Physics-informed ensemble learning for hierarchical fault diagnosis in quadruped robots

In this paper, an interacting multiple-model (IMM) based fault detection, diagnosis and reconfigurable control approach for discrete-time stochastic systems is proposed. Fault detection and diagnosis (FDD) is carried out using the IMM estimator. The linear quadratic regulator (LQR) and an eigenstructure assignment (EA) techniques have been used for nominal and reconfigurable control laws design, respectively. To achieve zero steady-state tracking error, a set of feedforward control gains is also designed using an input weighting approach. The paper has considered not only actuator and sensor faults, but also system component faults. To achieve fast and reliable fault detection, diagnosis and controller reconfiguration, new fault diagnosis and reconfiguration mechanisms have been proposed using appropriate combination of the information provided by the mode probabilities from the IMM algorithm and an index related to the closed-loop system performance. The proposed approach is evaluated using an aircraft example in the presence of system component, actuator and sensor faults.

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

An integrated fault detection, diagnosis, and reconfigurable control scheme based on interacting multiple model (IMM) approach is proposed. Fault detection and diagnosis (FDD) is carried out using an IMM estimator. An eigenstructure assignment (EA) technique is used for reconfigurable feedback control law design. To achieve steady-state tracking, reconfigurable feedforward controllers are also synthesized using input weighting approach. The developed scheme can deal with not only actuator and sensor faults, but also failures in, system components. To achieve fast and reliable fault detection, diagnosis, and controller reconfiguration, new fault diagnosis and controller reconfiguration mechanisms have been developed by a suitable combination of the information provided by the mode probabilities from the IMM algorithm and an index related to the closed-loop system performance. The proposed approach is evaluated using an aircraft example, and excellent results have been obtained.

The multi‐joint manipulator with vision sensors has been widely used in real applications. However, the fault detection and diagnosis accuracy are lowered and the time expense is increased for the increased number of sensors, as there are many factors that are relative with this problem. This paper is focused on the fault detection and diagnosis problem of multi‐joint manipulator, and the problem was divided into two sub‐problems. The first is that the position estimation strategy based on data fusion of visual sensor and the position sensor was designed to carry out the fault detection, and the whether the faults had happened or not were determined by the position estimation errors. The second was focused on the fault diagnosis problem, where the deep convolutional neural network (DCNN) fault diagnosis model based on time‐frequency mixed signal was constructed. The proposed DCNN uses the time and frequency domain information as its inputs and executes the classification tasks. The specific fault was determined through the output of DCNN model. The DCNN model was activated only when the first fault detection unit indicated that there was a fault, so the time expense was reduced from 5.3 to 2.6 s. The experiment based on the AUBO‐i5 manipulator was carried out to evaluate the proposed fault detection and diagnosis model, where 10 categories of data sets that represent different working conditions of manipulator were adopted. The experimental results showed that the proposed multi‐joint manipulator fault detection could improve the position estimation accuracy by 41.2%, and the fault diagnosis accuracy was improved by 20%.

The problem of the wind turbine gearbox fault diagnosis was investigated by employing the support vector machine (SVM), which is optimized by the improved fruit fly intelligent algorithm with decreasing steps. First of all, the fault characteristic value extracted by Hilbert transform envelope is presented. Secondly, the general SVM solution to the wind turbine gearbox fault diagnosis problem is presented, where the improved fruit fly intelligent algorithm is adopted to optimize the performance of the model of the fault detection and diagnosis. Thirdly, the effectiveness of three fault diagnosis models are compared, including the fault detection and diagnosis model by the traditional SVM, its improved model optimized by the partical swarm optimization algorithm, and improved model by the SVM optimized by the proposed approach. By using the proposed optimization algorithm, the accuracy of gearbox fault diagnosis is much better than other two models, which are validated by the extensive simulation results based on practical historical operation data.

Technological sophistication is growing day by day, that is embedded in the vehicle systems. A complex system interacts with its surrounding which executes many set of tasks by maintaining its performance. Any change in a system other than its acceptable performance or anything which is not tolerable is treated as a fault. So fault detection and diagnosis becomes necessary in system or component malfunction and thus decreasing the downtime of the system. Fault diagnosis includes necessary facts that are used to identify the faults encountered in the process. Various models have developed to map the hierarchical system-level fault diagnostic information consisting of causal dependencies between observable symptoms and failure modes that are associated with a system. In this paper, various areas are discussed that are involved in fault diagnosis and detection model related to the automobile domain. This survey mainly concentrate on which source is better for constructing fault diagnosis model. Text driven model is found to be reliable for creating D-matrix model to maintain accuracy.

Fundamentals.- Supervision and fault management of processes - tasks and terminology.- Reliability, Availability and Maintainability (RAM).- Safety, Dependability and System Integrity.- Fault-Detection Methods.- Process Models and Fault Modelling.- Signal models.- Fault detection with limit checking.- Fault detection with signal models.- Fault detection with process-identification methods.- Fault detection with parity equations.- Fault detection with state observers and state estimation.- Fault detection of control loops.- Fault detection with Principal Component Analysis (PCA).- Comparison and combination of fault-detection methods.- Fault-Diagnosis Methods.- Diagnosis procedures and problems.- Fault diagnosis with classification methods.- Fault diagnosis with inference methods.- Fault-Tolerant Systems.- Fault-tolerant design.- Fault-tolerant components and control.- Application Examples.- Fault detection and diagnosis of DC motor drives.- Fault detection and diagnosis of a centrifugal pump-pipe-system.- Fault detection and diagnosis of an automotive suspension and the tire pressures.

