Search

This study is devoted to the development of a logical system to improve the efficiency of wind turbines with the identification of patterns and relationships between wind speeds and directions in urban conditions. The results of the study made it possible to identify approaches to use wind energy efficiently to produce electricity with minimal losses for the power supply of urban facilities. Data Mining methods were applied, the dependencies between weather data were examined. Machine Learning algorithms have also been applied to forecast wind speed and direction in order to increase the efficiency of power generation. A thorough exploratory data analysis was carried out, including visualization of meteorological data and the study of their statistical indicators. To predict wind speed, numerical indicators such as air temperature, wind direction, pressure tendency and atmospheric pressure were taken. Linear Regression, Decision Tree and Support Vector Machine were taken as Machine Learning models. The results revealed that the Random Forest turned out to be the most effective, with a mean squared error of 0.302. According to the results of the research, it was found that the initial data on the characteristics of the wind are highly distorted when it circulates in the urban environment due to the influence of urban development. However, the use of a logical system made it possible to predict and adapt the operation of wind turbines to changes in wind characteristics with minimal losses. A pivotal outcome of this study is the creation of a real-time data processing system, enabling accurate predictions of wind characteristics from speed and direction measuring devices. Additionally, a MATLAB Simulink-based computer model was developed to explore the logical system's influence on wind turbine performance. The findings underscore the efficacy of employing forecasting within a logical system framework to harness wind energy efficiently, contributing to sustainable urban energy solutions. [ABSTRACT FROM AUTHOR]