Applications of machine learning to machine fault diagnosis: A review and roadmap

A fault diagnosis approach based on multi-sensor data fusion is a promising tool to deal with complicated damage detection problems of mechanical systems. Nevertheless, this approach suffers from two challenges, which are (1) the feature extraction from various types of sensory data and (2) the selection of a suitable fusion level. It is usually difficult to choose an optimal feature or fusion level for a specific fault diagnosis task, and extensive domain expertise and human labor are also highly required during these selections. To address these two challenges, we propose an adaptive multi-sensor data fusion method based on deep convolutional neural networks (DCNN) for fault diagnosis. The proposed method can learn features from raw data and optimize a combination of different fusion levels adaptively to satisfy the requirements of any fault diagnosis task. The proposed method is tested through a planetary gearbox test rig. Handcraft features, manual-selected fusion levels, single sensory data, and two traditional intelligent models, back-propagation neural networks (BPNN) and a support vector machine (SVM), are used as comparisons in the experiment. The results demonstrate that the proposed method is able to detect the conditions of the planetary gearbox effectively with the best diagnosis accuracy among all comparative methods in the experiment.

An Adaptive Multi-Sensor Data Fusion Method Based on Deep Convolutional Neural Networks for Fault Diagnosis of Planetary Gearbox.

Since the early 1980s, wind power technology has experienced an immense growth with respect to both the turbine size and market share. As the demand for large-scale wind turbines and lor operation & maintenance cost continues to raise, the interest on condition monitoring system has increased rapidly. The main components of wind turbines are the focus of all CMS since they frequently cause high repair costs and equipment downtime. However, vast quantities of their failures are caused due to a bearing failure. Therefore, bearing condition monitoring becomes crucial. This paper aims at providing a state-of-the-art review on wind turbine bearing condition monitoring techniques such as acoustic measurement, electrical effects monitoring, power quality, temperature monitoring, wear debris analysis and vibration analysis. Furthermore, this paper will present a literature review and discuss several technical, financial and operational challenges from the purchase of the CMS to the wind farm monitoring stage.

A review of wind turbine bearing condition monitoring: State of the art and challenges

Intelligent machinery fault diagnosis system has been receiving increasing attention recently due to the potential large benefits of maintenance cost reduction, enhanced operation safety and reliability. This paper proposes a novel deep learning method for rotating machinery fault diagnosis. Since accurately labeled data are usually difficult to obtain in real industries, data augmentation techniques are proposed to artificially create additional valid samples for model training, and the proposed method manages to achieve high diagnosis accuracy with small original training dataset. Two augmentation methods are investigated including sample-based and dataset-based methods, and five augmentation techniques are considered in general, i.e. additional Gaussian noise, masking noise, signal translation, amplitude shifting and time stretching. The effectiveness of the proposed method is validated by carrying out experiments on two popular rolling bearing datasets. Fairly high diagnosis accuracy up to 99.9% can be obtained using limited training data. By comparing with the latest advanced researches on the same datasets, the superiority of the proposed method is demonstrated. Furthermore, the diagnostic performance of the deep neural network is extensively evaluated with respect to data augmentation strength, network depth and so forth. The results of this study suggest that the proposed intelligent fault diagnosis method offers a new and promising approach.

Intelligent rotating machinery fault diagnosis based on deep learning using data augmentation

The task of condition monitoring and fault diagnosis of rolling element bearing is often cumbersome and labour intensive. Various techniques have been proposed for rolling bearing fault detection and diagnosis. The challenge however, is to efficiently and accurately extract features from signals acquired from these elements, particularly in the time–frequency domain. A new time–frequency technique, known as basis pursuit, was recently developed. This paper presents an application of this new basis pursuit method in the extraction of features from signals collected from faulty rolling bearings with inner race and outer race faults. Results obtained using this new technique were compared with discrete wavelet packet analysis (DWPA) and the matching pursuit technique. Basis pursuit represents features with very fine resolution and sparsity in the time–frequency domain thus rendering easier interpretation of the analysed results. The technique also improves the signal to noise ratio so that subsequent fault detection and identification can be conducted with confidence.

