For sudden water pollution incidents in rivers and lakes, the ability to quickly identify the pollution source is of great importance for providing early accident warning and implementing emergency control measures. Based on Bayesian reasoning, a variable-fidelity surrogate-differential evolution adaptive metropolis optimization(DREAM) optimization model for coupled inversion process is established in the posterior space of the pollution source..In order to verify the effectiveness of the algorithm, this paper takes lake A as the research area, and gives a hypothetical water pollution emergency, the pollution source location, release time and released mass of water pollutants suddenly released into water bodies were determined according to the method proposed in this paper. The results show that in the case of ensuring the accuracy of calculation, the algorithm can accelerate more than 200 times and effectively improves the computational efficiency of the traditional method for obtaining the source information of sudden water pollution events.

对于河流和湖泊中突然发生的水污染事件，快速识别污染源的能力对于提供早期事故预警和实施应急控制措施至关重要。在贝叶斯推理的基础上，在污染源的后部空间建立了耦合耦合反演的可变保真度-代-差分进化自适应都会优化（DREAM）优化模型。为了验证该算法的有效性，本文采用以A湖为研究区域，并给出了假设的水污染紧急情况，根据本文提出的方法确定了突然释放到水体中的水污染源的位置，释放时间和释放量。结果表明，在保证计算精度的情况下，