Analysis of factors affecting the electricity consumption of a delivery point cluster default provider

This paper considers data retrieval issues associated with the training and testing neural network algorithms for electric load short-term forecasting based on the analysis of factors affecting the behaviour of the time series of the electric load of default provider consumers. The study uses mathematical-statistical, correlation and factor analysis methods. In the context of district central heating and water supply, the paper considers the influence of time factors, meteorological conditions, reliability of power grid equipment and the operation mode of large electric energy consumers on the electric load of default provider consumers. The main factors having a decisive effect on the time-series behaviour of electric energy consumption are identified. In order to reduce the dimension of the space factor, several factors are combined into one. Issues concerning the sampling of statistical data necessary for short-term forecasting of electric energy consumption using artificial intelligence and machine learning tools are considered. The sample is then used as a basis for the training and testing of predictive models. The behaviour of electric energy consumption time series is mainly determined by temporal and meteorological factors. The following factors are also determining for the electric load time series of the delivery point cluster default provider: blackouts of 6-220 kV electric grid equipment, operation mode of consumers of 670-10000 kW electric energy, availability of central heating and water supply in a certain town or settlement. However, these factors are considered insignificant for the electrical load of the energy system as a whole. Having considered the group of meteorological factors, we found that hourly volumes of power consumption are affected not only by the temperature of the outside air, but also by the variations in the temperature regime having occurred in the previous period.

short-term load forecasting, factors, artificial intelligence, training sample, electric energy delivery point cluster, wholesale electricity market