Vibration Signal Analysis Research Articles

Anomaly Detection on Wind Turbines Based on a Deep Learning Analysis of Vibration Signals

ABSTRACT In this paper, we present a Semi-Supervised Deep Learning approach for anomaly detection of Wind Turbine generators based on vibration signals. The proposed solution is integrated into an IoT Platform as a real-time Workflow. The Workflow is responsible for the whole detection process when a new sample is inserted in the IoT Platform, performing data gathering, preprocessing, feature extraction, and classification. The chosen Semi-Supervised Deep Learning model is a DAE, which builds a normality model using healthy data. The classification consists of comparing the reconstruction error for the computed entry with a normality threshold. The normality threshold is selected through an F1-Score analysis of the reconstruction errors over labeled data. Finally, the Workflow can produce notifications to the users whenever unhealthy behavior is noticed. The ability of the proposed mechanism to detect abnormal behavior in wind turbines on an IoT Platform is evaluated using a case study of real-world healthy and unhealthy data from a Wind Turbine. The solution was able to correctly classify every unhealthy sample and presented a low false-positive rate. Moreover, Workflow results can be improved by conditioning alarm triggering with a windowed-based anomaly accumulation.

A Non-Contact Fault Diagnosis Method for Bearings and Gears Based on Generalized Matrix Norm Sparse Filtering.

Fault diagnosis of mechanical equipment is mainly based on the contact measurement and analysis of vibration signals. In some special working conditions, the non-contact fault diagnosis method represented by the measurement of acoustic signals can make up for the lack of contact testing. However, its engineering application value is greatly restricted due to the low signal-to-noise ratio (SNR) of the acoustic signal. To solve this deficiency, a novel fault diagnosis method based on the generalized matrix norm sparse filtering (GMNSF) is proposed in this paper. Specially, the generalized matrix norm is introduced into the sparse filtering to seek the optimal sparse feature distribution to overcome the defect of low SNR of acoustic signals. Firstly, the collected acoustic signals are randomly overlapped to form the sample fragment data set. Then, three constraints are imposed on the multi-period data set by the GMNSF model to extract the sparse features in the sample. Finally, softmax is used to as a classifier to categorize different fault types. The diagnostic performance of the proposed method is verified by the bearing and planetary gear datasets. Results show that the GMNSF model has good feature extraction ability performance and anti-noise ability than other traditional methods.

Time–frequency analysis of torsional vibration signals based on the improved complete ensemble empirical mode decomposition with adaptive noise, Robust independent component analysis, and Prony’s methods

The torsional vibration signals of rotating shafts are multimodal non-stationary noisy signals. The harmonics and attenuation characteristics of these non-stationary signals cannot be obtained effectively by the ordinary short-time Fourier transform algorithm. Although Prony analysis can accurately fit and identify the characteristic coefficients of such non-stationary signals, it is still sensitive to noise. In this article, we propose a system for denoising and identification of torsional vibration signals. In particular, the improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN), wavelet transform, and robust independent component analysis methods are used to denoise the torsional vibration signals, and then, Prony analysis is used to obtain the characteristic parameters of these signals. The proposed algorithm has good denoising performance and it can improve the identification accuracy and reduce the order of the Prony analysis.

Reconstruction and information entropy analysis of frictional vibration signals in running-in progress

Purpose This study aims to reconstruct the frictional vibration signal from noise and characterize the running-in process by frictional vibration. Design/methodology/approach There is a strong correlation between tangential frictional vibration and normal frictional vibration. On this basis, a new frictional vibration reconstruction method combining cross-correlation analysis with ensemble empirical mode decomposition (EEMD) was proposed. Moreover, the concept of information entropy of friction vibration is introduced to characterize the running-in process. Findings Compared with the wavelet packet method, the tangential friction vibration and the normal friction vibration reconstructed by the method presented in this paper have a stronger correlation. More importantly, during the running-in process, the information entropy of friction vibration gradually decreases until the equilibrium point is reached, which is the same as the changing trend of friction coefficient, indicating that the information entropy of friction vibration can be used to characterize the running-in process. Practical implications The study reveals that the application EEMD method is an appropriate approach to reconstruct frictional vibration and the information entropy of friction vibration represents the running-in process. Based on these results, a condition monitoring system can be established to automatically evaluate the running-in state of mechanical parts. Originality/value The EEMD method was applied to reconstruct the frictional vibration. Furthermore, the information entropy of friction vibration was used to analysis the running-in process.

