Metaheuristic Optimization Preliminary Design Process of Core-Form Autotransformers

Abstract

Because of the strong competition in the equipment market, the aim of the manufacturers is to reduce the product price as much as possible, fulfill the high-quality standards, and improve the reliability and efficiency of the transformers. The capitalization of the costs enables to handle the different economic and technological parameters together in the same cost function. Large power autotransformers are usually designed for specific economic and technical conditions specified by the customer. At the tender design stage, the most economical solution has to be found in a limited time, and a great number of feasible transformer designs have to be compared to find the optimal solution. Nowadays, the solution of this design process is mainly based on the experience of the transformer designer, but mathematical programming models can facilitate the process. It is well known that the design of shell-type power transformers can be optimized by geometric programming. The constructed models of geometric programming can be solved accurately with the interior-point-method-based solvers. According to the results, this kind of modeling is not applicable for the core-form transformers. However, with branch-and-bound search, the geometric programming method can also be used for these cases. This paper introduces a new algorithm, which can also handle the autotransformer designs with different positions of the regulating winding by applying a more general metaheuristic routine to perform the calculation of the short-circuit impedance.