件;所开发的S/O模型对于多污染源分别在稳定流和非稳定流下的反演均有非常高的精度,能够准确反演污染源位置及污染物排放浓度。

Remediation strategies can be successfully carried out only if the correctly identification of groundwater pollution source locations, concentrations and release history and this is a groundwater inverse issue. Based on the sufficient analysis for the structure of groundwater inverse issue, combined with groundwater flow model MODFLOW, solute transport model MT3 DMS and SCE-UA optimization algorithm, it designed a Simulation/Optimization(S/O) model. The case studies show that this S/O model can effectively and accurately determine the groundwater contaminant characteristics(1) a GT(Grids traversal) algorithm is proposed which can automatic search all possible combinations of pollution source locations;(2) comparing to the traditional identification model of groundwater pollution, the proposed S/O model can be applied both in steady-state flow and transient flow conditions;(3) case studies show that the proposed S/O model has high accuracy both for multiple pollution sources under steady-state flow and transient flow conditions and can effectively identify the pollution source locations and concentrations.