Session: 01-05: AI for Energy Sustainability V

Paper Number: 132526

132526 - Applications of Ai Methods in Cooling and Energy Management of Smart Data Centers 

Abstract: 

Abstract

Moving toward the digital world boosts the number of data centers all over the globe which intensifies the issue of cooling server rooms. Cooling server rooms is crucial for maintaining the optimal operation, reliability, and longevity of IT equipment while assuring data security. Proper cooling solutions should be carefully planned and implemented to create a stable and efficient environment for critical IT operations. The conventional heating ventilation and cooling (HVAC) systems are not adequate cooling systems for the server rooms. Advancing the conventional HVAC systems specifically for the data centers was the first step in cooling them. The new research ideas were developed by integrating different cooling systems and implementing optimization techniques. These days Artificial Intelligence (AI) techniques have drastically enhanced the technology and so can play a significant role in cooling server rooms by advancing the cooling systems and ensuring the reliable operation of IT infrastructure. Implementing the artificial neural network (ANN) methods to train the hardware and cooling systems is the comprehensive method for cooling the server rooms with minimum energy consumption. The present study presents the most recent investigation into the cooling and energy management of smart data centers through smart HVAC, smart server room control, and energy consumption optimization by servers. Smart server rooms consider managing data centers’ schedules as well as managing data center cooling by dynamic cooling control, environmental monitoring and prediction, adaptive vent, thermal mapping and zone control, and data-driven cooling. AI can integrate with the overall building management system and HVAC infrastructure to coordinate cooling efforts across the entire facility, ensuring efficient and coordinated cooling operations. The present study shows various approaches for cooling the server rooms by implementing ANN techniques through a comprehensive review of past studies. AI's capabilities are in monitoring, analyzing, predicting, and adapting to dynamic server room conditions, data center operators leading to higher cooling efficiency, reducing energy consumption, extending hardware lifespans, and enhancing overall server room reliability. These benefits contribute to more sustainable and cost-effective data center operations, which are essential in our increasingly digital and connected world. Results show AI has a great contribution to predicting performance in the server rooms by predicting the temperature and load of the server room. By knowing the load and temperature, they can be managed properly. Again for managing and balancing load and temperature a variety of AI techniques, which are presented in this study, have been employed to control the server rooms.

Presenting Author: Mychal Berlinski University of Pittsburgh, Bradford

Presenting Author Biography: Mychal is a 22 year-old undergraduate Energy Engineering Technology Student at the University of Pittsburgh at Bradford in Pennsylvania. Over the summer of 2023, Mychal participated in a new research program at his university and focused on artificial intelligence and its applications to more energy efficient cooling systems, specifically for a server room setting.

Authors:

Mychal Berlinski University of Pittsburgh, Bradford
Beth Rezaie University of Pittsburgh at Bradford