%0 Journal Article
%T DC-Link薄膜电容热分析与结温预测
DC-Link Film Capacitor Thermal Analysis and Junction Temperature Prediction
%A 李成夺
%A 王钦飏
%J Modeling and Simulation
%P 549-556
%@ 2324-870X
%D 2025
%I Hans Publishing
%R 10.12677/mos.2025.141051
%X DC-Link薄膜电容是电动汽车电驱系统中的一个重要组成部分，在反复充放电的过程中会导致电容发热，影响其使用寿命。本文基于ANSYS仿真软件对某型号DC-Link薄膜电容器进行温度场分析，结果表明，在高温环境中，电容器芯子中心处为温度最高点，而配备散热器后，最高温度点转移至远离散热器的外壳处，散热器能显著降低芯子温度。进一步分析揭示，电容瞬时温升主要由当前瞬时损耗和当前总体温升两个内在因素决定，基于此，采用LSTM神经网络对随机损耗下的电容器结温进行预测，并通过仿真数据验证了其预测精度。本文为电容器优化设计和动态工况下电容结温实时预测提供研究基础。
DC-Link film capacitor is an important part of electric vehicle electric drive system, and the repeated charging and discharging process will cause the capacitor to heat up, which affects its service life. This paper analyzes the temperature field of a certain model of DC-Link film capacitor based on ANSYS simulation software, and the results show that in a high-temperature environment, the highest temperature point is at the center of the capacitor core, and when equipped with a heat sink, the highest temperature point is shifted to the shell far away from the heat sink, and the heat sink can significantly reduce the core temperature. Further analysis reveals that the capacitor instantaneous temperature rise is mainly determined by two intrinsic factors, namely, the current instantaneous loss and the current total body temperature rise. Based on this, this paper adopts the LSTM neural network to predict the capacitor junction temperature under the stochastic loss, and verifies its prediction accuracy by simulation data. This paper provides a research basis for capacitor optimization design and real-time prediction of capacitor junction temperature under dynamic working conditions.
%K DC-Link薄膜电容，
%K 温度场，
%K LSTM神经网络，
%K 温度预测
DC-Link Film Capacitor
%K Temperature Field
%K LSTM Neural Network
%K Temperature Prediction
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=105161