3rd ICAI 2024

International Conference on Automotive Industry 2024

Mladá Boleslav, Czech Republic

The use of novel computational methods in forecasting the demand for electrical power ‒ starter battery production case study Marek Karkula 1 , Robert Mazur 2

AGH University of Krakow Faculty of Management 1, 2 ul. Gramatyka 10, Kraków, 30-067 Poland PPUH AUTOPART JACEK BĄK SP. Z O.O Al. Kwiatkowskiego 2A, Mielec, 39-300 Poland 2 e-mail: mkarkula@agh.edu.pl 1 , rmazur@agh.edu.pl 2

Abstract The current conditions of the global economy necessitate rapid decision-making by those managing production processes under conditions of uncertainty. The need to make numerous decisions regarding economic and energy efficiency requires the support of the decision-making process through modern computational methods, which allow for automation and minimization of the impact of risk on business. This publication attempts to formulate a model that supports decision-making in the area of ordering active electrical power capacity. The publication presents a detailed scheme of the proposed solution based on machine learning techniques, highlighting the key variables. The model highlighted in this publication is based on real data from a factory producing starter batteries, where there is significant variability in power consumption over time. Keywords: decision-making support, machine learning, energy efficiency, energy demand, production management JEL Classification: C450, C890, L620, Q410 1. Introduction In a modern dynamic world, companies need to continuously improve their operational strategies in order to remain both competitive and sustainable effectiveness. One of the key aspects of this improvement is the effective management of electricity load, which plays a key role in corporate operations. In the context of climate change, rising energy prices and increasing regulatory requirements, the optimal use of active power is becoming essential for economic sustainability and environmental protection. Recent years show a revival of the topic related to forecasting the value of electricity power used, as indicated by numerous applications such as; ESS (Energy Storage System), DERs (Distributed Energy Resources), DSM (demand side management), and EVs (Electric vehicles)(Malik et al., 2021). The modern approach to the problem of active power forecasting somehow necessitates the use of machine learning tools and artificial

78