Fault Detection And Diagnosis Method Research Articles

Fault detection and diagnosis of electric bus air conditioning systems incorporating domain knowledge and probabilistic artificial intelligence

The air conditioning systems in electric city buses usually operate in rapidly changing ambient conditions and are more likely to suffer from mechanical faults. Although many fault detection and diagnosis (FDD) methods have been developed for building air conditioning systems, they are difficult to be applied to bus air conditioners since its operation is highly dynamic and fault-free data are usually unavailable. Therefore, this paper proposes an FDD method for electric bus air conditioners to tackle the above issues. First, the method identifies faults in an unsupervised manner by comparing selected features among a group of peer systems. Then, considering the features are influenced by the operating conditions, Gaussian process regression (GPR) models are established to find the relationships between each feature and its influential parameters. The probabilistic nature of the GPR is used to differentiate predictions with large uncertainty, which are then excluded from FDD. In this way, robustness of the method is evidently improved. Finally, fault indexes are defined to detect and diagnose mechanical faults. We applied the method to a group of air conditioners in a city bus fleet. Results showed that it can effectively identify refrigerant undercharge and indoor and outdoor fan problems with low false positive/genitive rates. Also, the method is highly robust and not sensitive to the faulty systems in the bus fleet.

Fault detection and diagnosis of energy system based on deep learning image recognition model under the condition of imbalanced samples

Faults in energy systems impact the reliability of the energy supply and cause energy waste. Data-driven fault detection and diagnosis (FDD) methods can detect and diagnose system faults by mining historical operational data. However, the quantitative imbalance between fault and normal samples degrades the performance of the FDD method. Therefore, this study proposes a novel deep learning-based FDD method under the condition of imbalanced samples that converts the time-series signals into an image signal to extract the timing and coupling features, and then applies the improved conditional variational autoencoder-generative adversarial network (CVAE-GAN) to generate fault samples for balancing the training sample set. Subsequently, according to the framework of image recognition, a two-dimensional convolutional neural network was adopted to identify the image samples and achieve FDD. The experimental results showed that, under the condition of imbalanced samples, the proposed method could increase diagnosis accuracy by an average of 5.71% compared with other common data-driven methods. After the samples were augmented with the improved CVAE-GAN, the accuracy improved by an average of 3.79%. Consequently, the feasibility and superiority of the proposed method were demonstrated.

Robust and sparse canonical correlation analysis for fault detection and diagnosis using training data with outliers

A well-known shortcoming of the traditional canonical correlation analysis (CCA) is the lack of robustness against outliers. This shortcoming hinders the application of CCA in the case where the training data contain outliers. To overcome this shortcoming, this paper proposes robust CCA (RCCA) methods for the analysis of multivariate data with outliers. The robustness is achieved by the use of weighted covariance matrices in which the detrimental effect of outliers is reduced by adding small weight coefficients on them. The RCCA is then extended to the robust sparse CCA (RSCCA) by imposing the l1-norm constraints on canonical projection vectors to obtain the sparsity property. Based on the RCCA and RSCCA, a robust data-driven fault detection and diagnosis (FDD) method is proposed for industrial processes. A residual generation model is built using projection vectors of the RCCA or RSCCA. The robust squared Mahalanobis distance of the residual is used for fault detection. A contribution-based fault diagnosis method is developed to identify the faulty variables that may cause the fault. The performance and advantages of the proposed methods are illustrated with two case studies. The results of two case studies prove that the RCCA and RSCCA methods have high robustness against outliers, and the robust FDD method is able to yield reliable results even if using the low-quality training data with outliers.

Development and demonstration of a method to detect refrigerant charge level for variable refrigerant volume systems

Unlike traditional direct expansion systems, variable refrigerant flow (VRF) systems have longer routing, more complicated setup, and sophisticated controls and require much more refrigerant for proper operation. Proper refrigerant charge becomes more critical to their operations. Intensive studies have investigated refrigerant charge faults for VRF systems. Due to the daunting complexity, few studies use knowledge-based fault detection and diagnosis (FDD) methods. Most of them use data-driven FDD methods to study this fault. Although they seem promising, these methods can only handle single faults and are not scalable or interpretable. This paper proposes a decoupling-based FDD method to detect and diagnose improper refrigerant charge in VRF systems. This method was originally proposed by our previous work for rooftop units (RTUs) and split systems, which has been tested and validated by extensive studies. Due to the differences between VRF and RTU, the original method – virtual refrigerant charge sensor (VRC) for VRF is modified, analyzed, and investigated in detail. Data filter criteria were implemented to eliminate noisy data. The modified VRC, when combined with a data filter, was tested on two systems in the field. The results demonstrated its effectiveness in detecting an undercharged system, specifically a 30 % undercharge.

