Efficient strategies for reliability‐based design optimisation of a superconducting magnetic energy storage system‐based on reliability index approach

Abstract

This study presents an enriched reliability index approach for reliability-based design optimisation of electromagnetic devices in the presence of variability in design variables. To substantially improve computational efficiency, two new optimisation strategies consisting of a constraint indicator condition and a successive optimum search technique are incorporated into the existing method based on the reliability index approach. The constraint indicator decides whether reliability analysis is needed for each intermediate design during the reliability-based optimisation process. Employing the successive optimum search technique, a deterministic optimum is first sought from an initial design and then a probabilistic optimum is obtained by launching at the deterministic optimum. The proposed method is tested with a coil design problem for a superconducting magnetic energy storage system, and is thoroughly investigated by comparing the results with the conventional method based on the reliability index approach and the Monte Carlo simulation in terms of computational efficiency and accuracy. Finally, the results show that the method leads to a significantly decreasing computational cost.

Uncertainty identification method using kriging surrogate model and Akaike information criterion for industrial electromagnetic device

“…These uncertainties generally arise from the manufacturing process and material properties. To improve the robustness and ensure the reliability of electromagnetic (EM) devices under various uncertainties, several probabilistic design optimisation techniques, such as robust design optimisation (RDO) and reliability-based design optimisation (RBDO), have been developed [1][2][3][4]. The purpose of RDO is to find a less-sensitive optimal solution under uncertainties, which results in a minimal variance in performance, while satisfying all design constraints.…”

“…Although several probabilistic design optimisation techniques, such as RDO and RBDO, have been developed and applied to EM devices, they are usually based on the assumption that uncertainties follow a normal distribution [1][2][3][4][5]. In addition, the standard deviation of these uncertainties is usually determined based on the designer's experience.…”