Deep Learning-Based Data Center Cooling Control Method and System

Abstract

This invention discloses a data center cooling control method and system based on deep learning, relating to the field of cooling control technology. The method includes: using an infrared thermal imager to acquire high-temperature condensation points; acquiring the high-temperature condensation type and cooling time interval based on the location of ventilation openings and high-temperature condensation points; establishing a deep learning model; and performing real-time cooling control of the data center based on the high-temperature condensation type of the real-time condensation points. This invention addresses the problem in existing data center cooling control methods that cannot control the cooling process based on the location of high-temperature areas within the data center. This results in the cooling system consuming excessive resources and operating slowly when cooling the data center because it does not target high-temperature areas.

Description

Technical Field

[0001] This invention relates to the field of cooling control technology, specifically to a data center cooling control method and system based on deep learning. Background Technology

[0002] Data center cooling control refers to the effective control of temperature and humidity in a data center environment through various technologies and strategies to ensure that IT equipment operates in a suitable environment, thereby guaranteeing the stability and reliability of the data center. Cooling control is not only related to the stable operation of equipment, but also directly affects energy efficiency and environmental sustainability. The working principle of data center cooling control includes airflow management, cooling components, and monitoring and maintenance.

[0003] Existing methods for data center cooling control typically rely on pre-cooling models to obtain pre-cooling capacity within the data center. They then analyze the load and energy consumption of all equipment to determine the corresponding thermal power consumption and cooling load. This data is then combined with temperature measurement points to generate control strategies for cooling regulation. While this improved approach enhances the response speed and cooling capacity allocation accuracy of the cooling system, it cannot target the cooling process based on areas with higher temperatures within the data center. This means that even with improved cooling capacity allocation accuracy, the lack of targeted cooling for higher-temperature areas fails to address the problem of high resource consumption and slow cooling efficiency. For example, patent application CN120111852A discloses a data center cooling system with second-level adjustment... The proposed control method and system obtains optimal strategy feedback by establishing a pre-cooling model, a load energy consumption prediction model, and a cold load conversion model, and achieves iterative optimization of the model. Through task-aware pre-cooling, precise component-level control, and closed-loop feedback optimization, it improves the response speed, cold capacity allocation accuracy, and overall energy efficiency of the cooling system. Other improvements for data center cooling control are usually based on simulator-based cooling schemes and improvements in cooling regulation efficiency by selecting chillers. However, these improvements still cannot solve the problem of not being able to control the cooling process based on the high-temperature locations within the data center. This results in the cooling system consuming more resources and operating more slowly when cooling the data center because it does not target high-temperature areas. Therefore, it is necessary to improve the existing data center cooling control methods. Summary of the Invention

[0004] This invention aims to at least partially solve one of the technical problems in the prior art. By proposing a data center cooling control method and system based on deep learning, it addresses the issue that existing data center cooling control methods cannot regulate the cooling process based on the location of high temperature within the data center. This results in the cooling system consuming more resources and having slower cooling efficiency when cooling the data center because it does not target the high-temperature areas.

[0005] To achieve the above objectives, in a first aspect, this application provides a data center cooling control method based on deep learning, comprising the following steps:

[0006] When all equipment in the data center is operating normally, an infrared thermal imager is used to detect the temperature of the data center and obtain multiple high-temperature condensation points within the data center; multiple ventilation openings within the data center are also obtained, and based on the locations of the ventilation openings and the locations of the high-temperature condensation points, a cooling control analysis method is used to obtain the optimal cooling time and optimal cooling outlet for each high-temperature condensation point.

[0007] Based on the preferred cooling port of each high-temperature condensation point, all high-temperature condensation points are screened, and multiple high-temperature condensation types and cooling time intervals for each high-temperature condensation type are obtained based on the screening results.

[0008] Establish a deep learning model; store the refrigeration control analysis method into the deep learning model, and learn the process of obtaining the high-temperature condensation type from the high-temperature condensation point in the deep learning model;

[0009] Based on the infrared thermal imager, high-temperature condensation points in the data center are acquired in real time. When a new high-temperature condensation point is found, its location is entered into the deep learning model, and the high-temperature condensation type corresponding to the new high-temperature condensation point is obtained based on the deep learning model.

[0010] The latest high-temperature condensation point acquired by the infrared thermal imager is recorded as the real-time condensation point. The data center is then regulated in real-time based on the cooling time interval corresponding to the high-temperature condensation type of the real-time condensation point.

[0011] Furthermore, with all equipment in the data center operating normally, an infrared thermal imager is used to detect the temperature of the data center and identify multiple high-temperature condensation points within the data center, including:

[0012] Establish a spatial rectangular coordinate system, denoted as the spatial analysis coordinate system, where the units of the X-axis, Y-axis, and Z-axis of the spatial analysis coordinate system are all meters (m).

[0013] Based on the design drawings of the data center and the dimensional data of all equipment in the data center, a scaled model of the interior of the data center is established in the first quadrant of the spatial analysis coordinate system and denoted as the data three-dimensional model. The dimensional data includes height data, width data and length data.