A systematic review of Intelligent Fault-Tolerant Protection Scheme for Multi-terminal HVDC Grids


 
 
 Due to the advancement of power electronics devices and control techniques, the modular multilevel converter (MMC) has become the most attractive converter for multiterminal direct current (MTDC) grids thanks to its most relevant features, such as modularity and scalability. Despite their advantages, conventional MMCs face a major challenge with: i) fault-tolerant operation strategy; i) energy losses in conversion; iii) lack of DC fault handling capability. This paper provides a systematic review to identify the gaps in the literature about Intelligent Fault-Tolerant Protection Schemes for multi-terminal HVDC grids. Through the bibliometric analysis, it was possible to identify topics still to be developed within the four main clusters (Offshore wind farms, Wind turbines, Voltage Source Converters, and Wind power). The research topic opens three research paths: the first is the analysis of failures in HVDC (High Voltage Direct Current) grid equipment by the FDD (Fault Detection and Diagnosis) method; the second is failure analysis by the IFDD (Inverse Fault Detection and Diagnosis) method and the third is the possibility of interconnecting the different energy generation zones with different frequencies.
 
 

A novel building heat pump system semi-supervised fault detection and diagnosis method under small and imbalanced data

Faulty heat pump system operating can cause excessive energy consumption and an uncomfortable indoor environment. Data-driven method is one of the most widely studied Fault Detection and Diagnosis (FDD) methods in building heat pump systems. However, the fault data is difficult to collect and label, which makes it a significant challenge to apply a data-driven method. Therefore, a novel semi-supervised FDD method that could utilize small labelled and unlabelled fault data was proposed, considering unknown fault detection in unlabelled data. The fundamental model consisted of four parts: encoder, double discriminators, generator, and bagging classifiers, and all networks were integrated to train simultaneously. Additionally, batch nuclear-norm maximization and pseudo-label were utilized to improve the model. The proposed method could realize data augmentation, pseudo-label utilization, and FDD in one end-to-end model. Normal and fault data were collected from a variable refrigerant flow heat pump system to verify this method. The proposed method was compared with several methods in previous studies. The performance indexes showed that the proposed method had averaged 65.56%, 75.50%, 82.25%, 83.30%, and 86.26% overall accuracy under each training labelled group, respectively. The unlabelled data could improve the model effectively, and the averaged detection rate of unknown fault could reach 98.00%, 98.82%, 98.88%, 98.66%, 98.94%, and 98.82%, respectively, under different numbers of unlabelled data. Additionally, the analysis of unknown fault diagnosis results indicated that faults with different severities had similar operating characteristics.

A transformer-based approach for novel fault detection and fault classification/diagnosis in manufacturing: A rotary system application

Owing to the rapid development of Industry 4.0, new sensing and communication technologies have made vast amounts of untapped process data available. In order to transform such data assets into strong insights and knowledge that support manufacturing decisions, condition-based maintenance (CBM) and fault detection and diagnosis (FDD) have become effective ways to enhance equipment reliability and reduce costs. A successful data-driven FDD method must not only be capable of identifying the types of known faults, but also in detecting unseen or uncharacterized events during manufacturing system operation. To this end, this paper presents a Transformer-based classifier that can efficiently identify different known types and severity levels of fault conditions, in addition to novel fault detection. In this method, time-frequency spectrograms transformed from raw vibration signals are input to the classifier for known fault classification. Utilizing the advanced feature extracting performance of the classifier, a simple yet effective technique based on Mahalanobis distance is adopted to detect whether the fault comes from a previously unseen fault condition. When a novel condition is detected, the model is subsequently retrained using the novel data in an incremental learning manner. The proposed method is verified by an experimental case study with data collected from a testbed that has many features representative of common manufacturing equipment. The results demonstrated that the proposed method has superior performance in both fault diagnosis and novelty identification when compared with the baseline models and a cutting-edge model.

