Abstract
The research is aimed at solving the strategic task of reducing the energy intensity of a metallurgical enterprise by developing a universal, verifiable hybrid mathematical model. The model integrates energy balance, machine learning, and optimization programming techniques to accurately account for technological, stochastic, and operational factors. Practical testing was carried out on the data of the Aksu ferroalloy plant with a focus on ore-thermal furnaces). The result was a prototype of a digital decision support platform capable of generating operational recommendations to reduce specific energy consumption and identify hidden losses. The purpose of the research is to develop, test and implement a hybrid mathematical model for optimizing and predicting electricity consumption in key ferroalloy production units (ore-thermal furnaces). The novelty of the proposed approach lies in the integration of the following analytical components into a single structure: energy balance, optimization equations, regression models. The results obtained are intended for use in the following digital enterprise systems: digital energy monitoring systems, Energy Management class systems, Digital counterparts of production processes, Energy efficiency improvement Programs. The model provides an objective basis for the formation of energy saving programs, providing a confirmed reduction in specific electricity consumption by 5%, respectively, reducing the cost of production. Keywords: mathematical model, energy efficiency, digital platform, program, forecasting, machine learning,metallurgical enterprise.