The objective of this work is to develop techniques to automate the condition-based maintenance procedure. It is observed that vibration signals are capable of alarming the malfunctions in machineries. In order to overcome the shortcomings in the traditional vibration analysis using time-domain and frequency-domain features, two new approaches based on wavelet transform, artificial neural network and fuzzy rules are proposed for detecting and localizing defects in rolling element bearings. The two expert systems are developed and tested with the use of vibration signals collected from the bearing housing of an experimental setup. Experiment results show that the proposed approaches are sensitive and reliable in detecting defects on the outer race, inner race and rolling elements of bearings. The proposed approaches may be used for other fault diagnoses such as gear faults, coupling faults, belts in industries. It is also expected from the obtained results that the generalized defect detection will be easi...

Development of EBP-Artificial neural network expert system for rolling element bearing fault diagnosis

The objective of this paper is to present recent developments in the field of machine fault signature analysis with particular regard to vibration analysis. The different types of faults that can be identified from the vibration signature analysis are, for example, gear fault, rolling contact bearing fault, journal bearing fault, flexible coupling faults, and electrical machine fault. It is not the intention of the authors to attempt to provide a detailed coverage of all the faults while detailed consideration is given to the subject of the rolling element bearing fault signature analysis.

/pdf/machine-fault-signature-analysis-37twstnp96.pdf

Machine Fault Signature Analysis

Purpose – The objective of this paper is to provide a brief review of recent developments in the area of applications of ANN, Fuzzy Logic, and Wavelet Transform in fault diagnosis. The purpose of this work is to provide an approach for maintenance engineers for online fault diagnosis through the development of a machine condition‐monitoring system.Design/methodology/approach – A detailed review of previous work carried out by several researchers and maintenance engineers in the area of machine‐fault signature‐analysis is performed. A hybrid expert system is developed using ANN, Fuzzy Logic and Wavelet Transform. A Knowledge Base (KB) is created with the help of fuzzy membership function. The triangular membership function is used for the generation of the knowledge base. The fuzzy‐BP approach is used successfully by using LR‐type fuzzy numbers of wavelet‐packet decomposition features.Findings – The development of a hybrid system, with the use of LR‐type fuzzy numbers, ANN, Wavelets decomposition, and fuzz...

Application of ANN, Fuzzy Logic and Wavelet Transform in machine fault diagnosis using vibration signal analysis

Gears are critical element in a variety of industrial applications such as machine tool and gearboxes. An unexpected failure of the gear may cause significant economic losses. For that reason, fault diagnosis in gears has been the subject of intensive research. Vibration analysis has been used as a predictive maintenance procedure and as a support for machinery maintenance decisions. As a general rule, machines do not breakdown or fail without some form of warning, which is indicated by an increased vibration level. By measuring and analysing the machine's vibration, it is possible to determine both the nature and severity of the defect, and hence predict the machine's failure. The vibration signal of a gearbox carries the signature of the fault in the gears, and early fault detection of the gearbox is possible by analysing the vibration signal using different signal processing techniques. This paper presents a review of a variety of diagnosis techniques that have had demonstrated success when applied to rotating machinery, and highlights fault detection and identification techniques based mainly on vibration analysis approaches. The paper concludes with a brief description of a new approach to diagnosis using neural networks, fuzzy sets, expert system and fault diagnosis based on hybrid artificial intelligence techniques.

An investigation on gearbox fault detection using vibration analysis techniques: A review

A Fault Diagnosis System for Rotary Machinery Supported by Rolling Element Bearings

The authors propose a multilayer-feedforward-network-based machine state identification method, and represent certain fuzzy relationships between the fault symptoms and causes with high nonlinearity between the input and the output of the network. As a practical diagnosis example, the rolling bearing diagnosis problem has been studied. By collecting the vibration signals of its operation and using the diagnosis model, one can make a decision about the fault causes and fault degree. Simulation experiments have shown that the proposed diagnosis method achieves better performance consisting in high correct classification rate and good flexibility. >

Neural network and its application on machinery fault diagnosis

The authors present a method to efficiently implement the recursive instrumental variable (RIV) algorithm, which is often encountered in many practical applications. A sequential block-processing algorithm is derived, and its fast Fourier transform (FFT) implementation is discussed. Under several easily met conditions, this method can offer a remarkable computational advantage over the previous fast-RIV. To accommodate the slowly time-varying input situations, an exponentially windowed extension of the result is also given. >

http://ieeexplore.ieee.org/iel2/844/1812/00048618.pdf

A fast sequential block-processing instrumental variable algorithm and its FFT implementation

A version of recursive least-squares (RLS) adaptive algorithms, correlation-domain recursive least-squares (CDRLS) algorithms, is presented for identification applications using two correlation vectors to track the information. It is shown that in the CDRLS, various data windows can be manipulated simply by adjusting the correlation vectors in different ways, leading to the flexibility in exact initialization and real-time variation of the window length. In addition, the CDRLS is block-processing applicable and numerically stable. Two typical identification illustrations are provided. >

The correlation-domain recursive least-squares algorithms for identification

Two supervised learning algorithms for the continuous feedback neural network (CFNN) have been derived which can learn multiple patterns correctly. It is noted that, when the parameters of the CFNN are determined, its dynamic behavior is completely determined, so determining the weight coefficients of the CFNN is a crucial step for using the CFNN. For a CFNN used as associative memory, one wants the required patterns to be the equilibrium points of the CFNN. >

http://ieeexplore.ieee.org/iel2/669/6078/00236965.pdf

Zhen-Ya He

Papers

Neural network and its application on machinery fault diagnosis

A fast sequential block-processing instrumental variable algorithm and its FFT implementation

The correlation-domain recursive least-squares algorithms for identification

Supervised learning in continuous feedback neural network