Fault Diagnosis Part Research Articles

Multiple model fault diagnosis and fault tolerant control for the launch vehicle’s attitude control system

Abstract For the launch vehicle attitude control problem, traditional methods can seldom accurately identify the fault types, making the control method lack of pertinence, which largely affects the effect of attitude control. This paper proposes an active fault tolerant control strategy, which mainly includes fault diagnosis and fault tolerant control. In the fault diagnosis part, a small deviation attitude dynamics model of the launch vehicle is established, Kalman filters with different structures are designed to detect and isolate faults through residual changes, and the fault quantity of the actuator is further estimated. In the fault tolerant control part, the following control scheme is adopted according to the above diagnostic information: when the sensor fault is detected, the sensor measurement data is reconstructed; when the actuator fault is identified, the control allocation matrix is reconstructed. Simulation results show that the proposed method can effectively diagnose sensor fault and actuator faults, and significantly improve attitude tracking accuracy and control adjustment time.

Methodology and application of digital twin-driven diesel engine fault diagnosis and virtual fault model acquisition

Digital real-time fault diagnosis is an effective way to ensure the reliable long-term operation of the diesel engine, but there is still a lack of systematic methods with high integrity and practicability. Therefore, a digital twin-driven diesel engine fault diagnosis method based on the combination of the classification algorithm and the optimization algorithm is proposed and a case study of fuel injection system fault diagnosis is used to illustrate and verify the proposed method. This method closely links the physical system, virtual model, database, and diagnosis system through data transmission and the diagnostic process consists of three parts: classification, diagnosis, and decision. The fault classification part can preliminarily lock the possible types and degrees of faults, and point out the key classification features for each fault type by using classification algorithms such as Random Forest. The fault diagnosis part can diagnose and reproduce the diesel engine faults by using an optimization-simulation joint calculation model, where the virtual model variables and optimization algorithm are determined according to the possible fault types, and the optimization target depends on the importance of classification features. Then the maintenance decision can be made according to the fault detailed information. The proposed method reduces the requirement of covering the fault degree of the database, and the obtained fault model provides the possibility for subsequent online optimization and also facilitates the development of intelligent engine management.

Robust locally linear embedding and its application in analogue circuit fault diagnosis

During long-term operation of analogue circuits, fault diagnosis is important for preventing the occurrence of hazards. However, noise often accompanies sampled signals and makes the task of fault diagnosis more difficult. Therefore, developing a robust feature extraction technique is an indispensable part of fault diagnosis. The locally linear embedding (LLE) algorithm has recently emerged as a promising technique for dimensional reduction and feature extraction because it preserves linear neighborhoods, and it is quite effective when there is a locally linear dependent structure embedded in fault data. However, LLE is sensitive to noise. Therefore, the maximum correntropy criterion is adopted to resist non-Gaussian noise by seeking the optimal weight coefficient, and a half-quadratic optimization procedure is introduced to address the objective function. Moreover, softmax regression is applied to locate faults. Finally, two typical analogue circuit systems are used to demonstrate the robustness of the modified algorithm to non-Gaussian noise. The experimental results show that the robust LLE algorithm can outperform LLE in the extraction of fault features when there is non-Gaussian noise in the fault signals, and the proposed fault diagnosis method has a better effect in locating faults compared with other feature extraction methods.

Multi-Source Partial Discharge Fault Location with Comprehensive Arrival Time Difference Extraction Method and Multi-Data Dynamic Weighting Algorithm

The location of the partial discharge source is an important part of fault diagnosis inside power equipment. As a key step of the ultra-high frequency location method, the extraction of the time difference of arrival can generate large errors due to interference. To achieve accurate time difference extraction and further multi-source partial discharge location, a location method with comprehensive time difference extraction and a multi-data dynamic weighting algorithm is proposed. For time difference extraction, the optimized energy accumulation curve method applies wavelet transform and mode maximization calculations such that it overcomes the effect of interference signals before the wave peak. The secondary correlation method improves the interference capability by performing two rounds of correlation calculations. Both extraction methods are combined to reduce the error in time difference extraction. Then, the dynamic weighting algorithm effectively utilizes multiple data and improves the location accuracy. Experimental results on multi-source partial discharge locations performed in a transformer tank validate the accuracy of the proposed method.

