Influence Analysis Of Air Cooling System Parameters Of Energy-Fed Traction Power Supply Equipment Based On Data Driven

As an effective solution to the energy consumption of regenerative braking in urban rail transit,medium voltage energy-feed device has been widely used in urban rail transit system in China.Its core component,large capacity converter,has become the weak link of medium voltage energy-feed traction power supply system to cope with the high frequency and heavy load of energy transmission under different working conditions during its long worktime.It is of great significance to carry out life prediction and reliability estimation of IGBT and other components in converter to improve the intelligent level of equipment operation and maintenance and reduce the operation and maintenance cost.Aiming at the existing problems and key technical points of mining the correlation influence between medium voltage energy fed traction power supply device and air cooling system parameters,considering that the relationship between the converter lifetime and other parameters of the system is difficult to be deduced from the theoretical point of view,this paper proposes a method of converter life prediction and reliability estimation based on deep learning network using sensor measurement of historical operation data.The data processing and feature extraction methods of data-driven deep learning method in the scenario of urban railway energy-feed power supply system are studied,and a reference point temperature prediction model suitable for long-time prediction is proposed;The influence of life correlation factor of converter like air cooling parameters on equipment operation and life is studied and analyzed;A reliability estimation method of equipment and system based on Monte Carlo random and reliability block diagram is proposed.The main research problem and works are as follows(1)In order to obtain and analyze the IGBT junction temperature fluctuation of lowfrequency in long-time profile and solve the problem of stability and accuracy of temperature prediction in a long time scale,a thermoelectric coupling model and deep learning network are established about IGBT junction temperature and reference point temperature.Firstly,according to the thermoelectric equivalent theory,the Simulink simulation model of thermal resistance network is established according to the power loss calculation,and the junction temperature change of internal components in the chip is obtained through the thermoelectric coupling iterative simulation of measured data;Then,according to the prediction requirements of data-driven learning model,a deep neural network is established,and a network training and prediction framework with feedback regulation structure is proposed to predict the reference point temperature curve.The changes of reference point temperature and junction temperature curve parameters based on typical day are obtained,verified and analyzed.(2)In order to explore the influence of the temperature parameters of the converter air cooling system on the working state and service life of the equipment under complex working conditions,based on the actual measured data,the data processing and feature extraction process based on the data-driven method are carried out to form a data set,and the relationship between the starting parameters of different given air cooling system and the cumulative damage degree of power module IGBT is obtained and analyzed.Firstly,the status and requirements of device operation status acquisition and monitoring in traction power supply system are analyzed;Then,the data preprocessing and transformation of traction power supply system are carried out according to the historical data.The features are extracted and the data set is established through sliding window average and polynomial piecewise fitting;Then,considering the influence of the correlation between the parameters of the converter and the air-cooled system,the prediction is carried out under different given air-cooled system startup thresholds,and the variation relations of the parameters such as the thermal load and predicted life of the air-cooled system and IGBT are obtained and analyzed,so as to provide a basis for the optimization of relevant parameters of the equipment.(3)In order to obtain the overall reliability estimation of power devices and the system including air-cooled system,the lifetime estimation analysis method of power devices based on Monte Carlo process and the system reliability analysis method based on reliability block diagram are proposed.Firstly,the thermal load of the junction temperature curve is analyzed.Through the selection of life model and cumulative damage calculation,the damage estimation and life prediction results of IGBT power module in the system are obtained;Then,through the equivalent calculation of static parameters,the device life distribution model and failure probability curve obeying Weibull distribution are obtained based on Monte Carlo random process;Finally,the reliability is estimated from power unit to converter unit and air cooling system.Provide theoretical basis for the maintenance of medium voltage energy-feed traction power supply system,so as to ensure its operation reliability,and provide suggestions for equipment selection and maintenance scheme.