Optimization Study of Thermal Management of Domestic SiC Power Semiconductor Based on Improved Genetic Algorithm

作者：Feng Chen, Guoping Xu, Shangxi Li, Lan Yue, Hongjing Liang · 年份：2025 · DOI：10.1109/etae65337.2025.11089826 · 被引用次数：7 · 研究领域：Silicon Carbide Semiconductor Technologies、Photovoltaic System Optimization Techniques、Induction Heating and Inverter Technology

To enhance thermal management efficiency of domestic SiC power semiconductor devices under high heat density, a multi-parameter co-optimization model based on an improved genetic algorithm is proposed. This model integrates thermal-structural-fluid coupling among package structure, heat dissipation interface, and cooling system. A multi-objective fitness function, combined with the NSGA-II algorithm, minimizes thermal resistance, junction temperature, and response time. Based on thermal network modeling, structure simplification and parameter extraction are performed. Experimental validation shows the optimized system significantly reduces steady-state junction temperature and response delay, enhancing dynamic adaptability and thermal safety, with strong potential for engineering applications.