IGBT的寿命评估方法研究

发布时间：2018-11-14 08:56

【摘要】：随着半导体技术的不断发展,功率器件的研究日益成熟,其性能也在进行不断的完善,在诸多领域得到了广泛的应用。在非平稳工况下,IGBT在运行过程中因使用不当,工作周期长久而引发各种故障,加速器件的老化和失效程度,大大降低器件的使用寿命,影响整个装置的正常工作。因此,对IGBT的寿命进行评估显得十分的迫切。本文主要针对IGBT在加速老化实验条件下的失效问题,从IGBT的失效机理和寿命评估方法分析出发,在对IGBT的加速老化实验数据研究后,选取集电极-发射极关断电压尖峰值作为IGBT寿命评估的特征参数,利用Saber仿真模型分析该参数在关断过程的变化趋势,根据NASA PCoE研究中心提供的IGBT加速老化数据,对数据进行一次指数平滑和二次指数平滑处理,采用BP神经网络、MEA-BP以及NAR动态神经网络建立IGBT的寿命评估模型。根据预测结果采用评价指标均方误差MSE,决定系数R2对比模型的预测准确性。结果表明,数据经过二次指数平滑处理后的MEA-BP寿命评估模型的预测结果最好,从而为IGBT寿命评估的相关分析提供依据。本文的分析结果均以NASA PCoE研究中心提供的IGBT加速老化数据为基础,实验数据的分析结果表明本文所提出的方法能够运用于IGBT的寿命评估,具有一定的工程应用价值。
[Abstract]:With the continuous development of semiconductor technology, the research of power devices has become more and more mature, and its performance has been continuously improved, which has been widely used in many fields. Under the non-stationary working condition, the malfunction of IGBT is caused by improper use and long working cycle, which accelerates the aging and failure degree of the device, greatly reduces the service life of the device, and affects the normal operation of the whole device. Therefore, it is urgent to evaluate the life of IGBT. In this paper, aiming at the failure problem of IGBT under accelerated aging conditions, the failure mechanism and life evaluation method of IGBT are analyzed, and the experimental data of accelerated aging of IGBT are studied. The peak value of collector-emitter turn-off voltage is selected as the characteristic parameter of IGBT life evaluation. The change trend of this parameter in the turn-off process is analyzed by using Saber simulation model. According to the IGBT accelerated aging data provided by NASA PCoE Research Center, The data are processed by exponential smoothing and quadratic exponential smoothing. The IGBT life evaluation model is established by using BP neural network, MEA-BP and NAR dynamic neural network. According to the prediction results, the prediction accuracy of R ~ 2 model is compared with the evaluation index mean square error (MSE,) determination coefficient R2. The results show that the prediction results of MEA-BP life evaluation model after quadratic exponential smoothing are the best, which provides the basis for the analysis of IGBT life evaluation. The results of this paper are based on the accelerated aging data of IGBT provided by NASA PCoE Research Center. The experimental results show that the proposed method can be applied to the life evaluation of IGBT and has certain engineering application value.
【学位授予单位】：安徽理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN322.8

【参考文献】