Quantum parallel model predictive control for grid-connected solid oxide fuel cells

Abstract

The fuel utilization of the grid-connected solid oxide fuel cells (GSOFCs) controlled by model predictive control (MPC) and proportional–integral–derivative (PID) controller is 80.0%, which should be improved. MPC obtains the local optimal control strategy online when the power demand change. The predicted accuracy of MPC should be efficiently improved when the system is in a dynamic environment. This work proposes quantum parallel model predictive control (QPMPC) to solve the optimal local problem of MPC. The QPMPC contains one real-life quantum MPC and three virtual parallel quantum MPC controllers. The quantum revolving gate of the QPMPC can achieve more information about the optimal solution, greatly improving the solution efficiency and effectiveness. The MPC of the QPMPC obtains a current optimal strategy determined by online receding-horizon optimization for minimizing control errors. The QPMPC obtains higher prediction accuracy and better ability of power tracking. The QPMPC for GSOFC can save fuel consumption and improve the economy of renewable power generation. The QPMPC for GSOFC achieves higher fuel utilization of hydrogen (82.0%) than MPC (80.0%), PID (80.0%), reinforcement learning (80.0%), fuzzy control (80.0%) and sliding mode control (79.7%).