基于BP网络方法的长江水质综合评价

Comprehensive Evaluation of Water Quality of Yangtze River Based on BP Network

水质综合评价因子权重的确定方法使得评价结果带有较强的主观性,神经网络方法则可以有效地排除主观因素的干扰。文章构造了一个多因子水质综合评价的3层BP网络模型,以溶解氧(DO)、高锰酸盐指数(CODMn)、氨氮(NH3-N)为评价因子,对长江17个监测点的水质进行了综合评价。其中1个站点为Ⅰ类水质,7个站点为Ⅱ类水质,6个站点为Ⅲ类水质,2个站点为Ⅳ类水质,1个站点为V类水质。

The method determining the weighting factor of comprehensive evaluation of water quality makes the evaluation result very subjective, however, the neural network method can effectively rule out the disturbance of subjective factors. This paper constructs a three - layer BP net model which is a multi - factor comprehensive evaluation of water quality whose evaluation factors are potassium permanganate index, dissolved oxygen and ammonia -nitrogen. The comprehensive evaluation has been done in Yangtze River＇s 17 monitoring sites and the results indicates that the water quality of one site is Grade Ⅰ,seven sites are Grade Ⅱ ,six are Grade m ,two are Grade Ⅳand one is Grade Ⅴ.