Neural Network Fault Detection Research Articles

Convolutional Neural Networks for Fault Detection in Grid-Connected Photovoltaic Panels

Convolutional Neural Networks for Fault Detection in Grid-Connected Photovoltaic Panels

Protection Scheme based on Artificial Neural Network for Fault Detection and Classification in Low Voltage PV-Based DC Microgrid

With the expansion of the DC distribution market, protection, and operational concerns for Direct Current (DC) Microgrids have increased. Different systems have been investigated for detecting, finding, and isolating defects utilising a variety of protective mechanisms. It might be difficult to locate high-resistance faults and shorted DC faults on low-voltage DC (LVDC) microgrids. Therefore, in this study, a Field Transform Technique like Short-Time Fourier Transform (STFT) is proposed for detecting the Fault Current (FC). This method detects the faults Pole-ground (PG), pole-pole (PP), and Arc fault are the major fault types in the DC network with PG fault as the most common and less severe. One of the difficulties the DC system faces in the incidence of a malfunction is the protection of essential converters. During this fault, the diodes, being the most vulnerable component of the system, may encounter a substantial surge in current, which can potentially cause damage if the current surpasses double their specified capacity for withstanding. After the Fault detection (FD), a Taguchi-based ANN is presented to classify the detected faults. This method effectively classifies PV-based faults. Then, to safeguard the FC, the Improved Self-Adaptive Solid State Circuit Breaker (I-SSCB) is introduced. It safeguards the FC in the low-voltage PV-based DC microgrid (DCMG) and restricts FC in the DCMG. The suggested approach is evaluated using the Matlab software and the proposed method produces 400A current and 100 KW power during the PV temperature of 25°C. The output current of the ANN is then 1A for a duration of 0.3 to 0.4 seconds. The fault voltage and FC produced in this proposed work are 1900V and 1950A. Therefore, the proposed work's current and voltage values are 21 KV and 0.35 I. Therefore, the proposed method produces more power and limits the FC in the LV-DCMG. In future studies, the improved or modified neural network or machine learning (ML)-based techniques can be utilized which may improve the protection scheme of the work.

Application of non-Gaussian feature enhancement extraction in gated recurrent neural network for fault detection in batch production processes

The nonlinear, time correlation, and non-Gaussian features in data present significant challenges for effective fault detection. While the Gate Recurrent Unit (GRU) network is renowned for its capacity to manage time correlation, it falls short in capturing non-Gaussian features in process data, which can likely lead to suboptimal monitoring results. To address this limitation, the Enhancement Gate Recurrent Unit (ENGRU) is developed to perfect the fault detection accuracy of the network. Specifically, The ENGRU effectively extracts high order statistics information by employing the overcompleted independent component analysis method, thereby augmenting its ability to capture non-Gaussian properties. The extracted features information are then entered into the ENGRU model further to uncover additional hidden features beyond what the GRU can achieve. The ENGRU network, which is built upon the extracted characteristic information, even farther enhances the accuracy of the fault detection. The merits of the proposed model are demonstrated by comparing it with excellent fault detection algorithms on a benchmark platform.

An evolution strategy of GAN for the generation of high impedance fault samples based on Reptile algorithm

In a distribution system, sparse reliable samples and inconsistent fault characteristics always appear in the dataset of neural network fault detection models because of high impedance fault (HIF) and system structural changes. In this paper, we present an algorithm called Generative Adversarial Networks (GAN) based on the Reptile Algorithm (GANRA) for generating fault data and propose an evolution strategy based on GANRA to assist the fault detection of neural networks. First, the GANRA generates enough high-quality analogous fault data to solve a shortage of realistic fault data for the fault detection model’s training. Second, an evolution strategy is proposed to help the GANRA improve the fault detection neural network’s accuracy and generalization by searching for GAN’s initial parameters. Finally, Convolutional Neural Network (CNN) is considered as the identification fault model in simulation experiments to verify the validity of the evolution strategy and the GANRA under the HIF environment. The results show that the GANRA can optimize the initial parameters of GAN and effectively reduce the calculation time, the sample size, and the number of learning iterations needed for dataset generation in the new grid structures.

