Data center digital twin refrigeration method and system based on load sensing

Abstract

The invention belongs to the field of data center refrigeration, and provides a data center digital twin refrigeration method and system based on load sensing. The data center digital twin refrigeration method based on load sensing comprises the steps that the previous moment temperature and the current moment temperature of an air outlet of a server, the predicted power change condition of the server in a future set time period, a first temperature threshold value and a second temperature threshold value are obtained; wherein the second temperature threshold value is greater than the first temperature threshold value; the first temperature threshold value and the second temperature threshold value are both related to server attributes; based on the obtained information, a temperature control strategy is determined, and a control command is issued to the corresponding air conditioner so as to adjust the air outlet temperature of the corresponding air conditioner; the terminal air conditioner temperature control strategy can be obtained according to the current temperature of the air outlet of the server and the future power change condition, efficient and intelligent refrigeration is achieved, the refrigeration efficiency is improved, and local hot spots of the server are avoided.

Description

Technical Field

[0001] The present invention belongs to the field of data center refrigeration, and in particular relates to a data center digital twin refrigeration method and system based on load perception. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] As core facilities for data storage, processing and transmission, data centers are growing in size and complexity. Thousands of servers are usually deployed in data centers, which generate a lot of heat during operation and require an effective cooling system to maintain a suitable working environment.

[0004] Traditional refrigeration systems usually include two refrigeration modes: the first refrigeration mode is to adjust the module air conditioning based on a fixed temperature threshold or time period. This mode has a response lag and low efficiency. It cannot respond to changes in server load and temperature in real time, and cannot accurately cool according to the specific load and temperature requirements of the server. As a result, the refrigeration system may overwork in some periods, and may not cool enough in other periods, which may easily cause energy waste and even affect the performance and life of the server. The second refrigeration mode does not consider future load changes of the server, and only adjusts the outlet temperature of the terminal air conditioner according to the current temperature of the server. Although this method improves the refrigeration efficiency to a certain extent, due to the lack of a prediction mechanism for future load and temperature changes, the air conditioning temperature is adjusted too frequently, increasing the control complexity of the system. At the same time, frequent temperature adjustment not only increases the energy consumption of the refrigeration system, but may also cause additional wear and tear on the air conditioning equipment and shorten its service life. In addition, due to the inability to predict and respond to load changes in advance, some server areas may experience local overheating, affecting the overall heat dissipation effect. Summary of the invention

[0005] In order to solve the technical problems existing in the above-mentioned background technology, the present invention provides a digital twin cooling method and system for a data center based on load perception, which can obtain the terminal air-conditioning temperature control strategy according to the current temperature of the server outlet and future power changes, realize efficient and intelligent refrigeration, improve refrigeration efficiency, and avoid local hot spots of the server.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: A first aspect of the present invention provides a data center digital twin cooling method based on load sensing.

[0007] A digital twin cooling method for a data center based on load perception, comprising: Obtain the previous temperature and current temperature of the air outlet of the server, the predicted power change of the server within a future set time period, the first temperature threshold and the second temperature threshold; wherein the second temperature threshold is greater than the first temperature threshold; the first temperature threshold and the second temperature threshold are both related to the server attributes; Based on the above information, determine the temperature control strategy and send a control command to the corresponding air conditioner to adjust the air outlet temperature of the corresponding air conditioner; Among them, the temperature control strategy is: When the current temperature of the server outlet is higher than the temperature at the previous moment, the current temperature of the server outlet exceeds the first temperature threshold, and the power of the server is predicted to remain unchanged or increase within a future set time period, the temperature of the terminal air conditioner closest to the server is reduced; When the current temperature of the server outlet is higher than the previous temperature, and the power of the server is predicted to decrease in the future set time period, whether to reduce the temperature of the terminal air conditioner closest to the server is determined according to whether the current temperature of the server outlet exceeds the second temperature threshold; When the current temperature of the air outlet of the server is lower than the first temperature threshold, the temperature of the terminal air conditioner closest to the server is controlled to remain unchanged.

[0008] As an implementation of the first aspect, if the current temperature of the server air outlet exceeds the second temperature threshold, the temperature of the terminal air conditioner closest to the server is reduced; if the current temperature of the server air outlet is less than or equal to the second temperature threshold, the temperature of the terminal air conditioner closest to the server is controlled to remain unchanged.

[0009] As an implementation of the first aspect, the process of predicting the power change of the server within a future set time period is as follows: Based on the pre-built relationship model between the distance between the server and the terminal air conditioner and the air outlet temperature of the air conditioner, and the air inlet temperature of the server, the air inlet temperature of the server is calculated according to the distance between the server and the terminal air conditioner and the air outlet temperature of the air conditioner; Based on the server air outlet temperature and the calculated server air inlet temperature, the server power is calculated using a pre-built relationship model between the server power and the server air inlet and server air outlet temperatures; Based on the server power sequence within the set historical time period and the pre-trained power change prediction model, the power change of the server within the future set time period is predicted.

