Possible Types Of Faults Research Articles

FTT-NAS: Discovering Fault-tolerant Convolutional Neural Architecture

With the fast evolvement of embedded deep-learning computing systems, applications powered by deep learning are moving from the cloud to the edge. When deploying neural networks (NNs) onto the devices under complex environments, there are various types of possible faults: soft errors caused by cosmic radiation and radioactive impurities, voltage instability, aging, temperature variations, malicious attackers, and so on. Thus, the safety risk of deploying NNs is now drawing much attention. In this article, after the analysis of the possible faults in various types of NN accelerators, we formalize and implement various fault models from the algorithmic perspective. We propose Fault-Tolerant Neural Architecture Search (FT-NAS) to automatically discover convolutional neural network (CNN) architectures that are reliable to various faults in nowadays devices. Then, we incorporate fault-tolerant training (FTT) in the search process to achieve better results, which is referred to as FTT-NAS. Experiments on CIFAR-10 show that the discovered architectures outperform other manually designed baseline architectures significantly, with comparable or fewer floating-point operations (FLOPs) and parameters. Specifically, with the same fault settings, F-FTT-Net discovered under the feature fault model achieves an accuracy of 86.2% (VS. 68.1% achieved by MobileNet-V2), and W-FTT-Net discovered under the weight fault model achieves an accuracy of 69.6% (VS. 60.8% achieved by ResNet-18). By inspecting the discovered architectures, we find that the operation primitives, the weight quantization range, the capacity of the model, and the connection pattern have influences on the fault resilience capability of NN models.

Traveling wave-based protection for SVC connected transmission lines using game theory

This paper proposes a novel single-end protection scheme for the compensated transmission line based on traveling waves and Game Theory. Initially, this study investigates the effects of Static VAR Compensator (SVC) on Traveling Waves (TW) passing through SVC location. Subsequently, it evaluates the impacts of SVC switching on conventional protections based on traveling wave theory. Following that, a single-end method to locate faults on compensated transmission lines with SVC is presented in this study. Moreover, Game theory was utilized to discriminate extracted TWs in this paper. In order to investigate the sensitivity of the proposed scheme regarding the extraction techniques of TW, two techniques, namely Mathematical Morphology Filter (MMF) and Discrete Wavelet Transform (DWT) have been used to process the signals in this study. The simulation studies using PSCAD software indicate that the proposed protection scheme is reliable and accurate for all possible types of faults on transmission lines with and without SVC. In addition, the proposed technique presents noise immunity.

A novel approach to transformer fault diagnosis using IDM and naive credal classifier

Transformer fault diagnosis is important for improving the operational reliability of power systems. Despite great efforts to enhance the accuracy of fault diagnosis, the precise detection of transformer faults remains difficult. The major barriers are attributed to the lack of transformer fault records and related data, which may produce unconvincing diagnostic results, and to the fact that multiple latent faults in a transformer are difficult to detect distinctly. In this paper, an imprecise probability-based approach to transformer fault diagnosis by using the imprecise Dirichlet model and naive credal classifier is proposed. Instead of providing a single-valued probability, the approach calculates the interval probability for each possible type of transformer faults to explicitly indicate the uncertainty of the diagnosis caused by the insufficiency of historical fault records. Meanwhile, according to whether an overlap exists among the probability intervals of different types of faults, the proposed approach can output either one explicit fault or multiple probable latent faults that occurred in the transformer. In the proposed approach, the imprecise Dirichlet model (IDM) is used to evaluate the imprecise probabilistic relationships between each type of transformer fault and its corresponding symptom attributes. A naive credal classifier (NCC) is established to integrate the IDM estimation results and calculate the interval-valued probabilities of all types of possible faults. Then, the transformer fault diagnosis results can be obtained by the classification criteria of the NCC. The approach can compensate for the defects of conventional methods, including their inability to perform reliable diagnosis with insufficient fault samples or distinctly indicate the multiple latent transformer faults. The accuracy of diagnosis and the efficiency of transformer maintenance can be significantly improved by implementing the proposed approach. The effectiveness of the approach is verified through case studies.

