Neural Network Fault Diagnosis Research Articles

Deep balanced domain adaptation neural networks for fault diagnosis of planetary gearboxes with limited labeled data

Due to the time-varying working conditions of planetary gearboxes, the labeled data available are usually limited, current intelligent fault diagnosis methods cannot achieve satisfactory results in the case of limited labeled data. In this paper, a novel method called deep balanced domain adaptation neural network (DBDANN) is proposed for fault diagnosis of planetary gearboxes. First, the multiple convolutional layers were used to extract transferable features layer-by-layer from the raw vibration data of source and the target domain. Then, multi-layer balanced domain adaptation is applied to help training the model, which can reduce the discrepancy of the marginal probability distribution and the conditional probability distribution simultaneously. At last, the test data was fed into the model to evaluate the performance. Experiments with multiple cross-domain tasks under varying speeds and loads demonstrate the effectiveness of the DBDANN, and the proposed model obtained a better performance when compared with the state-of-the-art methods.

Application of neural network algorithm in fault diagnosis of mechanical intelligence

In recent years, mechanical fault diagnosis technology at home and abroad has developed rapidly, and its application has spread to various industrial fields. Due to the complex structure of rotating machinery, the ambiguity and complexity of fault characteristics and causes are common, and it is difficult to carry out fault diagnosis. Although many researches have been carried out and some research results have been obtained, the overall diagnostic level is not very high. High, which is extremely inconsistent with the status quo that is widely used in production. Therefore, it is of great significance to carry out fault diagnosis research on rotating machinery. In this paper, this paper briefly introduces the research and application of intelligent technology in equipment fault diagnosis, and gives the superiority of fuzzy neural network technology application in equipment fault diagnosis, and expounds the basics of fuzzy theory and neural network technology. Based on the principle, the advantages and disadvantages of the two in fault diagnosis are analyzed, and the necessity of combining the two is explained. Based on the previous research on the combination of fuzzy theory and neural network, a new combination method is proposed, and a fuzzy neural network model suitable for fault diagnosis is established. A fuzzy inference method based on fuzzy network is constructed, which realizes the knowledge of information through the extraction, optimization and screening of fuzzy rules. At the same time, the fuzzy neural network learning weights are transformed into case-based reasoning-based diagnostic guidance operators, which play an important role in the rapid extraction of knowledge and improve the diagnostic accuracy. The experimental results show that compared with the commonly used neural network and fuzzy theory fault diagnosis methods, this method can make up for the shortcomings of fuzzy theory and neural network alone. It has higher diagnostic accuracy and has a good application prospect in the field of rotating machinery fault diagnosis.

A simple data augmentation algorithm and a self-adaptive convolutional architecture for few-shot fault diagnosis under different working conditions

In the era of big data, various data-driven fault diagnosis algorithms, which are mainly based on traditional machine learning and deep learning, have been developed and successfully applied on several benchmark datasets. However, in the real world, there are two major obstacles that prevent existing data-driven algorithms from being applied in actual industrial diagnostics applications: a) few-shot learning with limited labelled data, and b) high requirement for model’s generalization ability to adapt different diagnosis circumstances. Two classic feature engineering methods of Order Tracking and Fast Fourier Transform give us inspirations to solve these problems. In this paper, we propose a data augmentation algorithm based on the core assumption of Order Tracking and present a self-adaptive convolutional neural network for fault diagnosis. The data augmentation algorithm utilizes resampling technique to simulate data under different rotating speeds and working loads, in which the Fast Fourier Transform is embedded alternately to calculate the frequency spectra of the expanded dataset. Based on the robust features in the spectra, the self-adaptive convolutional architecture is designed with much fewer Floating Points Operations (FLOPs) and trainable parameters than the deep counterparts, by which the extracted features are invariant for generalization and discriminative for classification. Experiments based on two bearing databases have been carried out and the results have verified the generalization ability and adaptability for few-shot learning of our proposed methods.

