Kalman-Augmented Estimation for Enhancing the Reliability Analysis of Power Electronic Converters via Sensor-Error Mitigation
Faculty Mentor
Dr. Masoud Davari
Location
Russell Union Ballroom
Type of Research
Completed
Session Format
Poster Presentation
College
Allen E. Paulson College of Engineering & Computing
Department
Electrical and Computer Engineering
Abstract
The lifetime and reliability of power electronic con- verters strongly depend on accurate thermal and electrical mon- itoring during operation. This paper presents a comprehensive reliability assessment of a buck converter system under sensor- induced uncertainties and introduces an augmented Kalman- based framework for error mitigation. A 12 kW, 300 VDC to 125 VDC buck converter was modeled in Simscape using a half-bridge IGBT–diode structure with detailed loss calculation and a Cauer-type thermal network. The monitoring subsystem records junction temperature, total losses, and load power amidst other measurements while enforcing a shutdown above 150°C. The study investigates four distinct cases: a baseline converter with ideal sensing, an error-injected model with bias, drift, and gain distortion, a basic Kalman filter for partial correction, and an augmented Kalman filter incorporating sensor fault dynamics. The converter was modeled in MATLAB/Simulink using a coupled electro-thermal structure to evaluate steady- state performance, junction temperature, and lifetime using the Arrhenius degradation model. Results show that sensor errors can elevate the mean junction temperature by over 16°C and artificially extend lifetime prediction by more than 160%. The augmented Kalman filter successfully restored temperature and reliability estimates to within 4% of the baseline, confirming its ability to compensate for both random and systematic sensor deviations. The proposed estimation approach demonstrates a physics-consistent pathway for improving converter health mon- itoring and reliability prediction in systems affected by sensor degradation.
Program Description
.
Start Date
4-23-2026 10:00 AM
End Date
4-23-2026 12:00 PM
Recommended Citation
Omoyiwola, Esther T., "Kalman-Augmented Estimation for Enhancing the Reliability Analysis of Power Electronic Converters via Sensor-Error Mitigation" (2026). GS4 Student Scholars Symposium. 49.
https://digitalcommons.georgiasouthern.edu/research_symposium/2026/2026/49
Kalman-Augmented Estimation for Enhancing the Reliability Analysis of Power Electronic Converters via Sensor-Error Mitigation
Russell Union Ballroom
The lifetime and reliability of power electronic con- verters strongly depend on accurate thermal and electrical mon- itoring during operation. This paper presents a comprehensive reliability assessment of a buck converter system under sensor- induced uncertainties and introduces an augmented Kalman- based framework for error mitigation. A 12 kW, 300 VDC to 125 VDC buck converter was modeled in Simscape using a half-bridge IGBT–diode structure with detailed loss calculation and a Cauer-type thermal network. The monitoring subsystem records junction temperature, total losses, and load power amidst other measurements while enforcing a shutdown above 150°C. The study investigates four distinct cases: a baseline converter with ideal sensing, an error-injected model with bias, drift, and gain distortion, a basic Kalman filter for partial correction, and an augmented Kalman filter incorporating sensor fault dynamics. The converter was modeled in MATLAB/Simulink using a coupled electro-thermal structure to evaluate steady- state performance, junction temperature, and lifetime using the Arrhenius degradation model. Results show that sensor errors can elevate the mean junction temperature by over 16°C and artificially extend lifetime prediction by more than 160%. The augmented Kalman filter successfully restored temperature and reliability estimates to within 4% of the baseline, confirming its ability to compensate for both random and systematic sensor deviations. The proposed estimation approach demonstrates a physics-consistent pathway for improving converter health mon- itoring and reliability prediction in systems affected by sensor degradation.