[0010] As an implementation manner of the first aspect, the first temperature threshold is set to T1, and its expression is: T1= T max +α1* (T min -Tmax ); Among them, T max is the upper temperature limit of the server; T min is the lower limit of server temperature; α1 is the first adjustment coefficient, 0<α1<1, which is determined by the server properties.

[0011] As an implementation manner of the first aspect, the second temperature threshold is set to T2, and its expression is: T2= T1+α2* (T min -T max ); Wherein, α2 is the second adjustment coefficient, 0<α2<1, which is determined by the server attributes.

[0012] A second aspect of the present invention provides a digital twin refrigeration system for a data center based on load sensing.

[0013] A digital twin cooling system for a data center based on load sensing, comprising: A server-related information acquisition module, which is used to obtain the previous and current temperatures of the server air outlet, the predicted power change of the server within a future set time period, the first temperature threshold and the second temperature threshold; wherein the second temperature threshold is greater than the first temperature threshold; the first temperature threshold and the second temperature threshold are both related to server attributes; A temperature control strategy determination module, which is used to determine the temperature control strategy based on the above-obtained information and issue a control command to the corresponding air conditioner to adjust the air outlet temperature of the corresponding air conditioner; Among them, the temperature control strategy is: When the current temperature of the server outlet is higher than the temperature at the previous moment, the current temperature of the server outlet exceeds the first temperature threshold, and the power of the server is predicted to remain unchanged or increase within a future set time period, the temperature of the terminal air conditioner closest to the server is reduced; When the current temperature of the server outlet is higher than the previous temperature, and the power of the server is predicted to decrease in the future set time period, whether to reduce the temperature of the terminal air conditioner closest to the server is determined according to whether the current temperature of the server outlet exceeds the second temperature threshold; When the current temperature of the air outlet of the server is lower than the first temperature threshold, the temperature of the terminal air conditioner closest to the server is controlled to remain unchanged.

[0014] As an implementation of the second aspect, in the temperature control strategy, if the current temperature of the server air outlet exceeds the second temperature threshold, the temperature of the terminal air conditioner closest to the server is reduced; if the current temperature of the server air outlet is less than or equal to the second temperature threshold, the temperature of the terminal air conditioner closest to the server is controlled to remain unchanged.

[0015] As an implementation of the second aspect, the first temperature threshold is set to T1, and its expression is: T1= T max +α1* (T min -T max ); Among them, T max is the upper temperature limit of the server; T min is the lower limit of server temperature; α1 is the first adjustment coefficient, 0<α1<1, which is determined by the server attributes; The second temperature threshold is set to T2, and its expression is: T2= T1+α2* (T min -T max ); Wherein, α2 is the second adjustment coefficient, 0<α2<1, which is determined by the server attributes.

[0016] The third aspect of the present invention provides another data center digital twin refrigeration system based on load sensing.

[0017] A digital twin refrigeration system for a data center based on load sensing, comprising: A data collection module is used to collect the server air outlet temperature, server air inlet temperature, server power and terminal air conditioner outlet temperature and store them in a database; A data analysis module is used to predict the power change of the server within a future set time period based on historical data in the database; and determine the temperature control strategy based on the previous temperature and current temperature of the server outlet, the predicted power change of the server within a future set time period, the first temperature threshold and the second temperature threshold; A decision control module, which is used to send a control command to the corresponding air conditioner according to the determined temperature control strategy to adjust the air outlet temperature of the corresponding air conditioner; Among them, the temperature control strategy is: When the current temperature of the server outlet is higher than the temperature at the previous moment, the current temperature of the server outlet exceeds the first temperature threshold, and the power of the server is predicted to remain unchanged or increase within a future set time period, the temperature of the terminal air conditioner closest to the server is reduced; When the current temperature of the server outlet is higher than the previous temperature, and the power of the server is predicted to decrease in the future set time period, whether to reduce the temperature of the terminal air conditioner closest to the server is determined according to whether the current temperature of the server outlet exceeds the second temperature threshold; When the current temperature of the air outlet of the server is lower than the first temperature threshold, the temperature of the terminal air conditioner closest to the server is controlled to remain unchanged.

[0018] A fourth aspect of the present invention provides a computer device.

[0019] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps in the load-aware digital twin cooling method for a data center are implemented as described above.

[0020] Compared with the prior art, the present invention has the following beneficial effects: (1) The present invention proposes a digital twin cooling method and system for a data center based on load perception, which takes into account the temperature of the server outlet at the previous moment and the current moment and the predicted power change of the server within a set future time period, and then determines the temperature control strategy and sends a control command to the corresponding air conditioner through the comparison result of the current moment temperature of the server outlet and the first temperature threshold and the second temperature threshold, thereby achieving the purpose of adjusting the outlet temperature of the corresponding air conditioner, realizing accurate perception and prediction of server load and temperature changes, thereby formulating a personalized temperature control strategy, improving cooling efficiency and reducing energy consumption.

