جستجوی مقالات مرتبط با کلیدواژه « fault detection » در نشریات گروه « فنی و مهندسی »

The design of a Spotted Hyena Optimization Algorithm-Variable Parameter Tilt Integral Derivative with Filter (SHO-VPTIDF) controller for improved performance and enhanced devaluation of harmonic components of grid-connected photovoltaic systems is the main objective of the suggested manuscript. The SHO-VPTIDF controller is proposed by reformulating Tilt Integral Derivative Controller with Filter (TIDCF). The TIDCF is characterized by longer simulation time, lower robustness, longer settling time, attenuated ability for noise rejection, and limited use. This research gap is addressed by replacing the constant gains of TIDCF by variable parameter tilt integral derivative with filter. The VPTIDF replaces the constant gains of TIDCF with error varying control parameters to improve the robustness of the system. The photovoltaic system with nonlinearities causes power quality issues and occasional faults, which can be detected by using Levenberg-Marquardt Algorithm (LMA) based machine learning technique. The novelties of the proposed manuscript including improved stability, better robustness, upgraded accuracy, better harmonic mitigation ability, and improved ability to handle uncertainties are verified in a Matlab simulink environment. In this manuscript, the SHO-VPTIDF and the Direct and Quadrature Control based Sinusoidal Pulse Width Modulation (DQCSPWM) method are employed for fault classification, harmonic diminishing, stability enhancement, better system performance, better accuracy, improved robustness, and better capabilities to handle system uncertainties.

In recent years, due to the widespread applications of DC power-based appliances, the researchers attention to the adoption of DC microgrids are continuously increasing. Nevertheless, protection of the DC microgrid is still a major challenge due to a number of protection issues, such as pole-to-ground and pole-to-pole faults, absence of a zero crossing signal, magnitude of the fault current during grid-connected and islanded mode, bidirectional behaviour of converters, and failure of the converters due to enormous electrical stress in the converter switches which are integrated in the microgrid.  Failure of the converter switches can interrupt the charging of the electrical vehicles in the charging stations which can affect transportation facilities. In addition to the above mentioned issues protection of the DC microgrid is more challenging when fault parameters are varying due to dissimilar grounding conditions and varying operational dynamics of the renewable sources of energy. Motivated by the above challenges a support vector machine and ensemble of k-nearest neighbor based protection scheme has been proposed in this paper to accurately detect and classify faults under both of the modes of operation. Results in the section 5 indicate that performance of the protection scheme is greater as compared to other algorithms.

Detecting and rectifying spark plug faults are pivotal in preventing engine-related issues that can have substantial operational and financial consequences. To improve the accuracy and robustness of spark plug fault diagnosis, this research introduces a novel Convolutional Transformer approach that leverages the strengths of Convolutional Neural Networks and Transformers, which effectively capture both local and extended temporal dependencies within spark plug acoustic signals. The results of this groundbreaking approach, as presented in accompanying tables and figures, demonstrate its superior performance, achieving an impressive 97.1% accuracy in a challenging 4-class classification scenario using solely acoustic signals. This achievement signifies a significant advancement in spark plug fault detection, potentially ushering in more reliable and precise diagnostic methods, ultimately contributing to the prevention of costly engine breakdowns and the extension of engine lifespan. Deep learning techniques such as Convolutional Transformers offer a promising way to improve the reliability and performance of internal combustion engines as the automotive industry continues to evolve, highlighting the importance of this research for future automotive developments.

Wound rotor induction machine (WRIM) has been extensively used in different applications such as medium-power wind turbines and traction systems. Since these machines work under harsh and difficult conditions, condition monitoring of such systems is crucial. Different electrical and mechanical signatures of machines were used for electrical and mechanical fault detection in electrical machines such as vibration, acoustic emission, stray flux, and stator current signature. In recent years, stator current signature analysis due to simplicity, cost-effectiveness, and availability has been considered for fault detection process in comparison with previous conventional methods such as acoustic and vibration. In this paper, a high-resolution technique based on the chirp-Z transform is used for rotor asymmetry fault (RAF) detection in induction machines through stator current signature analysis. In this regard, the Teager-Kaiser energy operator (TKEO) technique for demodulation fault characteristic frequency is used as a pre-processing stage to avoid leakage of the supply frequency. The method has better accuracy due to better spectral resolution and resolvability. Furthermore, computational complexity in the proposed method will be reduced in comparison to the previous conventional ones which have used the Fast Fourier transform (FFT). The proposed technique is tested through synthetic and experimental stator current of WRIM in healthy and faulty conditions with different rotational speeds and fault severities. The results show the validity of the proposed method in rotor asymmetry fault detection through the stator current signature of WRIM.

Fault detection and its restoration is the major challenge for the smooth functioning of the Multi-Level Inverter (MLI). In this paper, fault detection and its clearance scheme for an Open Circuit (OC) fault on a 3-level 5-level Cascaded H-Bridge Multi-Level Inverter (CHBMLI) has been developed and tested to improve the reliability and suitability of the system. An accurate and fast detection, isolation and bypassing of faulty bridges enhance the reliability, suitability, and acceptability of CHBMLI in any domestic, industrial drive applications. To reschedule the line voltage and current value close to the pre-fault level, a Neutral Point Shift (NPS) technique is presented in this paper. The desired output voltage is governed by Level Shift Pulse Width Modulation (LSPWM) technique. The proposed scheme is developed in MATLAB/Simulink environment and results are validated by using Opal-RT simulator. Simulation results has confirmed the performance and Opal-RT simulator results shows feasibility and applicability of the proposed scheme.

In spite of the numerous benefits over the traditional power distribution system, protection of the microgrid is a challenging and complex task. The varying fault resistances due to dissimilar grounding conditions can affect the performance of the protection scheme. Under such conditions, the magnitude of the fault current can vary from lower to higher level. In addition to the above, the dissimilar magnitude of fault current during grid connected and islanded mode demands a protection scheme that can easily discriminate the mode of operation. The magnitude of fault current in grid-connected and islanded modes needs a robust protection scheme. In this regard, an ensemble of subspace kNN based robust protection scheme has been proposed to detect the faulty conditions of the microgrid. The tasks of the mode detection, fault detection/classification as well as faulty line identification has been carried out in the proposed work. In the proposed protection scheme, discrete wavelet transform (DWT) has been used for processing of the data. After recording the voltage and current signals at bus-1, the protection scheme has been validated. The validation of the protection scheme in Section 6 reveals that the protection scheme is efficiently working.

Symmetrical nature of mean of electrical signals during normal operating conditions is used in the fault detection task for dependable, robust, and simple fault detector implementation is presented in this work. Every fourth cycle of the instantaneous current signal, the mean is computed and carried into the next cycle to discover nonlinearities in the signal. A fault detection task is completed using a comparison of two sub cycle means, and the same concept is extended to faulty phase classification. Under various fault and system operating situations, the suggested technique is assessed for regular faults, remote end faults, high resistive faults, and high impedance arcing faults. This paper's extensive case studies illustrate the suggested scheme's simplicity, computational flexibility, speed, and reliability. The suggested approach yields 100% consistent results in 4-8 msec detection time.