Multi-objective design optimization of hermetically sealed core-type distribution transformer considering current harmonics of power grid using NSGA III
Mohammad Hassan Hashemi,Ulas Kilic
DOI: https://doi.org/10.1016/j.jestch.2024.101745
IF: 5.155
2024-06-21
Engineering Science and Technology an International Journal
Abstract:In this paper, the design optimization of hermetically sealed oil-immersed distribution transformers is studied. For the first time in the design of this type of transformer, harmonics of the power grid affecting the transformer design, along with all design variables, are included, and their boundaries suitable for the production line are determined. The design probability (search space) for a total of 21 design variables is 1.32×1025 . The Brute Force method is not a time-efficient approach to finding the optimal result in this search space. Due to the mixed integer nonlinear programming (MINLP) nature of the transformer design problem, heuristic optimization algorithms are selected because of their shorter solution time and near-optimal results. Considering the principle of the "No Free Lunch" theorem, extensive evaluations were conducted on various heuristic algorithms to identify the most effective ones for addressing the transformer design problem. The most successful algorithms for solving the problem were identified after thorough evaluation, considering three distinct scenarios within a multi-objective optimization framework. These algorithms are namely MOEA/D, MOGWO, MOWOA, and NSGA III. In this study, an optimization tool is developed for the design optimization of hermetically sealed distribution transformers, covering a range from 50 kVA to 3150 kVA and encompassing all design variables. The developed optimization program is fully compatible with the production line and can be easily adjusted to meet all national and international standards, such as IEC and IEEE, as well as customer specifications.
engineering, multidisciplinary