Many assessment and detection methods are used to diagnose faults in machines. High accuracy in fault detection and diagnosis can be achieved by using numerical methods with noise-resistant properties. However, to some extent, noise always exists in measured data on real machines, which affects the identification results, especially in the diagnosis of early- stage faults. In view of this situation, a damage assessment method based on blind source separation and dynamic fuzzy neural network (DFNN) is presented to diagnose the early-stage machinery faults in this paper. In the processing of measurement signals, blind source separation is adopted to reduce noise. Then sensitive features of these faults are obtained by extracting low dimensional manifold characteristics from the signals. The model for fault diagnosis is established based on DFNN. Furthermore, on-line computation is accelerated by means of compressed sensing. Numerical vibration signals of ball screw fault modes are processed on the model for mechanical fault diagnosis and the results are in good agreement with the actual condition even at the early stage of fault development. This detection method is very useful in practice and feasible for early-stage fault diagnosis.

Quantum computing assisted deep learning for fault detection and diagnosis in industrial process systems

An inverse fault detection and diagnosis (IFDD) strategy for practical application on chiller product

Deep learning technology is increasingly used in the field of power system fault detection and diagnosis, and its powerful feature learning capability makes it play an important role in intelligent process control. In this paper, we propose a method for high resistance fault detection in power systems and design a CNN-Attention-LSTM fault diagnosis model using various deep learning models such as convolutional neural network. The model training and simulation experiments are carried out on the collected power fault dataset. The accuracy, reliability and security of the proposed power fault detection method for high resistance fault phase identification are 99.5%, 99.8% and 99.2%, respectively. The model can accurately classify cable faults in cable fault diagnosis, and also has better diagnostic effect on transformer faults in the power system, in which the diagnostic accuracy of harmonic faults is as high as 100%, showing better fault classification and diagnosis performance.

Fault identification and diagnosis in air compressor systems are critical for maintaining operational efficiency, reducing downtime, and minimizing maintenance costs. Traditional diagnostic methods often struggle with the complexity and variability of faults in these systems. This research paper explores the use of Radial Basis Function (RBF) neural networks for the identification and diagnosis of faults in air compressor systems. RBF neural networks, known for their powerful pattern recognition capabilities, are employed to enhance the accuracy and reliability of fault detection. The study details the structure and training of RBF neural networks, the methodology for fault diagnosis, and presents experimental results demonstrating the effectiveness of this approach. The findings show that RBF neural networks can significantly improve fault diagnosis accuracy, providing a robust and efficient framework for real-time fault detection in air compressor systems. This research contributes to the development of more reliable and efficient maintenance strategies for industrial applications.

Enhancing industrial machinery maintenance through advanced fault and novelty detection using variational autoencoder and hybrid transformer model

Utilizing the organic Rankine cycle (ORC) for waste heat recovery is an important energy conversion method. Some faults may occur in the ORC in actual operation, but few studies have focused on the fault detection and diagnosis of the whole ORC system. Fault detection detects whether a fault occurs in the system and fault diagnosis diagnoses where the fault is. This paper investigated a fault detection and diagnosis scheme of the ORC system for waste heat recovery based on machine learning. First, a thermodynamic ORC model was established. Three kinds of faults (expander fault, pump fault, and heat exchanger fault) and three kinds of algorithms [logistic regression, softmax regression, and support vector machines (SVMs)] were described. The data of four major important faults (fouling fault of the evaporator and of the condenser, looseness of the mechanical moving parts in the expander, and blocking of the pump) were generated from the thermodynamic ORC model and used to train the fault detection and diagnosis schemes. To evaluate the accuracy of the fault detection and diagnosis schemes, a set of experimental data was employed to test the schemes. The accuracy scores of fault detection using logistic regression and support vector machines were 77.42% and 96.77%, respectively. The accuracy scores of fault diagnosis using softmax regression and SVM were 91.78% and 94.52%, respectively. The test times of fault diagnosis using softmax regression and SVM were 0.0099 and 0.0085 s, respectively. The results demonstrated that machine learning–based fault detection and diagnosis schemes for the ORC have high accuracy and immediacy. Therefore, the proposed schemes are promising tools for fault detection and diagnosis of the ORC system for waste heat recovery.

Power systems frequently experience variations in their operation, which are mostly manifested as transmission line faults. Over the past decade, various techniques of fault diagnosis have been developed to ensure reliable and stable operation of power systems. This paper reviews the current literature on advanced application of fault diagnosis in power systems. Application of different fault diagnosis schemes is presented, with emphasis on reliable fault detection and classification of power system faults. The motivation behind applications of emerging process history, or pattern recognition, techniques in power system fault diagnosis has been reviewed. An extensive review of advanced mathematical techniques, in pattern recognition methods, involving wavelet transform, artificial neural networks and support vector machines has been presented. The paper also introduces a novel unsupervised technique of quarter-sphere support vector machine for power system fault detection and classification and reviews its application as future research in the developing area of fault diagnosis.