基于BP神经网络的大伙房水库水质综合评价

Comprehensive Evaluation of Water Quality in Dahuofang Reservoir Based on BP Neural Network

选取我国地表水环境质量标准(GB3838-2002)作为学习样本,建立了水质综合评价的3层BP神经网络模型,选取溶解氧、生化需氧量、高锰酸盐指数、氨氮、总氮、总磷6个指标为评价因子,在MATLAB平台中对大伙房水库7个主要监测断面进行水质综合评价。结果表明:入库河流中,浑河和苏子河水质级别为Ⅲ级,污染较严重,主要为氮素污染,社河断面水质虽为Ⅱ级,但有接近Ⅲ级的风险。库区内各断面及出库口水质均为Ⅱ级,总体水质较好,但总氮含量仍偏高。

Surface water quality standards(GB3838-2002) as learning samples to establish a three-layer BP neural network model for comprehensive evaluation of water quality,and had been selected six indicators, including dissolved oxygen, biochemical oxygen demand, permanganate index, ammonia nitrogen, total nitrogen, total phosphorus were selected as evalution factors to evalute water quality of seven major monitoring sections of Dahuofang reservior in MATLAB platform. The results showed that in storage rivers, water level of Hun River and Suzi River were grade Ⅲ, pollution was serious, mainly nitrogen pollution, water quality of She River section was grade Ⅱ, but had the risk to be nearly grade Ⅲ. The water quality of reservior area and the outbound mouth were grade Ⅱ. In general, water quality is good, but the total nitrogen content remains higher.