Improving the accuracy the forecasting volumes of electric grid construction in energy systems based on economic and statistical dynamic models

Abstract

The article deals with the issues related to increasing the accuracy of forecasting the volume of power grid construction (the total length of power lines and transformer substation capacity) for power systems A set of parameters that affects the volume of electric grid construction and reflects the level of development and structure of power systems has been identified. The parameters of a higher level of the hierarchy of forecasting the development of electric grids are selected, and, therefore, having more stable characteristics of changes in time in comparison with electric grid indicators. To consider the set of parameters in detail, a special method of multivariate statistical analysis – factor analysis, which allows relating the electric grid indicators and a large set of influencing parameters, was investigated. The novelty of the work lies in the proposed method, which allows to develop multifactor dynamic models of forecasting of power grid indicators (total length of power lines and power of transformer substations), in which the coefficients are functions of time and are based on extensive information on the development of electric grids, as well as including an additional independent variable that considers the previous value of the modeled indicator (autoregressive component). These models can be called dynamic factor-autoregressive ones, and they are more effective than the methods of direct optimization of network configuration, geometric modeling and extrapolation due to the lack of reliable information on the specific territorial characteristics of newly introduced electric grids for the long term (more than 5–7 years).