基于BP神经网络的分流制雨水管中旱流流量预测模型

Prediction model of separate discharge dry weather flow based on BP neural network

年径流总量的准确性会影响控制率的计算,而监测的年径流总量中存在旱流流量影响,准确预测雨水管中旱流流量并进行识别,能得到更加切合实际径流控制率的统计结果。将BP神经网络模型运用到旱流流量的预测上,可以提高其预测的准确性。以重庆市璧山区2017~2019年的典型下垫面雨水管出口旱流流量实测数据为例,选取月份、星期、小时、温度、湿度等指标作为预测参数,建立了BP神经网络预测模型,并运用该模型对各种典型下垫面的分流制雨水干管中的旱流流量进行预测与辨析。结果表明:模型预测值的相关性系数(R2)都在0.70以上,模型精度较高,能够准确地预测分流制雨水管出口中的旱流流量,可以提高雨水径流量计量的准确性。

The accuracy of annual total rainfall will affect the calculation of control rate.Dry weather flow effect the annual total rainfall monitored.The more accurate the dry weather flow in the rainwater was predicted and identified,the more realistic statistical results of runoff control rate can be obtained.Applying the BP neural network model to predict the dry weather flow can improve the accuracy of the prediction.This paper takes the measured data of the dry weather flow current of the typical underlying surface rainwater pipe from 2017to 2019in the Bishan District of Chongqing as an example.The BP neural network prediction model was established.It used the months,weeks,hours,temperature,humidity and other indicators as prediction parameters.And the model was also used to predict and analyze the dry weather flow in the separate discharge of various typical underlying surfaces.The result shows that the correlation coefficient(R2)of the model predictive value is above 0.70,indicating that the model has high precision and can accurately predict the dry flow in the outlet of the splitting rainwater pipe,which can improve the accuracy of the rainwater flow metering.