基于广义回归神经网络的地铁屏蔽门通风量预测研究

Research on Ventilation Forecast of Metro Platform Screen Door Based on Generalized Regression Neural Network

为了对地铁屏蔽门系统通风量变化进行实时预测,本文引入了广义回归神经网络模型对地铁屏蔽门通风量数值进行实时预测仿真,最后引入回归评价指标对预测结果进行模型评估。仿真结果表明:利用广义回归神经网络进行屏蔽门通风量预测效果准确有效,训练集与测试集平均准确率分别为99.45%和99.69%,这为地铁环控系统节能优化提供了理论参考。

In order to make real-time prediction of the changes in the ventilation volume of the subway screen door system,this paper introduces a generalized regression neural network model to simulate the real-time prediction and simulation of the ventilation volume of the subway screen door,and finally introduces regression evaluation indicators to evaluate the prediction results.The simulation results show that the use of generalized regression neural network to predict the ventilation volume of the screen door is accurate and effective.The average accuracy of the training set and the test set are 99.45%and 99.69%,respectively.This provides a theoretical reference for the energy-saving optimization of the subway environmental control system.