Neural network model of energy efficiency of the production process of the enterprise under the conditions of stabilization of electricity outages in Ukraine

DOI: 10.31673/2412-9070.2025.028328

Abstract

The issue of energy efficiency and functional stability of production processes is one of the key challenges for modern industry. Massive shelling of energy infrastructure facilities in Ukraine at the end of 2022 - early 2023 and starting from March 2024 had a significant negative impact on the functioning of industrial enterprises. Power outages have become frequent and prolonged, resulting in significant production shutdowns, reduced productivity, and financial losses in all sectors of the economy. The article analyzes the consequences of stabilization power outages on the ferrous metallurgy
and chemical industry, as well as developed recommendations for increasing the stability of the functioning of production processes in conditions of unstable power supply. A strategy of measures related to ensuring the stability of the enterprise's energy system as one of the key components of ensuring the functional stability of production processes has been developed. The need to increase the profitability of production makes the search for new tools and methods for optimizing energy consumption increasingly urgent. Neural networks, due to their capabilities for learning and prediction, are a powerful tool for analyzing energy consumption and improving the functional sustainability of production processes. The article analyzes modern approaches to the use of neural networks for energy saving at industrial enterprises, which allows to increase the efficiency and reliability of the functioning of production processes. The article examines the process of building the LSTM neural network and how to improve it for the forecast of energy consumption in production processes and further analysis of the efficiency of the use of energy carriers.

Keywords: energy infrastructure, neural networks, information technology, industrial enterprise, production process, LSTM, energy efficiency, model, algorithm.