Wavelet Packet Node Energy Research Articles

Robustness of wavelet features for damage detection with neural network and digital twin

Recent studies demonstrate that the wavelet energy–based features are highly sensitive to local damages, and accurate damage identification could be achieved by integrating such features with neural networks. However, the robustness of the features in practical applications and the feasibility of generating damage information with a digital twin to facilitate training of neural networks have not been systematically investigated. This paper aims to provide new insight into the practicality of using wavelet energy–based features for accurate damage detection of structures. A systematic investigation into the tolerance of the normalised wavelet packet node energy features with neural network (NWPNE-NN) approach against measurement noises, excitations uncertainties and limited frequency range in the measured responses is carried out, with consideration of data fusion from multiple measurement points. The feasibility of a digital twin to simulate damage scenarios and generate training data for neural networks is explored, and the use of relative WPNE as a new feature is proposed. The results show that the NWPNE-NN approach is capable of detecting and quantifying the main damage in a structure. The neural networks trained by data generated from a well-calibrated digital twin is capable of identifying damage in the actual structure.

Mechanomyography signals pattern recognition in hand movements using swarm intelligence algorithm optimized support vector machine based on acceleration sensors

On the basis of extracting mechanomyography (MMG) signal features, the classification of hand movements has certain application values in human-machine interaction systems and wearable devices. In this paper, pattern recognition of hand movements based on MMG signal is studied with swarm intelligence algorithms introduced to optimize support vector machine (SVM). Time domain (TD) features, wavelet packet node energy (WPNE) features, frequency domain (FD) features, convolution neural network (CNN) features were extracted from each channel to constitute different feature sets. Three novel swarm intelligence algorithms (i.e., bald eagle search (BES), sparrow search algorithm (SSA), grey wolf optimization (GWO)) optimized SVM is proposed to train the models and recognition of hand movements are tested for each MMG feature extraction method. Using GWO as the optimization algorithm, time consumption is less than using the other two swarm algorithms. Using GWO with TD+FD features can obtain the classification accuracy of 93.55 %, which is higher than other methods while using CNN to extract features can be independent of domain knowledge. The results confirm GWO-SVM with TD + FD features is superior to some other methods in the classification problem for tiny samples based on MMG.

Fault Diagnosis of Submersible Motor on Offshore Platform Based on Multi-Signal Fusion

As an important production equipment of the offshore platform, the operation reliability of submersible motors is critical to oil and gas production, natural gas energy supplies, and social and economic benefits, etc. In order to realize the health management and fault diagnosis of submersible motors, a motor fault-monitoring method based on multi-signal fusion is proposed. The current signals and vibration signals were selected as characteristic signals. Through fusion correlation analysis, the correlation between different signals was established to enhance the amplitude at the same frequency, so as to highlight the motor fault characteristic frequency components, reduce the difficulty of fault identification, and provide sample data for motor fault pattern identification. Furthermore, the wavelet packet node energy analysis and back propagation neural network were combined to identify the motor faults and realize the real-time monitoring of the operating status of the submersible motor. The genetic algorithm was used to optimize the parameters of the neural network model to improve the accuracy of motor fault pattern recognition. The results show that the combination of multi-signal fusion monitoring and an artificial intelligence algorithm can diagnose motor fault types with high confidence. This research originally proposed the fusion correlation spectrum technology, which solved the shortcomings of the small amplitude and complex composition of the single signal spectrum components in the fault diagnosis and improved the reliability of the fault diagnosis. It further combined the neural network to realize the automatic monitoring and intelligent diagnosis of submersible motors, which has certain application value and inspiration in the field of electrical equipment intelligent monitoring.

A PILOT STUDY OF MECHANOMYOGRAPHY-BASED HAND MOVEMENTS RECOGNITION EMPHASIZING ON THE INFLUENCE OF FABRICS BETWEEN SENSOR AND SKIN

Multi-channel mechanomyography (MMG) signals were acquired from the forearm when the subjects were performing eight classes of hand movements related to rehabilitation training. Ten time domain (TD) features and wavelet packet node energy (WPNE) features were extracted from each channel of MMG, and the hand movements were classified by support vector machine (SVM), extreme learning machine (ELM), linear discriminant analysis (LDA) and [Formula: see text]-nearest neighborhood (KNN) and the classifying results of three methods of collecting MMG (sensors directly on skin, sensors on cotton fabric and sensors on acrylic fiber) were compared. When all TD features were selected and SVM was adopted as the classifier, the total recognition rates of hand movements were 94.0%, 93.9% and 93.6%, respectively, of three collection methods. Using ELM can obtain similar results as SVM, with the recognition rates of 94.3%, 94.3% and 94.1%, respectively, better than using LDA (88.5%, 88.6% and 88.0%) or KNN (88.9%, 89.4% and 89.0%). For each algorithm, using TD features can acquire the highest recognition rates. Once the feature set and the classifier were selected, the total recognition rates were almost equally among three collection methods (especially for some feature sets, the differences are smaller than 1%). The results confirmed that satisfactory effects could be acquired even when the MMG was collected from sensors on fabrics with specific material, thus indicating that MMG has a unique potential value for developing wearable devices.