Weak Fault Diagnosis of Rolling Bearing Based on Improved Stochastic Resonance

Stochastic resonance can use noise to enhance weak signals, effectively reducing the effect of noise signals on feature extraction. In order to improve the early fault recognition rate of rolling bearings, and to overcome the shortcomings of lack of interaction in the selection of SR (Stochastic Resonance) method parameters and the lack of validation of the extracted features, an adaptive genetic random resonance early fault diagnosis method for rolling bearings was proposed. compared with the existing methods, the AGSR (Adaptive Genetic Stochastic Resonance) method uses genetic algorithms to optimize the system parameters, and further optimizes the parameters while considering the interaction between the parameters. This method can effectively extract the weak fault features of the bearing. In order to verify the effect of feature extraction, the feature signal extracted by AGSR method was input into the Fully connected neural network for fault diagnosis. the practicality of the algorithm is verified by simulation data and rolling bearing experimental data. the results show that the proposed method can effectively detect the early weak features of rolling bearings, and the fault diagnosis effect is better than the existing methods.

Comparisons of different deep learning-based methods on fault diagnosis for geared system

The running state of a geared transmission system affects the stability and reliability of the whole mechanical system. It will greatly reduce the maintenance cost of a mechanical system to identify the faulty state of the geared transmission system. Based on the measured gear fault vibration signals and the deep learning theory, four fault diagnosis neural network models including fast Fourier transform–deep belief network model, wavelet transform–convolutional neural network model, Hilbert-Huang transform–convolutional neural network model, and comprehensive deep neural network model are developed and trained respectively. The results show that the gear fault diagnosis method based on deep learning theory can effectively identify various gear faults under real test conditions. The comprehensive deep neural network model is the most effective one in gear fault recognition.

Fault Diagnosis Method of Photovoltaic Array Based on BP Neural Network

Photovoltaic arrays are prone to various failures due to long-term work. In order to quickly and accurately diagnose the type of failure of the PV array and implement online monitoring of the PV array, this paper proposes the BP neural network for PV array fault diagnosis, and proposes a network search method when training BP neural network. And the K-cross-validation method is used to select the number of hidden layer nodes. The BP neural network fault diagnosis model designed and trained by this method is proved to have high precision.

Application of simulated annealing genetic algorithm-optimized back propagation (BP) neural network in fault diagnosis

Optimal weights are usually obtained in neural network through a fixed network conformation, which affects the practicality of the network. Aiming at the shortage of conformation design and weight training algorithm in neural network application, the back propagation (BP) neural network learning algorithm combined with simulated annealing genetic algorithm (SAGA) is put forward. The multi-point genetic optimization of neural network topology and network weights is performed using hierarchical coding schemes and genetic operations. The simulated annealing mechanism is incorporated into the Genetic Algorithm (GA) to optimize the design and optimization of neural network conformation and network weights simultaneously. The SAGA takes advantage of GA excellent ability in grasping the overall ability of the search process, also uses the SA algorithm to control the convergence of the algorithm to avoid premature phenomenon. The fault diagnosis of one certain on-board electrical control box of helicopter and one certain flight control box of aircraft autopilot were used as a test platform to simulate the algorithm. The simulation conclusions reveal that the algorithm has good convergence rate and high diagnostic accurateness.

Application of Multi-Dimension Input Convolutional Neural Network in Fault Diagnosis of Rolling Bearings

Aiming at the problem of poor robustness of the intelligent diagnostic model, a fault diagnosis model of rolling bearing based on a multi-dimension input convolutional neural network (MDI-CNN) is proposed. Compared with the traditional convolution neural network, the proposed model has multiple input layers. Therefore, it can fuse the original signal and processed signal—making full use of advantages of the convolutional neural networks to learn the original signal characteristics automatically, and also improving recognition accuracy and anti-jamming ability. The feasibility and validity of the proposed MDI-CNN are verified, and its advantages are proved by comparison with the other related models. Moreover, the robustness of the model is tested by adding the noise to the test set. Finally, the stability of the model is verified by two experiments. The experimental results show that the proposed model improves the recognition rate, robustness and convergence performance of the traditional convolution model and has good generalization ability.

