3rd ICAI 2024

International Conference on Automotive Industry 2024

Mladá Boleslav, Czech Republic

neural networks methods. The traditional engineering approach is increasingly being displaced by the aforementioned techniques which is demonstrated by (Nti et al., 2020) in their work. In the study cited above, the authors indicate that close to 90% of the prediction models are based on AI solutions and 31.34% of the publications were from European countries. There is similar evidence in the analysis of the study of (Rodrigues et al., 2023), which also highlights aspects of model performance evaluation metrics in Short-Term Load Forecasting. The analysis presented here is based on empirical data and scientific literature in the field of energy management and planning using modern prediction methods. By exploring these issues, we aim to provide practical guidance for managers and engineers involved in energy management in enterprises where there is high variability in energy demand over time, so that they can effectively exploit the potential of electricity load planning to improve operational efficiency and achieve corporate strategic goals. In the following sections, the article will discuss the definition and importance of active power in an enterprise context, the benefits of effective active power planning, challenges, and barriers to implementation, selected computational methods and tools to support the planning process. By taking a holistic approach to the analysis of this issue, this article aims to contribute the knowledge of electric power planning in the working shift intervals and inspire further research into the optimal use of active power in enterprises. 2. Problem Formulation and Methodology Forecasting electrical active power demand plays a key role in efficient energy management, enabling businesses to predict demand patterns and make optimal use of resources. However, modern forecasting methods often face challenges in accurately predicting energy loads due to the dynamic nature of energy consumption patterns, which are influenced by factors such as weather fluctuations, socio-economic trends, technological advances, and industrial specification for process maintaining. Consequently, there is a critical need to develop advanced forecasting models that can address these complexities and provide more accurate forecasts without using time series methods. The essence of energy management as presented in ISO 50001, which is an internationally acknowledged standard for managing energy and improving the utilization method of its medium , determines the need to define energy receivers and plan their use over time without compromising the efficiency of the mentioned installations(António da Silva Gonçalves & Mil-Homens dos Santos, 2019). According to(Román-Collado & Economidou, 2021) the key factor of influence during both growth and recession is energy efficiency. The authors also emphasize in their paper that a detailed decomposition analysis is necessary to monitor energy efficiency. However, the costs of such a detailed analysis as well as the resources of the companies somehow limit the possibilities for detailed analyses of energy consumption over time and the impact of the used electrical load at the micro level on the energy distributors. An additional difficulty is the production characteristics of the lead-acid battery industry, from which the data for the model construction was obtained. The production process of the aforementioned batteries consists of numerous steps such as; production of lead oxide, active mass grids, electrodes, electrode seasoning,

79