Parallel Simulation of Power-System Electromagnetic Transients in the Laplace Domain

This paper presents a method based on the Laplace domain for the simulation of electromagnetic transients in power systems using a graphical processing unit (GPU). The proposed technique employs a parallel approach where all subsystems associated with different complex frequencies are solved simultaneously using the massive threads provided by the GPU architecture. Additionally, Numerical Laplace Transform (NLT) algorithms are parallelized using a CUDA library. To reduce computational costs, Kron's reduction method is used to calculate transient responses only at the nodes of interest in the network rather than across the entire network. Two case studies on an 18-node network are presented to demonstrate the performance and accuracy of the proposed method. Simulation results are compared with those of the conventional NLT implementation in MATLAB, as well as with those of the commercial software PSCAD. It is concluded that the proposed technique has potential applications for simulations in both modes, real-time and faster-than-real-time.