Priority Aware Under Frequency Load Shedding Via Particle Swarm Optimization for Low Inertia Island Power Systems

Island power grids like the Lombok system face severe stability constraints due to low rotational inertia and narrow reserve margins. This vulnerability is significantly exacerbated by the integration of a new 100 Megawatt coal fired generation complex comprising two 50 Megawatt units alongside a 10 percent share of batteryless solar photovoltaics. A sudden trip of these units risks catastrophic cascading blackouts, exposing the limitations of existing static Under Frequency Load Shedding (UFLS) schemes that often cause indiscriminate over shedding and hazardous frequency overshoots. To address this, this paper proposes a robust, five stage UFLS framework optimized via Particle Swarm Optimization (PSO). Utilizing DIgSILENT PowerFactory integrated with Python via an Application Programming Interface (API), the algorithm formulates load curtailment as a discrete optimization problem. It enforces a hierarchical load sorting mechanism to protect critical infrastructure while satisfying operational Grid Code constraints. Cross validation against extreme contingency scenarios (50 MW and 125 MW loss) proves the proposed optimized fixed step scheme successfully averts blackouts, maintains the frequency nadir above 48.25 Hz to provide a safety margin before reaching the 48.20 Hz islanding threshold, limits overshoots below 51.00 Hz, and minimizes unserved energy compared to the conventional eight stage scheme.