基于U-D分解的卡尔曼滤波法在地下水污染源识别中的应用

采用U-D分解的方法对常规卡尔曼滤波进行改进，以提高反演结果的有效性和可靠性.借助假想算例，引入模糊集合理论，将污染羽以模糊集的形式表示，利用U-D分解的卡尔曼滤波方法不断更新复合污染羽，通过复合污染羽和单个污染羽间模糊集的对比更新潜在污染源权重，进而达到真实污染源位置识别的目的.算例结果表明该方法与常规卡尔曼滤波方法相比具有更高的可靠性和有效性，能更好的达到污染源位置识别的目的.

Application of U-D factorization-based Kalman filter to identify the groundwater pollution source

A improved Kalman filter, which was factorized by U-D method, was proposed to enhance the validity and reliability of the inverse strategy. The fuzzy set theory was introduced to represent contaminant plume in a hypothetical case. Then the weight of the potential source location was updated through the comparison of the composite plume updated by U-D factorization-based Kalman filter and the individual plume. The case study indicated that the U-D factorization-based Kalman filter has better reliability and validity, which can identify the location of pollution source better.