Homogeneous ensemble model built from artificial neural networks for fault detection in navigation systems

In this paper, authors present a modern approach to the detection of malfunctioning sensory systems. The proposed solution is based on artificial neural networks. The application example uses a navigation system based on a 9-axis IMU (Inertial Measurement Unit), the signal fusion data is converted into quaternions. The form of quaternions is then analyzed along with sensor samples by an artificial neural network. If the network detects data processing inadequate to the pattern then we obtain information about the malfunction of a specific sensing axis from the sensors.The results compare fault detection capabilities using an ensemble structure built from three types of artificial neural networks: fully-connected, recurrent and convolutional. We provide a comprehensive analysis of all models; the proposed measures include RMSE (Root-Mean-Square Error), NRMSE (Normalized RMSE), t-SNE (t-distributed stochastic neighbor embedding) visualization, ROC (Receiver Operating Characteristic) curve, precision vs. recall curve, AUC (Area Under Curve) and F-score.

APGNN : Alarm Propagation Graph Neural Network for fault detection and alarm root cause analysis

Telecommunication network plays an important role in our daily life. Fault detection and alarm root cause analysis are the keys to ensure the normal operation of the network. To reduce the burden on operators, numerous methods are employed to analyse root cause of faults. However, there still remain a large amount of non-essential or transient alarms after root cause analysis. A simple Rule-based method may help ease the problems. But it needs prior expert knowledge and the diversity of alarm pattern makes the rules redundant and complicated. Moreover, it cannot accurately cover all true faults and need manual methods as complement. In this work, we propose Alarm Propagation Graph Neural Network(APGNN), a novel data-driven propagation-based root cause analysis and fault detection approach.It first associates alarms and extracts root-derived graph based on Bayesian Network. Then it constructs alarm propagation graphs(APG). We refine the repair orders to obtain actual fault information. At last, Graph Neural Network is used to extract features and learn the mapping from APG to the true fault. Our method not only detects the true fault from large volume of original alarms, but also analyses the root cause alarms. We evaluate our approach both on the offline and online environment of the real-world IP Radio Access Network. Experiments show that our model outperforms the state-of-art approach by 4.6% in F1-score on average.

Application of Convolutional and Feedforward Neural Networks for Fault Detection in Particle Accelerator Power Systems

High voltage converter modulators (HVCM) provide power to the accelerating cavities of the Spallation Neutron Source (SNS) facility. HVCMs experience catastrophic failures, which increase the downtime of the SNS and reduce beam time. The faults may occur due to different reasons including failures of the resonant capacitor, core saturation due to the magnetic flux, insulated-gate bipolar transistor (IGBT) failures, and others. We recently have setup a HVCM test stand to develop and test machine learning models for anomaly detection and fault prognostics. In this work, we propose binary classifiers and autoencoder architectures based on convolutional (CNN) and feedforward neural networks (FNN) to facilitate distinguishing normal from faulty waveforms coming from the HVCM during operation. The results indicate that the CNN binary classifier is the best model among the four showing very stable performance in the training and testing sets with impressive precision and recall metrics, reaching up to 99% with a very small uncertainty. The FNN classifier shows the least performance with a large uncertainty in its metrics. The performances of the two autoencoders based on CNN and FNN were in between, showing very good performance nonetheless.