STUDY OF TIME DEPENDENCE AND SPECTRAL COMPOSITION OF THE SIGNAL IN CIRCUIT OF AC ELECTRIC POINT MOTORS

Purpose. The paper is aimed to establish the dependence of changes in the time domain and spectral components of the current in the circuit of the AC electric point motor on its technical condition, to identify the common features for the same type of damage. It is necessary using the analysis of the received signals to carry out the remote diagnosis and determination of faults and defects of electric point motors. In addition it suggested to accelerate the process of the failure, malfunction and damage search. Authors propose the automated approach to the service of remote floor automation equipment, which is located in the envelope of trains. Reduction of the threat to life and health of staff by reducing the residence time in the zone of train movement. Reduce the impact of human factors n the result of service. Methodology. The paper studies the structure, parameters and characteristics, the operation and maintenance characteristics of the AC electric point motors. Determination of the main types of possible faults in the process depending on the operating conditions. Presentation of the electric motor as an object of diagnosis. Findings. The time dependences of the current in the circuit of electric point motor for its various states was obtained. The connection between the technical condition of electric point motor and the performance of current curve in time and spectral domains was established. The revealed deviations from the reference signal were justified. According to the obtained results it was made the conclusion. Originality. A method for diagnosing the state of the AC electric point motor by the time dependence and the spectral composition of the current in its circuit was proposed. The connection diagram to the motor windings based on non-infringement of electric parameters of connection circuit in the actual operating conditions was applied. Practical value. The obtained results suggest the possibility and feasibility of further development of the principles of remote diagnosis based on the analysis of temporal and spectral domains of the current curve in the circuit of the AC electric point motor. The implementation and practical application in operation as a separate system for the collection, analysis, processing and data communication is also real and rational. The given system makes possible tracking the changes in key parameters and forming the proposals on impact measures on the existing situation, as well as the ability to predict the state of the objects on the basis of the dependency determination of the previous changes.

A nonlinear quaternion-based fault-tolerant SINS/GNSS integrated navigation method for autonomous UAVs

The fault-tolerant integrated navigation methods using strapdown inertial navigation system (SINS) and global navigation satellite system (GNSS) for unmanned aerial vehicles (UAVs) are studied herein to address two types of possible faults, system fault and measurement fault. Two typical approximation-based nonlinear filters, extended Kalman filter (EKF) and divided difference Kalman filters (DDKFs), are modified to be fault-tolerant for accurate navigation even in the presence of these faults. Criteria are presented to detect the faults, and related covariance matrices of the predicted errors are enlarged using fault accommodation factors to make the faulty state or measurement components be normally treated. Navigation simulations using the nonlinear quaternion-based navigation error model are given to demonstrate the effectiveness of the proposed methods.

A random maintenance scheduling model to reduce fault diagnosis time

We present a random model for a situation in which some tests of fault detection are to be scheduled to diagnose what type of fault, out of some possible types of faults, has occurred. There are two variants of the model. In the first, the objective is average total diagnosis time. In the second, the objective is a linear combination of average and standard deviation of the diagnosis time, where standard deviation is multiplied with a positive weight. We give an exact solution method for the first case and a heuristic method for the second. A numerical experiment with randomly generated instances is done for the heuristic method. The methods appear to be suitable for practical applications.

A New Technique to Detect Faults in De-Energized Distribution Feeders—Part I: Scheme and Asymmetrical Fault Detection

Re-energizing an overhead distribution feeder safely is a major consideration for a utility's safe work practice. One way to improve the safety is to determine whether the feeder still experiences short circuits before it is energized. In this paper, a novel fault detection technique is proposed to detect if a de-energized distribution system still experiences short-circuit faults. The proposed method involves injecting a thyristor-generated-controllable signal into the de-energized feeder. The feeder voltage and current responses are analyzed to determine if a fault still exists. A thyristor gating control strategy and fault detection algorithm are also developed in this paper to detect all possible types of faults that can occur in a system. The effectiveness of the proposed method has been verified through theoretical analysis, computer simulations, and lab tests.

Classification and location of single line to ground faults in double circuit transmission lines using artificial neural networks

Fault detection, classification and location from the relay end in a double circuit transmission line are challenging tasks because of mutual coupling between the two circuits. In this paper, combined unsupervised and supervised neural networks-based fault detection, classification and distance location techniques are presented for a double circuit line. This technique does not require communication link to retrieve the remote end data. Zero sequence current compensation for healthy phases can also be avoided. Artificial neural network employs a reduced set of input features, i.e., the fundamental components of three phase voltages and the six phase currents of the two parallel lines at one end of the line only. The simulation results show that single phase-to-ground faults can be correctly classified and located after one cycle from the inception of fault. The complexity of the possible types of faults, fault locations, high path fault resistances, fault inception angles, mutual coupling effects and remote end in-feed do not affect the performance.

Fault diagnosis of pneumatic systems with artificial neural network algorithms

Pneumatic systems repeat the identical programmed sequence during their operation. The data was collected when the pneumatic system worked perfectly and had some faults including empty magazine, zero vacuum, inappropriate material, no pressure, closed manual pressure valve, missing drilling stroke, poorly located material, not vacuuming the material and low air pressure. The signals of eight sensors were collected during the entire sequence and the 24 most descriptive features of the data were encoded to present to the ANNs. A synthetic data generation process was proposed to train and test the ANNs better when signals are extremely repetitive from one sequence to other. Two artificial neural networks (ANN) were used for interpretation of the encoded signals. The tested ANNs were Adaptive Resonance Theory 2 (ART2), and Back propagation (Bp). ART2 correctly distinguished the perfect and faulty operations at all the tested vigilance values. It classified 11 faulty and 1 normal modes in seven or eight categories at the best vigilance values. Bp also distinguished perfect and faulty operations without even the slightest uncertainty. In less than 10 cases, it had difficulty identifying the 11 types of possible faults. The average estimation error of the Bp was better than 2.1% of the output range on the test data which was created by deviating the encoded values. The ART2 and Bp performance was found excellent with the proposed encoding and synthetic data generation procedures for extremely repetitive sequential data.