[0014] Furthermore, when all equipment in the data center is operating normally, using an infrared thermal imager to detect the temperature of the data center and identify multiple high-temperature condensation points within the data center also includes:

[0015] When all equipment in the data center is operating normally, use an infrared thermal imager to detect the internal temperature of the data center every kmin, and obtain the location of the highest temperature in the data center after the temperature detection, and mark it as the high temperature location; obtain all the high temperature locations obtained by the infrared thermal imager in the data center within a day, and mark all the high temperature locations in the data 3D model;

[0016] Obtain the temperature corresponding to the high-temperature location in the temperature detection when all high-temperature locations are marked, and record it as the high-temperature temperature of the high-temperature location; record all high-temperature locations with a high temperature greater than the maximum temperature threshold as high-temperature condensation points, where the maximum temperature threshold is the maximum value of the normal temperature range within the data center.

[0017] Furthermore, multiple ventilation openings within the data center are obtained, and based on the locations of these openings and the locations of high-temperature condensation points, a cooling control analysis method is used to determine the optimal cooling time and optimal cooling outlet for each high-temperature condensation point, including:

[0018] Within the 3D data model: acquire multiple ventilation openings within the data center and label them sequentially as ventilation opening TF1 to ventilation opening TF2. c Based on the equipment usage specifications of the data center, obtain the maximum and minimum number of ventilation openings that can be opened simultaneously, denoted as m and n respectively, where m and n are positive integers less than or equal to c and greater than or equal to 1.

[0019] Obtain all possible opening states for the ventilation vents within the data center, from n to m vents that are allowed to be opened simultaneously, and denote them as ventilation opening state TK1 to ventilation opening state TF. p ;

[0020] The locations of all high-temperature condensation points were obtained, and the optimal cooling time and optimal cooling port for each high-temperature condensation point were determined using the refrigeration control analysis method.

[0021] Furthermore, the refrigeration control analysis method includes:

[0022] For any high-temperature condensation point A: when the temperature of high-temperature condensation point A is detected by the infrared thermal imager to be equal to the high-temperature temperature of high-temperature condensation point A, the cooling system in the data center is activated and based on the ventilation status TK. q Open the ventilation openings in the data center, use an infrared thermal imager to obtain the temperature at the location of the high-temperature condensation point A in real time, and start the timer, where q is a positive integer less than or equal to p and greater than or equal to 1;

[0023] When the temperature of high-temperature condensation point A in the infrared thermal imager is less than or equal to the maximum temperature threshold, the timer is turned off, and the timer time is recorded as the cooling time.

[0024] Furthermore, the refrigeration control analysis method also includes:

[0025] When the temperature of high-temperature condensation point A is equal to the high temperature of high-temperature condensation point A, the cooling time corresponding to all ventilation opening conditions is obtained. The minimum cooling time is recorded as the preferred cooling time of high-temperature condensation point A, and all the vents opened in the ventilation opening conditions corresponding to the minimum cooling time are recorded as the preferred cooling vents of high-temperature condensation point A.

[0026] Obtain the optimal cooling time and optimal cooling port corresponding to all high-temperature condensation points.

[0027] Furthermore, based on the preferred refrigeration port for each high-temperature condensation point, all high-temperature condensation points are screened, and based on the screening results, multiple high-temperature condensation types and the refrigeration time interval for each high-temperature condensation type are obtained, including:

[0028] For any high-temperature condensation point A: High-temperature condensation points other than high-temperature condensation point A whose preferred refrigeration port is exactly the same as the preferred refrigeration port of high-temperature condensation point A are recorded as condensation points of the same type as high-temperature condensation point A, and high-temperature condensation point A and condensation points of the same type as high-temperature condensation point A are recorded as the same high-temperature condensation type.

[0029] Obtain multiple high-temperature condensation types corresponding to all high-temperature condensation points; for any high-temperature condensation type, denote the closed interval formed by the minimum and maximum values ​​of the preferred cooling times for all high-temperature condensation points corresponding to the high-temperature condensation type as the cooling time interval of the high-temperature condensation type.

[0030] Furthermore, the latest high-temperature condensation point acquired by the infrared thermal imager is recorded as the real-time condensation point. Real-time cooling control of the data center is then performed based on the cooling time interval corresponding to the high-temperature condensation type of the real-time condensation point, including:

[0031] When the real-time condensation point is obtained, the cooling system in the data center is started, and the preferred cooling ports of all high-temperature condensation points corresponding to the high-temperature condensation type of the real-time condensation point are started.

[0032] Maintain the start-up time of the cooling system within the data center and the opening time of all open vents greater than T1.

[0033] Furthermore, the latest high-temperature condensation point acquired by the infrared thermal imager is recorded as the real-time condensation point. Real-time cooling control of the data center based on the cooling time interval corresponding to the high-temperature condensation type of the real-time condensation point also includes:

[0034] When the start-up time of the cooling system in the data center and the opening time of all open vents are less than or equal to T2, and the real-time condensation point temperature is less than or equal to the maximum temperature threshold, the cooling system in the data center is shut down.

[0035] When the start-up time of the cooling system in the data center and the opening time of all open vents are greater than T2, and the temperature of the real-time condensation point is greater than the maximum temperature threshold, a temperature control anomaly report is made. Here, T1 is the minimum value in the cooling time interval of the high-temperature condensation type of the real-time condensation point, and T2 is the maximum value in the cooling time interval of the high-temperature condensation type of the real-time condensation point.

