In order to solve the problem of IGBT aging failure caused by the cyclic impact of thermal stress and electrical stress when working in a complex environment, a network model combining Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) was proposed to predict the IGBT life. Collector - emitter turn-off transient voltage is selected as the failure characteristic parameter, and a CNN-LSTM hybrid model is built. In order to accelerate the training speed of the network, the activation function uses the ELU function, and the Adam algorithm is used to train the network, so as to realize the prediction of the failure characteristic parameter data. Through experimental comparison with other time series prediction models, it is verified that the hybrid model in this paper can better realize the IGBT life prediction, and also provides a certain reference value for the life prediction of other power electronic devices.
IGBT Life Prediction Based on CNN1D-LSTM Hybrid Model
Lect. Notes Electrical Eng.
International Conference on Electrical and Information Technologies for Rail Transportation ; 2021 October 21, 2021 - October 23, 2021
Proceedings of the 5th International Conference on Electrical Engineering and Information Technologies for Rail Transportation (EITRT) 2021 ; Kapitel : 66 ; 660-668
2022-02-19
9 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
IGBT Life Prediction Based on CNN1D-LSTM Hybrid Model
British Library Conference Proceedings | 2022
|Aviation risk prediction based on Prophet–LSTM hybrid algorithm
Emerald Group Publishing | 2023
|LSTM LSTM-based future threat prediction method and apparatus
Europäisches Patentamt | 2021
|Traffic situation prediction method based on LSTM model
Europäisches Patentamt | 2024
|