[0021] (2) The present invention constructs a digital twin model between the distance between the server and the terminal air conditioner and the air outlet temperature of the air conditioner, the digital twin model between the two and the server air inlet temperature, and the digital twin model between the server power and the server air inlet and server air outlet temperatures, calculates the server power, and then combines the power change prediction model to predict the power change of the server in the future, providing an accurate data basis for subsequent temperature control strategies.

[0022] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0024] Figure 1 This is a flow chart of a data center digital twin cooling method based on load perception according to the first embodiment of the present invention; Figure 2 It is a process of predicting the power change of the server within a future set time period of an embodiment of the present invention; Figure 3 Schematic diagram of the structure of a digital twin refrigeration system for a data center based on load perception according to the second embodiment of the present invention; Figure 4 Schematic diagram of the structure of a digital twin refrigeration system for a data center based on load perception according to the third embodiment of the present invention; Figure 5 This is a working principle diagram of the load-aware data center digital twin refrigeration system of embodiment 3 of the present invention. DETAILED DESCRIPTION

[0025] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0026] It should be noted that the following detailed descriptions are all illustrative and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0027] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0028] Embodiment 1 Figure 1 This is a flow chart of a load-aware data center digital twin cooling method according to an embodiment of the present invention. Figure 1 , a data center digital twin cooling method based on load perception according to an embodiment of the present invention includes: S100: Obtain the previous temperature and current temperature of the server outlet, the predicted power change of the server within a future set time period, a first temperature threshold and a second temperature threshold; wherein the second temperature threshold is greater than the first temperature threshold; the first temperature threshold and the second temperature threshold are both related to server attributes.

[0029] S200: Based on the information obtained above, determine a temperature control strategy and send a control command to the corresponding air conditioner to adjust the air outlet temperature of the corresponding air conditioner.

[0030] In this embodiment, the temperature control strategy is: (1) When the current temperature of the server outlet is higher than the temperature at the previous moment, the current temperature of the server outlet exceeds the first temperature threshold, and the power of the server is predicted to remain unchanged or increase within a future set time period, the temperature of the terminal air conditioner closest to the server is reduced.

[0031] (2) When the current temperature of the server air outlet is higher than the temperature at the previous moment, and the power of the server is predicted to decrease within a set future time period, whether to reduce the temperature of the terminal air conditioner closest to the server is determined based on whether the current temperature of the server air outlet exceeds the second temperature threshold. If the current temperature of the server air outlet exceeds the second temperature threshold, reduce the temperature of the terminal air conditioner closest to the server. If the current temperature of the server air outlet is less than or equal to the second temperature threshold, control the temperature of the terminal air conditioner closest to the server to remain unchanged.

[0032] (3) When the current temperature of the server outlet is lower than the first temperature threshold, the temperature of the terminal air conditioner closest to the server is controlled to remain unchanged.

[0033] In step S100, Figure 2 As shown in the figure, the prediction process of the power change of the server in the future set time period is: S1001: Based on the pre-built distance between the server and the terminal air conditioner and the air outlet temperature of the air conditioner, and the relationship model between the two and the air inlet temperature of the server, the air inlet temperature of the server is calculated according to the distance between the server and the terminal air conditioner and the air outlet temperature of the air conditioner.

[0034] Specifically, based on the historical data of the distance between the server and the terminal air conditioner and the air outlet temperature of the air conditioner, and the air inlet temperature of the server, a relationship model of the distance between the server and the terminal air conditioner and the air outlet temperature of the air conditioner and the air inlet temperature of the server is fitted, and its expression can be: T in = h ( T air_con , d ).

[0035] in, T air_con Indicates the outlet air temperature of the terminal air conditioner. T in Indicates the server air inlet temperature. d Indicates the distance between the server and the terminal air conditioner. h Represents the relationship model between the distance between the server and the terminal air conditioner, the air outlet temperature of the air conditioner, and the air inlet temperature of the server; Relational Model h The expression is as follows:

[0036]

[0037] in, All are fitting coefficients.

[0038] S1002: Based on the server air outlet temperature and the calculated server air inlet temperature, the server power is calculated using a pre-built relationship model between the server power and the server air inlet and server air outlet temperatures.

[0039] Specifically, the historical data of server power and the temperatures of the server air inlet and server air outlet are used to construct a relationship model between the server power and the temperatures of the server air inlet and server air outlet.

[0040] The relationship model formula between the server power and the server air inlet and server air outlet temperatures is as follows: T out = f (P, T in ) in, T out Indicates the server outlet temperature. P Indicates the server power, T in Indicates the server air inlet temperature. f A model representing the relationship between server power and the temperature of the server air inlet and server air outlet; In actual environments, the relationship between the server power, air inlet temperature, and air outlet temperature is relatively complex and is affected by multiple factors such as heat conduction, air flow, and material properties. Therefore, a nonlinear function is used to fit the relationship. f The specific fitting formula is as follows:

[0041] in, , All are fitting coefficients.