Research on aero-engine fault diagnosis based on integrated neural network

In this paper, the fault diagnosis method of Integrated Neural Network based on oil and Vibration information fusion is studied and applied to Aero-engine Fault Diagnosis. Then, taking an CFM56-3 Aero-engine as an example, the application of Integrated Neural Network Fault Diagnosis method in bearing fault diagnosis of Aero-engine is studied by using the idea of Vibration information and oil information Fusion Diagnosis, and the diagnosis method is validated with specific data. The diagnosis results show that compared with the traditional single information source Fault Diagnosis method, the integrated neural network Fault Diagnosis method is more efficient, can detect more fault modes, and has lower misdiagnosis rate.

Physics-Based Convolutional Neural Network for Fault Diagnosis of Rolling Element Bearings

During the past few years, deep learning has been recognized as a useful tool in condition monitoring and fault detection of rolling element bearings. Although existing deep learning approaches are able to intelligently detect and classify the faults in bearings, they still face one or both of the following challenges: 1) most of these approaches rely exclusively on data and do not incorporate physical knowledge into the learning and prediction processes and 2) the approaches often focus on the fault diagnosis of a single bearing in a rotating machine, while in reality, a rotating machine may contain multiple bearings. To address these challenges, this paper proposes a novel approach, namely physics-based convolutional neural network (PCNN), for fault diagnosis of rolling element bearings. In PCNN, an exclusively data-driven deep learning approach, called CNN, is carefully modified to incorporate useful information from physical knowledge about bearings and their fault characteristics. To this end, the proposed approach 1) utilizes spectral kurtosis and envelope analysis to extract sidebands from raw sensor signals and minimize non-transient components of the signals and 2) feeds the information about the fault characteristics into the CNN model. With the capability to process signals from multiple sensors, the proposed PCNN approach is capable of concurrently monitoring multiple bearings and detecting faults in these bearings. The performance of PCNN in machinery fault diagnosis is compared with that of traditional machine learning- and deep learning-based approaches reported in the literature.

A minimum entropy deconvolution-enhanced convolutional neural networks for fault diagnosis of axial piston pumps

Condition monitoring of piston pumps has great significance to ensure the reliability and security of hydraulic systems. However, the complex working conditions of the integrated electromechanical systems make the fault mechanism unclear which is difficult for fault diagnosis by feature matching techniques. In this paper, a novel minimum entropy deconvolution (MED)-based convolutional neural network (CNN) is presented to classify faults in axial piston pumps. Firstly, the collected raw signals are preprocessed using the MED technique. Then, the filtered signals are used to construct training samples and testing samples. Finally, the constructed samples are fed into the CNN to classify the multi-faults of axial piston pumps. With the convolution and subsampling operations, the present model can automatically obtain data features via iterative learning processes, which is suitable for the unknown fault mechanism problems. The learned features are visualized by t-distributed stochastic neighbor embedding technique. A benchmark data simulation of mechanical transmission systems and an experimental data investigation of an axial piston pump are performed to manifest the superiority of the present method by comparing with the traditional CNN.

A Dynamic Adam Based Deep Neural Network for Fault Diagnosis of Oil-Immersed Power Transformers

This paper presents a Dynamic Adam and dropout based deep neural network (DADDNN) for fault diagnosis of oil-immersed power transformers. To solve the problem of incomplete extraction of hidden information with data driven, the gradient first-order moment estimate and second-order moment estimate are used to calculate the different learning rates for all parameters with stable gradient scaling. Meanwhile, the learning rate is dynamically attenuated according to the optimal interval. To prevent over-fitted, we exploit dropout technique to randomly reset some neurons and strengthen the information exchange between indirectly-linked neurons. Our proposed approach was utilized on four datasets to learn the faults diagnosis of oil-immersed power transformers. Besides, four benchmark cases in other fields were also utilized to illustrate its scalability. The simulation results show that the average diagnosis accuracies on the four datasets of our proposed method were 37.9%, 25.5%, 14.6%, 18.9%, and 11.2%, higher than international electro technical commission (IEC), Duval Triangle, stacked autoencoders (SAE), deep belief networks (DBN), and grid search support vector machines (GSSVM), respectively.

