• • Hybrid automata and hybrid Petri net models are structurally too broad for practical system analysis; switching and piecewise affine models are currently the most widely applied in power systems, as evidenced by their use in low-voltage ride-through switching sequences.
• • A second-order trajectory sensitivity-based parameter identification method for switching models is proposed, addressing the discontinuity in objective functions caused by discrete event triggering, which renders traditional gradient-based search ineffective.
• • For piecewise affine hybrid models, three parameter identification methods—algebraic geometry, data aggregation, and Bayesian—are presented, with their respective advantages and applicable scenarios discussed, enabling robust parameter estimation under discrete state changes.
• • A unit grouping method considering switching sequence similarity and continuous system dynamic similarity is proposed for hybrid equivalent modeling of renewable energy stations, along with discrete state and continuous system aggregation strategies, facilitating accurate station-level equivalents despite dispersed units and diverse discrete event states.