No abstract

The use of machine learning and data-driven methods for predictive analysis of power systems offers the potential to accurately predict and manage the behavior of these systems by utilizing large volumes of data generated from various sources. These methods have gained significant attention in recent years due to their ability to handle large amounts of data and to make accurate predictions. The importance of these methods gained particular momentum with the recent transformation that the traditional power system underwent as they are morphing into the smart power grids of the future. The transition towards the smart grids that embed the high-renewables electricity systems is challenging, as the generation of electricity from renewable sources is intermittent and fluctuates with weather conditions. This transition is facilitated by the Internet of Energy (IoE) that refers to the integration of advanced digital technologies such as the Internet of Things (IoT), blockchain, and artificial intelligence (AI) into the electricity systems. It has been further enhanced by the digitalization caused by the COVID-19 pandemic that also affected the energy and power sector. Our review paper explores the prospects and challenges of using machine learning and data-driven methods in power systems and provides an overview of the ways in which the predictive analysis for constructing these systems can be applied in order to make them more efficient. The paper begins with the description of the power system and the role of the predictive analysis in power system operations. Next, the paper discusses the use of machine learning and data-driven methods for predictive analysis in power systems, including their benefits and limitations. In addition, the paper reviews the existing literature on this topic and highlights the various methods that have been used for predictive analysis of power systems. Furthermore, it identifies the challenges and opportunities associated with using these methods in power systems. The challenges of using these methods, such as data quality and availability, are also discussed. Finally, the review concludes with a discussion of recommendations for further research on the application of machine learning and data-driven methods for the predictive analysis in the future smart grid-driven power systems powered by the IoE.

The present article studies the approach to building the distributed wireless sensor networks. It shows that, for RFID-based wireless sensor networks, the first thing to determine is the field it is intended to be used for, the far or the near one. The operating frequency value determines the size of the antenna, and the correlation is inverse. The operational contradiction analysis shows, that to increase the operating range of the tags, it makes sense to select the systems with an active power supply. But from the energy efficiency perspective, the passive tags are better to be used, and for increasing the range of data transmission, coordinators that collect and transmit the data to the central unit are preferable. Wireless sensor networks technology is the only wireless technology that can be used for solving some surveillance and control issues, where the time of processing of the sensor readings is critical. United into a wireless network with the suggested method, the sensors make up a spatially distributed self-organized system of collection, processing and transmission of information. A special attention is paid to the structural solution for temperature control in the wireless sensor networks elements. In the conclusion, recommendations on using radio frequency identification technology-based sensor systems are provided.

No abstract