A transfer convolutional neural network for fault diagnosis based on ResNet-50

With the rapid development of smart manufacturing, data-driven fault diagnosis has attracted increasing attentions. As one of the most popular methods applied in fault diagnosis, deep learning (DL) has achieved remarkable results. However, due to the fact that the volume of labeled samples is small in fault diagnosis, the depths of DL models for fault diagnosis are shallow compared with convolutional neural network in other areas (including ImageNet), which limits their final prediction accuracies. In this research, a new TCNN(ResNet-50) with the depth of 51 convolutional layers is proposed for fault diagnosis. By combining with transfer learning, TCNN(ResNet-50) applies ResNet-50 trained on ImageNet as feature extractor for fault diagnosis. Firstly, a signal-to-image method is developed to convert time-domain fault signals to RGB images format as the input datatype of ResNet-50. Then, a new structure of TCNN(ResNet-50) is proposed. Finally, the proposed TCNN(ResNet-50) has been tested on three datasets, including bearing damage dataset provided by KAT datacenter, motor bearing dataset provided by Case Western Reserve University (CWRU) and self-priming centrifugal pump dataset. It achieved state-of-the-art results. The prediction accuracies of TCNN(ResNet-50) are as high as 98.95% ± 0.0074, 99.99% ± 0 and 99.20% ± 0, which demonstrates that TCNN(ResNet-50) outperforms other DL models and traditional methods.

Fault Diagnosis of SOFC Stack Based on Neural Network Algorithm

To realize the commercialization of SOFC, it must be ensured that it can work efficiently and stably. SOFC fault diagnosis becomes an essential part of the research. Due to the strong coupling of faults in the stack, this paper uses neural network algorithm to detect and diagnose faults. The simulation results verify that through the diagnosis of the test sample, the recognition rate of the test sample by the network is found to be 95%, which explains the neural network fault diagnosis model established in this paper on identifying the normal working state of the stack, the electrode stacking of the stack, and the gas leakage fault of the stack has good effectiveness and accuracy.

Fault diagnosis and operation and maintenance of PV components based on BP neural network with data cloud acquisition

In order to improve the power generation efficiency of photovoltaic systems and to reduce the cost of manual maintenance, this paper proposes a fault diagnosis and operation and maintenance method based on BP neural network with data cloud acquisition. The causes of short circuit and abnormal aging failure of photovoltaic modules are analyzed. BP neural network fault diagnosis model is established in Matlab, combined with the mathematical model of photovoltaic components to simulate the output of various faults, performing sample training on the fault diagnosis model. The data collected by the distributed photovoltaic information processing and data analysis system platform is used as the input data of the neural network. According to the diagnosis result, reasonable suggestions for PV module operation and maintenance are provided from the aspects of component maintenance, photovoltaic panel cleaning, dip angle change, etc. Simulation results verify the correctness and effectiveness of the proposed method.

An application of ensemble empirical mode decomposition and correlation dimension for the HV circuit breaker diagnosis

ABSTRACTDuring the operation process of the high-voltage circuit breaker, the changes of vibration signals reflect the machinery states of the circuit breaker. The extraction of the vibration signal feature will directly influence the accuracy and practicability of fault diagnosis. This paper presents an extraction method based on ensemble empirical mode decomposition) and correlation dimension and a classification method with BP (back propagation) neural network. Firstly, original vibration signals are decomposed into a finite number of stationary intrinsic mode functions (IMFs). Secondly, correlation dimension of the top four IMFs by the G–P algorithm is calculated and the characteristic vector of the vibration signal of the circuit breaker is formed. At last, the classification of characteristic parameter is realized with a simple BP neural network for fault diagnosis. The experimentation without loads indicates that the method can easily and accurately diagnose breaker faults and exploit a new road for diagnosis of high-voltage circuit breakers.

