Numerical simulation of a precision air conditioner under variable working conditions

Abstract

Artificial intelligence (AI)-based intelligent working parameter setting of cooling equipment is a promising strategy for energy savings in large data centers. For this purpose, the performance curves of each piece of equipment must be obtained under variable working conditions. In the present study, a porous media computational fluid dynamics (CFD) model was established to investigate the flow and thermal performance of a finned-tube heat exchanger in a precision air conditioner (PAC) of a large data center with 3000 server racks. The temperature and velocity at the air outlet of the PAC were measured at 15 measurement points, and the results agreed well with the simulation results with a difference less than 10%. Under various inlet temperatures and flow rates of both chilled water and air, the heat transfer rate and pressure drop were studied. From the viewpoint of economical operation of the PAC, a critical flow rate was revealed for water and air, respectively. In addition, empirical formulas of the heat transfer rate of the PAC were obtained, which can facilitate energy savings by determining the optimized temperature of the chilled water and operating parameters of the pump and fan.