[0036] Secondly, this application also provides a data center cooling control system based on deep learning, including a high-temperature cooling analysis module, a high-temperature type screening module, and a deep real-time control module;

[0037] The high-temperature cooling analysis module is used to detect the temperature of the data center using an infrared thermal imager when all equipment in the data center is operating normally, and to obtain multiple high-temperature condensation points in the data center; to obtain multiple ventilation openings in the data center, and to obtain the optimal cooling time and optimal cooling outlet for each high-temperature condensation point based on the location of the multiple ventilation openings and the location of the high-temperature condensation points using the cooling control analysis method.

[0038] The high-temperature type screening module screens all high-temperature condensation points based on the preferred refrigeration port of each high-temperature condensation point, and obtains multiple high-temperature condensation types and the refrigeration time interval for each high-temperature condensation type based on the screening results.

[0039] The deep real-time control module is used to build a deep learning model; the refrigeration control analysis method is stored in the deep learning model, and the deep learning model learns the process of obtaining the high-temperature condensation type from the high-temperature condensation point;

[0040] Based on the infrared thermal imager, high-temperature condensation points in the data center are acquired in real time. When a new high-temperature condensation point is found, its location is entered into the deep learning model, and the high-temperature condensation type corresponding to the new high-temperature condensation point is obtained based on the deep learning model.

[0041] The latest high-temperature condensation point acquired by the infrared thermal imager is recorded as the real-time condensation point. The data center is then regulated in real-time based on the cooling time interval corresponding to the high-temperature condensation type of the real-time condensation point.

[0042] The beneficial effects of this invention are as follows: This application first uses an infrared thermal imager to acquire multiple high-temperature condensation points within a data center; based on the locations of multiple ventilation openings and high-temperature condensation points within the data center, a cooling control analysis method is used to obtain the optimal cooling time and optimal cooling outlet for each high-temperature condensation point. The advantage of this is that by acquiring the high-temperature condensation points, it is possible to identify the locations with higher temperatures within the data center during normal operation. Furthermore, by acquiring the optimal cooling time and optimal cooling outlet for each high-temperature condensation point based on the locations of the ventilation openings and high-temperature condensation points, it is possible to obtain the time when the cooling system in the data center achieves the fastest cooling efficiency for each high-temperature condensation point and the opening status of the ventilation openings within the cooling center at the time of the fastest cooling efficiency. This facilitates targeted cooling of areas with higher temperatures within the control center based on real-time condensation points during subsequent real-time cooling control, thereby reducing the resources consumed during cooling and improving cooling efficiency.

[0043] This application further filters all high-temperature condensation points based on the preferred cooling port for each high-temperature condensation point, and obtains multiple high-temperature condensation types and cooling time intervals for each high-temperature condensation type based on the filtering results. Finally, a deep learning model is established, and when a new high-temperature condensation point exists, the high-temperature condensation type corresponding to the new high-temperature condensation point is obtained based on the deep learning model. The latest high-temperature condensation point acquired by the infrared thermal imager is recorded as the real-time condensation point, and the data center is subjected to real-time cooling control based on the cooling time interval corresponding to the high-temperature condensation type of the real-time condensation point. The advantage of this is that by obtaining the cooling time interval, the cooling effect can be checked based on the cooling time interval after each cooling control, so as to promptly notify the staff to handle the cooling anomaly. By establishing a deep learning model, the efficiency of obtaining the high-temperature condensation type corresponding to the high-temperature condensation point after obtaining a new high-temperature condensation point can be improved, and the corresponding cooling time interval and preferred cooling port can be obtained in a timely manner for efficient cooling control. Attached Figure Description

[0044] Figure 1 This is a schematic diagram of the system of the present invention;

[0045] Figure 2 This is a flowchart illustrating the steps of the method of the present invention;

[0046] Figure 3 This is a flowchart of the steps of the refrigeration control analysis method of the present invention;

[0047] Figure 4 This is a schematic diagram of the electronic device of the present invention. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] Example 1, please refer to Figure 1 As shown, this application provides a data center cooling control system based on deep learning, including a high-temperature cooling analysis module, a high-temperature type screening module, and a deep real-time control module;

[0050] The high-temperature cooling analysis module is used to detect the temperature of the data center using an infrared thermal imager when all equipment in the data center is operating normally, and to obtain multiple high-temperature condensation points in the data center; to obtain multiple ventilation openings in the data center, and to obtain the optimal cooling time and optimal cooling outlet for each high-temperature condensation point based on the location of the multiple ventilation openings and the location of the high-temperature condensation points using the cooling control analysis method.

[0051] The high-temperature refrigeration analysis module includes a high-temperature condensation point analysis unit, which is equipped with a high-temperature condensation point analysis strategy. The high-temperature condensation point analysis strategy includes:

[0052] Establish a spatial rectangular coordinate system, denoted as the spatial analysis coordinate system, where the units of the X-axis, Y-axis, and Z-axis of the spatial analysis coordinate system are all meters (m).

[0053] Based on the design drawings of the data center and the size data of all equipment in the data center, a scale model of the interior of the data center is established in the first quadrant of the spatial analysis coordinate system and denoted as the data three-dimensional model. The size data includes height data, width data and length data.

