* If you want to update the article please login/register
Due to low maintenance costs, reduced transmission line losses, and network congestion, as well as minimal effect on climate change and global warming, distributed generation has gained a lot of attention. However, the distributed network that connected to the distribution network also poses a variety of security issues that cannot be addressed by conventional protection methods. The bi-directional power flow and the change in the fault current level during the topology change are two of these challenges. This paper provides a comprehensive review of the distributed network fault detection and protection techniques that have been embedded with distributed generation. This report also investigates the various fault detection techniques concerning types, communication protocols, operating mode, constraints, and benefits. The proposed research will hopefully be helpful to power system designers and researchers in the creation of distributed network fault and security architectures for the development of distributed network fault and security technologies for the installation and administration of new distributed generation systems.
Source link: https://doi.org/10.1109/ACCESS.2021.3121087
Flaring has always been an important component of oil production and exploration. Previously, waste gas collected from various areas of the facility was removed for safety or operational reasons, and combusted on top of a flare stack, but there was no way to treat or use this type of gas. The introduction of flare gas recovery equipment that have become increasingly popular have made treating flare gas more effective. Most solutions include a flare gas recovery system to an existing flare system. In this paper two designs of a gas control system are considered, with reliability chosen as the deciding factor. Following, a novel hybrid approach, the Stochastic Hybrid Tree Automaton, is used to simulate the operational conditions in which the system operates, with the intention of achieving a more realistic analysis and estimating the risk attributed to these failures.
Source link: https://doi.org/10.1109/ACCESS.2021.3069807
Transformers are one of the primary components of the power system, capable of transmitting and distributing the electricity produced by renewable energy sources. Dissolved Gas Analysis is one of the most common methods to detect early failures in oil-immersed transformers. Many researchers began AI algorithms to extract the attributes of DGA data based on the strong data mining capability and high availability of AI methods. Researchers select appropriate AI techniques or make necessary improvements to AI techniques to improve diagnostic results based on the DGA's characteristics and scale, which can vary with DGA's characteristics and scale. In addition to investigating the applications of these sophisticated technologies, the analysis of diagnostic methods in this journal, including the introduction of temporal parameters for comprehensive analysis of DGA results and the extraction of optimal features for DGA results, is also reviewed. This paper reviews and prospects the artificial intelligence techniques used by researchers in transformer fault diagnosis, as well as transformer fault diagnosis.
Source link: https://doi.org/10.3389/fenrg.2022.1006474
The Adaboost algorithm's weak classifier is used by the BP neural network, and many poor classifiers are made from a strong classifier with better classification capability to identify fault categories. The accuracy of classification is investigated; the algorithm's findings are analyzed by Matlab software; and the analysis results indicate that the updated BP-Adaboost algorithm has a decent classification result for multi-class aviation cable fault diagnosis.
Source link: https://doi.org/10.1177/16878132221125762
Artificial intelligence has a great deal of potential in intelligent grids. By compiling the most recent literature on insulator detection, three common application scenarios and research findings have been identified: the need for improved detection accuracy and real-time speed; inadequate image detection of intricate backgrounds and target occlusion; and multiscale object and small object detection advancements were discovered. At the same time, the best algorithms in the literature are comprehensively summarized, and performance evaluation indices of various algorithms are compared.
Source link: https://doi.org/10.3389/fenrg.2022.912453
This paper reviews the use of a fault diagnostics device for Switch-Mode Non-Isolated DC-DC converters, which is based on a questionnaire developed by this research. The proposed device simulates the operation of a Type II controller for a buck converter under various operating conditions and introduces the FDT simultaneously. The concepts of symmetry and asymmetry are implicit in the signal processing techniques used in the creation of the third module's FDTs. An algorithm for the construction of both the power section and the converter compensator is included in addition to the standard device.
Source link: https://doi.org/10.3390/sym14091886
Traditional protection schemes are difficult to address due to the increasing integration of distributed generations in distributed networks, their growth and operations are facing security challenges that traditional protection systems are unable to address. HHT is used first to extract energy information from the current signal. To separate the intrinsic mode function from the current signal's zero component, second variational mode decomposition is used. Then, the acquired energy feature, and intrinsic mode function are all contributing to the ensemble learning algorithm for fault detection and classification. The reliability of the ensemble bagged trees method is higher than those of the narrow-neural network, fine tree, quadratic SVM, fine-gaussian SVM, and wide-neural network. With or without DGs, the recommended method can identify and classify errors accurately in DN.
Source link: https://doi.org/10.3390/su141811749
A machine learning algorithm is used in this paper to develop a prediction scheme that captures real-time measurements of sensor nodes in a clinical setting. Using various technologies such as current sensors, a temperature and humidity sensor, an air quality sensor, ultrasonic sensor, and flame sensor, an IoT-based smart hospital environment has been created that controls and monitors appliances over the Internet. According to the provided statistics, the proposed fault prediction model was tested via decision tree, K-nearest neighbor, Gaussian naive Bayes, and random forest methods, but random forest had the highest accuracy over others. The results showed that the ML algorithms used over IoT-based sensors are very effective in monitoring this hospital automation process, and random forest was rated as the best with the highest accuracy of 94. 2 percent.
Source link: https://doi.org/10.3390/su141811667
Fault tree analysis is one of the most common analysis techniques used in various industries to determine and improve the reliability and safety of complex systems. From the SCIE and SSCI databases, 1469 FTA-related articles from the literature were retrieved to help understand the current state and growth trend of FTA study and to highlight FTA's future development directions. Informetric methods, such as co-authorship analysis, co-citation analysis, and co-occurrence analysis, were all adopted for investigating the collaboration between the FTA research community, literature base, research hotspots, and frontiers. The knowledge bases for FTA studies include dynamic fault tree analysis, fuzzy fault tree analysis, and FTA based on binary decision diagrams. Both the key research hotspots and frontiers can be quantitative fault tree analysis, dynamic fault tree analysis based on Bayesian networks and FTA, as well as management factors.
Source link: https://doi.org/10.3390/su141811430
First, an effective feature selection algorithm based on particle swarm optimization is suggested. The primary motivation for the use of the PSO algorithm is to delete unnecessary information and extract only the most relevant ones from raw data in order to improve classification task by using a neural networks classification algorithm. An improved PSO algorithm is suggested by Then, to solve the issue of premature convergence and local suboptimal areas when using the classical PSO optimization algorithm.
Source link: https://doi.org/10.3390/su141811195
* Please keep in mind that all text is summarized by machine, we do not bear any responsibility, and you should always check original source before taking any actions