自适应粒度指数积分算法用于电力电子变换器系统的高效仿真

Electromagnetic transient (EMT) simulation is increasingly important for complex power electronic converter system design and validation. This paper proposes an adaptive-grained exponential integrator (AGEI) algorithm designed to efficiently simulate power electronic networks. The AGEI algorithm relies on precomputation of carefully-chosen discretization steps to reduce memory burden. It then performs sequential intermediate integrations between discrete switching events to accelerate transient simulation. The variable time-step, high-order integration algorithm finds a balance between simulation run-time and computational overhead. The AGEI algorithm varies the number of the forcing function terms to meet desirable error thresholds. The proposed algorithm is L-stable and can be flexibly applied to stiff and non-stiff circuit systems, rendering it suitable for a wide arrange of power converter topologies and parameters. A hardware experimental study validates the AGEI algorithm's accuracy while simulation case studies demonstrate the AGEI algorithm significantly improves numerical efficiency. It is shown in the typical converter case studies that the proposed algorithm enables more than 8-fold and 3-fold simulation speedup compared to popular simulation tools, i.e., Simulink and PLECS, respectively. The numerical efficiency gains of the proposed algorithm become more apparent for power converter circuits with only-DC input sources or stiff circuit systems.