XFDDC: eXplainable Fault Detection Diagnosis and Correction framework for chemical process systems

Industry 4.0 process fault detection and diagnosis(FDD) is built on the foundations of Industrial Internet of Things(IIoT) for sensing and artificial intelligence for recognizing patterns. Despite the performance and application of various learning-based algorithms in multiple sectors, their black-box nature makes industrial experts skeptical. So, this paper proposes an eXplainable Fault Detection, Diagnosis, and Correction(XFDDC) Framework to create best-fit FDD models that are explainable. The XFDDC framework is designed to explain the FDD model predictions using eXplainable Artificial Intelligence(XAI) techniques. The proposed framework was applied to the bench-marked Tennessee Eastman Process(TEP) dataset. On evaluation, an XGBoost model yielded better Fault Detection Rate(FDR) and F1 score against popular transparent and complex models like Naive Bayes, K-Nearest Neighbors, Random Forest and a rule-based version of XGBoost. To explain the predictions of the XGBoost model, the XFDDC framework suggests the use of feature-based XAI techniques. So, the TreeSHAP algorithm is applied on the XGBoost model to generate local and global explanations as part of fault diagnosis. The proposed framework also recommends counterfactual explanations to provide action recommendations for correcting the fault situation. Thus, a best-fit explainable XGBoost Fault Detection, Diagnosis, and Correction (XGBoost-XFDDC) model is created.

A Feature Inherited Hierarchical Convolutional Neural Network (FI-HCNN) for Motor Fault Severity Estimation Using Stator Current Signals

Motors, which are one of the most widely used machines in the manufacturing field, take charge of a key role in precision machining. Therefore, it is important to accurately estimate the health state of the motor that affects the quality of the product. The research outlined in this paper aims to improve motor fault severity estimation by suggesting a novel deep learning method, specifically, feature inherited hierarchical convolutional neural network (FI-HCNN). FI-HCNN consists of a fault diagnosis part and a severity estimation part, arranged hierarchically. The main novelty of the proposed FI-HCNN is the special inherited structure between the hierarchy; the severity estimation part utilizes the latent features to exploit the fault-related representations in the fault diagnosis task. FI-HCNN can improve the accuracy of the fault severity estimation because the level-specific abstraction is supported by the latent features. Also, FI-HCNN has ease in practical application because it is developed based on stator current signals which are usually acquired for a control purpose. Experimental studies of mechanical motor faults, including eccentricity, broken rotor bars, and unbalanced conditions, are used to corroborate the high performance of FI-HCNN, as compared to both conventional methods and other hierarchical deep learning methods.

Real-time Online Fault Diagnosis of Rolling Bearings Based on KNN Algorithm

In order to realize the predictive maintenance of rolling bearings in industry, this paper proposes a real-time online fault diagnosis method for rolling bearings based on KNN algorithm. The method mainly includes two steps: fault diagnosis model training and real-time online fault diagnosis. Firstly, the vibration signal is preprocessed: data classification, data cleaning, data segmentation and feature parameter extraction, and then training and optimizing the fault diagnosis model. In the real-time online fault diagnosis part of the rolling bearing, the real-time online extraction of the characteristic parameters of the vibration signal is used to realize real-time online fault diagnosis through the fault diagnosis model. The results show that the fault diagnosis model based on KNN algorithm is better than the fault diagnosis model based on C4.5 algorithm and CART algorithm, which is more suitable for fault diagnosis of rolling bearings. Using this method to diagnose rolling bearings can help predictive maintenance before rolling bearing failures and reduce the economic losses caused by unplanned downtime of critical equipment.

An improved feature extraction method using texture analysis with LBP for bearing fault diagnosis

Bearings are one of the most widespread components used for energy transformation in machines. Mechanical wear and faulty bearings reduce the efficiency of rotating machines and thus increase energy consumption. The feature extraction process is an essential part of fault diagnosis in bearings. In order to diagnose the fault caused by the bearing correctly, it is necessary to determine an effective feature extraction method that best describes the fault.In this study, a new approach based on texture analysis is proposed for diagnosing bearing vibration signals. Bearing vibration signals were first converted to gray scale images. It can be understood from the images that the signals of different bearing failures form different textures. Then, using these images, LBP (Local Binary Pattern) and texture features were obtained. Using these features, different machine learning models and bearing vibration signals are classified. Three different data sets were created to test the proposed approach. For the first data set, the signals composed of very close velocities were classified. 95.9% success rate was observed for the first data set. The second data set consists of faulty signals at different parts of the bearing (inner ring, outer ring and ball) measured in the same RPM. The type of fault has been determined, and a 100% success rate was obtained for this data set. The final data set is composed of the fault size dimensions (mm) of different ratios. With the proposed approach, a 100% success rate was obtained in the classification of these signals. As a result, it was observed that the obtained feature had promising results for three different data types and was more successful than the traditional methods.