Fault Detection and Diagnosis for Thrust Drop of Launch Vehicles Against Disturbances

Rapid detection and accurate diagnosis for thrust drop fault are still open problems for launch vehicles. Existing studies have achieved fruitful results in fault detection and diagnosis (FDD) methods, leaving the complex disturbances not fully considered. The disturbances degrade FDD performance and are very hard to identify. In this paper, an FDD method based on multiple deep neural networks is proposed, which can deal with the complex disturbances. First, we split the FDD into two parts to reduce complexity. In detail, a fully connected neural network is selected to detect fault and diagnose fault degree, and a long short-term memory neural network is carefully tuned to diagnose fault mode. An analysis of the relationship between flight states and fault conditions is employed to design input of the fully connected neural network and the long short-term memory neural network. Then, a novel prediction and compensation method is proposed to deal with the unobservable and time-varying disturbances. Without extra information, data from the inertial measurement unit and navigation system are just enough to eliminate the disturbances, whereas existing methods may need more sensors. At last, illustrative simulations demonstrate the effectiveness and correctness of the proposed method.

An evolving learning-based fault detection and diagnosis method: Case study for a passive chilled beam system

Traditional fault detection and diagnosis (FDD) methods learn from training data obtained under limited operating conditions, after which they stop learning. In this study, we developed an evolving learning-based FDD method for HVAC systems, which learns as the performance of a building system and its components changes. Specifically, an evolving learning algorithm—growing Gaussian mixture regression—is used to construct both a data-driven model representing normal performance and a transfer function for fault diagnosis. The evolving learning-based FDD method was demonstrated for detecting and diagnosing common faults of passive chilled beam systems. We employ generalized performance indices, such as the deviations between predictions (expectations) and measurements, the differences between two parameters, and other features extracted from parameters. A novel feature selection method was developed for selecting fault signatures. An uncertainty threshold determining whether a performance index was within the range of normal operation influences false alarm rates. By increasing the uncertainty thresholds from zero to two standard deviations, false alarm rates for normal operations were reduced from 14.8% to 1.3% and the percentage of normal operation data categorized as an unknown operation was reduced from 25% to 0%. Eight known faults were detected and diagnosed with an accuracy of 100%. A new fault was first categorized as an unknown fault before evolving. After evolving the transfer function by updating the key parameters of the Gaussian components, the unknown fault was also accurately diagnosed. The evolving learning-based FDD method and novel feature selection method can be employed for detecting and diagnosing common faults of other systems or subsystems in the built environment.

Model-Based Deep Transfer Learning Method to Fault Detection and Diagnosis in Nuclear Power Plants

Deep learning–based nuclear intelligent fault detection and diagnosis (FDD) methods have been widely developed and have achieved very competitive results with the progress of artificial intelligence technology. However, the pretrained model for diagnosis tasks is hard in achieving good performance when the reactor operation conditions are updated. On the other hand, retraining the model for a new data set will waste computing resources. This article proposes an FDD method for cross-condition and cross-facility tasks based on the optimized transferable convolutional neural network (CNN) model. First, by using the pretrained model's prior knowledge, the model's diagnosis performance to be transferred for source domain data sets is improved. Second, a model-based transfer learning strategy is adopted to freeze the feature extraction layer in a part of the training model. Third, the training data in target domain data sets are used to optimize the model layer by layer to find the optimization model with the transferred layer. Finally, the proposed comprehensive simulation platform provides source and target cross-condition and cross-facility data sets to support case studies. The designed model utilizes the strong nonlinear feature extraction performance of a deep network and applies the prior knowledge of pretrained models to improve the accuracy and timeliness of training. The results show that the proposed method is superior to achieving good generalization performance at less training epoch than the retraining benchmark deep CNN model.

Octorotor PMSM 의 상호 다중 모델 필터 기반 고장 검출과 진단

Fault detection and diagnosis (FDD) is an essential algorithm applied to unmanned aerial vehicles (UAVs) performing low-level to high-level missions. A UAV consists of many mechanical and electrical components, and the failure of one component may affect others, potentially leading to fatal accidents. Herein, we propose a multiple filter-based FDD method that can classify component faults. An interacting multiple model (IMM) filter is combined with several filter algorithms, such as Kalman filter and extended Kalman filter to form a single multiple-filter structure. The suitability of the model can then be evaluated using the likelihood function and mode probability. The highest mode probability in the fault model suggests the existence of a fault. Thus, it is possible to identify the fault type. Because the IMM-based FDD method stochastically estimates the state of the system, it can flexibly cope with different fault situations. The proposed FDD method was applied to the permanent magnet synchronous motor (PMSM) of a UAV and validated via MATLAB/Simulink.

