Evaluation method for health state of power semiconductor device based on thermal impedance characteristic frequencies

Abstract

The invention provides an evaluation method for the health state of a power semiconductor device based on thermal impedance characteristic frequencies. The method comprises the following steps: extracting the thermal impedance characteristic frequencies of the power semiconductor device and radiators thereof; detecting the change conditions of the n thermal impedance characteristic frequencies; and comparing the change conditions of the thermal impedance characteristic frequencies with a thermal impedance characteristic frequency in a healthy state, and evaluating the healthy state of the power semiconductor device. According to the method, output heat flow of the power semiconductor device and temperature curves of junctions, shells and radiators are extracted to obtain thermal impedance characteristic frequencies of the power semiconductor device and the radiators thereof; and the health states of different positions inside and outside the power semiconductor device are reflected through a plurality of thermal impedance characteristic frequency change trends. According to the invention, the internal and external health states of the power semiconductor device can be efficiently and non-invasively detected at low cost.

Description

technical field [0001] The present invention relates to the technical field of power electronics, in particular to a method for evaluating the health status of a power semiconductor device based on the characteristic frequency of thermal impedance. Background technique [0002] Power devices are one of the most vulnerable components in converters. Monitoring the status of power semiconductor devices and judging their health status can effectively improve the reliability of power electronic systems. In recent years, the frequency domain thermal impedance model is gradually attracting people's attention, because compared with the traditional Foster and Cauer model, it can accurately predict the temperature characteristics of power semiconductor devices in the whole frequency band. However, the existing methods for reflecting the health status are all based on time-domain analysis. The application is relatively complex and needs to solve a large number of parameters. Using freq...

AI Technical Summary

Problems solved by technology

However, the existing methods for reflecting the health state are based on time-domain analysis. The application is relatively complex and requires solving a large number of parameters. Using frequency-domain analysis to evaluate the health state of semiconductor devices not only reduces the difficulty of parameter extraction, but also effectively reflects the health. state

Images