基于径向基函数模型的优化方法在地下水污染源识别中的应用

Optimization method of identification of groundwater pollution sources based on radial basis function model

采用一种基于径向基函数的替代模型代替地下水溶质运移模型,将其作为约束条件嵌入污染源识别的优化模型中,通过遗传算法对优化模型进行求解.最后通过一个假想例子评估优化模型的性能.研究表明:污染源泄漏量识别结果的平均绝对误差为1.00g/s,误差较小,计算时间为51min,耗时较少,因此,基于径向基函数模型的优化方法有效地避免了优化模型求解过程中多次调用模拟模型造成的巨大计算负荷,获得了较为准确的计算结果,是一种有效的地下水污染源识别方法,能够用来求解地下水污染源泄漏量.

In the process of optimization method for identification of groundwater pollution sources, computational load resulted from multiple invocations of the numerical simulation model of groundwater is huge. This paper used a surrogate model based on radical basis function to replace groundwater solute transport model, and the surrogate model was embedded in optimization model as a constraint. The optimization model was solved by a genetic algorithm. The performance of the model was evaluated in a hypothetical example. The mean absolute error of release rate of pollution sources was 1.00g/s and the calculation time was 51minutes, so the error and the time consumption were small. Therefore, the optimization method based on radical basis function model can effectively avoid the huge computational load and obtain more accurate results. It is an effective method for identification of groundwater pollution sources, which can be used to solve the release rate of groundwater pollution sources.