Of great significance is knowledge of failure modes of IGBT modules under power cycling test (PCT) in advance. It can not only precisely determine accurate utilization of physics-of-failure lifetime prediction methods but also help optimize IGBT designs. However, establishing an accurate and generic offline failure mode classification method for different IGBT modules remains a challenging problem due to the complex degradation process of IGBT modules. In this article, a data-driven convolutional neural network (CNN) based method is proposed for quad-classification of failure modes for different IGBT modules under different PCT conditions. First, in order to accurately characterize the mapping between the failure modes and the precursor parameters, a framework of precursor parameters is meticulously established. Then, mainly using collected training data from existing publications, the CNN classification model is developed by a newly dynamic tuning-multilevel particle swarm-back propagation optimization algorithm. Finally, experimental PCTs of various IGBT modules are performed. The obtained PCT data and additional testing data from existing publications are used to verify the generalizability and robustness of the proposed classification method. The superiority of the proposed method is well demonstrated through comparison with random CNN, the state-of-the-art particle swarm optimization CNN, and other intelligent algorithms.

The reliability of capacitors is one of the most important issues in power electronics. The health status of capacitors can be evaluated through the comparison of estimated C/ESR values with their original values. In this paper, a two-input artificial neural network (ANN) is proposed for C and ESR estimation of DC-link capacitors in Maglev choppers; combined with the existing voltage and current sensors that are used for protection and control, it provides a promising solution for the health condition monitoring of the Maglev chopper in the Maglev train. Compared with prior-art research, the actual capacitor degradation progress where both C and ESR degrade is considered in training. Moreover, ANN’s advantage of fitting nonlinear and complex relationships is explored by building an aggressive mapping between the voltage ripple at 5 kHz to C/ESR at 120 Hz, which cannot be described or analyzed by linear circuit equations. Thus, cross validation must be implemented to avoid an occasional poor fitting, and ensure the stability of ANN. Experimental results show that the proposed ANN outperforms the support vector regression in both ESR and C estimation. While C estimation suffered from over-fitting and instability, ESR estimation by the ANN is accurate and stable with a low average prediction error within 1%, showing great potential for condition monitoring of DC-link capacitors in Maglev choppers.