This study examines the negative effects of closed-loop control systems on fault detection in automated systems. Closed-loop controls are essential for maintaining robust performance through feedback mechanisms, even under various uncertainties, including disturbances and system changes. However, this unique advantage of closed-loop control may hinder timely fault detection. When a system changes due to faults, the closed-loop system under such conditions may behave similarly to when it is in healthy conditions. Often in practice, only the closed-loop system data is available, which includes the joint effects of the controller and the system and no additional system information is measured. However, the fault may not be directly observable from the measured information of the closed-loop system. In this case, reduced sensitivities to fault detection or delayed detection can result in severe faults, potentially compromising system reliability and safety. The paper presents theoretically how the corrective nature of closed-loop mechanisms may mask fault detection under different system physical parameter changes. Then a real-world case study of a 2-Degree-of-Freedom aircraft is presented, demonstrating the impact of closed-loop performance under six different types of propeller faults and how this affects fault detection.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

The Negative Impact of Closed-Loop Control on Fault Detection

  • Long Zhang,
  • Heng Yang,
  • Khalid Dandago,
  • Yunbo Li

摘要

This study examines the negative effects of closed-loop control systems on fault detection in automated systems. Closed-loop controls are essential for maintaining robust performance through feedback mechanisms, even under various uncertainties, including disturbances and system changes. However, this unique advantage of closed-loop control may hinder timely fault detection. When a system changes due to faults, the closed-loop system under such conditions may behave similarly to when it is in healthy conditions. Often in practice, only the closed-loop system data is available, which includes the joint effects of the controller and the system and no additional system information is measured. However, the fault may not be directly observable from the measured information of the closed-loop system. In this case, reduced sensitivities to fault detection or delayed detection can result in severe faults, potentially compromising system reliability and safety. The paper presents theoretically how the corrective nature of closed-loop mechanisms may mask fault detection under different system physical parameter changes. Then a real-world case study of a 2-Degree-of-Freedom aircraft is presented, demonstrating the impact of closed-loop performance under six different types of propeller faults and how this affects fault detection.