Adaptive management method for power consumption of data center

Abstract

The invention provides an adaptive management method for the power consumption of a data center. The method comprises the following steps of: partitioning a computer room according to functions of each part, arranging N+1 precision air conditioners in each function area, and sorting by priority according to the importance of a refrigerating action; adjusting the power consumption of a server in real time by using monitoring software according to a service load in the computer room; accomplishing generation of an air conditioner adjustment strategy with the help of a computational fluid dynamics (CFD) simulation computed result; and controlling the air conditioners of the computer room by using a fuzzy neural network control strategy. By the method, an optimum value can be achieved under the conditions of good radiation effect and low energy consumption.

Description

Technical field [0001] The invention relates to the field of cluster power consumption management, in particular to a data center power consumption adaptive management method. Background technique [0002] The energy composition of a data center is mainly divided into four parts, namely: power supply system, cooling system, IT equipment and other infrastructure (such as lighting, security system, etc.). The energy consumption of the power supply system mainly refers to the power loss of the power input of the data center at the UPS and PDU, which accounts for about 20% to 25% of the total energy consumption of the data center; the energy consumption of the cooling system mainly refers to the power consumed by the cooling system itself , Accounting for about 45% to 55% of the total energy consumption of the data center; lighting and security systems consume less power, accounting for about 2% to 3% of the total energy consumption of the data center. It can be seen that the energy ...

AI Technical Summary

Problems solved by technology

However, air conditioning units are usually installed near windows or walls, and the airflow temperature at the top return air vents does not accurately represent the ambient temperature of each rack, and often has large local deviations

[0007] 2. Due to the geometric complexity of the entire computer room space, as well as the unreasonable location and parameters of the air outlet of the air conditioner, the velocity field and pressure field of the airflow inside the computer room are extremely uneven, and there are many low-pressure and low-speed vortex areas, so The convective heat transfer state between the electronic components in the rack and the ambient airflow presents a serious inconsistency

Whether the cooling capacity adjusted by CRAC is appropriate, and after the cooling capacity is adjusted, what will happen to the airflow organization in the room, and whether new local hot spots and cold spots will be formed, etc. These are only determined by the sensors distributed in different locations of the data center. unsolvable

[0008] 3. The special air conditioner for the computer room has intelligent functions, but the "function" of the air conditioner for the reasonable control of the temperature and humidity range of the computer room is "insufficient", and the combined use of the "air conditioner group" is independent, which greatly reduces its use efficiency

Images