一种基于关断电压的IGBT模块结温监测方法
A method for monitoring junction temperature of IGBT module based on turn-off voltage
兰尉尹 1崔巍 2陈文轩 1邹盈巧 2李佳诚 3葛兴来2
作者信息
- 1. 西南交通大学 利兹学院,四川 成都 610031
- 2. 西南交通大学 电气工程学院,四川 成都 610031
- 3. 西南交通大学 物理科学与技术学院,四川 成都 610031
- 折叠
摘要
实现准确的结温监测对增强绝缘栅双极晶体管(IGBT)模块的可靠性、延长器件寿命至关重要.文章提出了一种基于关断电压的IGBT模块结温监测方法,具有不受负载电流干扰的特性.该方法首先验证了关断电压作为温敏电参数的合理性;其次,采用深度神经网络(DNN),排除关断电压对负载电流的依赖,达到不同工况下保持准确结温预测的目的;最后,通过单相脉宽调制(PWM)进行试验验证.结果显示,该结温监测方法误差在±5℃的范围内,这表明利用DNN优化结温监测是可行的.
Abstract
The precise junction temperature monitoring is of paramount importance for enhancing the reliability of insulated gate bipolar transistor(IGBT)modules and extending the lifespan of devices.This paper introduces a method for monitoring junction temper-ature of IGBT modules based on turn-off voltages(TOV),which highlights resistance to load currents.The study initially verified the ra-tionality of using TOVs as a temperature-sensitive electrical parameter.The deep neural network(DNN)technology was subsequently employed to eradicate the dependence of TOVs on load currents,facilitating accurate junction temperature prediction under varying oper-ational conditions.The proposed method was validated through a single-phase pulse width modulation(PWM)experiment.The findings reveal an error range of±5℃for this method,demonstrating the feasibility of optimizing junction temperature monitoring through DNN utilization.
关键词
绝缘栅双极晶体管/结温监测/深度神经网络/温敏电参数/关断电压Key words
insulated gate bipolar transistor(IGBT)/junction temperature monitoring/deep neutral network(DNN)/temperature-sensitive parameter/turn-off voltage(TOV)引用本文复制引用
基金项目
国家级大学生创新创业训练计划项目(202310613027)
出版年
2024
NETL
NSTL
万方数据