Fault diagnosis of rolling element bearings using basis pursuit

Condition monitoring of rotating machinery has attracted more and more attention in recent years in order to reduce the unnecessary breakdowns of components such as bearings and gears which suffer frequently from failures. Vibration based approaches are the most commonly used techniques to the condition monitoring tasks. In this paper, we propose a bearing fault detection scheme based on support vector machine as a classification method and binary particle swarm optimization algorithm (BPSO) based on maximal class separability as a feature selection method. In order to maximize the class separability, regularized Fisher's criterion is used as a fitness function in the proposed BPSO algorithm. This approach was evaluated using vibration data of bearing in healthy and faulty conditions. The experimental results demonstrate the effectiveness of the proposed method.

Bearing fault diagnosis using multiclass support vector machines with binary particle swarm optimization and regularized Fisher's criterion

This paper presents the application of new time frequency method, ensemble empirical mode decomposition (EEMD), in purpose to detect localized faults of damage at an early stage. EEMD is a self adaptive analysis method for non-linear and non-stationary signals and it was recently proposed by Huang and Wu to overcome the drawbacks of the traditional empirical mode decomposition (EMD). The vibration signal is usually noisy. To detect the fault at an early stage of its development, generally the residual signal is used. There exist different methods in literature to calculate the residual signal, in this paper we mention some of them and we propose a new method which is based on EEMD. The results given by the different methods are compared by using simulated and experimental signals.

/pdf/gearbox-fault-diagnosis-using-ensemble-empirical-mode-3cura5405h.pdf

Gearbox fault diagnosis using ensemble empirical mode decomposition (EEMD) and residual signal

The ensemble empirical mode decomposition (EEMD) was largely used in the diagnosis of the rotating machines, this method could detect the defect at an early stage in the case of non variable speed or slightly variable, but when the speed of the machine varies in acceleration or deceleration the use of the EEMD under these conditions shows a limitation with the detection of the impulses, that are influenced by the presence of the mode mixing, and the end effect. To detect the shocks due to the defect where the variation of speed is forced by the working conditions, we propose to use the Variational Mode Decomposition (VMD) which was recently proposed by Konstantin Dragomiretskiy. This method gave promising results in the detection of the defects on machine elements under non stationary conditions imposed by the variation of speed and torque. In this work, first we show by simulated signal the advantage of VMD compared to the EEMD in the detection of impulses in the case of variable speed and load. Then, we analyze vibration signals given by a dynamic modeling of a gear transmission in the case of non stationary load and speed, for healthy gear and two different of localized faults (early and advanced). The modes are extracted using VMD and followed by calculation of spectrogram and statistics values, which give more information about the defect and allow us to detect it at an early stage.

/pdf/detection-of-gear-faults-in-variable-rotating-speed-using-4qvd46zle4.pdf

Detection of gear faults in variable rotating speed using variational mode decomposition (VMD)

Currently, work on the automation of vibration diagnosis is mainly based on indicators extracted from Time sampled Acceleration signals. There are other attractive alternatives such as those based on Angle synchronized measurements, which can provide a considerable number of more relevant and diverse indicators and, thus, lead to better performance in gear fault classification. The diversity of angular measurements (Instantaneous Angular Speed, Transmission Error and Angular sampled Acceleration) represents potential sources of relevant information in fault detection and diagnosis systems. These complementary measurements of existing signals or new relevant signals allow the construction of Feature Vector (FV) offering robust and effective classification methods even for different or non-stationary running speed conditions. In this paper, we propose to build several FVs based on indicators derived from the angular techniques to compare them to the ones calculated from the time signals, proving their superior performance in detection and identification of gear faults. It will be a question to demonstrate the effectiveness of angular indicators in increasing classification performances, using a supervised classifier based on Artificial Neural Networks and thus determining the most suitable signals.

Ahmed Felkaoui

Papers

Bearing fault diagnosis using multiclass support vector machines with binary particle swarm optimization and regularized Fisher's criterion

Gearbox fault diagnosis using ensemble empirical mode decomposition (EEMD) and residual signal

Detection of gear faults in variable rotating speed using variational mode decomposition (VMD)

Contribution of angular measurements to intelligent gear faults diagnosis

Gear fault diagnosis under non-stationary operating mode based on EMD, TKEO, and Shock Detector