Dynamic-scale graph neural network for fault detection

Traditional graph-based dynamic fault detection methods describe the dynamic characteristic through constructing a single neighborhood graph at the current sample with some history samples. However, they ignore the diversity of dynamic properties of the variables in complex chemical processes. To overcome this problem, a novel neural network structure combining multiscale subgraphs is proposed, named dynamic-scale graph neural network (DSGNN), which divides variables into multiple groups according to their dynamic properties. DSGNN constructs a subgraph in each group. In traditional graph-based methods, the scale of the graph is usually manually designed. In DSGNN, the scale of each subgraph is decided by the dynamic properties of the variables in this subgraph. To aggregate the dynamic information, DSGNN utilizes convolution operations. The weights assigned to the neighbors in each subgraph are determined according to the similarity between the current data and its neighbors. Low-dimensional features are extracted through the back-propagation technique from the updated high-dimensional features produced by convolution operations. Two case studies on a multivariate dynamic process and the Tennessee Eastman process are conducted to show the superiority of DSGNN.

Resource Orchestration of Cloud-Edge–based Smart Grid Fault Detection

Real-time smart grid monitoring is critical to enhancing resiliency and operational efficiency of power equipment. Cloud-based and edge-based fault detection systems integrating deep learning have been proposed recently to monitor the grid in real time. However, state-of-the-art cloud-based detection may require uploading a large amount of data and suffer from long network delay, while edge-based schemes do not adequately consider the detection requirement and thus cannot provide flexible and optimal performance. To solve these problems, we study a cloud-edge based hybrid smart grid fault detection system. Embedded devices are placed at the edge of the monitored equipment with several lightweight neural networks for fault detection. Considering limited communication resources, relatively low computation capabilities of edge devices, and different monitoring accuracies supported by these neural networks, we design an optimal communication and computational resource allocation method for this cloud-edge based smart grid fault detection system. Our method can maximize the processing throughput of the system and improve resource utilization while satisfying the data transmission and processing latency requirements. Extensive simulations are conducted and the results show the superiority of the proposed scheme over comparison schemes. We have also prototyped this system and verified its feasibility and performance in real-world scenarios.

Genetic-Algorithm-Based Neural Network for Fault Detection and Diagnosis: Application to Grid-Connected Photovoltaic Systems

Modern photovoltaic (PV) systems have received significant attention regarding fault detection and diagnosis (FDD) for enhancing their operation by boosting their dependability, availability, and necessary safety. As a result, the problem of FDD in grid-connected PV (GCPV) systems is discussed in this work. Tools for feature extraction and selection and fault classification are applied in the developed FDD approach to monitor the GCPV system under various operating conditions. This is addressed such that the genetic algorithm (GA) technique is used for selecting the best features and the artificial neural network (ANN) classifier is applied for fault diagnosis. Only the most important features are selected to be supplied to the ANN classifier. The classification performance is determined via different metrics for various GA-based ANN classifiers using data extracted from the healthy and faulty data of the GCPV system. A thorough analysis of 16 faults applied on the module is performed. In general terms, the faults observed in the system are classified under three categories: simple, multiple, and mixed. The obtained results confirm the feasibility and effectiveness with a low computation time of the proposed approach for fault diagnosis.

Development of a multi-output feed-forward neural network for fault detection in Photovoltaic Systems

The main role of a fault detection and identification technique is to interpret what is​ causing a PVS’s real-time energy production to drop. To keep photovoltaic (PV) systems performing at their best, fault detection is crucial. Various fault detection systems have been proposed in the literature over the last decade. Some of the fault detection methods presented solely used data from I–V curves, while others relied on ANN techniques to find faults. This paper presents the development of a multi-output Artificial Neural Network (ANN), for fault detection and identification on the DC side of a Photovoltaic System (PVS). Measurements and benchmarking were carried out in order to validate the algorithm’s capacity to detect both normal operation at maximum power and the ability to determine and identify faults introduced during the experiment’s procedure. It is observed that the classification output, compared with the regression output, provides a better assessment of the operational status of the photovoltaic cell. The developed method has a 93.4 percent accuracy in detecting open-circuit, short-circuit, and mismatch faults, using the classification output, while the regression output provides an operational status estimation with mean absolute error (MAE) equal to 0.305. When compared to other implementations, the findings are encouraging, indicating that the approach could be used in PVS.

