Search

CN-122028392-A - Data center pre-refrigeration control method and system based on wind-liquid hybrid architecture

CN122028392ACN 122028392 ACN122028392 ACN 122028392ACN-122028392-A

Abstract

The application relates to the technical field of refrigeration control of a data center and discloses a method and a system for pre-refrigeration control of the data center based on a wind-liquid mixed architecture, wherein the method comprises the following steps of S1, acquiring operation data of the data center, and predicting the total heat production of the data center according to the operation data; S3, obtaining delay data of air cooling and liquid cooling, converting the heat generated by the air cooling and liquid cooling into heat dissipation according to the outdoor temperature and the delay data, and performing pre-cooling control on the data center according to the heat dissipation. The application can improve the prepositivity of judging the heat of the data center through the process of predicting the total heat production of the data center, and can more accurately acquire the control data of the air cooling system and the liquid cooling system according to the dynamic response delay of the air cooling system and the liquid cooling system through the process of converting the heat production into the heat dissipation, thereby providing accurate feedforward control and optimizing the dispatching input content.

Inventors

  • ZHOU HUAYI
  • YU SHENGLIANG
  • YU JIANWU

Assignees

  • 上海艾克森股份有限公司

Dates

Publication Date
20260512
Application Date
20260414

Claims (9)

  1. 1. The utility model provides a data center prefabricated cold control method based on wind-liquid mixed architecture which is characterized in that the method comprises the following steps: s1, acquiring operation data of a data center, and predicting the total heat production of the data center according to the operation data; s2, distributing air cooling and liquid cooling treatment shares according to the total heat generation amount; S3, acquiring delay data of air cooling and liquid cooling, converting heat generation quantity distributed by the air cooling and the liquid cooling into heat dissipation quantity according to outdoor temperature and the delay data, and performing prefabricated cold control on a data center according to the heat dissipation quantity.
  2. 2. The method for pre-cooling control of a data center based on a wind-liquid hybrid architecture according to claim 1, wherein the process of step S1 comprises: s11, calculating the heat generation quantity of the IT equipment according to the operation data of the data center; S12, calculating the heat transfer quantity of the maintenance structure according to the operation data of the data center; S13, calculating fresh air infiltration heat according to operation data of the data center; s12, taking the sum of the heat generation amount of IT equipment, the heat transfer amount of a maintenance structure and the fresh air permeation heat as a prediction result of the total heat generation amount.
  3. 3. The data center pre-cooling control method based on the wind-liquid mixing architecture according to claim 2, wherein the operation data comprises a task type and a task intensity L (t); The heat generation amount calculation process of the IT equipment comprises the following steps: determining a power consumption characteristic coefficient k according to the task type; By passing through Calculating to obtain the heating value of a single server ; For fixed loss, j represents the serial number of the hardware type, and the hardware type is CPU, GPU, memory and storage; Represents the idle heating value of the j-th hardware type, Representing the power consumption characteristic coefficient of the task type on the jth hardware type, Representing full power consumption of the jth hardware type, Representing idle power consumption of a j-th hardware type; Correcting power consumption for chip junction temperature; By passing through Calculating total heat productivity of single cabinet , The number of servers in a single cabinet; By passing through Calculating total heating value of data center 。
  4. 4. A data center pre-cooling control method based on a wind-liquid hybrid architecture according to claim 3, wherein the operation data comprises chip junction temperature; chip junction temperature correction power consumption The calculation process of (1) comprises: By passing through Calculating to obtain chip junction temperature correction power consumption Wherein, the method comprises the steps of, The junction temperature of the chip is used as the junction temperature of the chip, As a reference to the temperature of the liquid, As a characteristic temperature, the temperature of the material is, Is leakage power consumption of the chip at a reference temperature.
  5. 5. The data center pre-cooling control method based on the wind-liquid mixing architecture according to claim 2, wherein the operation data comprises an outdoor dry bulb temperature and an indoor reference temperature; the heat transfer amount calculation process of the enclosure structure comprises the following steps: By equation Calculating the heat transfer quantity of the enclosure structure , wherein, Is the temperature of the outdoor dry bulb, As the reference temperature in the room, Is the total thermal resistance of the enclosure structure, ; The heat transfer coefficient (W/(m2.K)) of each portion, For the area (m 2) of each part, Is the envelope time constant.
  6. 6. The data center pre-cooling control method based on the wind-liquid mixing architecture according to claim 2, wherein the operation data comprises fresh air power, outdoor dry bulb temperature, indoor reference temperature, outdoor moisture content and indoor moisture content; The calculation process of the fresh air infiltration heat comprises the following steps: By passing through Calculating fresh air infiltration heat , Is a coefficient of proportionality and is used for the control of the power supply, The fresh air power is used as the fresh air power, Is the specific heat of air, and the air is heated, Is the latent heat of vaporization of the water vapor, Is the temperature of the outdoor dry bulb, As the reference temperature in the room, For the outdoor moisture content of the water-soluble polymer, Is the indoor moisture content.
  7. 7. The method for pre-cooling control of a data center based on a wind-liquid mixing architecture according to claim 2, wherein the process of distributing the share of the wind cooling and the liquid cooling treatment comprises the following steps: By passing through Calculating the heat of the liquid cooling treatment , Is liquid cooling coverage rate; subtracting the heat of the liquid cooling treatment from the total heat Obtaining heat of air cooling treatment 。
  8. 8. The method for pre-cooling control of a data center based on a wind-liquid hybrid architecture according to claim 7, wherein the process of converting the generated heat of liquid cooling distribution into the dissipated heat comprises: By equation Liquid cooling heat dissipation capacity is obtained through calculation ; Is liquid cooling response time constant and meets , The maximum heat dissipation capacity of liquid cooling is achieved; By equation Calculating to obtain air-cooled heat dissipation capacity ; Is the response time constant of air cooling; In order to couple the items together, 。
  9. 9. A data center pre-cooling control system based on a wind-liquid mixing architecture, wherein the method adopts the data center pre-cooling control method based on the wind-liquid mixing architecture as claimed in any one of claims 1 to 8, and the method comprises the following steps: the parameter acquisition end is used for acquiring the operation data of the data center and the delay data of air cooling and liquid cooling; A prediction unit for predicting the total heat production of the data center according to the operation data; the distribution unit is used for distributing air cooling and liquid cooling treatment shares according to the total heat generation amount; The adjusting unit is used for converting the heat generation quantity distributed by air cooling and liquid cooling into heat dissipation quantity according to the outdoor temperature and the delay data; and the pre-cooling unit is used for carrying out pre-cooling control on the data center according to the heat dissipation capacity.

