Model Predictive Control-Enabled Primary Frequency Support for Variable-Speed Pumped Storage with Mechanical Constraints

Pumped hydro storage (PHS) systems, increasingly deployed in power systems with large shares of wind and solar generation, can play a key role in managing power system frequency. Variable-speed pumped hydro storage (VS-PHS) systems, in particular, have potential for rapid primary frequency response by enabling the quick release of machine rotor kinetic energy. However, using conventional proportional–integral (PI) control for converters and governors can result in large speed deviations and torque imbalance during fast system transients. This issue is intensified in PHS plants with slow hydraulic response, such as those with long penstocks or slow guide-vane adjustments, potentially violating mechanical operating constraints. This paper develops a model predictive control (MPC) strategy for coordinated governor and converter control, accounting for operational constraints. The proposed approach improves coordination of hydraulic and electrical systems, utilising DC-link storage and proactive guide-vane action for rapid power adjustments. Dynamic simulations using a complex nonlinear plant demonstrate that MPC redistributes energy extraction between the DC-link storage and the rotating mass while respecting their imposed limits. Furthermore, robustness tests indicate that MPC performance is sustained under plant nonlinearities and measurement noise. These results highlight the advantages of predictive control for supporting frequency response in VS-PHS systems.