Computational-Intelligence-Based Optimization of Missing Inertia in Low-Inertia Power Systems Using Real-Time Digital Simulation

Low-inertia power systems require scenario-dependent assessment of additional frequency support. This study develops a computational-intelligence-based framework for minimizing missing virtual inertia with conventional inertia fixed as an external input. Following a five-metaheuristic comparison, a whale optimization algorithm (WOA)-based search is coupled to real-time electrical simulation on RTDS NovaCor Light through RSCAD FX. Acquired trajectories determine candidate feasibility under event-direction rate of change of frequency (RoCoF), nadir, and zenith limits. A 70-case experiment combines conventional inertia of 1.0–4.0 s with signed demand changes of 5–25 MW. Of these cases, 56 yielded feasible outputs: 24 required no addition and 32 required 0.155–7.349 s on the system base. The remaining 14 cases at ±25 MW had no feasible candidate within the tested 0–10 s additional-inertia interval and budget. For positive-support cases, the nearest constraint shifted from RoCoF at 10–15 MW to nadir or zenith at 20 MW, revealing the need to assess frequency extrema and both disturbance directions. Their median optimization time was 28.95 min, supporting prospective scenario assessment. Virtual support was represented by an ideal equivalent generator-inertia increment; consequently, the best-found outputs inform battery energy storage and virtual synchronous machine design without demonstrating physical converter delivery.