A Data-driven Fault Detection Method Based on Dissipative Trajectories

The day-by-day increase in digital loads draws attention towards the need for an efficient and compatible distribution network. An LVDC distribution network has the capability to fulfill such digital load demands. However, the major challenge of an LVDC distribution network is its vulnerability during a fault. The need for a high-speed fault detection method is inevitable before it can be widely adopted. This paper proposes a new fault detection method which extracts the features of the current during a fault. The proposed fault detection method uses the merits of overcurrent, the first and second derivative of current, and signal processing techniques. Three different features are extracted from a time domain current signal through a sliding window. The extracted features are based upon the root squared zero, second, and fourth order moments. The features are then set with individual thresholds to discriminate low-, high-, and very high-resistance faults. Furthermore, a fault is located through the superimposed power flow. Moreover, this study proposes a new method based on the vector sum of positive and negative pole currents to identify the faulty pole. The proposed scheme is verified by using a modified IEEE 13 node distribution network, which is implemented in Matlab/Simulink. The simulation results confirm the effectiveness of the proposed fault detection and identification method. The simulation results also confirm that a fault having a resistance of 1 m Ω is detected and interrupted within 250 μ s for the test system used in this study.

Fault detection based on robust characteristic dimensionality reduction

In this study, the use of instantaneous symmetrical components (ASB) for mechanical fault detection in an inverter controlled induction motor has been investigated. The proposed fault detection method is based on the calculation of the ASBs of the stator currents. The spectral density (PSD) of the positive sequence power component is estimated using the least squares method (EKY). Then, mechanical fault detection is performed using the generalized likelihood ratio test (GLRT). Modulated signals from both stator currents and closed-loop have been shown to be successful in detecting faults. The simulation results in the analytical model of an inverter controlled induction motor give the effectiveness of this proposed approach. This results in an effective fault detection method for an induction motor load torque oscillation.

This paper presents a simple fault detection method and fault-tolerant strategy based on the model predictive control theory. The proposed fault detection method can detect and locate a failed switch with current error signals dedicated to each switch, based on the position of space vectors. Following the fault detection, the proposed fault-tolerant control strategies are developed to produce three-phase balanced sinusoidal output voltages/currents with only normally functioning bidirectional switches. Therefore, the developed control technique for the matrix converter can preserve continuous operation even after the failure of switches and improve reliability of the drives. Simulation results are provided to validate the effectiveness of the applied detection and control strategy.

This paper proposes a new on-line fault detection method to detect open-switch faults in a Matrix Converter (MC). The proposed fault detection method is based on a two-step algorithm. The first step identifies the occurrence of the fault by monitoring the measured load currents. The second step locates the faulty switch by measuring the response of the MC to a set of predefined switching states. The proposed method allows fast detection and isolation of a faulty switch, independent of the modulation strategy, the voltage transfer ratio, or the output frequency, and without any requirement of additional voltage sensors. The paper also presents a Model Predictive Control (MPC)-based post-fault strategy that enables the MC to continue its operation during faulty conditions. Time-domain simulation studies are performed in the MATLAB/SIMULINK environment to evaluate the performance of the proposed fault detection method and the effectiveness of the MPC as a post-fault control strategy.

As the growth of Industry 4.0, online fault detection plays a crucial role in ensuring the manufacturing quality. Generally, the fault detection methods can be classified into model-based and data-driven methods. There are advantages/disadvantages between two methods. In this study, we integrated both methods in order to develop an efficient fault detection method for non-Gaussian industrial processes. The data-driven method, independent component analysis (ICA) is used to extract non-Gaussian information and dimensionality reduction. Meanwhile, the model-based method, generalized likelihood ratio (GLR) test is adopted as the charting statistic. The proposed ICA-GLR method has advantages of 1) detecting a wide range of process changes, 2) estimating the change points and 3) needless prior parameters to be specified by practitioner. The efficiency of the proposed ICA-GLR fault detection method will be verified via implementing one simulated non-Gaussian process and two real manufacturing processes: Tennessee Eastman process and semiconductor manufacturing process. Results demonstrate that the proposed ICA-GLR method has superior fault detectability when compared to traditional methods, such as principal component analysis and ICA.

Manifold regularized stacked autoencoders-based feature learning for fault detection in industrial processes

Data-driven fault detection and identification methods are important in large-scale chemical processes. However, some traditional methods often fail to show superior performance owing to the self-limitations and the characteristics of process data, such as nonlinearity, non-Gaussian distribution, and multi-operating mode. To cope with these issues, the k-NN (k-Nearest Neighbor) fault detection method and extensions have been developed in recent years. Nevertheless, these methods are primarily used for fault detection, and few papers can be found that examine fault identification. In this paper, in order to extract effective fault information, the relationship between various faults and abnormal variables is studied, and an accurate “fault–symptom” table is presented. Then, a novel fault identification method based on k-NN variable contribution and CNN data reconstruction theories is proposed. When there is an abnormality, a variable contribution plot method based on k-NN is used to calculate the contribution index of each variable, and the feasibility of this method is verified by contribution decomposition theory, which includes a feasibility analysis of a single abnormal variable and multiple abnormal variables. Furthermore, to identify all the faulty variables, a CNN (Center-based Nearest Neighbor) data reconstruction method is proposed; the variables that have the larger contribution indices can be reconstructed using the CNN reconstruction method in turn. The proposed search strategy can guarantee that all faulty variables are found in each sample. The reliability and validity of the proposed method are verified by a numerical example and the Continuous Stirred Tank Reactor system.

