【摘要】 湔江流域的西河水库是四川省彭州市的主要灌溉和饮用水源,该流域非点源污染已严重影响了彭州市的饮用水安全。为正确地掌握西河水库污染物来源,合理评价西河水库水质及水环境质量对保障彭州市居民饮水安全、维护彭州市社会稳定有着重要意义。采用Elman神经网络模型对西河水库进行水质与水环境质量评价。结果表明:①西河水库水质呈现由库尾到坝前递减的趋势,说明西河水库的水质自净能力明显。②西河水库2014～2017年大部分断面的水质等级均有所提高,其中断面1-1(主坝右坝肩)的水质最好,是西河水库理想的取水口位置。③西河水库在水文年内呈现枯水期水质好、丰水期水质差的规律,其中4～9月的水质普遍较差,这也说明水库上游来水携带的污染物是西河水库水环境恶化的主因。④构建的Elman模型适用性较高,可以满足西河水库水质等级评价。该模型计算时间耗费不到10 s,运算效率远高于传统的灰色聚类法,具有一定的参考意义。更多还原

【Abstract】 Xihe Reservoir is the main irrigation and drinking water source for Pengzhou City, Sichuan Province, and the non-point source pollution in the Lujiang River Basin has seriously affected the drinking water security of Pengzhou City. Correctly figuring out the pollutants source in Xihe Reservoir and reasonably evaluating the water quality and water environment of Xihe Reservoir are of great significance for ensuring the drinking water security and maintaining the social stability of Pengzhou City. In this paper, the Elman neural network model was used to evaluate the water quality and water environment quality of Xihe Reservoir. The results showed that:(1) The water quality of Xihe Reservoir decreased from the end of the reservoir to the front of the dam, indicating that the water quality self-purification ability of Xihe Reservoir was obvious.(2) The water quality levels of most sections of Xihe Reservoir had been improved from 2014 to 2017. The section 1-1(right abutment of main dam) had the best water quality and was the ideal intake location of Xihe Reservoir.(3) The water quality of Xihe Reservoir was good in dry season and bad in wet season, and the water quality from April to September was generally poor, which also showed that the pollutants carried by the upstream water of Xihe Reservoir were the main cause of the deterioration of the water environment of Xihe Reservoir.(4) The Elman model constructed in this paper had high applicability and could satisfy the water quality level evaluation of Xihe Reservoir. In addition, the calculation time of the model was less than 10 seconds, and the calculation efficiency was much higher than that of the traditional grey clustering method.更多还原