A method, device and storage medium for predicting the residual life of a high-power thyristor

Abstract

This invention discloses a method, device, and storage medium for predicting the remaining lifetime of high-power thyristors, belonging to the technical field of high-power thyristor remaining lifetime prediction. It addresses the problem of how to more comprehensively predict the health status of thyristors. The method involves collecting degradation test data of the thyristor's on-state voltage drop and reverse recovery charge; calculating an acceleration factor based on the test data; establishing a binary degradation trajectory model of the thyristor based on the degradation data of the thyristor's on-state voltage drop and reverse recovery charge; estimating the unknown parameters of the model using the Markov chain Monte Carlo method; and evaluating the remaining lifetime of the thyristor using the lifetime prediction model. This invention uses two-dimensional data of the thyristor's on-state voltage drop and reverse recovery charge to evaluate the thyristor's lifetime and models their correlation. The resulting thyristor lifetime prediction results can comprehensively reflect the health status of the thyristor, overcoming the difficulty of accurately predicting the thyristor's lifetime due to a small number of failure samples, and improving the accuracy of thyristor lifetime prediction.

Images