D26.9 - Online Monitoring and Control of Power Converter Junction Temperature in Wave Energy Systems Using Artificial Neural-Network-Based Model Predictive Control

Thursday, March 26, 2026

12:00 PM - 1:45 PM CT

Location: Hemisfair Ballroom

Technical Dialogue Presenter(s)

Qiang Mu

University of North Carolina at Charlotte

Presentation Paper

Co-Author(s)

KP

Kamal Paul

The University of North Carolina at Charlotte
Zaheen Mustakin

University of North Carolina at Charlotte
JZ

Jiale Zhou

University of North Carolina at Charlotte
LP

Lucas Pereira

University of North Carolina at Charlotte
BP

Babak Parkhideh

University of North Carolina at Charlotte
TZ

Tiefu Zhao

University of North Carolina at Charlotte

This paper proposes an online junction temperature monitoring and control approach based on FCS-MPC. Additionally, an ANN approach is combined with FCS-MPC to improve the output voltage quality. Meanwhile, the junction temperature of power devices is controlled in a closed-loop manner, effectively minimizing the overall junction temperature in three-phase SiC inverters. The results demonstrate that the proposed method enables real-time monitoring of the junctiontemperatures of all SiC MOSFETs in a three-phase inverter and achieves closed-loop regulation of their overall junction temperature while maintaining low THD. Furthermore, a more detailed reliability assessment based on the ANN-FCS-MPC framework, incorporating enhancements to the Rds-on sensor board and further optimization of the ANN, will be carried out in future work.