[0054] When all equipment in the data center is operating normally, use an infrared thermal imager to detect the internal temperature of the data center every kmin, and obtain the location of the highest temperature in the data center after the temperature detection, and mark it as the high temperature location; obtain all the high temperature locations obtained by the infrared thermal imager in the data center within a day, and mark all the high temperature locations in the data 3D model;

[0055] In the specific implementation process, the value of k can be set according to the temperature detection interval allowed when the infrared thermal imager is actually used. To ensure more comprehensive temperature detection in the data center, the value of k should be as small as possible. In this embodiment, the value of k is set to 10; that is, the infrared thermal imager is used to detect the temperature inside the data center every 10 minutes.

[0056] Obtain the temperature corresponding to the high temperature location in the temperature detection when all high temperature locations are marked, and record it as the high temperature of the high temperature location; record all high temperature locations with a high temperature greater than the maximum temperature threshold as high temperature condensation points, where the maximum temperature threshold is the maximum value of the normal temperature range in the data center;

[0057] In the specific implementation process, the maximum temperature threshold can be set according to the operating status of the equipment in the data center and the maximum temperature that it can withstand. For example, if it is found during a data analysis that the ambient temperature of the server in the data center should be maintained between 22°C and 26°C when it is operating normally, then the maximum temperature threshold can be set to 26°C.

[0058] Within the 3D data model: acquire multiple ventilation openings within the data center and label them sequentially as ventilation opening TF1 to ventilation opening TF2. c Based on the equipment usage specifications of the data center, obtain the maximum and minimum number of ventilation openings that can be opened simultaneously, denoted as m and n respectively, where m and n are positive integers less than or equal to c and greater than or equal to 1.

[0059] In the specific implementation process, marking the location of the vents in the three-dimensional data model is beneficial for subsequent data processing in the three-dimensional data model, thereby improving the efficiency of obtaining the preferred cooling vents.

[0060] Obtain all possible opening states for the ventilation vents within the data center, from n to m vents that are allowed to be opened simultaneously, and denote them as ventilation opening state TK1 to ventilation opening state TF. p ;

[0061] In the specific implementation process, for example, during a data analysis, if three ventilation openings in the data center are allowed to be opened at the same time and four ventilation openings are allowed to be opened at the same time, then all the situations corresponding to three ventilation openings being opened at the same time and all the situations corresponding to four ventilation openings being opened at the same time can be recorded as ventilation opening situations in turn.

[0062] The locations of all high-temperature condensation points were obtained, and the optimal cooling time and optimal cooling port for each high-temperature condensation point were determined using the refrigeration control analysis method.

[0063] Refrigeration control analysis methods include: Please refer to: Figure 3 As shown, for any high-temperature condensation point A: when the temperature of high-temperature condensation point A is detected by the infrared thermal imager to be equal to the high-temperature temperature of high-temperature condensation point A, the cooling system in the data center is activated and the ventilation status TK is adjusted accordingly. qOpen the ventilation openings in the data center, use an infrared thermal imager to obtain the temperature at the location of the high-temperature condensation point A in real time, and start the timer, where q is a positive integer less than or equal to p and greater than or equal to 1;

[0064] When the temperature of high-temperature condensation point A in the infrared thermal imager is less than or equal to the maximum temperature threshold, the timer is turned off and the timer time is recorded as the cooling time.

[0065] When the temperature of high-temperature condensation point A is equal to the high temperature of high-temperature condensation point A, the cooling time corresponding to all ventilation opening conditions is obtained. The minimum cooling time is recorded as the preferred cooling time of high-temperature condensation point A, and all the vents opened in the ventilation opening conditions corresponding to the minimum cooling time are recorded as the preferred cooling vents of high-temperature condensation point A.

[0066] In the specific implementation process, for example, in a data analysis, the number of ventilation openings is 7. Under the 7 ventilation opening conditions, the time for the temperature of high temperature point A to change from high temperature to less than or equal to the maximum temperature threshold is 1 min, 1.5 min, 1.1 min, 0.9 min, 1.2 min, 1.3 min and 2 min respectively. Then, all the ventilation openings in the ventilation opening condition corresponding to 0.9 min can be recorded as the preferred cooling openings of high temperature point A, so as to obtain the opening conditions of the ventilation openings with the fastest cooling efficiency for high temperature point A.

[0067] Obtain the optimal cooling time and optimal cooling port corresponding to all high-temperature condensation points.

[0068] The high-temperature type screening module screens all high-temperature condensation points based on the preferred refrigeration port of each high-temperature condensation point, and obtains multiple high-temperature condensation types and the refrigeration time interval for each high-temperature condensation type based on the screening results.

[0069] The high-temperature type screening module includes a high-temperature condensation point screening unit, which is configured with a high-temperature condensation point screening strategy. The high-temperature condensation point screening strategy includes:

[0070] For any high-temperature condensation point A: High-temperature condensation points other than high-temperature condensation point A whose preferred refrigeration port is exactly the same as the preferred refrigeration port of high-temperature condensation point A are recorded as condensation points of the same type as high-temperature condensation point A, and high-temperature condensation point A and condensation points of the same type as high-temperature condensation point A are recorded as the same high-temperature condensation type.

[0071] In the specific implementation process, the high-temperature condensation point is analyzed to obtain the high-temperature condensation type. The high-temperature condensation point can be classified based on the preferred refrigeration port. This is beneficial for subsequent analysis, where the corresponding high-temperature condensation type can be obtained directly through the high-temperature condensation point based on the deep learning model. This makes it easier to obtain the corresponding preferred refrigeration port more quickly, thereby improving refrigeration efficiency.

