基于数据驱动的功率MOSFET可靠性预测综述  

Data-driven Reliability Prediction Review of Power MOSFETs

作　　者：高乐 任默 刘超铭 GAO Le;REN Mo;LIU Chaoming(School of Astronautics,Harbin Institute of Technology,Harbin 150001,P.R.China;Equipment Department of Military Space Force,Beijing 100094,P.R.China)

机构地区：[1]哈尔滨工业大学航天学院,哈尔滨150001 [2]军事航天部队装备部,北京100094

出　　处：《微电子学》2024年第4期644-652,共9页Microelectronics

基　　金：国家自然科学基金(12275061)。

摘　　要：随着大数据和计算技术的发展,数据驱动的可靠性预测方法在电子和电力系统领域正被越来越广泛地应用。对国内外功率场效应晶体管(MOSFET)数据驱动的可靠性预测方法进行介绍和分析,揭示该方法从经典统计方法到先进机器学习方法的演变过程,对于统计学方法,介绍了高斯过程回归、自回归积分移动平均模型等经典统计学方法,以及不断优化和扩展模型以进行改进的统计学方法;对于机器学习方法,集中探讨了如支持向量机、人工神经网络以及当前不断发展的深度学习模型,最后,总结发展趋势并探讨未来研究方向。This study provides a comprehensive review and analysis of reliability prediction methods for power metal oxide semiconductor field effect transistors(MOSFETs)both domestically and globally,elucidating the evolution from classical statistical methods to advanced machine learning techniques.Statistical methodologies,such as gaussian process regression,autoregressive integrated moving average(ARIMA)models,and other classical statistical methods,were investigated,with an emphasis on continuous model optimization and extension.Regarding machine learning approaches,the investigation focused on techniques such as support vector machines(SVM),artificial neural networks(ANN),and continuously evolving deep learning models.Finally,development trends were analyzed,and potential future research directions discussed.

关 键 词：功率MOSFET 可靠性预测 机器学习 数据驱动 

分 类 号：TN432[电子电信—微电子学与固体电子学]