International Scientific Journals
of Scientific Technical Union of Mechanical Engineering "Industry 4.0"

The paper presents possibility of fault detection and isolation in rotation machinery using analytical redundancy. It outlines the most important techniques of model-based residual generation using parameter identification and state estimation methods with emphasis the problems of reliability. A solution to the fundamental problem of fault detection providing the maximum achievable effectiveness by using condition-based maintenance system, reducing downtime, decreasing maintenance cost, and increasing machine availability is given. With the aim of synthesizing and providing the information of researcher`s community, this paper attempts to summarize and classify the recent published techniques in diagnosis and prognosis of rotating machinery. Furthermore, it also discusses the opportunities as well as the challenges for conducting advance research in the field of remain useful life prognosis.

The results are very important for robust instrument fault detection, component fault detection and actuator fault detection. Finally we discuss the approach of fault diagnosis using a combination of analytical and knowledge-based redundancy.

The active fault-tolerant control approach relies heavily on the occurred faults. Higher performances and more rigorous security requirements have invoked an ever increasing demand to develop real time fault detection and isolation system. The problem of fault diagnosis using analytical redundancy (model-based) methods has received increasing attention during recent years due to the rapid growth in available computer power. The main objective is to design and maintenance a fault-tolerant control system which guarantees a high overall system reliability and dependability both in nominal operation and in the presence of faults. Such an objective is achieved by a control performance index, which is proposed based on system reliability analysis. The methods involve generation and evaluation of signals that are accentuated by faults that have actually occurred. The procedures for generating such signals, called residuals, are based on two main distinct approaches. Direct approach consists in the elimination of all the unknown variable , keeping input-output relations involving only observable variables. Indirect approach estimates states, outputs or parameters in order to generate discrepancy signals obtained by the difference between the actual variables and their estimates.