Marcelo D. Silva, PostDoc researcher at Uppsala University and former member of the SEC doctoral network, will be presenting a webinar spanning his PhD journey and his current research on machine learning methods for electric motor design.

The session will open with reflections on the experience of being a PhD student within the SEC network, working at the intersection of academia and industry in close collaboration with leading companies in heavy-duty road and rail-bound vehicles. The session will then cover the design and experimental validation of spoke-type Permanent Magnet Synchronous Machines using ferrite magnets as a rare earth element-free solution for electromobility, including prototype results obtained in collaboration with Scania and Alstom.

Finally, the webinar will explore how artificial intelligence and machine learning techniques, in particular Bayesian Optimisation and Gaussian Process Regression, can dramatically accelerate the design and optimisation of electrical machines, cutting computation times from days down to minutes.

Biography

Marcelo D. Silva received his M.Sc. in Electrical Engineering from the University of Porto in 2018, with a thesis on torque ripple mitigation in induction machines developed in collaboration with WEG’s R&D department in Brazil, where he subsequently worked as a Research and Development Engineer. He then pursued a Ph.D. at Uppsala University, Sweden, which he completed in 2025 with a thesis on the design and optimization of spoke-type permanent magnet synchronous machines as a rare-earth-element-free solution for electromobility.

During his doctoral studies, through a Swedish Electromobility Centre (SEC) project, he collaborated with Scania and Alstom on the development and experimental validation of ferrite-based electric motors, achieving high performance with the rare-earth-element-free solution developed. He also conducted a research visit at the e-Motion Laboratory at University College London (UCL), where he investigated Bayesian optimisation and Gaussian process regression for electrical machine meta-modelling. His research interests include the design of ferrite-magnet electrical machines, meta-modelling techniques, and the optimisation of electrical machines for electromobility applications.

Date            1 April 2026
Time           13:00-14:00
Location    Online, Teams – the Teams link will be sent to registered participants the day before the webinar.

Please click HERE to register for the webinar.
Register by 12:00 on 30 March.