Disposable soft manipulators provide a promising solution for robot-assisted surgeries in low- and middle-income countries due to their cost advantages. The surgical applications are of high reliability requirement, while the low-cost materials and mechanical structures impair robot reliability, and therefore, some measures should be taken to enhance the reliability of disposable soft manipulators, for example, fault detection, identification and tolerant control. This paper proposes a new fault detection algorithm for locked-motor faults of soft manipulators. By leveraging the projection Jacobian matrix, it requires less actuators than conventional methods. In terms of the uncertainty introduced by the projection Jacobian, an adaptive algorithm is proposed to compensate for it. Dedicated experiments are conducted to validate our proposed fault detection method and results prove its efficacy for monitoring system health conditions and controlling the system under faulty modes.

Long–short-term memory encoder–decoder with regularized hidden dynamics for fault detection in industrial processes

In this paper, we address the problem of real-time fault detection in wind turbines. Starting from a data-driven fault detection method, the contribution of this paper is twofold. First, a sensor selection algorithm is proposed with the goal to reduce the computational effort of the fault detection method. Second, an analysis is performed to reduce the data acquisition time needed by the fault detection method, that is, with the goal of reducing the fault detection time. The proposed methods are tested in a benchmark wind turbine where different actuator and sensor failures are simulated. The results demonstrate the performance and effectiveness of the proposed algorithms that dramatically reduce the number of sensors and the fault detection time.

In this paper, a simple and fast open-phase fault (OPF) detection and fault tolerant control method is proposed for three-phase induction machine. Instead of using integrators or low pass filter, this method utilizes the phase current magnitude and trajectory to detect open phase fault in three-phase motor drive. This significantly increases the speed of fault detection, allowing an OPF to be detected and compensated within several sampling cycles, which is advantageous for safety-critical applications. In addition, a double synchronous reference frame PI (DSRF-PI) controller method is adopted as the fault tolerant controller. Compared to previous methods that utilizes feedforward voltage compensation, the DSRF-PI does not require prior knowledge on the machine parameters and operating condition, and therefore is easier to implement. The proposed fault detection and fault tolerant control method are simulated using MATLAB Simulink simulation where the results confirms the high speed and robustness of the proposed methods.

A Data-Driven Fault Detection and Diagnosis by NSGAII-t-SNE and Clustering Methods in the Chemical Process Industry

Three-phase induction motors are among the most frequently used motors in industrial areas due to their simple structure, low cost, power specifications, etc. Electric energy consumption from these motors accounts for 68% of the energy used by all motors. The faults that occur in these motors over time decrease motor efficiency and result in significant energy consumption. In this study, a new method based on spectral subtraction (SS) was suggested for determining broken rotor bar (BRB) faults in these motors. The traditional method of motor current signature analysis via Fast Fourier Transform (FFT) is hard to use to diagnose BRB faults because the sideband characteristics used as a fault indicator are also seen in the healthy state at low amplitude levels. In the proposed fault detection method, the FFT of both healthy case and faulty case current signals were calculated and then the SS signal is obtained by subtracting the FFT of the healthy motor from the faulty motor for each time step. In the residual SS signal, fault detection was performed by examining the amplitude levels of the harmonic component of the BRB fault. Experimental results indicate that BRB faults can be successfully detected in squirrel-cage rotor induction motors using the suggested method.

The increasing importance of power electronic converters in supplying electrical energy to utility grids places a higher priority to detect and protect against fault conditions. Fault detection and isolation are particularly important for inverters that provide black-start recovery for microgrids since these converters provide the energy source for restoration after a power outage. This paper presents a new fault detection and location method for Cascaded B-Bridge (CHB) multilevel inverters. The new fault detection method is based on monitoring the output voltage of each cell and output current directions along with each switch's state. By monitoring each cell's output voltage and current direction, the faulty cell can be detected and isolated. After the faulty cell is detected, the faulty switch can be located by comparing the current direction with the switching states. This technique is implemented with Level-Shifted Pulse Width Modulation (LS-PWM) in order to maintain acceptable total harmonic distortion (THD) levels at the converter. The proposed method can be implemented for a CHB with any number of cells, can operate with nonlinear loads, and offers very fast detection times. Simulation and experimental results verify the performance of this method.