arXiv subjects
Publications and source records attributed to Andreas Klaedtke.
This paper proposes the utilization of a periodic Parareal with a periodic coarse problem to efficiently perform adjoint sensitivity analysis for the steady state of time-periodic nonlinear circuits. In order to implement this method, a modified formulation for adjoint sensitivity analysis based on the transient approach is derived.
This paper presents a transient forward harmonic adjoint sensitivity analysis (TFHA), which is a combination of a transient forward circuit analysis with a harmonic balance based adjoint sensitivity analysis. TFHA provides sensitivities of quantities of interest from time-periodic problems w.r.t. many design parameters, as used in the design process of power-electronics devices. The TFHA shows advantages in applications where the harmonic balance based adjoint sensitivity analysis or finite difference approaches for sensitivity analysis perform poorly. In contrast to existing methods, the TFHA can be used in combination with arbitrary forward solvers, i.e. general transient solvers.
This paper presents a parallel-in-time adjoint sensitivity analysis which combines a transient adjoint sensitivity analysis with the parareal approach in order to significantly speed up the simulation. The adjoint method is the method of choice to calculate the sensitivities in a many-parameter setting. In order to obtain sensitivity information that is time-dependent, multiple adjoint problems must be solved. This slows down the simulation wall-clock time and leaves a large optimization potential for the analysis. The parareal is applied to the adjoint solution, significantly speeding up the adjoint solution for every timestep respectively.
Parasitic extraction is a powerful tool in the design process of electromechanical devices, specifically as part of workflows that check electromagnetic compatibility. A novel scheme to extract impedances from CAD device models, suitable for a finite element implementation, is derived from Maxwell's equations in differential form. It provides a foundation for parasitic extraction across a broad frequency range and is able to handle inhomogeneous permittivities and permeabilities, making it more flexible than existing integral equation approaches. The approach allows for the automatic treatment of multi-port models of arbitrary conductor geometry without requiring any significant manual user interaction. This is achieved by computing a connecting source current density that supplies current to the model's terminals, whatever their location in the model, subsequently using this current density to compute the electric field, and finally calculating the impedance via a scalar potential. A mandatory low-frequency stabilization scheme is outlined, ensuring a stable evaluation of the model at low frequencies as well. Two quasistatic approximations and the special case of perfect electric conductors are treated theoretically. The magnetoquasistatic approximation is validated against an analytical model in a numerical experiment. Moreover, the intrinsic capability of the method to treat inhomogeneous permittivities and permeabilities is demonstrated with a simple capacitor-coil model including dielectric insulation and magnetic core materials.