Fast and Scalable Simulation Framework for Intensive Power Electronics Simulation through Advanced Computing Techniques

Abstract

Power electronics are widely used in power and energy systems, such as electrified transportation and renewable energy systems. These systems are increasing related simulations' size and complexity, resulting in longer computation times. Advanced computing techniques offer new tools for efficient simulation while designing and simulating complex energy systems. This paper introduces high-performance computing for intensive power electronics simulation and demonstrates resultant simulation speedup in a quantified and scalable manner. First, a quantitative study is performed to compare a slower-than-real-time (STRT) simulation benchmark and the proposed faster-than-real-time (FTRT) simulation through a single power electronics building block (PEBB) case study. The impact of switching frequencies in the range of tens to hundreds of kHz considering wide bandgap (WBG) power semiconductors is also investigated. The simulation speed is observed to be accelerated by a factor of 43.8 when using high-performance computing techniques compared to the sequential-based simulation benchmark. Next, a scalable simulation framework is proposed for expanding a single PEBB to an energy system consisting of multiple PEBBs. The framework leverages the high-performance programming language Julia with multi-threaded parallel computing capabilities to reduce the computational burden of power system simulation. The performance gains from the case study demonstrate an average speedup of 2540 times in a 15.0 s multi-PEBB simulation case study compared to its baseline version, with maintained simulation accuracy and ensured scalability.