Research On A Forecasting Model Of Metro Thermal Environment Based On Wavelet Analysis And Artificial Neural Network

With the development of metro network, the requirement of operating environment for metro operation is also getting higher and higher. Metro thermal environment is an important part of metro operating environment. Monitoring and predicting metro environment with appropriate control measures is important to improve the safety of metro operation and the comfort of passengers, so the metro environmental control has attracted the attention of operation and management department.In this paper, a forecasting model of metro thermal environment based on wavelet analysis and artificial neural networks is studied. The main contents are as following:1. Analysis of metro thermal environment and its change characteristics. The characteristics and regularity of metro thermal environment are analyzed; the parameters and data characteristics of the metro thermal environment are summarized.2. Research on trend forecasting model of metro thermal environment. According to the parameters of metro thermal environment and its data characteristics, a trend forecasting model of metro thermal environment based on wavelet analysis is proposed. Firstly, the data of metro thermal environment is preprocessed by wavelet de-noising. Then, the trend forecasting model of metro thermal environment is built in use of artificial neural networks.3. Case studies. Firstly, the trend forecasting model of metro thermal environment system in a big city based on the model presented in the paper is introduced. Then, the model of this paper is analyzed and verified in use of metro thermal environment data of the big city, which is acquired by the acquisition module of the system. The practicability and validity of the model are illustrated by the relevant indicators.Trend prediction of the metro thermal environment parameters based on wavelet neural networks has been applied to a certain station in the big city. At present, the operating condition is working well, and the prediction precision is high, it has been recognized by the operator. At present, some new methods are being tried to improve the forecasting efficiency and analysis ability of the model, in order to provide a better basis for metro environmental control and analysis.