Vibration Behavior during Underground Drilling Based on an Innovative Measurement Method and the Application

Pressure relief drilling is one of the most common techniques to reduce the impact of rock burst, but the useful dynamic phenomena in the drilling process are ignored due to the lack of corresponding technical methods. Based on the fact, an innovative measurement method for vibration signals during the drilling construction by pneumatic drill rig is presented in this study. The acquisition and analysis of vibration signals show that the vibration information can reflect the drilling depth, the amplitude of different depth can reflect the pressure, and the vibration events of different depths can reflect the pressure relief effect. The method proposed in the study is of great significance to the supervision of the relief work quantity, the disclosure of the pressure situation, and the evaluation of the pressure relief effect.

Analysis and Application of Vibration Signals of Aerospace Servo Actuator

This paper analyzes the vibration signals of aerospace servo actuator and studies their applications. In the signal pre-processing, an adaptive filter based on genetic algorithm is added to improve the signal-to-noise ratio. Time domain analysis is used to monitor the vibration signals and determine the main vibration directions. Frequency domain analysis is used to extract the main vibration frequencies and perform preliminary fault analysis. The WT method with optional window function is used for time-frequency analysis, which can describe the relationship between vibration signal frequency and time, and decompose the signal characteristics required for the analysis. In the research, LabVIEW and MATLAB are used for data processing and analysis. The research lays the foundation of aerospace servo actuator vibration signal analysis and processing system.

Application of multi-layer denoising based on ensemble empirical mode decomposition in extraction of fault feature of rotating machinery

Aiming at the problem that the weak features of non-stationary vibration signals are difficult to extract under strong background noise, a multi-layer noise reduction method based on ensemble empirical mode decomposition (EEMD) is proposed. First, the original vibration signal is decomposed by EEMD, and the main intrinsic modal components (IMF) are selected using comprehensive evaluation indicators; the second layer of filtering uses wavelet threshold denoising (WTD) to process the main IMF components. Finally, the virtual noise channel is introduced, and FastICA is used to de-noise and unmix the IMF components processed by the WTD. Next, perform spectral analysis on the separated useful signals to highlight the fault frequency. The feasibility of the proposed method is verified by simulation, and it is applied to the extraction of weak signals of faulty bearings and worn polycrystalline diamond compact bits. The analysis of vibration signals shows that this method can efficiently extract weak fault characteristic information of rotating machinery.

Vibration signature analysis of coffee beans processing Machineries

It is practical that Major research in the analysis of rolling elements was carried out in the case of one-point defects. For multiple and generalized defect diagnosis, there is a lot of flexibility. The vibration signals for various machines with faulty rolling component bearings are composed & analyzed using vibration signature analysis techniques in this study. Multiple defects on bearing components are used to test the test bearings. The use of vibration signature analysis methods to identify faults in rolling aspect bearings by means of time & frequency domain procedures is investigated in order to determine the best method for fault discovery in the case of several faults in bearing machinery. The feature extracted in this study is expected to be useful in the development of automated fault diagnostics systems for bearing components with multiple defects. In This work we have applied vibration signature analyses on the ball bearings of coffee processing machines and evaluated the results.

Fault detection and diagnosis using vibration signal analysis in frequency domain for electric motors considering different real fault types

PurposeElectric motors are present in most industries today, being the main source of power. Thus, detection of faults is very important to rise reliability, reduce the production cost, improving uptime and safety. Vibration analysis for condition-based maintenance is a mature technique in view of these objectives.Design/methodology/approachThis paper shows a methodology to analyze the vibration signal of electric rotating motors and diagnosis the health of the motor using time and frequency domain responses. The analysis lies in the fact that all rotating motor has a stable vibration pattern on health conditions. If the motor becomes faulty, the vibration pattern gets changed.FindingsResults showed that through the vibration analysis using the frequency domain response it is possible to detect and classify the motors in several induced operation conditions: healthy, unbalanced, mechanical looseness, misalignment, bent shaft, broken bar and bearing fault condition.Originality/valueThe proposed methodology is verified through a real experimental setup.