Description

Data center pre-refrigeration control method and system based on wind-liquid hybrid architecture Technical Field The application relates to the technical field of refrigeration control of data centers, in particular to a data center pre-refrigeration control method and system based on a wind-liquid mixed architecture. Background With the rapid development of applications such as artificial intelligence and high-performance computing, the power density of a single cabinet of a data center has been increased from traditional 5-10 kW to more than 30 kW, and even an ultra-high density scene that the power density of the single cabinet exceeds 100 kW appears. The traditional air cooling system is limited by the physical limits of specific heat capacity of air and power consumption of a fan, and the problems of insufficient heat dissipation capacity, prominent local hot spots, rapid increase of cooling energy consumption and the like generally exist when the traditional air cooling system is used for coping with such high-heat-density loads. The wind-liquid hybrid architecture is a mainstream technical scheme for newly building and reforming a high-density data center because the wind-liquid hybrid architecture can efficiently take away chip-level heat and simultaneously maintain the stable environment of a machine room. However, the refrigeration control of the air-liquid mixing architecture is far more complex than the traditional air cooling, on one hand, the liquid cooling system has high response speed and directly absorbs heat of main heat sources such as a CPU/GPU (Central processing Unit/graphics processing Unit), but the heat dissipation capacity of the liquid cooling system is limited by the temperature, the flow and the outdoor weather conditions of cooling water, on the other hand, the air cooling system has slower response, the two systems have obvious differences in time scale, thermal inertia and control targets, if mutually independent PID (proportion integration differentiation) control is adopted, the problems of coupling imbalance such as air cooling overcompensation or improper liquid cooling water supply temperature setting and the like are extremely easy to occur, so that energy waste and even potential safety hazards are caused, and therefore, how to provide accurate input for feedforward control and optimal scheduling of a data center cooling system is the fundamental problem to be solved. Disclosure of Invention In order to provide accurate input for feedforward control and optimal scheduling of a cooling system of a data center, the application provides a data center pre-refrigeration control method and system based on a wind-liquid mixing architecture. In a first aspect, the application provides a data center pre-refrigeration control method based on a wind-liquid mixing architecture, which adopts the following technical scheme: a data center pre-refrigeration control method based on a wind-liquid hybrid architecture comprises the following steps: s1, acquiring operation data of a data center, and predicting the total heat production of the data center according to the operation data; s2, distributing air cooling and liquid cooling treatment shares according to the total heat generation amount; S3, acquiring delay data of air cooling and liquid cooling, converting heat generation quantity distributed by the air cooling and the liquid cooling into heat dissipation quantity according to outdoor temperature and the delay data, and performing prefabricated cold control on a data center according to the heat dissipation quantity. Optionally, the process of step S1 includes: s11, calculating the heat generation quantity of the IT equipment according to the operation data of the data center; S12, calculating the heat transfer quantity of the maintenance structure according to the operation data of the data center; S13, calculating fresh air infiltration heat according to operation data of the data center; s12, taking the sum of the heat generation amount of IT equipment, the heat transfer amount of a maintenance structure and the fresh air permeation heat as a prediction result of the total heat generation amount. Optionally, the running data includes a task type and a task intensity L (t); The heat generation amount calculation process of the IT equipment comprises the following steps: determining a power consumption characteristic coefficient k according to the task type; By passing through Calculating to obtain the heating value of a single server;For fixed loss, j represents the serial number of the hardware type, and the hardware type is CPU, GPU, memory and storage; Represents the idle heating value of the j-th hardware type, Representing the power consumption characteristic coefficient of the task type on the jth hardware type,Representing full power consumption of the jth hardware type,Representing idle power consumption of a j-th hardware type; Correcting power consumption for chip ju