Real-time simulation of power electronic systems based on step-size adaptive algorithm

In a real-time power system simulation, it is essential to output data strictly at equally spaced sampling points. Therefore, accurate synchronization of data exchanges between the target machine and external devices is critical. When switching events occur between these sampling points, the trigger time for the event is typically delayed to the next available sampling point. This issue can reduce the accuracy of the simulation. Interpolation methods are commonly employed for asynchronous switching events to mitigate this issue. However, iterative computations following interpolation always introduce a temporal offset relative to the sampling points, and additional interpolation is needed to resynchronize the simulation results. To address this problem, this paper proposes a step-size adaptive algorithm that dynamically adjusts the integration step size to ensure that simulation results are accurately output at sampling points. Unlike traditional numerical integration methods, which necessitate frequent recalculations of the admittance matrix during variable step-size computations, the proposed algorithm introduces a parameter adapting to the step size. Our approach allows the admittance matrix to remain unchanged and significantly reduce computational overhead. Simulation experiments validate that the proposed algorithm achieves high accuracy and stability in variable step-size integration while effectively suppressing numerical oscillations in non-state variables caused by switching events.