基于拟牛顿法的二维水污染扩散源溯源定位

Tracing the Origin of Two-dimensional Water Pollution Diffusion Source Based on Quasi-Newton Method

针对传统水污染溯源方法精度较低、耗时较长等缺点,提出了一种基于多节点数据融合的二维水污染源溯源定位方法,即使用多节点采集电导率数据,通过拟合出的污染物电导率和浓度关系将观测点的电导率值转换为浓度值,根据节点位置建立理论值模型,并结合观测值与理论值构建的差值函数,采用拟牛顿法求得污染源关键参数,实现污染源的定位追踪与污染物浓度分布模型的重构。算法仿真与实际测试表明,所提算法定位误差约为4%,平均响应时间约为1.6s,性能优于已有算法。

Aiming at the shortcomings of lower precision and time-consuming,this paper proposes a two-dimensional method on the basis of multi-node data fusion for the positioning and tracking of water pollution source.According to the relationship between conductivity and concentration of pollutants,the conductivity data collected by multiple nodes will be converted to concentration values.Based on the difference function between the observed values and the theoretical values,the key parameters of pollution source are solved using quasi-Newton method.And then the distribution model of pollutant concentration is reconstructed.Experimental simulation and practical test show that the performance of the proposed method outperforms the existing algorithms due to the fact that the relative localization error approximates 4%and the average response time is about 1.6 s.