A NOVEL INTELLIGENT FAULT OBSERVER TO DIAGNOSE ACTUATOR FAULT AND SENSOR NOISE BASED ON PROBABILITY DISTRIBUTIONS

作     者：AlBayati, Ahmad Hussain 

作者机构：Department of Software Computer Science and Information Technology University of Kirkuk Iraq 

出 版 物：《International Journal of Robotics and Automation》 (Int J Rob Autom)

年 卷 期：2024年第39卷第10期

核心收录：

学科分类：0810[工学-信息与通信工程] 12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 0808[工学-电气工程] 0809[工学-电子科学与技术（可授工学、理学学位）] 08[工学] 0837[工学-安全科学与工程] 0714[理学-统计学（可授理学、经济学学位）] 0701[理学-数学] 

主　　题：fault prognosis (FP) Fault tolerant control (FTC) intelligent optimal actuator fault diagnosis observer (IAFO) optimal actuator fault diagnosis observer (AFO) sequential sampling filter (SSF) 

摘      要：This paper aims to bridge the gap between classic model base observer and intelligent observer based on the model base and probability distributions to monitor the plant which is affected by actuator fault and sensor noise simultaneously. The paper includes the implementing and designing of two new optimal actuator observers which also relies on two optimal theories. The first new optimal actuator fault observer is a model base observer based on Lyapunov conditions while, the novelty of the second optimal actuator observer comes from introducing a fault diagnosis algorithm based on a sequential sampling filter (SSF) as a prominent part of the algorithm to filter the diagnosed fault in using the rules of the algorithm. The filters depend on the likelihood ratio, which consists of two hypothetical distributions;the target distribution and the sampling distribution. The new intelligent observer has given rise to innovative algorithms in the field of fault detection and diagnose observer. Moreover, the algorithms verify Lyapunov conditions. In addition, to demonstrate the integrity of the observers, a comparison has been made in the presence of white sensor noise, which is the main reason for inaccurate measurement in addition to two types of actuator faults;Gaussian and non-Gaussian. The study used to verify the performance, where two physical and relative error criteria have been used to evaluate the performance. However, the simulation results appear the flexibility of the second observer to diagnose the actuator fault with sensor noise together due to the number of parameters in diagnosing rules and parameters of the two probabilities. © 2024 Acta Press. All rights reserved.