Concrete piezoceramic smart module pairs-based damage diagnosis of hydraulic structure

As a main reaction on the ageing behavior of hydraulic concrete structure, the structural damage is harm to structural safety. With the dynamic damage experiment of concrete gravity dam model under ambient excitation, a concrete piezoceramic smart module (CPSM) pairs-based approach is studied to implement the dynamic characteristics identification and damage diagnosis of hydraulic structure. Based on the piezoelectric coupling property and elastic-plastic contact property of CPSM, which are analyzed by use of numerical simulation software, the optimal shape of CPSM is given. A self-made CPSM-based monitoring platform is developed to obtain the information on structural damage. A wavelet packet decomposition method of CPSM signal is introduced. By analyzing the relation between CPSM signal amplitude and wavelet packet node energy, an index, which is defined with wavelet packet node energy, is proposed to describe the structural damage. The dynamic experimental result shows that above defined index changes obviously with the structural developing damage, and can well reflect the structural damage status. The feasibility and effectiveness of self-made CPSM-based approach are verified to identify the dynamic characteristics and diagnose the hidden damage for hydraulic structure.

Condition monitoring of induction motor bearing based on bearing damage index

Abstract The rolling element bearings are used broadly in many machinery applications. It is used to support the load and preserve the clearance between stationary and rotating machinery elements. Unfortunately, rolling element bearings are exceedingly prone to premature failures. Vibration signal analysis has been widely used in the faults detection of rotating machinery and can be broadly classified as being a stationary or non-stationary signal. In the case of the faulty rolling element bearing the vibration signal is not strictly phase locked to the rotational speed of the shaft and become “transient” in nature. The purpose of this paper is to briefly discuss the identification of an Inner Raceway Fault (IRF) and an Outer Raceway Fault (ORF) with the different fault severity levels. The conventional statistical analysis was only able to detect the existence of a fault but unable to discriminate between IRF and ORF. In the present work, a detection technique named as bearing damage index (BDI) has been proposed. The proposed BDI technique uses wavelet packet node energy coefficient analysis method. The well-known combination of Hilbert transform (HT) and Fast Fourier Transform (FFT) has been carried out in order to identify the IRF and ORF faults. The results show that wavelet packet node energy coefficients are not only sensitive to detect the faults in bearing but at the same time they are able to detect the severity level of the fault. The proposed bearing damage index method for fault identification may be considered as an ‘index’ representing the health condition of rotating machines.

Bearing Fault Diagnosis Based on Improved Locality-Constrained Linear Coding and Adaptive PSO-Optimized SVM

A novel bearing fault diagnosis method based on improved locality-constrained linear coding (LLC) and adaptive PSO-optimized support vector machine (SVM) is proposed. In traditional LLC, each feature is encoded by using a fixed number of bases without considering the distribution of the features and the weight of the bases. To address these problems, an improved LLC algorithm based on adaptive and weighted bases is proposed. Firstly, preliminary features are obtained by wavelet packet node energy. Then, dictionary learning with class-wise K-SVD algorithm is implemented. Subsequently, based on the learned dictionary the LLC codes can be solved using the improved LLC algorithm. Finally, SVM optimized by adaptive particle swarm optimization (PSO) is utilized to classify the discriminative LLC codes and thus bearing fault diagnosis is realized. In the dictionary leaning stage, other methods such as selecting the samples themselves as dictionary and K-means are also conducted for comparison. The experiment results show that the LLC codes can effectively extract the bearing fault characteristics and the improved LLC outperforms traditional LLC. The dictionary learned by class-wise K-SVD achieves the best performance. Additionally, adaptive PSO-optimized SVM can greatly enhance the classification accuracy comparing with SVM using default parameters and linear SVM.

ANN based Performance Evaluation of BDI for Condition Monitoring of Induction Motor Bearings

One of the critical parts in rotating machines is bearings and most of the failure arises from the defective bearings. Bearing failure leads to failure of a machine and the unpredicted productivity loss in the performance. Therefore, bearing fault detection and prognosis is an integral part of the preventive maintenance procedures. In this paper vibration signal for four conditions of a deep groove ball bearing; normal (N), inner race defect (IRD), ball defect (BD) and outer race defect (ORD) were acquired from a customized bearing test rig, under four different conditions and three different fault sizes. Two approaches have been opted for statistical feature extraction from the vibration signal. In the first approach, raw signal is used for statistical feature extraction and in the second approach statistical features extracted are based on bearing damage index (BDI). The proposed BDI technique uses wavelet packet node energy coefficients analysis method. Both the features are used as inputs to an ANN classifier to evaluate its performance. A comparison of ANN performance is made based on raw vibration data and data chosen by using BDI. The ANN performance has been found to be fairly higher when BDI based signals were used as inputs to the classifier.

