Heat Transfer Coefficient Optimization Using Artificial Intelligence Algorithms: Accuracy and Computational Efficiency Analysis

Abstract

This paper presents the results of a study on the reconstruction of the heat transfer coefficient under boundary conditions of the fourth kind, which is of key importance for casting processes. Understanding and optimizing this coefficient has a direct impact on the efficiency and quality of metallurgical production processes, which can contribute to significant material savings when reconstructing casting conditions. Using swarm algorithms, such as bee and ant algorithms, to estimate the coefficient is an innovative approach to the problem. These artificial intelligence methods are known for their efficiency in solving complex optimization problems, demonstrating the potential of implementing modern technology in traditional industries. The research includes a detailed analysis of how different levels of noise (0%, 1%, 3%, 6%) and algorithm parameters, i.e., the number of individuals (20, 40, 60) or the number of iterations (10, 14, 20), affect the accuracy of the results. This analysis improves our understanding of the impact of the aforementioned variables on the results and enables us to optimize them for improved accuracy and efficiency. Running the simulation six times for each iteration increases the reliability of the results, as it allows for estimating the variance and confidence in the results. This is an important aspect of scientific research that ensures the robustness and reproducibility of the results obtained. The study's findings provide concrete guidance for engineers and scientists involved in modeling thermal processes, which can lead to improved design and management of foundry processes. In this field, the application of artificial intelligence opens up new opportunities for innovation and improvement. The paper's authors raised important questions about the risks associated with increasing the number of iterations or individuals in a population. This is important for the practical application of results in real industrial settings, where computational and time resources are often limited. The results presented in the article provide important insights for engineers and scientists involved in modeling thermal processes in foundry operations, offering a new perspective on the use of advanced artificial intelligence techniques. Modern technologies and research methodologies can help the metallurgical industry advance technologically and economically.