Wavelet-based Synchroextracting Transform: An effective TFA tool for machinery fault diagnosis

In the literature of fault diagnosis for machinery, mechanical vibration signal analysis serves one of the most important approaches. By extracting features of vibration signal, such as instantaneous frequency (IF), instantaneous amplitude or spectral kurtosis, the fault of machinery can be effectively diagnosed. As an efficient tool of analyzing vibration signal, time–frequency (TF) analysis (TFA) technology has been widely employed in this area. Restricted by Heisenberg uncertainty theory, the TF resolution of the traditional linear TF analysis technique may not be optimal. To overcome this problem, in this paper, inspired by wavelet-based synchrosqueezing transform (WSST), synchroextracting transform (SET) and the ideal TFA principle, we present a novel wavelet-based TFA approach, which is named wavelet-based synchroextracting transform (WSET) and acts as a TF post-processing technology. The core idea of WSET is that we only extract wavelet transform TF spectrum of signal in scale correspond to IF, and get rid of burry TF energy. This proposed method is capable of enhancing the concentration of TF representation. In the current study, firstly, the theory of the WSET is analytically deduced. Secondly, the validity of WSET is proved through processing the nonstationary and multi-component bat signal. Finally, we employ two benchmarks of rotor and rolling bearing to verify the effectiveness of WSET to extract the failure features for malfunction diagnosis.

Impact of selected operational parameters on measures of technical condition of rolling bearings in means of transport built based on analysis of vibration signals

This article presents an innovative method of diagnostics of rolling bearings used in the bearing nodes of motor vehicles, with the use of a prototype specialist stand. The tests were carried out based on a developed research plan, which included the impact of damage to the bearing and tyre of the vehicle, as well as the vehicle speed. Vibration accelerations were recorded in three measurement axes. Signal spectra were created based on the time courses of the vibration signals and were further analysed. The presented method is aimed at detecting excessive wear of rolling bearings in wheels from its early period.

Fault diagnosis and severity analysis of rolling bearings using vibration image texture enhancement and multiclass support vector machines

Fault detection and diagnosis of its severity for machine health monitoring can be stated as a nested classification problem. For a faulty bearing, each fault location whether belonging to inner race, outer race or the ball can be seen as multiclass classification with three classes while the varying degree of severity in each class can be viewed as a sub classification task. The peculiar vibration patterns generated from the flaws in different bearing parts and with varying degree of distortion can be classified into various classes and subclasses for analysis of vibration signatures. This paper proposes a multiclass support vector machines (MSVMs) based fault classification approach for fault diagnosis of ball bearings. The one dimensional vibration signals are converted to two dimensional gray scale images resulting in textural patterns which are then enhanced using the wave atom transform. Features such as semivariance, skewness and entropy are extracted from the texture images and the MSVM is then trained using feature matrices generated from feature vectors. The MSVM is trained in two phases; in the first phase, the classifier categorizes the location of the fault and in the second phase the classifier does the diagnosis regarding the size of the fault at that particular location. Simulation results show that the proposed technique is highly robust in locating the fault and its severity.

Automatic representation and detection of fault bearings in in-wheel motors under variable load conditions

The wear fault of the inner and outer race of bearing in an in-wheel motor is vital as the performance of bearing effects the transmission efficiency. Unlike the commonly used vibration analysis-based monitoring methods, the accelerometer cannot be installed inside the in-wheel motor since it would damage the structural reliability. The electricity-related signals sourcing from the motor controller may be a feasible way. In this paper, a load demodulation and normalization method under the condition of no vibration sensor is proposed to solve the problem of fault detection and diagnosis of the in-wheel motor transmission component under the condition of variable load. First, the stator current is selected to reflect the load changes according to the mathematical model of in-wheel motor in the three-phase stationary coordinates. Second, the order tracking is applied to synchronize the sampled signals on time domain with synchronous resampling, which converts the equally spaced signals into the angularly spaced signals. The merits are improving the definition of time–frequency representations (TFRs) of wear fault and avoiding the difficulty to determine the resolution of TFRs caused by load fluctuations. The short-time Fourier transform (STFT) is introduced to convert the angularly spaced signals into the TFRs. Finally, an adapted convolutional neural network (CNN) with random weight initialization and dropout strategy is employed to classify the wear of inner race and outer race. The proposed framework is verified on the in-wheel motor simulation platform. The experimental results show that the proposed method has a higher fault detection precision than the other methods. The core contribution reveals that this paper does not use the vibration-free sensor. Instead, the electrical signals of the controller are used to realize the fault detection and diagnosis of the inner and outer wear of the in-wheel motor bearing under the variable load conditions. Compared with the commonly used fault diagnosis method based on vibration signal analysis, the proposed method is more suitable for the condition monitoring system of the in-wheel motor transmission mechanism.

