Thermostatically Controlled Loads as Virtual Energy Storage Systems: Study, Modeling, Comfort-Aware Frequency Regulation and Sensitivity Studies

High renewable penetration reduces system inertia and stresses frequency stability, motivating demand-side flexibility solutions beyond conventional storage. This paper explores aggregated thermostatically controlled loads (TCLs), particularly inverter air conditioners (IACs), as virtual energy storage systems (VESSs) to provide scalable frequency regulation. A comprehensive review of TCL modeling frameworks, including equivalent thermal parameter, capacity-oriented, and reduced-order approaches, is presented. Different control and optimization strategies, ranging from fixed-droop to active PI-based VESS controllers, have been studied and explored. Analytical integration within a two-area automatic generation control (AGC) framework demonstrated the capability of aggregated IACs to emulate synthetic inertia and damping. Simulation case studies validated the performance across step and ramp disturbances, which showed improved frequency nadir, faster settling, and enhanced interarea oscillation damping. Comfort-aware regulation was incorporated using thermostat deadbands, ASHRAE-55 adaptive bands, and predicted mean vote/predicted percentage of dissatisfied indices, confirming compliance without compromising grid support. Sensitivity analysis through systematic delay sweeps and Monte Carlo parameter perturbations quantified robustness against latency and model uncertainty, revealing bounded degradation up to ∼0.5–0.6 s delay and ±40% parameter spread. Results highlight TCL aggregation as a low-cost, occupant-compliant, and delay-resilient demand-side resource for ancillary services in high-renewable power systems.