Wavelet Packet Decomposition-Based Fault Diagnosis Scheme for SRM Drives With a Single Current Sensor

Power converters are a key, but vulnerable component in switched reluctance motor (SRM) drives. In this paper, a new fault diagnosis scheme for SRM converters is proposed based on the wavelet packet decomposition (WPD) with a dc-link current sensor. Open- and short-circuit faults of the power switches in an asymmetrical half-bridge converter are analyzed in details. In order to obtain the fault signature from the phase currents, two pulse-width modulation signals with phase shift are injected into the lower-switches of the converter to extract the excitation current, and the WPD algorithm is then applied to the detected currents for fault diagnosis. Moreover, a discrete degree of the wavelet packet node energy is chosen as the fault coefficient. The converter faults can be diagnosed and located directly by determining the changes in the discrete degree from the detected currents. The proposed scheme requires only one current sensor in the dc link, while conventional methods need one sensor for each phase or additional detection circuits. The experimental results on a 750-W three-phase SRM are presented to confirm the effectiveness of the proposed fault diagnosis scheme.

Fault diagnosis of rolling element bearing based on S transform and gray level co-occurrence matrix

Time-frequency analysis is an effective tool to extract machinery health information contained in non-stationary vibration signals. Various time-frequency analysis methods have been proposed and successfully applied to machinery fault diagnosis. However, little research has been done on bearing fault diagnosis using texture features extracted from time-frequency representations (TFRs), although they may contain plenty of sensitive information highly related to fault pattern. Therefore, to make full use of the textural information contained in the TFRs, this paper proposes a novel fault diagnosis method based on S transform, gray level co-occurrence matrix (GLCM) and multi-class support vector machine (Multi-SVM). Firstly, S transform is chosen to generate the TFRs due to its advantages of providing frequency-dependent resolution while keeping a direct relationship with the Fourier spectrum. Secondly, the famous GLCM-based texture features are extracted for capturing fault pattern information. Finally, as a classifier which has good discrimination and generalization abilities, Multi-SVM is used for the classification. Experimental results indicate that the GLCM-based texture features extracted from TFRs can identify bearing fault patterns accurately, and provide higher accuracies than the traditional time-domain and frequency-domain features, wavelet packet node energy or two-direction 2D linear discriminant analysis based features of the same TFRs in most cases.

Reliable fault diagnosis for incipient low-speed bearings using fault feature analysis based on a binary bat algorithm

In this paper, we propose a highly reliable fault diagnosis scheme for incipient low-speed rolling element bearing failures. The scheme consists of fault feature calculation, discriminative fault feature analysis, and fault classification. The proposed approach first computes wavelet-based fault features, including the respective relative wavelet packet node energy and entropy, by applying a wavelet packet transform to an incoming acoustic emission signal. The most discriminative fault features are then filtered from the originally produced feature vector by using discriminative fault feature analysis based on a binary bat algorithm (BBA). Finally, the proposed approach employs one-against-all multiclass support vector machines to identify multiple low-speed rolling element bearing defects. This study compares the proposed BBA-based dimensionality reduction scheme with four other dimensionality reduction methodologies in terms of classification performance. Experimental results show that the proposed methodology is superior to other dimensionality reduction approaches, yielding an average classification accuracy of 94.9%, 95.8%, and 98.4% under bearing rotational speeds at 20revolutions-per-minute (RPM), 80RPM, and 140RPM, respectively.

Seismic Target Classification Using a Wavelet Packet Manifold in Unattended Ground Sensors Systems

One of the most challenging problems in target classification is the extraction of a robust feature, which can effectively represent a specific type of targets. The use of seismic signals in unattended ground sensor (UGS) systems makes this problem more complicated, because the seismic target signal is non-stationary, geology-dependent and with high-dimensional feature space. This paper proposes a new feature extraction algorithm, called wavelet packet manifold (WPM), by addressing the neighborhood preserving embedding (NPE) algorithm of manifold learning on the wavelet packet node energy (WPNE) of seismic signals. By combining non-stationary information and low-dimensional manifold information, WPM provides a more robust representation for seismic target classification. By using a K nearest neighbors classifier on the WPM signature, the algorithm of wavelet packet manifold classification (WPMC) is proposed. Experimental results show that the proposed WPMC can not only reduce feature dimensionality, but also improve the classification accuracy up to 95.03%. Moreover, compared with state-of-the-art methods, WPMC is more suitable for UGS in terms of recognition ratio and computational complexity.

Damage Localization of Offshore Platform Based on the Virtual Impulse Response Function

Because the ambient excitation is difficult to test, it is necessary to study the damage detection method only with structure responses. In this paper, two node structure responses under white noise are used to calculate the virtual impulse response function, the amplitude of the virtual impulse response function is decomposed by wavelet packet to calculate the node energy. The wavelet packet node energy change pre and post damage is used as the damage characteristic vector, and the pattern classification function of BP neural network is employed to determine the structure damage location. The numerical simulation and model experiment results of the offshore platform show the effectives of the method, whereas which is easy to be influenced by the noises.