[0042] S1003: Predict the power change of the server in a future set time period (eg, the next hour, etc.) based on the server power sequence in a set historical time period and a pre-trained power change prediction model.

[0043] For time series prediction, some lag features need to be generated to capture time dependencies. The power change of the server at the current moment may be affected by its own power level in the previous few minutes or tens of minutes and related environmental factors (such as temperature, etc.). Lag features mainly introduce data from different time points in the past as additional features into the analysis model to characterize this temporal dependency.

[0044] Specifically, in the process of training the power change prediction model, first, collect the historical power data of the server, and the collected data includes time points and power values. For example, the power is collected every 10 minutes, and a data set is constructed based on the time points and power values; then use the constructed data set to train the power change prediction model.

[0045] In some optional embodiments, the power change prediction model may be implemented using an ARIMA (Auto Regressive Integrated Moving Average) model.

[0046] In other embodiments, the power change prediction model may also be implemented using an existing neural network model, which will not be described in detail here.

[0047] In step S100, the first temperature threshold is set to T1, which is expressed as: T1= T max +α1* (T min -T max ); Among them, T max is the upper temperature limit of the server; T min is the lower limit of the server temperature; α1 is the first adjustment coefficient, 0<α1<1, which is determined by the server attributes. The server attributes include but are not limited to the server model and workload characteristics.

[0048] For example, if the server is relatively sensitive to heat dissipation and it is hoped that measures can be taken as soon as possible to deal with the temperature rise, α1 can take a smaller value, such as 0.1, which means that T1 is set to a value higher than the upper limit of the normal operating temperature T max 10% higher (T min -T max ) temperature difference; if the server has better heat dissipation performance and is more tolerant to temperature changes, α1 can take a relatively larger value, such as 0.3.

[0049] The second temperature threshold is set to T2, and its expression is: T2= T1+α2* (T min -T max ); Wherein, α2 is the second adjustment coefficient, 0<α2<1, which is determined by the server attributes.

[0050] T2 is generally larger than T1, for example, it can be in the range of 0.2 - 0.5, which means that T2 is higher than T1 by a certain percentage of the normal operating temperature range of the server, so as to reflect the response strategy under higher temperature risks. At the same time, combined with the temperature change of the server when the load is suddenly reduced after a period of high load (through experimental testing to record the outlet temperature change curve under different load switching scenarios), observe the rate of temperature drop and whether it is easy to have abnormal conditions such as overheating, so as to further adjust the value of α2 and the specific value of T2.

[0051] T2 is set above the upper limit of the server's normal operating temperature range to ensure rapid response and avoid overheating risks when the server heat dissipation is extremely high. For example, if T1 is set at 25°C, T2 may be set at 27°C or higher.

[0052] It should be noted here that the setting values ​​of T1 and T2 are not fixed, but need to be adjusted and optimized according to actual conditions. In addition, the setting of these threshold temperatures also needs to match the performance and control strategy of the refrigeration system to ensure that the system can operate stably and achieve the expected energy-saving effect.

[0053] Embodiment 2 like Figure 3 As shown, an embodiment of the present invention provides a data center digital twin refrigeration system based on load perception, including: The server-related information acquisition module 301 is used to obtain the previous temperature and current temperature of the server outlet, the predicted power change of the server within a future set time period, the first temperature threshold and the second temperature threshold; wherein the second temperature threshold is greater than the first temperature threshold; the first temperature threshold and the second temperature threshold are both related to the server attributes.

[0054] The temperature control strategy determination module 302 is used to determine the temperature control strategy based on the above-obtained information and issue a control command to the corresponding air conditioner to adjust the air outlet temperature of the corresponding air conditioner.

[0055] In this embodiment, the temperature control strategy is: (1) When the current temperature of the server outlet is higher than the temperature at the previous moment, the current temperature of the server outlet exceeds the first temperature threshold, and the power of the server is predicted to remain unchanged or increase within a future set time period, the temperature of the terminal air conditioner closest to the server is reduced.

[0056] (2) When the current temperature of the server air outlet is higher than the temperature at the previous moment, and the power of the server is predicted to decrease within a set future time period, whether to reduce the temperature of the terminal air conditioner closest to the server is determined based on whether the current temperature of the server air outlet exceeds the second temperature threshold. If the current temperature of the server air outlet exceeds the second temperature threshold, reduce the temperature of the terminal air conditioner closest to the server. If the current temperature of the server air outlet is less than or equal to the second temperature threshold, control the temperature of the terminal air conditioner closest to the server to remain unchanged.

[0057] (3) When the current temperature of the server outlet is lower than the first temperature threshold, the temperature of the terminal air conditioner closest to the server is controlled to remain unchanged.

[0058] In the temperature control strategy, if the current temperature of the server air outlet exceeds the second temperature threshold, the temperature of the terminal air conditioner closest to the server is reduced; if the current temperature of the server air outlet is less than or equal to the second temperature threshold, the temperature of the terminal air conditioner closest to the server is controlled to remain unchanged.

