Model predictive control of the high-pressure side of simple supercritical CO2 cycle

Control of non-condensing non-ideal-gas power cycles is challenging because of their complex nonlinear dynamics and the presence of state constraints. This article presents a first-principles model predictive control (MPC) strategy for these systems. We demonstrate this methodology on the high-pressure side of a simple supercritical CO 2 cycle power block. We develop the control model using timescale-separation arguments, then implement MPC by linearizing the model online at each sampling instant. Closed-loop simulations with a gas-dynamics truth model demonstrate the effectiveness of the controller. This article demonstrates the suitability of MPC for the supercritical CO 2 cycle, and provides a pathway to implementing MPC for more complex cycle variants such as the recuperated and recompression cycle.