Multi-objective optimization of SOFC systems

For solid oxide fuel cell (SOFC) development, maximizing its electrical efficiency and minimizing its cost are two important optimization objects. A new optimization strategy is proposed in this work, which can maximize the SOFC electrical efficiency and minimize the cost in the case of an air leakage fault. The proposed optimization method involves a fault diagnosis module, a switching module and two backup optimizers. The fault diagnosis part is employed to identify the SOFC current fault type, and the switching module is used to select the appropriate backup optimizer. For the efficiency and cost are two conflicting objectives, the multi-objective optimization strategy based on a non-dominated sorting particle swarm optimization algorithm is applied to determine the trade-off solutions. The optimization results show the proposed method can achieve the maximum efficiency and the minimum cost in the case of SOFC normal, and even in the air leakage fault.

Feature extraction of rotor fault based on EEMD and curve code

Feature extraction is an important part of fault diagnosis. Rotor faults are of various types with little differences between each other and rely on effective methods of feature extraction to be recognized correctly. There are many methods applied to feature extraction at present. However, fault features computed are generally based on fixed samples, which may lead to an unsatisfactory result for other samples. To solve this problem, a novel feature extraction method based on Ensemble Empirical Mode Decomposition (EEMD) and Curve Code (CC) is presented in this paper. The noisy vibration signals are first denoised using wavelet thresholding-based de-noising technology. Then EEMD is applied to decompose the de-noised signals to obtain all Intrinsic Mode Functions (IMFs), and the corresponding Standard Deviations (SDs) are calculated. Finally, the curve plotted by the SDs is coded into a digital sequence called CC as a fault feature. Results of comparative experiments, designed using a large number of samples from rotor laboratory bench, prove the validity of the proposed method and show that the proposed method has higher and more stable Fault Discrimination Rate (FDR) for rotor fault recognition.

Optimal fault-tolerant control strategy of a solid oxide fuel cell system

For solid oxide fuel cell (SOFC) development, load tracking, heat management, air excess ratio constraint, high efficiency, low cost and fault diagnosis are six key issues. However, no literature studies the control techniques combining optimization and fault diagnosis for the SOFC system. An optimal fault-tolerant control strategy is presented in this paper, which involves four parts: a fault diagnosis module, a switching module, two backup optimizers and a controller loop. The fault diagnosis part is presented to identify the SOFC current fault type, and the switching module is used to select the appropriate backup optimizer based on the diagnosis result. NSGA-II and TOPSIS are employed to design the two backup optimizers under normal and air compressor fault states. PID algorithm is proposed to design the control loop, which includes a power tracking controller, an anode inlet temperature controller, a cathode inlet temperature controller and an air excess ratio controller. The simulation results show the proposed optimal fault-tolerant control method can track the power, temperature and air excess ratio at the desired values, simultaneously achieving the maximum efficiency and the minimum unit cost in the case of SOFC normal and even in the air compressor fault.

An Efficient Simplified Physical Faulty Model of a Permanent Magnet Synchronous Generator Dedicated to Stator Fault Diagnosis Part II: Automatic Stator Fault Diagnosis

This paper proposes an automatic and intelligent stator fault diagnosis system for permanent magnet synchronous generators. The system is based on the use of a feedforward multilayer perception artificial neural network (ANN) performed by the back-propagation training algorithm. From a thorough study and analysis of the behavior of the negative-sequence voltage of the machine under different stator faults and different operating conditions, two new robust indicators of faults are selected as the ANN inputs. The two fault indicators are the phase angle of the negative sequence voltage and the frequency of the machine voltages. After successfully training and testing the ANN with specific and meaningful databases, the ANN behavior is validated by using an experimental database acquired from a real machine. The accurate results provided by the ANN prove the reliability of the proposed system to automatically diagnose different stator faults under variable speed, variable fault current, and variable load conditions overcoming different disturbances, such as the noisy signals and the harmonics in the machine.