[0059] In the server related information acquisition module 301, as Figure 2 As shown in the figure, the prediction process of the power change of the server in the future set time period is: S1001: Based on the pre-built distance between the server and the terminal air conditioner and the air outlet temperature of the air conditioner, and the relationship model between the two and the air inlet temperature of the server, the air inlet temperature of the server is calculated according to the distance between the server and the terminal air conditioner and the air outlet temperature of the air conditioner.

[0060] Specifically, based on the historical data of the distance between the server and the terminal air conditioner and the air outlet temperature of the air conditioner, and the air inlet temperature of the server, a relationship model of the distance between the server and the terminal air conditioner and the air outlet temperature of the air conditioner and the air inlet temperature of the server is fitted, and its expression can be: T in = h ( T air_con , d ).

[0061] in, T air_con Indicates the outlet air temperature of the terminal air conditioner. T in Indicates the server air inlet temperature. d Indicates the distance between the server and the terminal air conditioner. h Represents the relationship model between the distance between the server and the terminal air conditioner, the air outlet temperature of the air conditioner, and the air inlet temperature of the server; Relational Model h The expression is as follows:

[0062]

[0063] in, All are fitting coefficients.

[0064] S1002: Based on the server air outlet temperature and the calculated server air inlet temperature, the server power is calculated using a pre-built relationship model between the server power and the server air inlet and server air outlet temperatures.

[0065] Specifically, the historical data of server power and the temperatures of the server air inlet and server air outlet are used to construct a relationship model between the server power and the temperatures of the server air inlet and server air outlet.

[0066] The relationship model formula between the server power and the server air inlet and server air outlet temperatures is as follows: T out = f (P, T in ) in, T out Indicates the server outlet temperature. P Indicates the server power, T in Indicates the server air inlet temperature. f A model representing the relationship between server power and the temperature of the server air inlet and server air outlet; In actual environments, the relationship between the server power, air inlet temperature, and air outlet temperature is relatively complex and is affected by multiple factors such as heat conduction, air flow, and material properties. Therefore, a nonlinear function is used to fit the relationship. f The specific fitting formula is as follows:

[0067] in, , All are fitting coefficients.

[0068] S1003: Predict the power change of the server in a future set time period (eg, the next hour, etc.) based on the server power sequence in a set historical time period and a pre-trained power change prediction model.

[0069] For time series prediction, some lag features need to be generated to capture time dependencies. The power change of the server at the current moment may be affected by its own power level in the previous few minutes or tens of minutes and related environmental factors (such as temperature, etc.). Lag features mainly introduce data from different time points in the past as additional features into the analysis model to characterize this temporal dependency.

[0070] Specifically, in the process of training the power change prediction model, first, collect the historical power data of the server, and the collected data includes time points and power values. For example, the power is collected every 10 minutes, and a data set is constructed based on the time points and power values; then use the constructed data set to train the power change prediction model.

[0071] In some optional embodiments, the power change prediction model may be implemented using an ARIMA (Auto Regressive Integrated Moving Average) model.

[0072] In other embodiments, the power change prediction model may also be implemented using an existing neural network model, which will not be described in detail here.

[0073] The first temperature threshold is set to T1, and its expression is: T1= T max +α1* (T min -T max ); Among them, T max is the upper temperature limit of the server; T min is the lower limit of the server temperature; α1 is the first adjustment coefficient, 0<α1<1, which is determined by the server attributes. The server attributes include but are not limited to the server model and workload characteristics.

[0074] For example, if the server is relatively sensitive to heat dissipation and it is hoped that measures can be taken as soon as possible to deal with the temperature rise, α1 can take a smaller value, such as 0.1, which means that T1 is set to a value higher than the upper limit of the normal operating temperature T max 10% higher (T min -T max ) temperature difference; if the server has better heat dissipation performance and is more tolerant to temperature changes, α1 can take a relatively larger value, such as 0.3.

[0075] The second temperature threshold is set to T2, and its expression is: T2= T1+α2* (T min -T max ); Wherein, α2 is the second adjustment coefficient, 0<α2<1, which is determined by the server attributes.

[0076] T2 is generally larger than T1, for example, it can be in the range of 0.2 - 0.5, which means that T2 is higher than T1 by a certain percentage of the normal operating temperature range of the server, so as to reflect the response strategy under higher temperature risks. At the same time, combined with the temperature change of the server when the load is suddenly reduced after a period of high load (through experimental testing to record the outlet temperature change curve under different load switching scenarios), observe the rate of temperature drop and whether it is easy to have abnormal conditions such as overheating, so as to further adjust the value of α2 and the specific value of T2.

[0077] T2 is set above the upper limit of the server's normal operating temperature range to ensure rapid response and avoid overheating risks when the server heat dissipation is extremely high. For example, if T1 is set at 25°C, T2 may be set at 27°C or higher.

[0078] It should be noted here that the setting values ​​of T1 and T2 are not fixed, but need to be adjusted and optimized according to actual conditions. In addition, the setting of these threshold temperatures also needs to match the performance and control strategy of the refrigeration system to ensure that the system can operate stably and achieve the expected energy-saving effect.

