基于BP人工神经网络的磨盘山水库水质评价  

Evaluation of water quality based on BP artificial neural for the Mopanshan reservoir

作　　者：庄清 杨旭[1] 杜崇[1] 王敏[1] ZHUANG Qing;YANG Xu;DU Chong;WANG Min(School of Hydraulic and Electric Power,Heilongjiang University,Harbin 150080,China)

机构地区：[1]黑龙江大学水利电力学院,哈尔滨150080

出　　处：《黑龙江大学工程学报》2022年第3期21-24,共4页Journal of Engineering of Heilongjiang University

基　　金：黑龙江省省属高等学校基本科研业务费项目(2020-KYYWF-1020)。

摘　　要：磨盘山水库水环境质量对哈尔滨市有重要意义。利用Matlab BP人工神经网络,并考虑到人工神经网络中的不足对其进行了改进。应用Powell-Beale共轭梯度法和traincgb算法进行网络训练,选取COD(化学需氧量)、pH值、溶解氧、NH_(3)-N、电导率5种评价因子,对水库水质进行评价。结果表明磨盘山水库水质等级为Ⅰ级。改进后的算法提高了网络的稳定性和精准性。The water environment quality of the Mopanshan reservoir is of great significance to Harbin.Matlab BP artificial neural network is used and the deficiencies in the artificial neural network is improved.The Prolak-Ribiere conjugate gradient method and the traincgb algorithm are used for network training and five evaluation factors of COD(chemical oxygen demand),pH value,dissolved oxygen,NH_(3)-N,and electrical conductivity are selected to evaluate the water quality of the reservoir.The results show that the water quality of Mopanshan Reservoir is gradeⅠ.It is worthy of reference in the water quality problems of similar reservoirs.

关 键 词：BP神经网络 共轭梯度法 评价因子 水质评价 磨盘山水库 

分 类 号：X824[环境科学与工程—环境工程]