Stability-improved network partition based on a small-step synthesis model for electromagnetic transient simulation

Electromagnetic transient simulation offers high fidelity for modern power systems with high penetration of power electronics, but it suffers from low efficiency due to small time step requirements. To improve efficiency, network partition based on energy storage elements (i.e., capacitors and inductors) has gained attention, yet its numerical stability is limited by the simulation time step. This paper proposes a stability-improved network partition method using a small-step synthesis model. First, the semi-implicit leapfrog method-based network partition formulation is derived. Second, the small-step synthesis model-based network partition method is proposed. The method improves the numerical stability of the simulation by applying a small-step synthesis model to the network partition branches. Finally, the stability boundaries for the time step and synthesis order are derived based on the Lyapunov criterion and spectral radius analysis. The IEEE 14-bus system and a wind farm model are used to validate the proposed method. The results demonstrate that the proposed method enhances stability without sacrificing accuracy or efficiency. • A semi-implicit leapfrog-based network partition method is derived. • The proposed small-step synthesis model removes the traditional time step limitation. • Lyapunov-based stability analysis confirms this method improves simulation stability. • Case studies confirm stability improvement without sacrificing simulation efficiency.