Robust Fault Detection and Estimation of Sensor Fault for Closed-loop Control Systems
| Autor: | Yang Zhang, Jian Shi, Shaoping Wang |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
0301 basic medicine
Scheme (programming language) Observer (quantum physics) Computer science Estimation theory 030106 microbiology Fault (power engineering) Fault detection and isolation Compensation (engineering) 03 medical and health sciences 030104 developmental biology Control theory Servo control system Random vibration computer computer.programming_language |
| Zdroj: | AIM |
| Popis: | This paper presents a robust fault detection and estimation (RFDE) scheme exclusively for the feedback sensor of closed-loop dynamic systems. An unknown input observer (UIO) is used to generate robust residuals to minimize the effects of disturbance on the RFDE accuracy. Different from other UIOs, this UIO uses the output of feedback controller and a virtual system output instead of the actual sensor measurement as the inputs to generate sensitive residuals in the presence of feedback sensor fault. To further improve the accuracy of fault estimation, a model-based compensation is also employed to modify the estimated real-time residuals, which possess a specific physical meaning and the modified residuals can be used to identify the fault parameter. A case study of electrohydraulic servo control system is conducted to illustrate the proposed RFDE scheme. Results show that in the presence of random vibration, the proposed method can detect the sensor fault accurately and timely during the steady-state stage of closed-loop system. Meanwhile, the fault parameter estimation has a high accuracy, with no more than ±2.64% when the sensor gain varies within ±20%, which demonstrates the validity of the proposed RFDE scheme. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|