RideNN: A New Rider Optimization Algorithm-Based Neural Network for Fault Diagnosis in Analog Circuits

Fault diagnosis in electronic circuits is an emerging area of research, where fully automated diagnosis systems are being developed for the investigation of the circuits. Developing test methods for the diagnosis of faults in analog circuits is still a complex task. Consequently, a technique for the fault diagnosis in analog circuits is designed by proposing a new optimization algorithm, named, rider optimization algorithm (ROA). The development of ROA is based on a group of riders, racing toward a target location. Moreover, a classifier, termed RideNN, is developed by including the proposed algorithm as the training algorithm for the neural network (NN). RideNN, along with the orthogonal transformation and Bhattacharyya coefficient, is applied for the fault diagnosis of analog circuits. The proposed technique is experimented using three basic circuits, such as triangular wave generator (TWG), low noise bipolar transistor amplifier (BTA), and differentiator (DIF) and an application circuit, solar power converter (SPC). The performance is evaluated using two evaluation metrics, namely, accuracy (ACC) and false alarm ratio (FAR). The analysis results show that the proposed technique attains an ACC of 99.9% in TWG, 99.9% in BTA, 99% in DIF, and 95% in SPC without noise.

A New Snapshot Ensemble Convolutional Neural Network for Fault Diagnosis

Deep learning (DL) has been widely applied in fault diagnosis field. The ensemble of DL can significantly improve the generalization ability of DL. Snapshot ensemble learning, which uses the cyclic learning rate scheduler (CLR) and then combines the local minima in once training to form an ensemble method, has shown powerful performance. However, the learning rate range of CLR should be pre-defined by experiences, which may limit its prediction accuracy. In this paper, a new snapshot ensemble convolutional neural network (SECNN) is proposed, which can find the proper range of learning rate for SECNN automatically when facing a new dataset. First, a max–min cosine cyclic learning rate scheduler (MMCCLR) is designed to avoid learning rate range being affected by other parameters. Then, a new learning rate testing (logLR Test) is applied to estimate the proper learning rate range for MMCCLR. Finally, the SECNN with MMCCLR is developed. The SECNN is tested on three famous datasets, including bearing dataset of Case Western Reserve University, self-priming centrifugal pump dataset, and bearing dataset provided by the Machinery Failure Prevention Technology. The results have been improved a lot by the proposed SECNN methods. The SECNN has also achieved the start-of-the-art prediction accuracies by comparing with other DL and traditional methods.

An SDP Characteristic Information Fusion‐Based CNN Vibration Fault Diagnosis Method

This study proposes a symmetrized dot pattern (SDP) characteristic information fusion‐based convolutional neural network (CNN) fault diagnosis method to resolve issues of high complexity, nonlinearity, and instability in original rotor vibration signals. The method was used to conduct information fusion of real modal components of vibration signals and SDP image identification using CNN in order to achieve vibration fault diagnosis. Compared with other graphic processing methods, the proposed method more fully expressed the characteristics of different vibration signals and thus presented variations between different vibration states in a simpler and more intuitive way. The proposed method was experimentally investigated using simulation signals and rotor test‐rig signals, and its validity and advancements were demonstrated using experimental analysis. By using CNN through deep learning to adaptively extract SDP characteristic information, vibration fault identification was ultimately realized.

Fault Diagnosis of Reciprocating Compressor Based on Convolutional Neural Networks with Multisource Raw Vibration Signals

Reciprocating compressors are widely used in petroleum industry. Due to containing complex nonlinear signal, it is difficult to extract the fault features from its vibration signals. This paper proposes a new method named Convolutional Neural Network based on Multisource Raw vibration signals (MSRCNN). The proposed method uses multisource raw vibration signals collected by several sensors as input and uses the designed CNN to operate both the feature extraction and classification. The gas valve signals of reciprocating compressor in different states are used as the experimental data. In order to test the effectiveness of the proposed method, it is compared with the traditional BP (Back‐Propagation) neural network fault diagnosis method based on power spectrum energy and wavelet packet energy. In order to further test the antinoise performance of the proposed method, some noisy signals with different signal‐to‐noise ratios were constructed by adding white noise into sampled signals for testing. The results show that the MSRCNN model has higher fault recognition rate than the traditional methods. This indicates that the MSRCNN method not only has good fault recognition effect, but also has certain antinoise performance.