Insulated-gate bipolar transistor （IGBT） modules, as the core power components of wind turbine power converters, are crucial for achieving high operational efficiency and cost-effectiveness. Optimizing the electrical conductivity and manufacturing cost of these modules is essential. However, the demanding operational environments often cause performance degradation, and traditional design optimization methods incur high computational costs, making rapid iterative optimization difficult. To address these limitations, this study proposes a rapid iterative design optimization method for wind turbine converter IGBTs based on an improved NSGA-Ⅱ algorithm （Non-dominated Sorting Genetic Algorithm Ⅱ）. Firstly, the optimization problem is formulated with electrical conductivity and manufacturing cost as dual objectives. A parametric thermoelectric coupling model is developed to enable accurate and efficient evaluation of conductivity under varying design conditions. To further enhance the optimization process, an improved NSGA-Ⅱ algorithm based on kernel density estimation is proposed, significantly improving computational efficiency and optimization accuracy compared to traditional optimization algorithms. Additionally, a Kriging-based surrogate model is employed to construct high-fidelity mappings between design variables and optimization objectives, thereby reducing computational burdens and enabling rapid iterative optimization. Numerical experiments confirm the effectiveness and robustness of the proposed method, demonstrating reductions in electrical losses of up to 20.00% and decreases in manufacturing costs by as much as 27.63%. This study provides a practical and efficient design framework for IGBT modules, offering valuable insights into multi-objective optimization in the field of power electronics. By integrating advanced optimization algorithms with surrogate modeling, the proposed method addresses key challenges in the design and performance enhancement of wind turbine power systems.

IGBT  /  wind turbines  /  power converters  /  NSGA-Ⅱ algorithm  /  kernel density estimation  /  optimization algorithm