Joint frequency and voltage support from inverter air-conditioner fleets through network-aware predictive control
Year: 2026
Type: preprint
Abstract: Inverter-based air-conditioner (IAC) fleets can provide frequency support through compressor active-power modulation and voltage support through inverter reactive-power exchange. Coordinating these services requires joint decisions that account for feeder conditions, thermal dynamics, and shared inverter capacity. This paper develops a network-aware fleet model predictive controller (MPC) that combines frequency and thermal prediction with feeder-minimum voltage feedback, updated active/reactive-power sensitivity, and an apparent-power constraint. The resulting convex optimization selects representative per-IAC commands for aggregate fleet dispatch. For the IEEE 13-node test feeder with a 1200 kW/600 kvar load event, the complete 600 s scenario yields 1100 kvar peak support and a minimum feeder voltage of 0.9499 pu, compared with 0.9274 pu without support and 0.9383 pu with active-power MPC plus local Volt-Var control. The proposed controller retains the latter's 59.910 Hz frequency nadir. Reducing inverter rating from 4.0 to 2.6 kVA lowers peak reactive support to 990 kvar and reduces the active-power ceiling by up to 0.92 kW per IAC, identifying the effect of scarce headroom. Across 26 scenarios using National Institute of Standards and Technology measured-data thermal calibration, mean frequency nadir improves from 59.805 to 59.901 Hz and mean minimum fleet-weighted voltage from 0.9399 to 0.9704 pu, with no comfort violations during the tested short events. Nonlinear frequency, transmission--distribution, and converter simulations assess dynamic response and command tracking. The results establish how feeder-informed active/reactive coordination adds voltage support while preserving compressor frequency response, and how inverter rating and disturbance location determine the attainable service.
Source: Research Square
Institution: Skolkovo Institute of Science and Technology
Language: English
Topic: Smart Grid Energy Management
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7224956649