Data-Driven Digital Twin for Foundry Production Process: Facilitating Best Practice Operations Investigation and Impact Analysis

In the context of increasing environmental concerns, the iron and steel industry faces large pressure to reduce its energy consumption and carbon footprint while maintaining economic viability. This paper explores the implementation of best practice operations within foundry processes, specifically induction furnace melting, to enhance energy and cost efficiency and reduce CO₂ emissions. A digital twin model is developed integrating discrete event simulation, system dynamics modeling, and symbolic regression to simulate the foundry production process and evaluate the impact of various operational practices. A large Danish foundry is used as a case study, providing data for induction furnace production incorporating various electricity market data sources. Symbolic regression models are deployed to accurately predict melt temperatures and energy requirements. Results indicate that adopting best practices can lead to significant savings - up to 21% in electricity costs and 14.2% in CO₂ emissions - while improving productivity. The study also highlights a point of diminishing returns at 65% adherence to best practices due to existing production schedules. Furthermore, the study demonstrates the digital twin’s potential as a decision-support tool in optimizing industrial process operations.