Artificial Intelligence-Based Prediction of Molten Steel Temperature of Induction Furnace in Computational Fluid Dynamics (CFD) Environment

Abstract

The efficient and precise control of molten steel temperature in an induction furnace is paramount in modern steelmaking processes. This master's thesis presents a novel approach to address this critical challenge that combines the power of Artificial Intelligence (AI) with Computational Fluid Dynamics (CFD) to predict and regulate the molten steel temperature accurately. The primary objective of this study is to develop an AI-based predictive model that can anticipate the molten steel temperature in real-time, enhancing process control, optimizing energy consumption, and ultimately improving product quality. The research methodology encompasses four main stages: data collection and preprocessing, AI model selection, CFD simulation, and model validation. The first phase focuses on creating a comprehensive CFD environment to simulate the induction furnace's thermal behavior, incorporating thermal boundary conditions and leveraging data augmentation techniques to generate a substantial dataset. In the second phase, several AI models are evaluated, and the most suitable one is selected based on performance metrics. The final phase entails training and validating the chosen AI model using the simulated data. The thesis introduces an AI model to address the complexities and dynamics of the induction furnace environment. This model harnesses the power of machine learning algorithms, enabling it to capture intricate patterns and non-linear relationships in the data, leading to more accurate temperature predictions. The study's results showcase the effectiveness and reliability of the proposed AI-based predictive model, demonstrating a notable improvement in temperature prediction accuracy compared to traditional methods. The thesis also explores real-world case studies, validating the model's applicability and efficiency in practical steelmaking scenarios. iv In conclusion, this research successfully establishes a novel approach for predicting molten steel temperature in an induction furnace using AI within a CFD environment. The outcomes offer valuable insights into the steel industry, fostering more intelligent decision-making, reduced energy wastage, and enhanced process control. Additionally, this work sets a foundation for further exploration of AI integration in steelmaking processes and opens avenues for future research in related domains.