基于小波神经网络的天然水体溶解氧预测研究

Study on DO Prediction of Natural Water Body Based on Wavelet Neural Network

建立基于小波神经网络的预测模型,以不同时间滞差和影响因子组合作为输入变量,对海河流域四个监测断面的溶解氧浓度进行短期预测.结果表明,基于溶解氧历史数据的小波神经网络预测模型精度更高,可用于天然水体的水质预测,为水质管理提供更客观的参考和依据.

The prediction models based on wavelet neural network were established and applied to four certain monitoring sections of Haihe basin. Different combinations of water quality parameters were set as input variables to predict dissolved oxygen. The results demonstrate that models with historical data of the target variable are better fitting with the real data, and have a higher accuracy. Therefore, it will provide significant decision support for water protection and water environment treatment.