汉江水质综合评价的BP网络方法

以往的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排除主观因素的干扰.构造了一个多因子水质综合评价的3层BP网络模型,以生化耗氧量、氨氮和总磷为评价因子,对汉江武汉段水质的富营养化状况进行了评价.结果表明,该方法评价结果更为客观,符合实际情况.

ANN method for comprehensive evaluation of water quality in Hanjiang River

In the comprehensive evaluation of water quality, the results obtained by using those traditional methods for determining the weighting factors are normally quite subjective. The artificial neural network(ANN) method now can effectively exclude the interference of some subjective factors in determining the weighting factors. A three-layer back-propagation(BP) ANN model is constructed for the comprehensive evaluation of the multi-factor water qulity. Taking BOD_5, NH_3-N and TP as the evaluation factors, the application of this model to the eutrophication evaluation in Wuhan section of the Hanjiang River shows that the results of this ANN method indeed are more objective than those obtained by the traditional methods and are very consistent with the observation.