An active fault-tolerant MPC for systems with partial actuator failures

An active fault-tolerant model predictive control (FTMPC) scheme for systems with partial actuator failures is presented. The faulty system is first transformed into a linear system with input uncertainties, and then a minimum variance unbiased (MVU) filter is introduced to estimate both the state and the input uncertainties. To obtain a reliable fault information, a closed loop identification based fault detection and isolation approach is proposed. Offset-free control can be achieved by using the proposed FTMPC in the presence of partial actuator faults and/or unmeasured disturbances. Several example are given to demonstrate the effectiveness of the proposed approach.