Performance Comparison of Multiple Neural Networks for Fault Detection of Sensors Array in Oil Heating Reactor

Fault detection is an important issue for early failure revelation and machine components preserving before the damage. The processes of fault detection, diagnosis and correction especially in oil heating reactor sensors are among the most crucial steps for reliable and proper operation inside the reactor.The fault detection in sensors array of heating reactor is considered as an important tool to guarantee that the controller can take the best possible action to insure the quality of the output. In this paper, fault detection for the temperature sensor in oil heating reactor using different types of faults with different levels is addressed. Multiple approaches based on Neural Network (NN)s such as the classical Fully Connected Neural Network (FCNN), Bidirectional Long Short Term Memory network (BiLSTM) based on Recurrent Neural Network (R.N.N.) and Convolutional Neural Network (CNN) are suggested for this purpose. The suggested networks are trained and tested on real dataset sequences taken from sensors array readings of real heating reactor in Egypt. The performance comparison of the suggested networks is evaluated using different metrics such as “confusion matrix”, accuracy, precision, etc. The various NN are simulated, trained and tested in this paper using MATLAB software 2021 and the advanced tool of “DeepNetworkDesigner”. The simulation results prove that CNN outperforms the other comparative networks with classification accuracy reached to 100% with different levels and different types of faults.

Rolling bearing fault diagnosis based on wireless sensor network data fusion

With the continuous development of intelligent manufacturing, mechanical equipment is developing in the direction of large-scale, integration, precision and intelligence. The coupling between different equipment in the operation of the system also makes the mechanical equipment increasingly become a whole, which puts forward higher requirements for the fault diagnosis of mechanical equipment, especially rolling bearings. With the rapid development of wireless network technology, rolling bearing fault diagnosis has become a reality. In this paper, the wireless sensor network will be innovatively used to collect the data of key parts of mechanical equipment, so as to improve the problem of insufficient and accurate collected data in the traditional convolution neural network fault detection algorithm, convert the corresponding time-domain signal into image signal, and combine the global average pool layer on the basis of the combination of nonlinear convolution layers, So as to optimize the corresponding wireless network structure, and finally realize the adaptive extraction and fault diagnosis of rolling bearing fault features. The experimental results show that the fault detection accuracy of the optimized convolution neural network algorithm is about 99.8%, and has good robustness and practical value.

Technology Focus: Data Analytics (October 2021)

