| 符东,吴雪菲,易珍言,陈永灿.沱江水质模糊综合评价及主要污染物的预测研究[J].农业环境科学学报,2020,39(12):2844-2852. |
| 沱江水质模糊综合评价及主要污染物的预测研究 |
| Fuzzy comprehensive assessment of water quality and prediction of main pollutants in the Tuo River |
| 投稿时间：2020-06-28 |
| DOI：10.11654/jaes.2020-0730 |
| 中文关键词: 沱江 模糊综合评价 BP神经网络 |
| 英文关键词: Tuo River fuzzy comprehensive assessment BP neural network |
| 基金项目:国家自然科学基金项目（51809219）；四川省科技计划项目（2018JZ0001，2019YFG0143）；水沙科学与水利水电国家重点实验室开放基金项目（2019-B-02） |
|
| 摘要点击次数: 1737 |
| 全文下载次数: 1539 |
| 中文摘要: |
| 为准确掌握沱江水质状况，探明沱江主要污染物，对沱江水质进行了模糊综合评价和BP神经网络预测。使用沱江31个监测断面2018年1月—2019年10的逐月水质数据，通过筛选优化评价因子，对各断面水质进行模糊综合评价。对沱江水质进行主成分分析以确定主要污染源和主要污染因子，并构建了BP神经网络对主要污染因子进行预测。研究发现，沱江有9个断面水质符合Ⅰ类水质标准，其余22个断面水质均为Ⅴ类水且在沱江上游、中游和下游均有分布。TN浓度在所有监测断面中均超过了Ⅳ类水质标准，其中27个断面的TN浓度超过了Ⅴ类水质标准。使用上游断面水质数据构建的BP神经网络预测下游断面的TN浓度，平均相对误差为2.041%。研究表明，沱江受TN的污染较为严重，其主要污染源为农业面源和工业废水，同时根据构建沱江其他断面的BP神经网络模型可实现对沱江TN浓度的准确预测。 |
| 英文摘要: |
| To accurately investigate the water quality of the Tuo River and to predict its main pollutants, the fuzzy comprehensive assessment model and the BP neural network were used, respectively. By selecting and optimizing the evaluation factors, a fuzzy comprehensive assessment of the water quality was conducted using the monthly water quality data of 31 monitored sections of the Tuo River from January 2018 to October 2019. Principal component analysis of water quality in the Tuo River was carried out to identify the main pollution sources and pollutants, and BP neural network was constructed to predict the main pollution factors. The results showed that the water quality of 9 sections of the Tuo River met Class Ⅰ water quality standards, and the remaining 22 sections were of Class Ⅴ water quality, and were distributed along the upper, middle, and lower reaches of the Tuo River. The concentration of TN exceeded Class Ⅳ water quality standards in all monitoring sections, of which 27 sections exceeded Class Ⅴ water quality standards. BP neural network constructed using the water quality data of the upstream section successfully predicted the TN concentration of the downstream section, with an average relative error of 2.041%. The results implied that the Tuo River was significantly polluted by TN, with non-point agricultural and industrial wastewater being the main sources of pollution. Additionally, according to this work, BP neural network of other sections of the Tuo River can be built to effectively predict the TN concentration in the Tuo River. Our findings can provide a reference for the comprehensive management and pollution control of the Tuo River basin. |
| HTML
查看全文
查看/发表评论 下载PDF阅读器 |
|
|
|
