摘要
以水动力计算为基础,依据正向浓度概率密度与逆向位置概率密度的耦合关系,构建了以污染源位置和释放时间为参数的相关性优化模型,并利用微分进化算法实现了模型的求解。在此基础上,根据污染物正向浓度概率密度函数构建优化模型确定了污染物强度。耦合概率密度方法将突发水污染溯源问题转换为两个最小值优化问题,原理简单,求解方便,实现了污染源参数的解耦。在求解优化模型时,将梯度概念引入到微分进化算法中,提高了搜索效率。将所提出的溯源方法运用于仿真实验案例以及南水北调中线应急示范工程实例,结果表明,模型溯源效果好、模拟精度高,溯源结果对于河渠突发水污染事件应急处置具有一定指导意义。
With the coupling relationship between forward concentration probability density and backward lo-cation probability density,a correlation optimization model is established for contamination source locationand event initial time after the hydrodynamic computation,and the differential evolution algorithm is ap-plied in solving the optimization model. Following that,another optimization model is established to deter-mine the contamination mass. The coupled probability density function method succeeds in transforming thesource identification problem into two minimum optimization problems that are easy to understand and conve-nient to solve. In addition,it decouples the relationship between the contamination mass and source loca-tion or initial time. With the＂gradient＂concept,the differential evolution algorithm is improved for a bet-ter search efficiency in this paper. Applied in an artificial testing case and the emergency demonstrationproject in the middle line of South-North Water Diversion Project,the model produces good source identifi-cation results with high precision. The results provide important guidance to the emergency management inriver sudden water contamination accidents.
出处
《水利学报》
EI
CSCD
北大核心
2015年第11期1280-1289,共10页
Journal of Hydraulic Engineering
基金
国家自然科学基金资助项目(51409282)
国家科技重大专项项目(2012ZX07205)
关键词
耦合概率密度函数
溯源
微分进化算法
参数解耦
优化模型
梯度
coupled probability density function
source identification
differential evolution algorithm
parameter decoupling
optimization model
gradient