[0079] Embodiment 3 like Figure 4 As shown, this embodiment provides a data center digital twin refrigeration system based on load perception, which includes: The data collection module 401 is used to collect the server outlet temperature, server inlet temperature, server power and terminal air conditioner outlet temperature and store them in the database.

[0080] For example, the server power, server air inlet and outlet temperatures are monitored through the server BCM interface, and the terminal air conditioning outlet temperature is collected at the same time.

[0081] The data analysis module 402 is used to predict the power change of the server within a future set time period based on the historical data in the database; and determine the temperature control strategy and the temperature control strategy based on the previous temperature and current temperature of the server outlet, the predicted power change of the server within a future set time period, the first temperature threshold and the second temperature threshold. Figure 5 shown.

[0082] like Figure 2 As shown in the figure, the prediction process of the power change of the server in the future set time period is: S1001: Based on the pre-built distance between the server and the terminal air conditioner and the air outlet temperature of the air conditioner, and the relationship model between the two and the air inlet temperature of the server, the air inlet temperature of the server is calculated according to the distance between the server and the terminal air conditioner and the air outlet temperature of the air conditioner.

[0083] Specifically, based on the historical data of the distance between the server and the terminal air conditioner and the air outlet temperature of the air conditioner, and the air inlet temperature of the server, a relationship model of the distance between the server and the terminal air conditioner and the air outlet temperature of the air conditioner and the air inlet temperature of the server is fitted, and its expression can be: T in = h ( T air_con , d ).

[0084] in, T air_con Indicates the outlet air temperature of the terminal air conditioner. T in Indicates the server air inlet temperature. d Indicates the distance between the server and the terminal air conditioner. h Represents the relationship model between the distance between the server and the terminal air conditioner, the air outlet temperature of the air conditioner, and the air inlet temperature of the server; Relational Model h The expression is as follows:

[0085]

[0086] in, All are fitting coefficients.

[0087] S1002: Based on the server air outlet temperature and the calculated server air inlet temperature, the server power is calculated using a pre-built relationship model between the server power and the server air inlet and server air outlet temperatures.

[0088] Specifically, the historical data of server power and the temperatures of the server air inlet and server air outlet are used to construct a relationship model between the server power and the temperatures of the server air inlet and server air outlet.

[0089] The relationship model formula between the server power and the server air inlet and server air outlet temperatures is as follows: T out = f (P, T in ) in, T out Indicates the server outlet temperature. P Indicates the server power, T in Indicates the server air inlet temperature. f A model representing the relationship between server power and the temperature of the server air inlet and server air outlet; In actual environments, the relationship between the server power, air inlet temperature, and air outlet temperature is relatively complex and is affected by multiple factors such as heat conduction, air flow, and material properties. Therefore, a nonlinear function is used to fit the relationship. f The specific fitting formula is as follows:

[0090] in, , All are fitting coefficients.

[0091] S1003: Predict the power change of the server in a future set time period (eg, the next hour, etc.) based on the server power sequence in a set historical time period and a pre-trained power change prediction model.

[0092] For time series prediction, some lag features need to be generated to capture time dependencies. The power change of the server at the current moment may be affected by its own power level in the previous few minutes or tens of minutes and related environmental factors (such as temperature, etc.). Lag features mainly introduce data from different time points in the past as additional features into the analysis model to characterize this temporal dependency.

[0093] Specifically, in the process of training the power change prediction model, first, collect the historical power data of the server, and the collected data includes time points and power values. For example, the power is collected every 10 minutes, and a data set is constructed based on the time points and power values; then use the constructed data set to train the power change prediction model.

[0094] In some optional embodiments, the power change prediction model may be implemented using an ARIMA (Auto Regressive Integrated Moving Average) model.

[0095] In other embodiments, the power change prediction model may also be implemented using an existing neural network model, which will not be described in detail here.

[0096] The first temperature threshold is set to T1, and its expression is: T1= T max +α1* (T min -T max ); Among them, T max is the upper temperature limit of the server; T min is the lower limit of the server temperature; α1 is the first adjustment coefficient, 0<α1<1, which is determined by the server attributes. The server attributes include but are not limited to the server model and workload characteristics.

[0097] For example, if the server is relatively sensitive to heat dissipation and it is hoped that measures can be taken as soon as possible to deal with the temperature rise, α1 can take a smaller value, such as 0.1, which means that T1 is set to a value higher than the upper limit of the normal operating temperature T max 10% higher (T min -T max ) temperature difference; if the server has better heat dissipation performance and is more tolerant to temperature changes, α1 can take a relatively larger value, such as 0.3.

[0098] The second temperature threshold is set to T2, and its expression is: T2= T1+α2* (T min -T max ); Wherein, α2 is the second adjustment coefficient, 0<α2<1, which is determined by the server attributes.

