Parameterization of the statistical model for electrical energy efficiency control

Abstract

Purpose. Justification of a structural construction and parameters of a regression model for the normalization of specific energy consumption when controlling the production process energy efficiency. Methodology. Analysis of the peculiarities of energy efficiency control of the production process in conditions of frequent and significant changes in specific energy consumption, followed by the determination of the structure and parameters of the regression model. Findings. Based on the presence of frequent and significant changes in the energy efficiency control of the production process, the reasonableness of normalizing the specific energy consumption by using the regression model with a variable structure is substantiated. The actual daily specific energy consumption indicators, obtained during the month to control energy consumption efficiency and build the regression model of the variable structure, are used. The limited possibilities for the formation of voluminous statistical samples with homogeneous data, and the complexity and laboriousness of measuring a significant number of influence parameters make it necessary to reduce the number of explanatory variables of the regression model. The feasibility of using the value of the output volume, as a comprehensive characteristic of the level of energy consumption, is proven. The acceptability of the application of linear and non-linear univariate regression dependencies is determined. The nonlinear model, as a result of reducing the linear model of energy consumption to a nonlinear form characteristic of the values of its specific consumption, is obtained. Originality. For the first time, the use of the regression model of the variable structure for the normalization of specific energy consumption in conditions of frequent and significant changes in the energy efficiency of the production process, which helps to increase the accuracy of their determination, is proposed. The need to reduce the number of explanatory variables of the regression model is proven. The expediency of using linear or non-linear one-factor regression dependencies in the given conditions of energy efficiency control, which helps to simplify the procedure of registering the initial data for their construction, is confirmed. Practical value. The scientific results of the performed studies allow for taking into account the peculiarities of the production conditions when determining the structure and parameters of the regression model for normalizing the specific energy consumption. This contributes to increasing the accuracy and energy efficiency control of the production process.