A Rotating Machinery Fault Diagnosis Method Based on Feature Learning of Thermal Images

Abstract

The rotating machinery plays a vital role in industrial systems, in which unexpected mechanical faults during operation can lead to severe consequences. For fault prevention, many fault diagnostic methods based on vibration signals are available in the literature. However, the vibration signals are obtained by using different types of sensors, which can cause sensor installation issues and damage the rotating machinery. In addition, this kind of data acquisition through vibration signal induces a large amount of signal noise during machine operation, which will challenge the later fault diagnosis. A recent fault detection method based on infrared thermography (IRT) for rotating machinery avoids these issues. However, the corresponding literature is limited by the fact that the characteristics of the manual design cannot characterize the fault completely so that the diagnostic accuracy cannot exceed the diagnostic method based on the vibration signals. This paper introduces a popular image feature extraction method into the fault diagnosis of rotating machinery based on IRT for the first time. First, capturing the IRT images of the rotating machinery in different states, and then two popular feature extraction methods for IRT images, bag-of-visualword, and convolutional neural network, are tested in turn. Finally, the extracted features are classified to implement the automatic fault diagnosis. The developed method is applied to analyze the experimental IRT images collected from bearings, and the results demonstrate that the developed method is more effective than the traditional methods based on vibration signals.INDEX TERMS Fault diagnosis, infrared thermography, convolutional neural network, bag-of-visual-words, feature recognition.

A Review on Deep Learning Applications in Prognostics and Health Management

“…Depending on the complexity of a concrete problem, using a pretrained network or data augmentation techniques may not be necessary. Jia et al trained a CNN with only 450 images per class to diagnose nine faulty states of rolling bearings, attaining nearly 100% accuracy [143]. Likewise, Li et al trained a CNN with only 1400 samples per class to diagnose five module defects in photovoltaic farms, again with highly accurate results [144].…”

“…Strictly speaking, compound faults (several faults occur simultaneously) that are not included in the training set should also be considered unhappened. For this reason, several studies meticulously collected compound faults related data and incorporated them into their training set [104], [129], [143], [162]. However, this type of study is restricted in the sense that the combinatorial explosion of many faulty types prevents us from collecting sufficiently labelled data to train an all-in-one diagnostic model.…”

Multi-Band Frequency Window for Time-Frequency Fault Diagnosis of Induction Machines

“…The proposed method addresses this problem, filling the whole spectrogram with useful diagnostic contents. It relies on frequency shifting the Gaussian analysing window, as in (13). If the current signal is low-pass filtered with a cut-off frequency f b (using, for example, a frequency filter as in Reference [45]), then its spectrogram, built with a Gaussian window (5), will be non-zero only in the frequency band of diagnostic interest [0 − f b ].…”

“…A way to reduce these risks is the continuous monitoring of the machine's condition, in order to detect the presence of a fault at an early stage, when corrective measures can be implemented or maintenance works scheduled. Diverse quantities have been proposed in the technical literature for implementing condition based maintenance systems (CBMS) [1,2], such as the analysis of the stator currents [3][4][5][6][7], machine vibrations [8][9][10], fluxes [11,12], thermal images [13] or acoustic signals [14,15]. These techniques have been applied to detect different types of faults not only of the IM, such as stator inter-turn short circuits [11,16], broken bars [17,18], rotor asymmetries [19], eccentricity [20], bearing faults [6,21], but also of the inverter drive [3] or the mechanical coupling to the load, as gearboxes and pulleys [3,22].…”

A New Intelligent Fault Diagnosis Method of Rotating Machinery under Varying‐Speed Conditions Using Infrared Thermography

“…Monitoring the health condition, prognostics the possible faults can be of importance for the safety operation of rotating machinery. To achieve this purpose, vibration signal is the most commonly used signal resource and the most widely applied method [1][2][3][4][5][6][7]. The recent progress of the vibrationbased fault diagnosis can be reviewed from Ref.…”