Comparative study of methods for optimization of electromagnetic devices with uncertainty

Abstract

Purpose

This paper compares six reliability-based design optimization (RBDO) approaches dealing with uncertainties for a simple mathematical model and a multidisciplinary optimization problem of a safety transformer to highlight the most effective.

Design/methodology/approach

The RBDO and various approaches to calculate the probability of failure are is presented. They are compared in terms of precision and number of evaluations on mathematical and electromagnetic design problems.

Findings

The mathematical example shows that the six RBDO approaches have almost the same results except the approximate moment approach that is less accurate. The optimization of the safety transformer highlights that not all the methods can converge to the global solution. Performance measure approach, single-loop approach and sequential optimization and reliability assessment (SORA) method appear to be more stable. Considering both numerical examples, SORA is the most effective method among all RBDO approaches.

Originality/value

The comparison of six RBDO methods on the optimization problem of a safety transformer is achieved for the first time. The comparison in terms of precision and number of evaluations highlights the most effective ones.

Citation

Deng, S., Brisset, S. and Clénet, S. (2018), "Comparative study of methods for optimization of electromagnetic devices with uncertainty", COMPEL - The international journal for computation and mathematics in electrical and electronic engineering, Vol. 37 No. 2, pp. 704-717. https://doi.org/10.1108/COMPEL-11-2016-0502