摘要
近年来,政府相关部门虽然对地表水加大了治理力度,基本遏制河流水质恶化的势头,但是,突发环境污染事件仍然时有发生,对人体健康、生态安全造成了重要影响。利用水质在线监测仪虽然可以实时监测水质变化,但是智能化程度低,为此亟需采用先进手段实现对河流水质的预测预警并提前进行防范,最大程度降低类似藻类暴发等事件带来的损失。文中以嘉兴市河道水质为主要对象,开展水质预测模型研究,具有一定的实际应用价值。采用基于ARIMA自回归积分滑动平均模型与改进的BP神经网络算法相结合的方法进行水质预测的建模,研究水质数据和气象数据包含的线性关系和非线性关系,建立水质预测组合模型,并通过模型进行水质电导率、溶解氧、总磷、总氮、高锰酸盐、氨氮的预测;通过理论分析及试验对比,基于ARIMA自回归积分滑动平均模型与BP神经网络算法构建的模型,在水质预测方面比单纯使用传统的ARIMA模型具有更高的精度,各指标的MRE(平均百分比误差)、RMSE(均方根误差)均有很大程度的减小,提供了更科学、更准确的河流水质指数预测方法。
In recent years,although relevant government departments have strengthened the treatment of surface water and basically stopped the deterioration of river water quality,sudden environmental pollution incidents still occur from time to time,which has an important impact on human health and ecological safety.Although online water quality monitor can monitor water quality changes in real time,it has a low level of intelligence.Therefore,water quality prediction and pollution alarming is necessary to be applied to prevent deterioration of river water quality and minimize the loss caused by events such as algae outbreaks.The paper takes water quality of Jiaxing River as the main object to carry out the research of water quality prediction model.This research has certain practical application value.In the paper,a method based on the combination of autoregressive integrated moving average model(ARIMA)and improved BP neural network algorithm is used to model water quality prediction.The linear relationship and nonlinear relationship included water quality data and meteorological data are studied.The combined model of water quality prediction is established.Water quality variables predicted by the model including conductivity,dissolved oxygen,total phosphorus,total nitrogen,permanganate,ammonia nitrogen.The integration methods of ARIMA and BP-NN show a better prediction accuracy than ARIMA alone in terms of theorical analysis and experimental comparison,and MRE(mean percentage error)and RMSE(root mean square error)of each index have been greatly reduced,providing a more scientific and accurate prediction method of river water quality index.
作者
顾杰
王嘉
邓俊晖
王荣昌
GU Jie;WANG Jia;DENG Junhui;WANG Rongchang(Zhejiang Jiake Information Technology Co.,Ltd.,Jiaxing 314000,China;Key Laboratory of Yangtze Aquatic Environment〈MOE〉,College of Environmental Science and Engineering,Tongji University,Shanghai 200092,China)
出处
《净水技术》
CAS
2020年第6期73-82,共10页
Water Purification Technology
基金
国家水体污染控制与治理科技重大专项(2017ZX07206-001)
“水体污染控制与治理”科技重大专项成果
嘉兴市水污染协调控制与水源地质量改善项目
区域水环境质量改善综合调控系统与平台建设课题。
关键词
BP
神经网络算法
ARIMA
自回归积分滑动平均模型
水质预测
BP neural network algorithm
ARIMA autoregressive integrated moving average model
water quality forecasting