[0072] Obtain multiple high-temperature condensation types corresponding to all high-temperature condensation points; for any high-temperature condensation type, the closed interval formed by the minimum and maximum values ​​of the preferred cooling times of all high-temperature condensation points corresponding to the high-temperature condensation type is denoted as the cooling time interval of the high-temperature condensation type.

[0073] In the specific implementation process, for example, if the preferred cooling times for all high-temperature condensation points in a high-temperature condensation type are 1 min, 1.5 min, 1.1 min, and 2 min, the cooling time interval can be set to [1 min, 2 min]. This means that when the vent is opened based on this high-temperature condensation type, it will take at least 2 minutes to reduce the temperature of all areas in the data center to below the maximum temperature threshold. Therefore, in actual cooling, if there are still areas with temperatures higher than the maximum temperature threshold after the vent is opened for 2 minutes based on this high-temperature condensation type, it indicates that there is an abnormality in the cooling system, and a cooling abnormality report should be made.

[0074] The deep real-time control module is used to build a deep learning model; the refrigeration control analysis method is stored in the deep learning model, and the deep learning model learns the process of obtaining the high-temperature condensation type from the high-temperature condensation point;

[0075] Based on the infrared thermal imager, high-temperature condensation points in the data center are acquired in real time. When a new high-temperature condensation point is found, its location is entered into the deep learning model, and the high-temperature condensation type corresponding to the new high-temperature condensation point is obtained based on the deep learning model.

[0076] The latest high-temperature condensation point acquired by the infrared thermal imager is recorded as the real-time condensation point. The data center is then cooled and controlled in real time based on the cooling time interval corresponding to the high-temperature condensation type of the real-time condensation point.

[0077] The deep real-time control module includes a real-time cooling control unit, which is configured with a real-time cooling control strategy, including:

[0078] When the real-time condensation point is obtained, the cooling system in the data center is started, and the preferred cooling ports of all high-temperature condensation points corresponding to the high-temperature condensation type of the real-time condensation point are started.

[0079] Maintain the start-up time of the cooling system within the data center and the opening time of all open vents greater than T1;

[0080] When the start-up time of the cooling system in the data center and the opening time of all open vents are less than or equal to T2, and the real-time condensation point temperature is less than or equal to the maximum temperature threshold, the cooling system in the data center is shut down.

[0081] When the start-up time of the cooling system in the data center and the opening time of all open vents are greater than T2, and the temperature of the real-time condensation point is greater than the maximum temperature threshold, a temperature control anomaly report is made. Here, T1 is the minimum value in the cooling time interval of the high-temperature condensation type of the real-time condensation point, and T2 is the maximum value in the cooling time interval of the high-temperature condensation type of the real-time condensation point.

[0082] Example 2, please refer to Figure 2 As shown, this application also provides a data center cooling control method based on deep learning, including the following steps:

[0083] Step S1: When all equipment in the data center is operating normally, use an infrared thermal imager to detect the temperature of the data center and obtain multiple high-temperature condensation points in the data center; obtain multiple ventilation openings in the data center, and use the cooling control analysis method to obtain the optimal cooling time and optimal cooling outlet for each high-temperature condensation point based on the location of the multiple ventilation openings and the location of the high-temperature condensation points.

[0084] Step S1 includes: Step S101, establishing a spatial rectangular coordinate system, denoted as the spatial analysis coordinate system, wherein the units of the X-axis, Y-axis and Z-axis of the spatial analysis coordinate system are all meters;

[0085] Step S102: Based on the design drawings of the data center and the size data of all equipment in the data center, establish a scale model of the interior of the data center in the first quadrant of the spatial analysis coordinate system and record it as the data three-dimensional model. The size data includes height data, width data and length data.

[0086] Step S103: When all equipment in the data center is operating normally, use an infrared thermal imager to detect the temperature inside the data center every kmin, and obtain the location of the highest temperature in the data center after the temperature detection, and mark it as a high temperature location; obtain all high temperature locations obtained by the infrared thermal imager in the data center within a day, and mark all high temperature locations in the data three-dimensional model;

[0087] Step S104: Obtain the temperature corresponding to the high temperature location in the temperature detection when all high temperature locations are marked, and record it as the high temperature of the high temperature location; record all high temperature locations with a high temperature greater than the maximum temperature threshold as high temperature condensation points, where the maximum temperature threshold is the maximum value of the normal temperature range in the data center.

[0088] Step S105, within the 3D data model: obtain multiple ventilation openings within the data center, and label them sequentially as ventilation opening TF1 to ventilation opening TF2. cBased on the equipment usage specifications of the data center, obtain the maximum and minimum number of ventilation openings that can be opened simultaneously, denoted as m and n respectively, where m and n are positive integers less than or equal to c and greater than or equal to 1.

[0089] Step S106: Obtain all opening states corresponding to the following: n vents are allowed to be opened simultaneously, n+1 vents are allowed to be opened simultaneously, ... up to m vents are allowed to be opened simultaneously, and record them as ventilation opening state TK1 to ventilation opening state TF respectively. p ;

[0090] Step S107: Obtain the location of all high-temperature condensation points and use the refrigeration control analysis method to obtain the optimal refrigeration time and optimal refrigeration port for each high-temperature condensation point.

