Research on data center load distribution based on dung beetle optimization algorithm

With the rapid growth of data center power density and size, operators often reduce the energy consumption of cooling infrastructure by increasing the supply air temperature of cooling system. However, when implementing energy-saving measures in a complex environment, it is necessary to assess the "temperature risk" and effectively manage it to prevent possible hot spot effects. In view of the above problems, this paper studies the server power consumption model and heat transfer model in a given data center room by means of data fitting and CFD simulation, and discusses the influence of server load distribution on its own power and temperature distribution in the room. Finally, taking minimum variance of entrance temperature among nodes as optimization objective, node load and upper temperature limit as constraint conditions, an optimization control method based on dung beetle optimization algorithm is proposed to find optimal node load distribution under uniform entrance temperature distribution. The experimental results show that the control based on dung beetle optimization algorithm can reduce the entrance temperature of the equipment room nodes by 1.2℃ at most, which is 4.8% lower than that of the control based on dung beetle optimization algorithm, and promote the balanced distribution of the equipment room temperature, thus reducing local hot spots and improving the utilization rate of the cooling system and the reliability of the equipment.