河流突发污染的污染物浓度动态校正方法

Dynamic pollutant concentration correction method for river sudden pollution

为降低河流突发污染事故的影响,提高下游污染物预测精度、提高预测实时性,结合一维水质模型、卡尔曼滤波及改进的网格寻优算法,综合考虑支流的影响,研究河流突发污染事件中污染物扩散情况的动态预测方法.分析一种改进的网格寻优算法并利用历史数据校正模型参数;借助水质模型构造状态方程引入污染物浓度观测值;运用卡尔曼滤波动态校正预测结果,并在预测过程中考虑支流的影响.在理论研究的基础上,设计基于风浪水槽的污染物模拟扩散实验,对比分析采用不同预测方法的污染物峰现时间、峰值浓度及相对误差.实验结果表明,不同的预测方法所求得的峰现时间相对误差总体相当;采用多步动态校正预测和考虑了支流影响的校正预测方法预测峰值浓度得到的相对误差明显降低.

A dynamic pollutant concentration correction forecasting method for sudden pollution accidents in rivers was explored,in order to reduce the influence of sudden pollution accidents and predict pollutant concentrations in the downstream accurately and timely.The one-dimensional water quality model,Kalman filter algorithm and improved grid optimization algorithm were investigated considering the influence of tributaries.An improved grid optimization algorithm using the historical monitoring data was developed to correct the model parameters;the state equation based on one-dimensional water quality model was established;the Kalman filter algorithm was developed to correct the predicted concentration based on the updated observed data considering the influence of tributaries.On the basis of theoretical research,experiments based on the wave flume were set up to simulate the dispersion of pollutants,and to analyze peak time,peak concentrations and the relative errors of different predict methods.Theoretical and experimental results show that the relative errors of predicted peak-time values are at similar levels by different prediction algorithms;the dynamic correction forecasting method contributes to improve the prediction accuracy and promote the prediction capability of algorithms.