AI based Multi Criterion Optimization Solution for Magnetic Design

Wickramasinghe, Gayan, Wamakulasuriya, Kapila orcid iconORCID: 0000-0002-1617-3083, Yapa, Ruchira orcid iconORCID: 0000-0002-1617-3083, Porawagamage, Gayashan and Dhanapala, Vajira (2025) AI based Multi Criterion Optimization Solution for Magnetic Design. In: PCIM Conference 2025; International Exhibition and Conference for Power Electronics, Intelligent Motion, Renewable Energy and Energy Management. Verband der Elektrotechnik (VDE), pp. 457-464. ISBN 978-3-8007-6541-6

[thumbnail of AAM] PDF (AAM) - Accepted Version
Restricted to Repository staff only

567kB

Official URL: https://doi.org/10.30420/566541055

Abstract

Designing magnetic components is a complex task involving over forty parameters with numerous possible values, leading to countless design permutations. This complexity often hampers the pursuit of optimal designs, as results can vary based on different expert opinions. Traditional computational methods struggle to manage this challenge. To address this, an advanced AI-based autonomous approach that uses sophisticated algorithms to evaluate all design parameters comprehensively was developed. This method generates optimized designs that surpass the capabilities of experienced human designers, paving the way for significant advancements in magnetic design. This paper details the innovative methodology and its advantages, showcasing how AI-driven optimization can reduce design time by over 95% while exploring parameter combinations that human designers might overlook due to time constraints.


Repository Staff Only: item control page