With a moderate- to low-oil-price environment being the new normal, improving process efficiency, thereby leading to hydrocarbon recovery at reduced costs, is becoming the need of the hour. The oil and gas industry generates vast amounts of data that, if properly leveraged, can generate insights that lead to recovering hydrocarbons with reduced costs, better safety records, lower costs associated with equipment downtime, and reduced environmental footprint. Data analytics and machine-learning techniques offer tremendous potential in leveraging the data. An analysis of papers in OnePetro from 2014 to 2020 illustrates the steep increase in the number of machine-learning-related papers year after year. The analysis also reveals reservoir characterization, formation evaluation, and drilling as domains that have seen the highest number of papers on the application of machine-learning techniques. Reservoir characterization in particular is a field that has seen an explosion of papers on machine learning, with the use of convolutional neural networks for fault detection, seismic imaging and inversion, and the use of classical machine-learning algorithms such as random forests for lithofacies classification. Formation evaluation is another area that has gained a lot of traction with applications such as the use of classical machine-learning techniques such as support vector regression to predict rock mechanical properties and the use of deep-learning techniques such as long short-term memory to predict synthetic logs in unconventional reservoirs. Drilling is another domain where a tremendous amount of work has been done with papers on optimizing drilling parameters using techniques such as genetic algorithms, using automated machine-learning frameworks for bit dull grade prediction, and application of natural language processing for stuck-pipe prevention and reduction of nonproductive time. As the application of machine learning toward solving various problems in the upstream oil and gas industry proliferates, explainable artificial intelligence or machine-learning interpretability becomes critical for data scientists and business decision-makers alike. Data scientists need the ability to explain machine-learning models to executives and stakeholders to verify hypotheses and build trust in the models. One of the three highlighted papers used Shapley additive explanations, which is a game-theory-based approach to explain machine-learning outputs, to provide a layer of interpretability to their machine-learning model for identification of identification of geomechanical facies along horizontal wells. A cautionary note: While there is significant promise in applying these techniques, there remain many challenges in capitalizing on the data—lack of common data models in the industry, data silos, data stored in on-premises resources, slow migration of data to the cloud, legacy databases and systems, lack of digitization of older/legacy reports, well logs, and lack of standardization in data-collection methodologies across different facilities and geomarkets, to name a few. I would like to invite readers to review the selection of papers to get an idea of various applications in the upstream oil and gas space where machine-learning methods have been leveraged. The highlighted papers cover the topics of fatigue dam-age of marine risers and well performance optimization and identification of frackable, brittle, and producible rock along horizontal wells using drilling data. Recommended additional reading at OnePetro: www.onepetro.org. SPE 201597 - Improved Robustness in Long-Term Pressure-Data Analysis Using Wavelets and Deep Learning by Dante Orta Alemán, Stanford University, et al. SPE 202379 - A Network Data Analytics Approach to Assessing Reservoir Uncertainty and Identification of Characteristic Reservoir Models by Eugene Tan, the University of Western Australia, et al. OTC 30936 - Data-Driven Performance Optimization in Section Milling by Shantanu Neema, Chevron, et al.

Switched time delay control based on neural network for fault detection and compensation in robot

Switched time delay control based on neural network for fault detection and compensation in robot

Switched time delay control based on neural network for fault detection and compensation in robot

Fault detection in robotic manipulators is necessary for their monitoring and represents an effective support to use them as independent systems. This present study investigates an enhanced method for representation of the faultless system behavior in a robot manipulator based on a multi-layer perceptron (MLP) neural network learning model which produces the same behavior as the real dynamic manipulator. The study was based on generation of residue by contrasting the actual output of the manipulator with those of the neural network; Then, a time delay control (TDC) is applied to compensate the fault, in which a typical sliding mode command is used to delete the time delay estimate produced by the belated signal in order to obtain strong performances. The results of the simulations performed on a model of the SCARA arm manipulator, showed a good trajectory tracking and fast convergence speed in the presence of faults on the sensors. In addition, the command is completely model independent, for both TDC and MLP neural network, which represents a major advantage of the proposed command.

Switched time delay control based on artificial neural network for fault detection and compensation in robot\xa0manipulators

This work proposes a switched time delay control scheme based on neural networks for robots subjected to sensors faults. In this scheme, a multilayer perceptron (MLP) artificial neural network (ANN) is introduced to reproduce the same behavior of a robot in the case of no faults. The reproduction characteristic of the MLPs allows instant detection of any important sensor faults. In order to compensate the effects of these faults on the robot’s behavior, a time delay control (TDC) procedure is presented. The proposed controller is composed of two control laws: The first one contains a small gain applied to the faultless robot, while the second scheme uses a high gain that is applied to the robot subjected to faults. The control method applied to the system is decided based on the ANN detection results which switches from the first control law to the second one in the case where an important fault is detected. Simulations are performed on a SCARA arm manipulator to illustrate the feasibility and effectiveness of the proposed controller. The results demonstrate that the free-model aspect of the proposed controller makes it highly suitable for industrial applications.