214 POSTER Characterization of a recombinant, fully human monoclonal antibody directed against the human insulin-like growth factor-1 receptor

The fault-tolerant integrated navigation methods using strapdown inertial navigation system (SINS) and global navigation satellite system (GNSS) for unmanned aerial vehicles (UAVs) are studied herein to address two types of possible faults, system fault and measurement fault. Two typical approximation-based nonlinear filters, extended Kalman filter (EKF) and divided difference Kalman filters (DDKFs), are modified to be fault-tolerant for accurate navigation even in the presence of these faults. Criteria are presented to detect the faults, and related covariance matrices of the predicted errors are enlarged using fault accommodation factors to make the faulty state or measurement components be normally treated. Navigation simulations using the nonlinear quaternion-based navigation error model are given to demonstrate the effectiveness of the proposed methods.

A Systematic Stochastic Petri Net Based Methodology for Transformer Fault Diagnosis and Repair Actions

Transformer fault diagnosis and repair is a complex task that includes many possible types of faults and demands special trained personnel. Moreover, the minimization of the time needed for transformer fault diagnosis and repair is an important task for electric utilities, especially in cases where the continuity of supply is crucial. In this paper, Stochastic Petri Nets are used for the simulation of the fault diagnosis process of oil-immersed transformers and the definition of the actions followed to repair the transformer. Transformer fault detection is realized using an integrated safety detector, in case of sealed type transformer that is completely filled with oil, while a Buchholz relay and an oil thermometer are used, in case of transformer with conservator tank. Simulation results for the most common types of transformer faults (overloading, oil leakage, short-circuit and insulation failure) are presented. The proposed Stochastic Petri Net based methodology provides a systematical determination of the sequence of fault diagnosis and repair actions and aims at identifying the transformer fault and estimating the duration for transformer repair.

A novel digital distance relaying technique for transmission line protection

Parallel transmission lines often pose more difficult protection problems than single lines. In the case of parallel lines, a number of problems arise when using the distance protection. The paper discusses problems associated with parallel line distance relaying schemes and presents a novel technique to overcome these problems. Two relays instead of four are proposed for the double lines, One relay is located at the beginning and another one at the end. Each relay is fed by three voltage and six current signals from the two lines. The suggested technique is based on the comparison of the measured impedance of corresponding phases. So, the complexity of the possible types of faults, high path fault resistance, mutual effects, current in-feed, inter-system faults are solved. Moreover, 100% of line is protected and the problem of balance-point locations is avoided. Alternative Transient Program models the power system and simulates different fault conditions.

A Novel Digital Distance Relaying Technique for Transmission Line Protection

Parallel transmission lines often pose more difficult protection problems than single lines. In the case of parallel lines, a number of problems arise when using the distance protection. This paper discusses problems associated with parallel line distance relaying schemes and presents a novel technique to overcome these problems. Two relays instead of four are proposed for the double lines. One relay is located at the beginning and another one at the end. Each relay is fed by three voltage and six current signals from the two lines. The suggested technique is based on the comparison of the measured impedance of corresponding phases. So, the complexity of the possible types of faults, high path fault resistance, mutual effects, current in-feed, inter-system faults are solved. Moreover, 100% of line is protected and the problem of balance-point locations is avoided. Alternative transient program models the power system and simulates different fault conditions.

Fundamental Basis for Distance Relaying on 3-Phase Systems

In the preceding investigation the quantities upon which distance relays must depend for their operation have been determined for all possible types of faults occurring at a single location. It has been shown that the quantities obtained by the use of earlier distance relay connections for both line and ground faults do not give the closest approach to the desired result, a constant indication for all types of faults at a given location. It has been shown further that, by the use of different connections and certain auxiliary apparatus, a constant relaying indication may be obtained by the line relays for all types of line faults and by the ground relays for all types of faults involving ground, if the fault resistance is zero, and, if the fault resistance is not zero, that smaller deviations from the normal value will be obtained for a given fault resistance. The analysis has been based upon the assumptions of sine wave currents and voltages, constant circuit impedances, and negligible distributed capacitance in the faulted sections of line. It has been assumed further that the fault impedance is resistance only and that it is effectively constant throughout each cycle. All of these assumptions, except perhaps the latter, are reasonably valid as far as relaying is concerned, although some consideration must be given to the d-c component of the short-circuit currents with very high-speed relays.