Online Condition-Monitoring of Power Semiconductor Devices using Acoustic Time-frequency Spectral Signature

The main goal in this research is to apply acoustic method for aging detection of power semiconductor switches. The idea is new and the main plan is to have a proof of concept so it can be considered for patent. Up to now, components monitoring and aging detection are limited to either using thermal images or using measuring current and voltage probes. The issue is that in both cases the power switch needs to be removed from the package and therefore it cannot be utilized in every product. However, using generated acoustic emission of power switches it could be possible to not only monitor the device but also easily apply diagnosis and more importantly prognosis, which can save huge costs. This is of special importance in applications such as wind turbines where condition important has significant importance and diagnosis causes significant cost.