Authors: Girish Kishor Ingavale
Abstract: The exponential growth in computational power and industrial processes has led to an increased demand for efficient cooling solutions in data centers. Traditional air-cooling systems are becoming inadequate due to their limitations in managing high thermal loads and their high energy consumption. In response to these challenges, Direct-to-Chip Liquid Cooling Systems (D2C LCS) have emerged as a promising alternative for thermal management in high-density computing environments. This study investigates the application of Artificial Intelligence (AI) and Machine Learning (ML) technologies to optimize the performance of D2C LCS in rack-mounted data center setups. The primary objective of this research is to develop and implement AI-driven models that can predict temperature and fluid flow within D2C LCS, thereby enabling the optimization of cooling strategies. By leveraging advanced algorithms such as Linear Regression and Support Vector Machine, the study aims to enhance thermal efficiency and reduce the energy consumption of data centers. Experimental data was collected from a simulated data center environment equipped with D2C LCS. The data was used to train and validate ML models, ensuring their accuracy and reliability in real-world applications. The results demonstrate that AI-optimized cooling strategies can achieve a 15% reduction in temperature and a 20% decrease in energy consumption compared to traditional air-cooling systems. The findings of this study highlight the significant benefits of integrating AI and ML technologies with D2C LCS for thermal management in data centers. The predictive models and optimized cooling strategies presented herein provide a robust framework for improving the efficiency and sustainability of data center operations. Future research directions include the development of more advanced AI models and the implementation of real-time monitoring systems to further enhance the performance of D2C LCS.
DOI: http://doi.org/10.5281/zenodo.18125704
