Research on the Application Effects of AI Technology in Data Center Energy Saving

Current data centers face urgent demands for energy conservation and efficiency improvements. Integrating AI energy-saving systems into the data center infrastructure allows for the prediction of operational trends and enables second-level adjustments to the return air temperature setpoints of the air conditioning units. This proactive regulation of the AC equipment's operational state results in real-time alignment of cooling capacity with the IT power consumption trends. The study selected an existing air-cooled micro-module data center and a chilled water terminal air-conditioned data center. After over six months of application, a substantial amount of data related to IT power consumption, air conditioning power consumption, and data center temperatures were collected. Comparative analyses were conducted on the energy consumption and PUE under conditions with and without AI technology during the operation of the data centers. The results indicate that AI technology can effectively reduce the energy consumption and PUE of the data centers. AI energy-saving technology not only demonstrates efficacy in existing data centers but also holds significant potential for energy savings in newly constructed communication and data centers, offering high cost-effectiveness and considerable promotion value.