Optimal Energy Converter Design Using Math

PASIROM: Research project started by universities and industrial partners

2018/04/05 by

Finding the optimal design for energy converters with mathematical methods: The recently launched research project PASIROM, involving three universities and two application partners, is dedicated to this task. etit professor Sebastian Schöps plays an important role in the research project.

Photomontage of a design draft and the field line image of an electric machine. Image: simonkr / iStock

In times of energy transition, the design of electromechanical energy converters, which are used, for example, in household appliances or drives for electric mobility, is becoming increasingly important: What form should the individual components have in order to generate, convert and distribute energy as efficiently and sustainably as possible? How can uncertainties in production, for example in the form of geometric deviations or material fluctuations, be compensated in such a way that there are no equipment breakdowns or a drop in performance?

A research team of Technische Universität Darmstadt, the University of Wuppertal and the University of Hamburg, together with application partners of Robert Bosch GmbH from Stuttgart and the software manufacturer Computer Simulation Technology (CST) GmbH from Darmstadt, is dedicated to such questions. The “PASIROM” project is in line with the new high-tech strategy of the German government.

Without mathematical methods of modeling, simulation and optimization, the search for optimal and particularly robust designs of electromechanical energy converters cannot progress today. In modeling, the behavior of electrical machines is expressed by mathematical equations. In addition to physical effects such as rotation, induction and heat generation, the researchers also take into account uncertainties in the manufacturing process in order to increase the description quality of the models. This is where the “Modelling” work package of professors Sebastian Schöps and Stefan Ulbrich from TU Darmstadt comes in.

These models then enable computer-aided simulations, which in turn provide conclusions about the real problem. To reduce simulation times, researchers rely on parallel processes and efficient use of current high-performance computers with sophisticated system architecture. Dr. Stephanie Friedhoff (Bergische Universität Wuppertal) and Professor Sebastian Schöps are responsible for this work package “Simulation”. In addition, the researchers use mathematical methods of model reduction and adaptation to optimize electrical machines directly on the computer. Professor Michael Hinze (University of Hamburg) and Professor Stefan Ulbrich are responsible for the sub-project “Optimization”.

Faster and better informed design decisions

Result of a parallel time domain simulation of an asynchronous machine with the software getDP. Graphic: PASIROM

All sides benefit from the collaborative research work. By using mathematical methods as early as possible in the product development process, application partners can make much faster and better informed design decisions without having to resort to costly prototypes and experiments. This opens up opportunities for the application partners to produce even more environmentally friendly and with an even lower risk of error, and to bring new, more powerful products onto the market more quickly, which can help to further reduce energy consumption. For the scientists involved, basic research is also about developing a solid mathematical basis for new approaches in the field of robust optimization and uncertainty quantification.

The research association PASIROM – “Parallel simulation and robust optimization of electromechanical energy converters with uncertainties” was selected as one of ten projects within the highly competitive funding priority “Mathematics for Innovations”. The project builds on the results of the predecessor project SIMUROM, which was part of the identical funding programme of the Federal Ministry of Education and Research from 2013 to 2016.