Machine learning methods play a key role in solving forecasting challenges in the electric power industry, including those related to electricity consumption. However, improving the efficiency and robustness of machine learning models requires more research besides hyperparameter tuning and architecture optimization. This study aims to develop a model that can predict the daily load schedule with corrections. For this, an ensemble adaptive boosting algorithm was used. The algorithm was implemented using the Scikit-learn machine learning library in the Python programming language. The study focuses on the central power system of Mongolia. Actual energy consumption data and meteorological factors from January 1, 2019, to October 10, 2024 were used as the initial data. The models were tested during the trial operation of the developed system from December 1 to December 10, 2024. The adaptive boosting model was found to enable forecasting of the daily load schedule with an average error of 31.4 MW(2.4
Digital technical guidance documents of electric power industry, especially in the field of relay protection and automation, significantly enhance the development and examination of design documentation. The present paper considers an approach to the development and examination of digital technical guidance documents for relay protection and automation to generate the requirements for design solutions using artificial intelligence (AI).
The present paper describes a method for identifying increased dynamic eccentricity of high-voltage induction motors for power plant auxiliaries. The method is based on spectral analysis of the external magnetic field during motor run-down. Diagnostic signs of increased dynamic eccentricity that can be detected in the spectrum of the external magnetic field are presented. A method for signal processing that ensures sufficient reliability of results is proposed; conditions for method application are described. The method is tested on simulation models of high-voltage induction motors, as well as on several real electrical machines. The obtained test results confirming the possibility of determining increased eccentricity during operation are consistent with other studies conducted on induction motors.
The article examines the specific features of conducting laboratory experiments aimed at estimating overflow and splash discharge per unit width under the impact of design waves on port hydraulic structures (PHS) in conditions of sea level rise due to global climate change. The analysis of the findings reported in climatological studies enabled us to determine that for ports located in the Arctic and Far Eastern regions of the Russian Federation, sea level rise may exceed one meter by the end of the estimated service life of the facilities under the SSP3-7.0 and SSP5-8.5 socioeconomic scenarios used in the sixth assessment report of the Intergovernmental Panel on Climate Change (IPCC).We have developed and tested a new technique for conducting laboratory experiments. It enables a quantitative evaluation of the sea level rise impact on the PHS operating conditions.
The article describes a fundamentally new resource-saving technology in the form of a location method and equipment for monitoring damage and ice on the wires of overhead power lines with a voltage of 35 – 330 kV, first implemented at substations in Russia and having no analogues in the world. The article shows a technique for identifying reflected pulses among high-frequency communication interference, as well as methods for identifying the most accident blackspot of the line under icy conditions. Examples of manufactured hardware and software location complexes for monitoring overhead power lines are given. The scheme of placement of six location complexes in Russia with control of their operability from the Monitoring Center located in Kazan (KSPEU) is analyzed. The advantages of location monitoring of overhead power lines over currently monitoring methods are shown.
