Objective 1: Renewing the HF modeling of electromechanical devices

To solve complex problems, we can either make as few modeling assumptions as possible and try to resolve the problems using brute force HPC, or try to perform intelligent approximations and techniques to reduce the need of computational power. We should be able to choose the ideal type of modeling approach for each problem. This objective is twofold: On one hand, we will develop a computationally efficient FE modeling framework for the HF modeling of windings. Compared to conventional magneto-quasistatic FE simulations, we have recently demonstrated up to 97% savings in computation times through subdomain pre- processing, model order reduction (MOR) and dimensionality reduction allowed by clever exploitation of symmetry properties. We will now create the theoretical basis and build the numerical tools for applying similar techniques in HF electromagnetic modeling, simultaneously accounting for fine winding geometries, eddy currents and capacitive coupling between conductors. On the other hand, we will also develop n ew HPC techniques customized for EMs to get the most out of modern computing architectures.

Objective 2: Enabling the design and implementation of safe and reliable HiECSs

We aim to bring the reliability of HiECSs to the next level by building aself-learning systemlevel modeling environment capable of accurately describing and predicting the electromechanical and mechanical subsystem interactions online during HS operation. New stochastic techniques will be developed in order to integrate consideration of uncertainty propagation from manufacturing tolerances and variations in the operating conditions into the system behavior. This modeling environment will be used as a basis for creating new adaptive control solutions for attenuating rotor vibrations up to supercritical speeds as well as pre-emptive fault diagnosis methods for ensuring safe operation.

Objective 3: Setting a new benchmark forthe operational limits of HiECSs

This objective brings together the methods developed in Objectives 1 and 2 to demonstrate how the acquired knowledge can be used to renew EM design practices for producing prototypes of the next generation of HiECSs, ambitiously aiming to exceed the specific power of 20 kW/kg. Firstly, by combining the improved understanding of the HF power loss mechanisms to innovative liquid cooling solutions, we aim to reach world-record current densities above 50 A/mm2 in copper. Secondly, exploiting the possibilities offered by WBG switching technology, we will push the fundamental frequency of HS machines to themultiple-kHz range. This not only allows increasing the operating speeds to 100 000 rpm, but also increasing the pole-pair number to reduce the length of flux paths and thus the volume of the stator. Thirdly, we aim for smoother torque and increased redundancy by introducing multiphase windings to HS machines, avoiding the limitations of conventional three-phase systems. Finally, we will explore the suitability of advanced manufacturing techniques for producing high-strength rotors capable of withstanding the HS operation. Three fully operating prototype systems will be implemented to validate the models as well as to demonstrate the feasibility of the machine designs and controlalgorithms.