[0091] The cooling control analysis method includes: Step S1071, for any high-temperature condensation point A: when the temperature of high-temperature condensation point A is detected by the infrared thermal imager to be equal to the high-temperature temperature of high-temperature condensation point A, the cooling system in the data center is started and based on the ventilation status TK q Open the ventilation openings in the data center, use an infrared thermal imager to obtain the temperature at the location of the high-temperature condensation point A in real time, and start the timer, where q is a positive integer less than or equal to p and greater than or equal to 1;

[0092] Step S1072: When the temperature of high-temperature condensation point A in the infrared thermal imager is less than or equal to the maximum temperature threshold, turn off the timer and record the timer time as the cooling time.

[0093] Step S1073: When the temperature of high temperature condensation point A is equal to the high temperature of high temperature condensation point A, the cooling time corresponding to all ventilation opening conditions is obtained. The minimum cooling time is recorded as the preferred cooling time of high temperature condensation point A, and all the vents opened in the ventilation opening conditions corresponding to the minimum cooling time are recorded as the preferred cooling vents of high temperature condensation point A.

[0094] Step S1074: Obtain the preferred cooling time and preferred cooling port corresponding to all high-temperature condensation points.

[0095] Step S2: Based on the preferred cooling port of each high-temperature condensation point, all high-temperature condensation points are screened, and multiple high-temperature condensation types and the cooling time interval of each high-temperature condensation type are obtained based on the screening results.

[0096] Step S2 includes: Step 201, for any high-temperature condensation point A: record the high-temperature condensation points other than high-temperature condensation point A whose preferred refrigeration port is exactly the same as the preferred refrigeration port of high-temperature condensation point A as the same type of condensation point of high-temperature condensation point A, and record high-temperature condensation point A and the same type of condensation point of high-temperature condensation point A as the same high-temperature condensation type.

[0097] Step 202: Obtain multiple high-temperature condensation types corresponding to all high-temperature condensation points; for any high-temperature condensation type, the closed interval formed by the minimum and maximum values ​​of the preferred cooling times of all high-temperature condensation points corresponding to the high-temperature condensation type is recorded as the cooling time interval of the high-temperature condensation type.

[0098] Step S3: Establish a deep learning model; store the refrigeration control analysis method into the deep learning model, and learn the process of obtaining the high-temperature condensation type from the high-temperature condensation point in the deep learning model;

[0099] Based on the infrared thermal imager, high-temperature condensation points in the data center are acquired in real time. When a new high-temperature condensation point is found, its location is entered into the deep learning model, and the high-temperature condensation type corresponding to the new high-temperature condensation point is obtained based on the deep learning model.

[0100] The latest high-temperature condensation point acquired by the infrared thermal imager is recorded as the real-time condensation point. The data center is then cooled and controlled in real time based on the cooling time interval corresponding to the high-temperature condensation type of the real-time condensation point.

[0101] Step S3 includes: Step S301, when the real-time condensation point is obtained, start the cooling system in the data center and start the preferred cooling port of all high-temperature condensation points corresponding to the high-temperature condensation type of the real-time condensation point.

[0102] Step S302: Ensure that the start-up time of the cooling system in the data center and the opening time of all open vents are greater than T1.

[0103] Step S303: When the start-up time of the cooling system in the data center and the opening time of all open vents are less than or equal to T2 and the temperature of the real-time condensation point is less than or equal to the maximum temperature threshold, shut down the cooling system in the data center.

[0104] Step S304: When the start-up time of the cooling system in the data center and the opening time of all open vents are greater than T2 and the temperature of the real-time condensation point is greater than the maximum temperature threshold, a temperature control anomaly report is made. Here, T1 is the minimum value in the cooling time interval of the high-temperature condensation type of the real-time condensation point, and T2 is the maximum value in the cooling time interval of the high-temperature condensation type of the real-time condensation point.

[0105] Example 3, please refer to Figure 4 As shown, Figure 4A schematic diagram of an electronic device is provided, which may include a processor, a communication interface, a memory, and a communication bus. The processor, communication interface, and memory communicate with each other via the communication bus. The memory stores computer-readable instructions, and the processor can call these instructions. When the processor executes a computer-readable instruction, it performs steps similar to those in a deep learning-based data center cooling control method to achieve the following functions: First, an infrared thermal imager is used to acquire multiple high-temperature condensation points within the data center. Based on the locations of multiple ventilation openings and the high-temperature condensation points, a cooling control analysis method is used to obtain the optimal cooling time and optimal cooling outlet for each high-temperature condensation point. Then, based on the optimal cooling outlet for each high-temperature condensation point, all high-temperature condensation points are screened, and multiple high-temperature condensation types and cooling time intervals for each high-temperature condensation type are obtained based on the screening results. Finally, a deep learning model is established, and when a new high-temperature condensation point exists, the high-temperature condensation type corresponding to the new high-temperature condensation point is obtained based on the deep learning model. The latest high-temperature condensation point acquired by the infrared thermal imager is recorded as the real-time condensation point, and real-time cooling control of the data center is performed based on the cooling time interval corresponding to the high-temperature condensation type of the real-time condensation point.