Deep Learning Approach for Vibration Signals Applications.

This study discusses convolutional neural networks (CNNs) for vibration signals analysis, including applications in machining surface roughness estimation, bearing faults diagnosis, and tool wear detection. The one-dimensional CNNs (1DCNN) and two-dimensional CNNs (2DCNN) are applied for regression and classification applications using different types of inputs, e.g., raw signals, and time-frequency spectra images by short time Fourier transform. In the application of regression and the estimation of machining surface roughness, the 1DCNN is utilized and the corresponding CNN structure (hyper parameters) optimization is proposed by using uniform experimental design (UED), neural network, multiple regression, and particle swarm optimization. It demonstrates the effectiveness of the proposed approach to obtain a structure with better performance. In applications of classification, bearing faults and tool wear classification are carried out by vibration signals analysis and CNN. Finally, the experimental results are shown to demonstrate the effectiveness and performance of our approach.

The analysis of tool vibration signals by spectral kurtosis and ICEEMDAN modes energy for insert wear monitoring in turning operation

Surface finish quality is becoming even more critical in modern manufacturing industry. In machining processes, surface roughness is directly linked to the cutting tool condition; a worn tool generally produces low-quality surfaces, incurring additional costs in material and time. Therefore, tool wear monitoring is critical for a cost-effective production line. In this paper, the feasibility of a vibration-based approach for tool wear monitoring has been checked for turning process. AISI 1045 unalloyed carbon steel has been machined with TNMG carbide insert twenty-one times for a total of 27 min of machining, which was a necessary amount of time to exceed (300 μm) as a flank wear threshold. Vibration signals have been acquired during the operation and then processed in order to extract a correlation between the surface roughness, tool wear level, and vibration comportment. First, spectral kurtosis has been calculated for the twenty-one performed runs signals; this step has allowed the locating of the optimal frequency band that contains the machining vibration signature, yet it did not give significant information about wear evolution. The signals have then been decomposed with ICEEMDAN and the energy of the high-frequency modes has been calculated. It has been found that the energy of the optimal frequency ICEEMDAN modes has increased in proportion to the increase of surface roughness degradation and thus, to tool wear increase. Therefore, IMF’s energy can be used for tool wear condition monitoring.

Multisource Fault Signal Separation of Rotating Machinery Based on Wavelet Packet and Fast Independent Component Analysis

The vibration signal of rotating machinery compound faults acquired in actual fields has the characteristics of complex noise sources, the strong background noise, and the nonlinearity, causing the traditional blind source separation algorithm not be suitable for the blind separation of rotating machinery coupling fault. According to these problems, an extraction method of multisource fault signals based on wavelet packet analysis (WPA) and fast independent component analysis (FastICA) was proposed. Firstly, according to the characteristic of the vibration signal of rotating machinery, an effective denoising method of wavelet packet based on average threshold is presented and described to reduce the vibration signal noise. In the method, the thresholds of every node of the best wavelet packet basis are acquired and averaged, and then the average value is used as a global threshold to quantize the decomposition coefficient of every node. Secondly, the mixed signals were separated by using the improved FastICA algorithm. Finally, the results of simulations and real rotating machinery vibration signals analysis show that the method can extract the rotating machinery fault characteristics, verifying the effectiveness of the proposed algorithm.

