Enhancing operational efficiency in energy-intensive industries through AI: a case study of electricity demand forecasting in foundries

Accurate electricity demand forecasts are crucial for energy-intensive industries that procure electricity via energy trading markets, as they minimize procurement risks and unplanned costs. However, in manufacturing companies, the dependencies of electricity demand on various factors are often unclear, making accurate forecasting challenging. This paper introduces a data-driven approach for predicting monthly electricity demand using historical consumption data, production planning parameters, and production output. Artificial Intelligence (AI) offers various methods for predictive modeling that can be applied to electricity demand forecasting. Following a literature review of existing forecast models, a framework for training and validating AI models is proposed. To validate this approach, a case study is conducted on a foundry’s electricity demand. The study includes a thorough industry-specific analysis, examining electricity consumption, and preprocessing relevant data. Identified models are then trained and evaluated in the context of the foundry’s operational energy demand planning. Comparative analysis of the models provides insights into their performance, operational suitability, key predictor variables, and production planning parameters. The findings aid in selecting optimal electricity demand prediction models and offer insights applicable to similar challenges in other industries, enhancing operational efficiency and cost management.