Fault-Tolerant Flight Control Design with Explicit Consideration of Reconfiguration Transients

Abstract

FAULT-TOLERANT control (FTC) has drawn considerable attention over the last three decades, due to its important role in preventing system breakdown via configured system redundancy [1,2]. FTC designmethodologies can generally be divided into either passive FTC or active FTC [3]. A passive FTC is synthesized under both normal and faulty cases, and a fixed controller is always commissioned, regardless of whether or not prescribed malfunctions are present. On the other hand, an active FTC reacts to specific faults by reconfiguring the controller based on the up-to-date information obtained from a fault detection and diagnosis (FDD) unit. Various design approaches have been exploited for active faulttolerant flight control (FTFC) [1,2,4]. Two notable issues exist in an active FTFC system: 1) The mismatch between the previous and current control signals at the reconfiguration/switching instance usually induces unexpected transients of aircraft outputs [5]. Transients are potentially dangerous to post-failure aircraft, when the healthy actuators attempt to compensate for failed actuators. As pointed out in [1], “However, how to manage or reduce these transients during a controller reconfiguration is still an open issue.” In the latest literature [6], a set of reconfigurable controllers is designed in response to the anticipated fault cases, while a hysteresis supervisory approach is adopted to eliminate switching bumps. 2) Another issue is that healthy actuators are easily saturated due to unexpected transients and improper FTFC design. Saturation of healthy actuators can induce secondary damage to the faulty aircraft and may even jeopardize the safety of the aircraft. In [7–9], the concept of “graceful performance degradation” is applied for the active FTFC design under actuator faults. Healthy actuators’ saturation can be prevented by decreasing the requirements for both transient and steady-state performance. Most of the existing studies address the following notions individually: 1) fault accommodation without explicitly considering the impact of reconfiguration switch actions [1,2], 2) bumpless transfer using dynamic output feedback in fault-free systems [10–12], and 3) controller design within actuator limitations [13–16]. From a safety standpoint, it would be highly desirable if one can design a FTFC scheme to ensure that the process of reconfiguration switching is smooth, while the limits of the remaining actuators are not violated. Motivated by this fact, an adaptation mechanism is developed to prevent abrupt transients and actuator saturation by mitigating the deviation between the aircraft and the target model. Consequently, the active FTFC system, including the reconfigurable controller and the adaptationmechanism, is capable of compensating for the adverse effects resulting from actuator abnormalities and a reconfigurable controller switch. The main contribution is that the proposed FTFC scheme can guarantee a graceful reconfiguration process and respect remaining actuators’ capabilities.