SiC MOSFET结温的机械应力波监测方法及系统
By extracting multidimensional acoustic feature vectors and establishing a junction temperature prediction model using machine learning algorithms, the problems of sensor thermal degradation and single feature in SiC MOSFET junction temperature monitoring are solved, realizing high-precision, electrically isolated online junction temperature monitoring that can adapt to complex operating conditions.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HUNAN UNIV
- Filing Date
- 2026-04-08
- Publication Date
- 2026-07-17
AI Technical Summary
In existing SiC MOSFET junction temperature monitoring technologies, sensor thermal degradation leads to signal distortion, and the single feature model cannot decouple the nonlinear coupling effect of current and temperature, resulting in large junction temperature estimation errors that are difficult to meet the needs of engineering applications.
By collecting mechanical stress wave signals under different preset junction temperatures and turn-off currents, multi-dimensional acoustic feature vectors such as root mean square, peak-to-peak value, and kurtosis are extracted. Combined with machine learning algorithms, a junction temperature prediction model is established to achieve non-invasive monitoring.
It achieves high-precision, electrically isolated online monitoring of SiC MOSFET junction temperature, adapts to complex operating conditions, reduces monitoring errors, and ensures the safe and reliable operation of devices.
Smart Images
Figure CN121978500B_ABST