Fault Tolerant Control System Design with Explicit Consideration of Performance Degradation

To design fault tolerant control systems (FTCS), one of the important issues to consider is whether to recover the original system performance/functionality completely or to accept some degree of performance degradation after occurrence of a fault. What are the consequences if the performance degradation is not taken into consideration and how to take such performance degradation into account in the design process? Most of the earlier work on FTCS design is centered around the philosophy to recover the prefault system performance as much as possible [8, 10, 12, 14, 16]. In practice, however, as a result of an actuator fault, the degree of the system redundancy and the available actuator capabilities could be significantly reduced. If the design objective is still to maintain the original system performance, this may force the remaining actuators to work beyond the normal duty to compensate for the handicaps caused by the fault. This situation is highly undesirable in practice due to physical limitations of the actuators. The consequence of the so-designed FTCS may lead to actuator saturation, or worse still, to cause further damage. Therefore, trade-off between achievable performance and available actuator capability should be carefully considered in all FTCS designs. Designing an FTCS against actuator faults to achieve specified degraded performance without violating the actuator limits is therefore the main focus of the work presented here. In a control system, there are two aspects of performance: dynamic and steady-state. In FTCS, both types should be considered as well. To represent the degradation in dynamic performance, one could use a performance-degraded reference model with a model-following control principle. In general, at least two different models: one for normal and one or more for impaired systems need to be used. Furthermore, to avoid actuator saturation, adjustment to the system command input levels is often necessary in the event of actuator failure. One way to achieve this is through reference governor/management [1, 4, 5], or command limiting in the context of flight control [2, 9]. The command management can be designed separately from the feedback controller. An adjustment strategy has been proposed here to provide an appropriate command input at both the steady-state and during the initial period of controller reconfiguration. The paper is organized as follows. Modeling of actuator faults, the concept of performance degradation in FTCS and the overall structure of the proposed FTCS are presented in Section II. A scheme for selecting a degraded reference model and a strategy for managing the command input in the presence of actuator faults are proposed in Section III. Detailed design process and associated algorithms are