Fault-tolerant control of a ship propulsion system using model predictive control

Abstract

Recently, it has been shown how model predictive control (MPC) can adapt to faults in certain circumstances. This paper describes how MPC was successfully implemented as a fault-tolerant controller for a single engine/propeller model of a ship propulsion system. It is shown that the MPC controller can be tuned to be robust to internal faults that develop in the ship propulsion system, even in the absence of any fault detection and isolation (FDI) information for the internal faults. For the case of sensor faults, it is assumed that FDI information is available and it is shown how the MPC controller, in combination with a Kalman estimator, can drastically improve the tracking response of the system in the presence of sensor faults. The paper concludes that MPC is a very good candidate for a fault-tolerant controller for the ship propulsion system, requiring re-configuration only at the supervisory level, without the need for additional re-configuration in the lower-level control systems.