Efficient Simplified Physical Faulty Model of a Permanent Magnet Synchronous Generator Dedicated to the Stator Fault Diagnosis Part I: Faulty Model Conception

This paper proposes an original, simple, and fast permanent magnet synchronous generator model implemented with a new conception on the graphic interface of MATLAB/Simulink environment. The obtained physical model suggests a new way to easily carry out different types of stator faults such as interturns short-circuit faults, phase-to-phase faults, and phase-to-ground faults, as well as simultaneous multifaults. This is handled by connecting the two desired points in the stator on the graphic interface of Simulink exactly as they can be carried out in the experimental tests. This model has the originality to take into account the arrangement of the stator windings of a multipole machine by differentiating between the windings under the same pole-pair and those under different pole-pairs. A comparison between the simulated and the experimental results is done to verify the behavior of the proposed model in both healthy and faulty modes under different operating speeds. The good comparison results lead to validate the correct behavior of the proposed model with a satisfactory accuracy.

Fault tolerance control of SOFC systems based on nonlinear model predictive control

Heat management and load tracking are two crucial tasks for solid oxide fuel cell (SOFC) systems development. In literature, plenty of temperature controllers and load tracking controllers have been successfully designed for SOFC systems. However, previous researches are limited to control design in the case of SOFC normal conditions. For a SOFC system, faults can occur on any parts at any time, thus a controller with a specific design must be used. In this work a new control strategy that tolerates the SOFC system faults is proposed, which includes a fault diagnosis module, a decision-making part and four backup controllers. The fault diagnosis part is used to classify the SOFC system current faults (normal, fuel leakage fault, air compressor fault, or both fuel leakage and air compressor faults). Based on the diagnosis results, the decision-making part is designed to select the appropriate backup controller. Four nonlinear model predictive controllers based on back propagation (BP) neural networks are respectively built to follow loads and maintain appropriate temperatures in the case of fuel leakage fault, air compressor fault, both fuel leakage and air compressor faults, and SOFC normal condition. The results show that the proposed fault tolerance control strategy can track the voltage and make the SOFC temperature steady near the enactment value in working in faulty conditions, which may result in both lifetime and performance improvement for the SOFC systems.

A Survey on Wind Turbine Condition Monitoring and Fault Diagnosis—Part I: Components and Subsystems

This paper provides a comprehensive survey on the state-of-the-art condition monitoring and fault diagnostic technologies for wind turbines (WTs). The Part I of this survey briefly reviews the existing literature surveys on the subject, discusses the common failure modes in the major WT components and subsystems, briefly reviews the condition monitoring and fault diagnostic techniques for these components and subsystems, and specifically discusses the issues of condition monitoring and fault diagnosis for offshore WTs.

A Survey on Wind Turbine Condition Monitoring and Fault Diagnosis—Part II: Signals and Signal Processing Methods

This paper provides a comprehensive survey on the state-of-the-art condition monitoring and fault diagnostic technologies for wind turbines. The Part II of this survey focuses on the signals and signal processing methods used for wind turbine condition monitoring and fault diagnosis.

A Setup for Active Fault Diagnosis

A setup for active fault diagnosis (AFD) of parametric faults in dynamic systems is formulated in this note. It is shown that it is possible to use the same setup for both open loop systems, closed-loop systems based on a nominal feedback controller as well as for closed-loop systems based on a reconfigured feedback controller. This will make the proposed AFD approach very useful in connection with fault tolerant control (FTC). The setup will make it possible to let the fault diagnosis part of the fault tolerant controller remain unchanged after a change in the feedback controller. The setup for AFD is based on the Youla-Jabr-Bongiorno-Kucera (YJBK) parameterization of all stabilizing feedback controllers and the dual YJBK parameterization. It is shown that the AFD is based directly on the dual YJBK transfer function matrix. This matrix will be named the fault signature matrix when it is used in connection with AFD

POSSIBILISTIC CAUSAL DIAGNOSIS: APPLICATION TO ENGINE DYNO TEST BENCHES

A causal diagnosis solution relying on experts' knowledge concentrates on the different malfunctions that may disturb the data acquisition process and tells the engine test bench tuning engineer which malfunction has occurred or which malfunctions are most likely suspected. The use of fuzzy sets and possibility theory provides better feedback and knowledge representation. The general architecture of the system is described, and a prototype of the fault-diagnosis part of this sytem is presented. It concerns the implementation of an (off-line) automatic knowledge formalization system and the implementation of the (on-line) possibilistic causal diagnosis process.