Statistical characterization of semi-supervised neural networks for fault detection and diagnosis of air handling units

Air handling units have been widely adopted in modern buildings for indoor air regulation and circulation tasks. The accurate and reliable fault detection and diagnosis (FDD) of air handling units (AHU) are of great significance to maintain indoor environment while ensuring the energy efficiency in building operations. Data-driven FDD approaches have gained great popularity due to their excellence and flexibilities for practical applications. Given sufficient labeled data, existing studies have validated the value of various supervised learning algorithms for FDD tasks. However, it can be very challenging, expensive, time-consuming and labor-intensive to obtain data labels for faulty operations, making it impractical to fully realize the potential of advanced supervised learning algorithms. To tackle this problem, this study proposes a novel semi-supervised FDD method using neural networks. The method adopts the self-training strategy for semi-supervised learning and has been tested for two practical applications, i.e., fault diagnosis and unseen fault detection. A number of data experiments have been conducted to statistically characterize the influence of key learning parameters, including the labeled data availabilities, the maximum semi-supervised learning iterations, the threshold and learning rate for pseudo-label data utilization. The results indicate that the method can effectively enhance model generalization performance by utilizing large amounts of unlabeled data. The insights obtained are helpful for developing advanced data-driven tools for smart building system fault detection and diagnosis.

Deadlock Control and Fault Detection and Treatment in Reconfigurable Manufacturing Systems Using Colored Resource-Oriented Petri Nets Based on Neural Network

A reconfigurable manufacturing system (RMS) means that it can be reconfigured and become more complex during its operation. In RMSs, deadlocks may occur because of sharing of reliable or unreliable resources. Various deadlock control techniques are proposed for RMSs with reliable and unreliable resources. However, when the system is large-sized, the complexity of these techniques will increase. To overcome this problem, this paper develops a four-step deadlock control policy for the detection and treatment of faults in an RMS. In the first step, a colored resource-oriented timed Petri net (CROTPN) is designed for rapid and effective reconfiguration of the RMS without considering resource failures. In the second step, “sufficient and necessary conditions” for the liveness of a CROTPN are introduced to guarantee that the model is live. The third step considers the problems of failures of all resources in the CROTPN model and guarantees that the model is reliable by designing a common recovery subnet and adding it to the obtained CROTPN model at the second step. The fourth step designs a new hybrid method that combines the CROTPN with neural networks for fault detection and treatment. A simulation is performed using the GPenSIM tool to evaluate the proposed policy under the RMS configuration changes and the results are compared with the existing approaches in the literature. It is shown that the proposed approach can handle any complex RMS configurations, solve the deadlock problem in an RMS, and detect and treat failures. Furthermore, is simpler in its structure.

Tree Search Fuzzy NARX Neural Network Fault Detection Technique for PV Systems with IoT Support

The photovoltaic (PV) panel’s output energy depends on many factors. As they are becoming the leading alternative energy source, it is essential to get the best out of them. Although the main factor for maximizing energy production is proportional to the amount of solar radiation reaching the photovoltaic panel surface, other factors, such as temperature and shading, influence them negatively. Moreover, being installed in a dynamic and frequently harsh environment causes a set of reasons for faults, defects, and irregular operations. Any irregular operation should be recognized and classified into faults that need attention and, therefore, maintenance or as being a regular operation due to changes in some surrounding factors, such as temperature or solar radiation. Besides, in case of faults, it would be helpful to identify the source and the cause of the problem. Hence, this study presented a novel methodology that modeled a PV system in a tree-like hierarchy, which allowed the use of a fuzzy nonlinear autoregressive network with exogenous inputs (NARX) to detect and classify faults in a PV system with customizable granularity. Moreover, the used methodology enabled the identification of the exact source of fault(s) in a fully automated way. The study was done on a string of eight PV panels; however, the paper discussed using the algorithm on a more extensive PV system. The used fuzzy NARX algorithm in this study was able to classify the faults that appeared in up to five out of the eight PV panels and to identify the faulty PV panels with high accuracy. The used hardware could be controlled and monitored through a Wi-Fi connection, which added support for Internet of Things applications.