A Multistage Cutting Tool Fault Diagnosis Algorithm for the Involute form Cutter Using Cutting Force and Vibration Signals Spectrum Imaging and Convolutional Neural Networks

In a machining system, tool condition monitoring systems are required to get a high-quality product and to prevent the downtime of machine tools due to tool failures. For this purpose, tool condition monitoring systems have become very important during the years since the mechanical faults can cause high cost. This study introduces a multistage cutting tool fault diagnosis method to detect the presence and level of the involute form cutter faults on the by the cutting force and vibration signal analysis. Therefore, different fault levels (low, medium and high) were generated on the involute form cutter as a tool breakage. During the experiments, the cutting force, vibration and acoustic signals were gathered with three different feed rates for each fault level. The gathered signals were processed by a multistage signal processing algorithm developed in the MATLAB environment. As an initial step, the continuous wavelet transform of the obtained signals was taken and saved as an image by the developed algorithm. After that, a convolutional neural network model is trained and tested by using the obtained images. The developed algorithm firstly checks the presence of the cutting tool fault. Once the algorithm labels the cutting tool is damaged, it then checks the damage level of the cutting tool fault. It is observed from the results, cutting force analysis is sufficient for the detection of cutting tool fault. On the other hand, the cutting force signal analysis is insufficient to detect the damage level of the cutting tool. Therefore, the vibration signal analysis is required to detect the damage level of the cutting tool. Results prove that, by the vibration analysis, the developed algorithm could detect not only the presence of the damage on the cutting tool but also the damage level. The results of the algorithm for each stage and signal are given in the results section.

A Method Combining Fractal Analysis and Single Channel ICA for Vibration Noise Reduction

Aiming at the problem that real engineering vibration signals are interfered by strong noise, this paper proposes a method combining single channel-independent component analysis (SCICA) and fractal analysis (FD) to reduce the effect of noise on the time-frequency analysis of vibration signals. First, phase space reconstruction is performed on the vibration signal to make the proper input for ICA algorithm. The original is then decomposed into several component signals. The fractal dimension of each component signals is calculated to determine whether the signal should be considered noise. Noisy component signals are then processed by wavelet denoising. Finally, the output signal after noise reduction is reconstructed using the filtered “right” component signals. This paper uses the method to analyze real noisy vibration signal. Experimental results show the effectiveness of the proposed method.

Identification of fault size of rolling element bearing on outer race from vibration signature analysis

Prognostics deals with the estimation of the remaining useful life (RUL) of machine, which is predicated based on their present condition and their future operating condition. Rolling element bearing failure is one of the commonly explained reason for machine breakdown. At the point when machine is in running condition and fault occur in bearing, it is extremely difficult to identify fault size in bearing. Thus, for safety of machine it is necessary to change bearing before fault size increases to a significant level. Further, the replacement of bearing might be extremely costly and on other hand chances can’t be taken with safety aspects. Therefore, it is necessary to find the RUL of bearing by identifying the spall size of fault from vibration signal. Fault size estimation can be done by decomposition of vibration signal by using discrete wavelet transform and also introducing the autoregressive method, minimum entropy deconvolution and Hilbert transform. The decomposed signal is divided into peak corresponding to the ball enter into the fault and exit from the fault. Experiment conducted for various size of faults present on the 25 mm inner diameter polyamide cage deep groove ball bearing. The minimum size of 0.3 mm is detected in the present work for outer race defect with only radial load. Further, fault size estimation is investigated for 1mm defect present at outer race subjected to both radial and axial load

Entropy approach in the analysis of vibration and partial discharge signals

The development of new methods of technical diagnostics is an important task aimed at improving the efficiency and safety of the operation of industrial equipment. Such methods are complexes of jointly used methods for obtaining diagnostic information, methods for extracting from it the actual data on the technical state of the diagnosed objects, and methods for organizing diagnostic processes. At the same time, the criteria of importance and value, as a rule, are different for the developers of these methods, developers of diagnostic systems, and direct operators of equipment, which requires the search for solutions that are a compromise for all interested parties and meet certain optimality requirements. One of the new approaches in the field of identifying and controlling new diagnostic information is the entropy method for evaluating data. Testing the applicability of this approach to the analysis of signals of vibration and partial discharges has shown the possibility of obtaining data on changes in the state of the diagnosed object. This information can be used to develop new diagnostic features designed for earlier detection of developing defects, as well as for checking and confirming technical diagnoses obtained by classical methods.