[0106] Furthermore, when the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0107] Example 4: This application also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it performs the steps of the data center cooling control method based on deep learning described above to achieve the following functions: First, an infrared thermal imager is used to acquire multiple high-temperature condensation points within the data center; based on the locations of multiple ventilation openings and high-temperature condensation points within the data center, a cooling control analysis method is used to obtain the optimal cooling time and optimal cooling outlet for each high-temperature condensation point; then, based on the optimal cooling outlet for each high-temperature condensation point, all high-temperature condensation points are screened, and multiple high-temperature condensation types and cooling time intervals for each high-temperature condensation type are obtained based on the screening results; finally, a deep learning model is established, and when a new high-temperature condensation point exists, the high-temperature condensation type corresponding to the new high-temperature condensation point is obtained based on the deep learning model; the latest high-temperature condensation point acquired by the infrared thermal imager is recorded as the real-time condensation point, and real-time cooling control of the data center is performed based on the cooling time interval corresponding to the high-temperature condensation type of the real-time condensation point.

[0108] Based on the above description of the embodiments, the embodiments of the present invention can be provided as methods, systems, or computer program products. Based on this understanding, the above technical solutions, in essence or in terms of their contribution to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or certain parts of the embodiments.

[0109] In the embodiments provided in this application, it should be understood that the disclosed system or method can be implemented in other ways. The embodiments described above are merely illustrative. For example, the division of modules or units is only a logical functional division, and there may be other division methods in actual implementation. Furthermore, multiple modules or units may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the coupling or direct coupling or communication connection shown or discussed may be through some communication interfaces. The indirect coupling or communication connection between systems, modules, and units may be electrical, mechanical, or other forms.

[0110] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A data center cooling control method based on deep learning, characterized in that, Includes the following steps: When all equipment in the data center is operating normally, an infrared thermal imager is used to detect the temperature of the data center and obtain multiple high-temperature condensation points within the data center; multiple ventilation openings within the data center are also obtained, and based on the locations of the ventilation openings and the locations of the high-temperature condensation points, a cooling control analysis method is used to obtain the optimal cooling time and optimal cooling outlet for each high-temperature condensation point. Based on the preferred cooling port of each high-temperature condensation point, all high-temperature condensation points are screened, and multiple high-temperature condensation types and cooling time intervals for each high-temperature condensation type are obtained based on the screening results. Build deep learning models; The refrigeration control analysis method is stored in a deep learning model, and the process of obtaining the high-temperature condensation type from the high-temperature condensation point is learned in the deep learning model. Based on the infrared thermal imager, high-temperature condensation points in the data center are acquired in real time. When a new high-temperature condensation point is found, its location is entered into the deep learning model, and the high-temperature condensation type corresponding to the new high-temperature condensation point is obtained based on the deep learning model. The latest high-temperature condensation point acquired by the infrared thermal imager is recorded as the real-time condensation point. The data center is then regulated in real-time based on the cooling time interval corresponding to the high-temperature condensation type of the real-time condensation point.

2. The data center cooling control method based on deep learning according to claim 1, characterized in that, With all equipment in the data center operating normally, an infrared thermal imager is used to detect the temperature of the data center and identify multiple high-temperature condensation points within the data center, including: Establish a spatial rectangular coordinate system, denoted as the spatial analysis coordinate system, where the units of the X-axis, Y-axis, and Z-axis of the spatial analysis coordinate system are all meters (m). Based on the design drawings of the data center and the dimensional data of all equipment in the data center, a scaled model of the interior of the data center is established in the first quadrant of the spatial analysis coordinate system and denoted as the data three-dimensional model. The dimensional data includes height data, width data and length data.

3. The data center cooling control method based on deep learning according to claim 2, characterized in that, When all equipment in the data center is operating normally, using an infrared thermal imager to detect the temperature of the data center and identify multiple high-temperature condensation points within the data center also includes: When all equipment in the data center is operating normally, use an infrared thermal imager to detect the internal temperature of the data center every kmin, and obtain the location of the highest temperature in the data center after the temperature detection, and mark it as the high temperature location; obtain all the high temperature locations obtained by the infrared thermal imager in the data center within a day, and mark all the high temperature locations in the data 3D model; Obtain the temperature corresponding to the high-temperature location in the temperature detection when all high-temperature locations are marked, and record it as the high-temperature temperature of the high-temperature location; record all high-temperature locations with a high temperature greater than the maximum temperature threshold as high-temperature condensation points, where the maximum temperature threshold is the maximum value of the normal temperature range within the data center.

4. The data center cooling control method based on deep learning according to claim 3, characterized in that, Multiple ventilation openings within the data center are identified, and based on the locations of these openings and high-temperature condensation points, a cooling control analysis method is used to determine the optimal cooling time and preferred cooling outlet for each high-temperature condensation point. Within the 3D data model: acquire multiple ventilation openings within the data center and label them sequentially as ventilation opening TF1 to ventilation opening TF2. c Based on the equipment usage specifications of the data center, obtain the maximum and minimum number of ventilation openings that can be opened simultaneously, denoted as m and n respectively, where m and n are positive integers less than or equal to c and greater than or equal to 1. Obtain all possible opening states for the ventilation vents within the data center, from n to m vents that can be opened simultaneously, and denote them as ventilation opening state TK1 to ventilation opening state TF. p ; The locations of all high-temperature condensation points were obtained, and the optimal cooling time and optimal cooling port for each high-temperature condensation point were determined using the refrigeration control analysis method.

