Event-Triggered Decentralized Model Predictive Control for Low-Speed Two-Stroke Marine Main Engine Power Systems

For a low-speed two-stroke marine main engine (LTMME) power system consisting of several coupled subsystems, an event-triggered decentralized model predictive control (MPC) scheme is presented to reduce communication and computational burdens effectively. Firstly, an optimization problem of the LTMME power system with input constraints is formulated to ensure steady operation. Secondly, a constraint is introduced to provide robustness to dynamic couplings, where the inference of system coupling is regarded as external disturbance and the optimization problem is converted into a decentralized control problem. Thirdly, the event-triggering condition is set up for the LTMME power system control problem. An event-triggered decentralized MPC scheme is established to save communication and computational resources. The optimization problem is solved only when the event-triggering condition is satisfied. The feasibility and stability of the proposed algorithm is analyzed. Finally, a simulation between the event-triggered decentralized MPC and time-triggered one is carried out to show the advancement of the proposed algorithm.