[0099] T2 is generally larger than T1, for example, it can be in the range of 0.2 - 0.5, which means that T2 is higher than T1 by a certain percentage of the normal operating temperature range of the server, so as to reflect the response strategy under higher temperature risks. At the same time, combined with the temperature change of the server when the load is suddenly reduced after a period of high load (through experimental testing to record the outlet temperature change curve under different load switching scenarios), observe the rate of temperature drop and whether it is easy to have abnormal conditions such as overheating, so as to further adjust the value of α2 and the specific value of T2.

[0100] T2 is set above the upper limit of the server's normal operating temperature range to ensure rapid response and avoid overheating risks when the server heat dissipation is extremely high. For example, if T1 is set at 25°C, T2 may be set at 27°C or higher.

[0101] It should be noted here that the setting values ​​of T1 and T2 are not fixed, but need to be adjusted and optimized according to actual conditions. In addition, the setting of these threshold temperatures also needs to match the performance and control strategy of the refrigeration system to ensure that the system can operate stably and achieve the expected energy-saving effect.

[0102] In this embodiment, the temperature control strategy is: (1) When the current temperature of the server outlet is higher than the temperature at the previous moment, the current temperature of the server outlet exceeds the first temperature threshold, and the power of the server is predicted to remain unchanged or increase within a future set time period, the temperature of the terminal air conditioner closest to the server is reduced.

[0103] (2) When the current temperature of the server air outlet is higher than the temperature at the previous moment, and the power of the server is predicted to decrease within a set future time period, whether to reduce the temperature of the terminal air conditioner closest to the server is determined based on whether the current temperature of the server air outlet exceeds the second temperature threshold. If the current temperature of the server air outlet exceeds the second temperature threshold, reduce the temperature of the terminal air conditioner closest to the server. If the current temperature of the server air outlet is less than or equal to the second temperature threshold, control the temperature of the terminal air conditioner closest to the server to remain unchanged.

[0104] (3) When the current temperature of the server outlet is lower than the first temperature threshold, the temperature of the terminal air conditioner closest to the server is controlled to remain unchanged.

[0105] The decision control module 403 is used to send a control command to the corresponding air conditioner according to the determined temperature control strategy to adjust the air outlet temperature of the corresponding air conditioner.

[0106] The advantages of the present invention are mainly reflected in the following four aspects: Precision cooling: The present invention fully considers the current power and operating temperature of the server and predicts the future power changes of the server. According to the future power changes of the server, a temperature control strategy of the refrigeration system is formed to control the air outlet temperature of the terminal air conditioner without causing additional system control energy consumption. At the same time, it realizes precise cooling based on load perception, which greatly improves the cooling efficiency.

[0107] High efficiency and energy saving: By accurately controlling the outlet temperature and inlet and return water temperature of the terminal air conditioner, the present invention realizes accurate cooling based on load perception, greatly improving the cooling efficiency. At the same time, since the temperature control frequency is moderate, it will not cause additional system control energy consumption, thus achieving the goal of high efficiency and energy saving.

[0108] Stable and reliable: The present invention uses a digital twin model to monitor and predict server load and temperature in real time, which can promptly detect and respond to potential overheating risks, ensure a stable data center environment, and protect server performance and data security.

[0109] Intelligent management: The present invention introduces intelligent management means, which can make adaptive adjustments according to the actual situation of the data center, thereby improving the flexibility and response speed of the data center in dealing with different load and temperature requirements.

[0110] Embodiment 4 This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps in the load-aware data center digital twin cooling method as described in the first embodiment are implemented.

[0111] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) containing computer-usable program code.

[0112] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0113] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0114] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A digital twin cooling method for a data center based on load perception, characterized in that: include: Obtain the previous temperature and current temperature of the air outlet of the server, the predicted power change of the server within a future set time period, the first temperature threshold and the second temperature threshold; wherein the second temperature threshold is greater than the first temperature threshold; the first temperature threshold and the second temperature threshold are both related to the server attributes; Based on the above information, determine the temperature control strategy and send a control command to the corresponding air conditioner to adjust the air outlet temperature of the corresponding air conditioner; Among them, the temperature control strategy is: When the current temperature of the server outlet is higher than the temperature at the previous moment, the current temperature of the server outlet exceeds the first temperature threshold, and the power of the server is predicted to remain unchanged or increase within a future set time period, the temperature of the terminal air conditioner closest to the server is reduced; When the current temperature of the server outlet is higher than the previous temperature, and the power of the server is predicted to decrease in the future set time period, whether to reduce the temperature of the terminal air conditioner closest to the server is determined according to whether the current temperature of the server outlet exceeds the second temperature threshold; When the current temperature of the air outlet of the server is lower than the first temperature threshold, the temperature of the terminal air conditioner closest to the server is controlled to remain unchanged.

2. The load-aware data center digital twin cooling method according to claim 1, characterized in that: If the current temperature of the server outlet exceeds the second temperature threshold, the temperature of the terminal air conditioner closest to the server is lowered; if the current temperature of the server outlet is less than or equal to the second temperature threshold, the temperature of the terminal air conditioner closest to the server is controlled to remain unchanged.

