Plugin required: in order to see this object, your browser has to support files of type image/svg+xml. Download

The chair of Computational Electromagnetics (CEM) is part of the institute TEMF and the Centre for Computational Engineering. Teaching and research focus on the third pillar of understanding: computer simulation. Besides theory and observation, it can give answers to questions from engineering and science.

Mission statement

Electrotechnical systems are becoming more and more complex. Innovative devices are designed closely to what is technically and physically possible. Consequently, the theory required to analyze the corresponding systems is becoming increasingly involved as well. Experimental investigations are often too complex, too risky, or too costly and the presence of test probes might corrupt the experiment data. Computational Electromagnetics is in those cases the most appropriate way to gain knowledge.

Computer model (image based on iStock.com/simonkr)
Computer model (image based on iStock.com/simonkr)

Computer-based modeling, analysis, simulation, and optimization are a cost-effective and efficient alternative to investigate real-world applications and to engineer new technical solutions. The digital models (`virtual prototypes') support research, development, design, construction, evaluation, production, and give further insight into the operation of devices like semiconductors, filters, antennas or electrical machines. It allows us to find optimal strategies which address key issues in future technical developments both for the economy and for society in areas such as energy, health, safety, and mobility.

The main subject of research and teaching at the chair is the modeling and simulation of electromagnetic and multiphysical phenomena by means of numerical solutions of partial differential equations and in particular of Maxwell's equations. We are working on all the development stages, mainly on modeling and the development of numerical algorithms, but also on real-world applications. For this work existence, uniqueness and differentiability of solutions, robustness, convergence, and scalability of the algorithms are as important as their efficient implementation, e.g., acceleration of numerical linear algebra by Graphics Processing Units (GPUs).

Recent Preprints

  • 1907.12626 Efficient simulation of DC-AC power converters using Multirate Partial Differential Equations
  • 1907.06404 Robust Optimization of a Permanent Magnet Synchronous Machine Considering Uncertain Driving Cycles
  • 1907.03719 Multirate PWM balance method for the efficient field-circuit coupled simulation of power converters
  • 1906.09151 Uncertainty Modeling and Analysis of the European XFEL Cavities Manufacturing Process
  • 1906.00785 Bembel: The Fast Isogeometric Boundary Element C++ Library for Laplace, Helmholtz, and Electric Wave Equation
  • 1905.13076 An efficient steady-state analysis of the eddy current problem using a parallel-in-time algorithm
  • 1905.06879 Multigrid-reduction-in-time for Eddy Current problems
  • 1902.08277 Accelerated Steady-State Torque Computation for Induction Machines using Parallel-In-Time Algorithms