Fault Detection and Isolation of a Pressurized Water Reactor Based on Neural Network and K-Nearest Neighbor

Nuclear power plants (NPPs) are complex dynamic systems with multiple sensors and actuators. The presence of faults in the actuators and sensors can deteriorate the system’s performance and cause serious safety issues. This calls for the development of fault detection and diagnosis systems for detection and isolation of such faults. In this study, fault detection and diagnosis (FDD) based on neural networks (NN) and K-nearest neighbour (KNN) algorithm is applied to a pressurized water reactor (PWR). Fault detection is first determined based on the NN. Second, the KNN algorithm is used to classify the faults. The proposed approach is capable of classifying a variety of actuator faults, sensor faults, and multiple simultaneous actuator and sensor faults. A set of simulation results is provided to demonstrate the accuracy of the FDD method. The classifier performance is further compared with other machine learning techniques.

Multivariate fault detection for residential HVAC systems using cloud-based thermostat data, part II: Case studies

Fault detection and diagnosis (FDD) using smart thermostat data is a relatively new field of research, but one with immediate practical application to residential indoor climate control. In this two-part report, a statistics-based FDD method is proposed to identify anomalous residential air conditioning systems using aggregated thermostat data. The proposed method extracts features that characterize system operational behaviors, and determines outliers by comparing the features between thousands of systems. Those outliers are subsequently classified into several categories of anomalous operational behaviors, such as near-continuous operation, poor control, and so on, which are closely related to soft mechanical faults. While the first part of the report details the methodology, this second part provides five case studies to demonstrate the effectiveness of the proposed method. Although these cases should not be considered a complete representation of all kinds of anomalous behaviors, the examples illustrate typical problems, including performance degradation faults with or without affecting occupant comfort, control problems with slight degradation, and oversized systems. Each case study exemplifies one type of behavior that is common among a group of systems, rather than a single outlier that behaves much differently from all others. The case studies illustrate how the proposed fault detection and diagnosis method is capable of identifying malfunctioning systems from a population of residential air conditioning systems.

Fault detection and diagnosis with a novel source-aware autoencoder and deep residual neural network

The capability of deep learning (DL) techniques for dealing with non-linear, dynamic and correlated data has paved the way for DL-based fault detection and diagnosis (FDD). Among them, autoencoders (AEs) have shown their potential to serve as the fault detection network. However, misclassifying faulty samples that share similar patterns to normal samples is a common drawback of AEs. In this work, a source-aware autoencoder (SAAE) is proposed as an extension of AEs to incorporate faulty samples in the training stage. In SAAE, flexibility in tuning recall and precision trade-off, ability to detect unseen faults and applicability in imbalanced data sets are achieved. Bidirectional long short-term memory (BiLSTM) with skip connections SAAE is designed as the structure of the fault detection network. Further, a deep network with BiLSTM and residual neural network (ResNet) is proposed for the subsequent fault diagnosis step to avoid randomness imposed by the order of the input features. A framework for combining fault detection and fault diagnosis networks is also presented without the assumption of having a perfect fault detection network. A comprehensive comparison among relevant existing techniques in the literature and SAAE-ResNet is also conducted on the Tennessee-Eastman process, which shows the superiority of the proposed FDD method.

Fault detection and diagnosis in multivariate systems using multiple correlation regression

Correlations among variables are inherent features of a multivariate system. Making a good use of variable correlations is important to fault detection and diagnosis (FDD). In this paper, a novel FDD method is proposed using multiple correlation regression (MCR). MCR builds a regression model using the multiple correlation of a variable with other variables. MCR is the best linear regression that produces residuals with the minimal variance. Based on the MCR residuals of variables, a T2 statistic is defined for fault detection. This T2 statistic consists of a set of independent components, with each component corresponding to the MCR residual of a variable. To reduce the amplification and masking effects caused by fault-free components, three improvements to the residual-based T2 statistic are proposed, called the intersecting T2 (IT), weighted T2 (WT) and enhanced T2 (ET) statistics. In particular, the WT and ET statistics and their control limits vary with samples to adapt to dynamic changes of the system. An online fault detection strategy is proposed by combining the WT and ET statistics. The WT statistic is firstly used to detect the occurrence of faults. After faults were detected, the ET statistic is used to replace the WT statistic in order to enhance the capability of detecting subsequent faulty samples. A contribution-based fault diagnosis method is also developed. Faulty variables are identified by comparing variable contributions to the key faulty component of the residual-based T2 statistic. The implementation and advantages of the proposed FDD method are illustrated with a simulation example and an industrial case study.

