Optimization of Hydrogenerator Pole Shapes Using a Sensitivity-Scaled Multi-Objective Level-Set Method

This paper investigates a bi-objective magnetic pole shape optimization problem for a hydrogenerator, targeting the fundamental component of the no-load induced voltage and iron loss. To address the imbalance of gradient magnitudes in gradient-based multi-objective optimization, a level-set-based topology optimization method with sensitivity scaling is proposed. The proposed method enables effective control of the search direction based on designer-defined weighting factors. Numerical experiments demonstrate that the method allows flexible transitions between voltage-oriented optimization and iron-loss reduction. The results confirm that the proposed framework is effective in improving performance and providing useful design insights for hydrogenerator magnetic pole optimization.