KITE

AI-basedtopology optimizationof electricalmachines - KITE

The increasing market penetration of electromobility and the resulting growth in unit numbers mean that the material costs for central components in electric machines, such as magnets, and the energy efficiency of electric machines are becoming critical both economically and ecologically, and the ability to optimize these is proving to be highly relevant to competition.

The aim of the KITE joint project is to develop highly efficient motor topologies for electric drive machines. A model based on artificial intelligence (AI), which maps the large solution space of possible topologies and proposes new, highly efficient topologies, is intended to help with this. On the one hand, the project aims to substantially accelerate the development process by efficiently evaluating a large number of possible topologies with regard to various quality criteria in an early development phase. On the other hand, the design itself is to be supported by AI methods.

At the start of the project, requirements are first analyzed and specifications defined in order to determine a concept for the development of the electric drive machines. The next development steps include AI modeling, expanding the database and setting up the intelligent tool chain. One of the main tasks will be to train the algorithm for topology optimization. The project will be completed with the prototypical implementation of various drive machines and their testing on engine test benches.

Project duration: 01.09.2021 to 31.08.2024

• Volkswagen AG
• Merantix Labs GmbH
• Department of Intelligent Embedded Systems, Prof. Dr. rer. nat. Bernhard Sick, University of Kassel
• Department of Vehicle Systems and Fundamentals of Electrical Engineering, Prof. Dr. rer. nat. Ludwig Brabetz, University of Kassel