基于PSO算法优化的RBF神经网络水厂混凝投药控制  

Coagulation dosing control in waterworks based on RBF neural network optimized by PSO algorithm

作　　者：庹婧艺 徐冰峰[1] 徐悦[1] 喻岚 王雪颖 郭露遥 TUO Jingyi;XU Bingfeng;XU Yue;YU Lan;WANG Xueying;GUO Luyao(Faculty of Civil Engineering and Mechanics,Kunming University of Science and Technology,Kunming 650500,China)

机构地区：[1]昆明理工大学建筑工程学院,昆明650500

出　　处：《工业安全与环保》2022年第9期83-86,共4页Industrial Safety and Environmental Protection

基　　金：国家自然科学基金(41801211)。

摘　　要：在PSO算法优化后的RBF神经网络的基础上,建立了误差反向传播的非线性高维映射水厂投药量动态模型。优化后的RBF模型,平均相对误差降低25.04%,最大相对误差降低1.101,相较于原始模型具有更快的迭代收敛速度,对丰水期和枯水期的源水水质都有较高的模拟精度,减少了水厂在每个处理单元的机理模拟,直接得出投药与源水水质的映射关系,对不同工艺的水厂也具有较好的适应性。A nonlinear high-dimensional mapping water plant dosing dynamic model with error back propagation was established on the basis of the RBF neural network optimized by PSO algorithm. The optimized RBF model had an average relative error reduction of 25.04% and the maximum relative error reduction of 1.101,which had a faster iterative convergence speed than the original model,and had a high simulation accuracy for the raw water quality during the flood and dry periods,reduced simulation of the mechanism of each processing unit of the waterworkst,and directly derived the mapping relationship between dosing and raw water quality,which also had good adaptability for water plants with different processes.

关 键 词：水厂投药 RBF神经网络 PSO粒子群 给水处理 

分 类 号：TU991.22[建筑科学—市政工程]