基于BP神经网络的串联塘水质综合评价模型

介绍BP神经网络模型的结构原理和算法,分析了在建模过程中可能出现的问题并提出了解决方案,根据国家环保总局发布的地表水环境质量标准和串联塘修复水质的实测数据,建立了具有较好泛化能力的三层BP神经网络,为评价污水净化效率及串联塘的合理规划提供了科学的依据.同时表明,由于BP神经网络模型高度非线性以及输出结果以连续函数形式表达,其对水质进行的综合评价更客观.

THE MODELS OF COMPREHENSIVE ASSESSMENT OF SERIES PONDS WATER QUALITY BASED ON BP NEURAL NETWORK

The construction,theory and method of BP neural network were introduced in this paper.The difficulties during modeling were analyzed and their solutions were thus put forward.Based on Surface Water Environment Quality Standard recommended by Environmental Regional Central of China,using the measured data of series ponds,three-layer BP network possessed the capacity of higher generalization was set up.It supplied references to evaluate the efficiency to purify slops and properly position these series ponds.Because of the nolinear of BP network and the output by continuous function,the results showed that this way was more objective.