Multi-Objective Optimization Method for Power Transformer Design Based on Surrogate Modeling and Hybrid Heuristic Algorithm

• DOI: 10.3390/electronics14061198
• Publication Date: 2025-03-18T15:02:31Z
• Authors: Baidi Shi; Wei Xiao; Liangxian Zhang; Tao Wang; Yongfeng Jiang; Jingyu Shang; Zixing Li; Xinfu Chen; Meng Li
• Keywords: Surrogate model

In response to the increasing demands for energy conservation and pollution reduction, optimizing transformer design reduce operational losses minimize raw material usage has become crucial. This paper introduces an innovative methodology that combines ensemble learning models with hybrid multi-objective optimization heuristic algorithms optimize leakage impedance deviation, on-load loss, consumption in power transformers. The stacking model uses support vector machines, linear regression, decision tree K-nearest neighbors as base learners, extreme machine serving meta-learner re-learn outputs from first-level learners. Given significant impact of hyperparameters on prediction performance models, improved particle swarm method is proposed effective hyperparameter optimization. To assess uncertainty model, a Kriging surrogate model-based analysis outlined. Moreover, powerful algorithm integrates grey wolf (MOGWO) non-dominated sorting genetic algorithm-III (NSGA3) presented approach demonstrates superior compared mainstream algorithms. effectiveness this further validated through engineering tests two real cases. can accommodate various requirements and, under given constraints, achieve transformers, ensuring optimal different scenarios.