3. The load-aware data center digital twin cooling method according to claim 1, characterized in that: The prediction process of the server power change within the future set time period is as follows: Based on the pre-built relationship model between the distance between the server and the terminal air conditioner and the air outlet temperature of the air conditioner, and the air inlet temperature of the server, the air inlet temperature of the server is calculated according to the distance between the server and the terminal air conditioner and the air outlet temperature of the air conditioner; Based on the server air outlet temperature and the calculated server air inlet temperature, the server power is calculated using a pre-built relationship model between the server power and the server air inlet and server air outlet temperatures; Based on the server power sequence within the set historical time period and the pre-trained power change prediction model, the power change of the server within the future set time period is predicted.

4. The load-aware data center digital twin cooling method according to claim 1, characterized in that: The first temperature threshold is set to T1, and its expression is: T1= T max +α1* (T min -T max ); Among them, T max is the upper temperature limit of the server; T min is the lower limit of server temperature; α1 is the first adjustment coefficient, 0<α1<1, which is determined by the server properties.

5. The load-aware digital twin cooling method for a data center according to claim 4, characterized in that: The second temperature threshold is set to T2, and its expression is: T2= T1+α2* (T min -T max ); Wherein, α2 is the second adjustment coefficient, 0<α2<1, which is determined by the server attributes.

6. A digital twin cooling system for a data center based on load perception, characterized in that: include: A server-related information acquisition module, which is used to obtain the previous and current temperatures of the server air outlet, the predicted power change of the server within a future set time period, the first temperature threshold and the second temperature threshold; wherein the second temperature threshold is greater than the first temperature threshold; the first temperature threshold and the second temperature threshold are both related to server attributes; A temperature control strategy determination module, which is used to determine the temperature control strategy based on the above-obtained information and issue a control command to the corresponding air conditioner to adjust the air outlet temperature of the corresponding air conditioner; Among them, the temperature control strategy is: When the current temperature of the server outlet is higher than the temperature at the previous moment, the current temperature of the server outlet exceeds the first temperature threshold, and the power of the server is predicted to remain unchanged or increase within a future set time period, the temperature of the terminal air conditioner closest to the server is reduced; When the current temperature of the server outlet is higher than the previous temperature, and the power of the server is predicted to decrease in the future set time period, whether to reduce the temperature of the terminal air conditioner closest to the server is determined according to whether the current temperature of the server outlet exceeds the second temperature threshold; When the current temperature of the air outlet of the server is lower than the first temperature threshold, the temperature of the terminal air conditioner closest to the server is controlled to remain unchanged.

7. The load-aware digital twin refrigeration system for a data center according to claim 6, characterized in that: In the temperature control strategy, if the current temperature of the server air outlet exceeds the second temperature threshold, the temperature of the terminal air conditioner closest to the server is reduced; if the current temperature of the server air outlet is less than or equal to the second temperature threshold, the temperature of the terminal air conditioner closest to the server is controlled to remain unchanged.

8. The load-aware digital twin refrigeration system for a data center according to claim 6, characterized in that: The first temperature threshold is set to T1, and its expression is: T1= T max +α1* (T min -T max ); Among them, T max is the upper temperature limit of the server; T min is the lower limit of server temperature; α1 is the first adjustment coefficient, 0<α1<1, which is determined by the server attributes; The second temperature threshold is set to T2, and its expression is: T2= T1+α2* (T min -T max ); Wherein, α2 is the second adjustment coefficient, 0<α2<1, which is determined by the server attributes.

9. A digital twin cooling system for a data center based on load perception, characterized in that: include: A data collection module is used to collect the server air outlet temperature, server air inlet temperature, server power and terminal air conditioner outlet temperature and store them in a database; A data analysis module is used to predict the power change of the server within a future set time period based on historical data in the database; and determine the temperature control strategy based on the previous temperature and current temperature of the server outlet, the predicted power change of the server within a future set time period, the first temperature threshold and the second temperature threshold; A decision control module, which is used to send a control command to the corresponding air conditioner according to the determined temperature control strategy to adjust the air outlet temperature of the corresponding air conditioner; Among them, the temperature control strategy is: When the current temperature of the server outlet is higher than the temperature at the previous moment, the current temperature of the server outlet exceeds the first temperature threshold, and the power of the server is predicted to remain unchanged or increase within a future set time period, the temperature of the terminal air conditioner closest to the server is reduced; When the current temperature of the server outlet is higher than the previous temperature, and the power of the server is predicted to decrease in the future set time period, whether to reduce the temperature of the terminal air conditioner closest to the server is determined according to whether the current temperature of the server outlet exceeds the second temperature threshold; When the current temperature of the air outlet of the server is lower than the first temperature threshold, the temperature of the terminal air conditioner closest to the server is controlled to remain unchanged.

10. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps in the load-aware data center digital twin cooling method as described in any one of claims 1 to 5 are implemented.

Images