A framework to automate fault detection and diagnosis based on moving window principal component analysis and Bayesian network

Through Condition-Based Maintenance strategy, planners can monitor the health of the machinery and recommend actions based on the information obtained. Nevertheless, this approach depends on the successful establishment of Fault Detection and Diagnosis (FDD) processes. Although FDD is a research area in full growth with the development of several methods and heuristics, the availability of data from systems under a fault condition is still scarce in many applications, mainly related to complex systems. In many circumstances, only data from the system in healthy conditions is available and the applied FDD method should be able to detect variations in system conditions and diagnose faults without the need for previous labeled fault data. In this context, this article proposes a hybrid framework to automate FDD based on Moving Window Principal Component Analysis (MWPCA) and Bayesian Network (BN). First, the knowledge base on technical systems is organized to support the next steps of the framework. Then, the detection and diagnosis processes are performed sequentially through MWPCA and BN. The framework was implemented in the analysis of a simplified model of a hydrogenerator, considering real and simulated data. The results showed that the proposed method was able to detect and diagnose several simulated failures.

Fault detection and diagnosis for multiple faults of VAV terminals using self-adaptive model and layered random forest

Fault detection and diagnosis (FDD) can reduce a large amount of energy wasted by faults in variable air volume (VAV) systems. This paper presents a FDD method for detecting and diagnosing multiple faults of VAV terminals using a hybrid method. A self-adaptive zone air-temperature model is presented and used to detect faults of the VAV terminals by comparing the residuals between predicted and measured values with corresponding fault detection thresholds. The model parameters, which are normally constant and determined according to experience from short-period data, are periodically adjusted to improve the model prediction accuracy on the basis of operating data by using a chaos particle swarm optimization algorithm. Moreover, a two-layered random forest-based fault diagnosis method is presented to isolate multiple faults of VAV terminals according to the complexity of the fault symptoms. An expert rule-based fault diagnosis layer is used to isolate VAV terminal faults with simple fault symptoms. A random forest-based fault diagnosis layer is utilized to diagnose VAV terminal faults with complex fault symptoms by adopting a binary decision tree-based multi-label technique. The proposed FDD method is validated using operating data from real VAV systems involving 13 single-fault cases and 6 simultaneous-fault cases. The validation results indicate that the proposed FDD method can accurately detect and diagnose simultaneous multiple faults with only a few independent fault samples and a minimal number of simultaneous-fault samples.

Fault detection and diagnosis for Air Handling Unit based on multiscale convolutional neural networks

This paper proposes a novel fault detection and diagnosis (FDD) method using multiscale convolutional neural networks (MCNNs) for Air Handling Unit (AHU) in Heating, Ventilation, and Air conditioning (HVAC) system. In existing works, it is challenging to achieve high diagnosis performance on multiscale monitoring signals from AHU system since the feature extraction methods in these works are not powerful enough. Although the single-scale convolutional neural networks (CNNs) have been adopted to improve the ability of feature extraction in FDD, it remains difficult to extract strong discriminative feature from multiscale monitoring signals only using single-scale kernels. In this paper, a novel MCNNs-based FDD method is proposed with three different scale kernels to improve the ability of feature extraction and the end-to-end learning strategy is adopted to optimize the model of MCNNs. With strong representation ability, the proposed method can capture highly discriminative features, which can help to improve the diagnostic performance of AHU. The proposed method is compared with other five commonly used methods using the measured data from our AHU experiment platform. The comparison results demonstrate that the proposed MCNNs-based FDD method outperforms other FDD methods.

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.

Deep Learning-Based Fault Diagnosis of Photovoltaic Systems: A Comprehensive Review and Enhancement Prospects

Photovoltaic (PV) systems are subject to failures during their operation due to the aging effects and external/environmental conditions. These faults may affect the different system components such as PV modules, connection lines, converters/inverters, which can lead to a decrease in the efficiency, performance, and further system collapse. Thus, a key factor to be taken into consideration in high-efficiency grid-connected PV systems is the fault detection and diagnosis (FDD). The performance of the FDD method depends mainly on the quality of the extracted features including real-time changes, phase changes, trend changes, and faulty modes. Thus, the data representation learning is the core stage of intelligent FDD techniques. Recently, due to the enhancement of computing capabilities, the increase of the big data use, and the development of effective algorithms, the deep learning (DL) tool has witnessed a great success in data science. Therefore, this paper proposes an extensive review on deep learning based FDD methods for PV systems. After a brief description of the DL-based strategies, techniques for diagnosing PV systems proposed in recent literature are overviewed and analyzed to point out their differences, advantages and limits. Future research directions towards the improvement of the performance of the DL-based FDD techniques are also discussed. This review paper aims to systematically present the development of DL-based FDD for PV systems and provide guidelines for future research in the field.