5. The data center cooling control method based on deep learning according to claim 4, characterized in that, Refrigeration control analysis methods include: For any high-temperature condensation point A: when the temperature of high-temperature condensation point A is detected by the infrared thermal imager to be equal to the high-temperature temperature of high-temperature condensation point A, the cooling system in the data center is activated and based on the ventilation status TK. q Open the ventilation openings in the data center, use an infrared thermal imager to obtain the temperature at the location of the high-temperature condensation point A in real time, and start the timer, where q is a positive integer less than or equal to p and greater than or equal to 1; When the temperature of high-temperature condensation point A in the infrared thermal imager is less than or equal to the maximum temperature threshold, the timer is turned off, and the timer time is recorded as the cooling time.

6. The data center cooling control method based on deep learning according to claim 5, characterized in that, Refrigeration control analysis methods also include: When the temperature of high-temperature condensation point A is equal to the high temperature of high-temperature condensation point A, the cooling time corresponding to all ventilation opening conditions is obtained. The minimum cooling time is recorded as the preferred cooling time of high-temperature condensation point A, and all the vents opened in the ventilation opening conditions corresponding to the minimum cooling time are recorded as the preferred cooling vents of high-temperature condensation point A. Obtain the optimal cooling time and optimal cooling port corresponding to all high-temperature condensation points.

7. The data center cooling control method based on deep learning according to claim 6, characterized in that, All high-temperature condensation points are screened based on the preferred refrigeration port for each high-temperature condensation point, and multiple high-temperature condensation types and their respective refrigeration time intervals are obtained based on the screening results: For any high-temperature condensation point A: High-temperature condensation points other than high-temperature condensation point A whose preferred refrigeration port is exactly the same as the preferred refrigeration port of high-temperature condensation point A are recorded as condensation points of the same type as high-temperature condensation point A, and high-temperature condensation point A and condensation points of the same type as high-temperature condensation point A are recorded as the same high-temperature condensation type. Obtain multiple high-temperature condensation types corresponding to all high-temperature condensation points; for any high-temperature condensation type, denote the closed interval formed by the minimum and maximum values ​​of the preferred cooling times for all high-temperature condensation points corresponding to the high-temperature condensation type as the cooling time interval of the high-temperature condensation type.

8. The data center cooling control method based on deep learning according to claim 7, characterized in that, The latest high-temperature condensation point acquired by the infrared thermal imager is recorded as the real-time condensation point. Real-time cooling control of the data center is then performed based on the cooling time interval corresponding to the high-temperature condensation type of the real-time condensation point, including: When the real-time condensation point is obtained, the cooling system in the data center is started, and the preferred cooling ports of all high-temperature condensation points corresponding to the high-temperature condensation type of the real-time condensation point are started. Maintain the start-up time of the cooling system within the data center and the opening time of all open vents greater than T1.

9. The data center cooling control method based on deep learning according to claim 8, characterized in that, The latest high-temperature condensation point acquired by the infrared thermal imager is recorded as the real-time condensation point. Real-time cooling control of the data center based on the cooling time interval corresponding to the high-temperature condensation type of the real-time condensation point also includes: When the start-up time of the cooling system in the data center and the opening time of all open vents are less than or equal to T2, and the real-time condensation point temperature is less than or equal to the maximum temperature threshold, the cooling system in the data center is shut down. When the start-up time of the cooling system in the data center and the opening time of all open vents are greater than T2, and the temperature of the real-time condensation point is greater than the maximum temperature threshold, a temperature control anomaly report is made. Here, T1 is the minimum value in the cooling time interval of the high-temperature condensation type of the real-time condensation point, and T2 is the maximum value in the cooling time interval of the high-temperature condensation type of the real-time condensation point.

10. A deep learning-based data center cooling control system, used to implement the deep learning-based data center cooling control method according to any one of claims 1-9, characterized in that, It includes a high-temperature refrigeration analysis module, a high-temperature type screening module, and a deep real-time control module; The high-temperature cooling analysis module is used to detect the temperature of the data center using an infrared thermal imager when all equipment in the data center is operating normally, and to obtain multiple high-temperature condensation points in the data center; to obtain multiple ventilation openings in the data center, and to obtain the optimal cooling time and optimal cooling outlet for each high-temperature condensation point based on the location of the multiple ventilation openings and the location of the high-temperature condensation points using the cooling control analysis method. The high-temperature type screening module screens all high-temperature condensation points based on the preferred refrigeration port of each high-temperature condensation point, and obtains multiple high-temperature condensation types and the refrigeration time interval for each high-temperature condensation type based on the screening results. The deep real-time control module is used to build deep learning models; The refrigeration control analysis method is stored in a deep learning model, and the process of obtaining the high-temperature condensation type from the high-temperature condensation point is learned in the deep learning model. Based on the infrared thermal imager, high-temperature condensation points in the data center are acquired in real time. When a new high-temperature condensation point is found, its location is entered into the deep learning model, and the high-temperature condensation type corresponding to the new high-temperature condensation point is obtained based on the deep learning model. The latest high-temperature condensation point acquired by the infrared thermal imager is recorded as the real-time condensation point. The data center is then regulated in real-time based on the cooling time interval corresponding to the high-temperature condensation type of the real-time condensation point.

Images