Nonlinear Model Predictive Control of a Variable-Speed Pumped-Storage Power Plant

Optimal operation and control of (variable-speed) pumped-storage power plants (PSPPs) is essential to meet the growing demands on the dynamics for the stabilization of power distribution grids with an increasing amount of renewable energy sources. Existing work on the control of PSPPs is typically based on rather simplified system models, in particular of the (long) pipeline system. In this article, a nonlinear model predictive control (NMPC) strategy is proposed, which enables fast closed-loop dynamics while keeping all system constraints, including the pressure constraints along the pipeline system. The control strategy is based on a physics-based model, which enables easy parameterization and application to other plant sizes or topologies. To ensure real-time capability of the model predictive control (MPC) strategy, a number of measures are outlined in this article. Finally, the feasibility of the proposed control strategy is demonstrated by detailed simulation studies. An accurate tracking for fast changing desired grid powers as well as a high robustness with respect to parameter uncertainties is demonstrated.

Automated summary

This article introduces a nonlinear model predictive control strategy for pumped-storage power plants, which achieves fast dynamic adjustment while meeting system constraints, and proposes measures to ensure real-time capability. Through simulation studies, the feasibility and robustness of the control strategy in tracking fast-changing grid powers and parameter uncertainties are demonstrated.