Method and Device for Adjusting Temperature of Computer Room Air Conditioner, and Electronic Equipment

Abstract

The present application discloses a method and device for adjusting the temperature of a computer room air conditioner, belonging to the technical field of automatic control. The method includes: in response to adjusting the temperature of the air conditioner in the target computer room, obtaining the target temperature of the target computer room and the real-time data of the preset influence factors, wherein the preset influence factors include: the air conditioner operation parameter factor, the IT equipment load factor, and the environmental factor; according to the real-time data and the target temperature, solving the target control temperature of the air conditioner through the multi-dimensional temperature influence model of the computer room; and adjusting the temperature of the air conditioner based on the target control temperature. This method realizes the automatic control of the temperature of the computer room air conditioner by combining multi-dimensional factors affecting the temperature of the computer room, improves the real-time performance and temperature adjustment accuracy of the computer room air conditioner temperature adjustment, and helps to reduce the energy consumption of the computer room air conditioner.

Description

Technical Field

[0001] This application relates to the field of automatic control technology, and particularly to a method and device for adjusting the temperature of a computer room air conditioner, as well as an electronic device and a computer-readable storage medium. Background Art

[0002] The temperature of traditional computer room air conditioners is adjusted manually. For example, manually adjust the set value of the computer room air conditioner temperature periodically according to experience, or adjust the computer room air conditioner temperature temporarily according to the real-time temperature in the computer room collected by sensors installed in the computer room. When adjusting the temperature of the computer room air conditioner according to manual experience, the air conditioner temperature is usually adjusted to a fixed value according to the periodic changes of seasons and outdoor air temperatures. For example, the temperature of the computer room air conditioner is set to 24°C in summer and 26°C in winter. The temperature set value is single, and there is a defect of excessive energy consumption of the computer room air conditioner. In the prior art, the scheme of manually adjusting the temperature of the computer room air conditioner according to the indoor temperature collected by sensors requires a large number of sensors to be installed in the computer room, with high hardware costs. Moreover, it cannot be adjusted in time and also has the defect of excessive energy consumption.

[0003] It can be seen that the method for adjusting the temperature of the computer room air conditioner in the prior art still needs to be improved. Summary of the Invention

[0004] The embodiments of this application provide a method and device for adjusting the temperature of a computer room air conditioner, which can improve the real-time performance and temperature adjustment accuracy of the computer room air conditioner temperature adjustment, and help reduce the energy consumption of the computer room air conditioner.

[0005] In a first aspect, the embodiments of this application disclose a method for adjusting the temperature of a computer room air conditioner, including:

[0006] In response to adjusting the temperature of the air conditioner in the target computer room, obtain the target temperature of the target computer room and the real-time data of the preset influencing factors, where the preset influencing factors include: air conditioner operation parameter factors, IT equipment load factors, and environmental factors;

[0007] Use the real-time data and the target temperature as the input of a pre-trained multi-dimensional temperature influence model for the computer room, and solve the target control temperature of the air conditioner corresponding to the real-time data and the target temperature through the multi-dimensional temperature influence model for the computer room;

[0008] Adjust the temperature of the air conditioner based on the target control temperature.

[0009] Optionally, after using the real-time data and the target temperature as the input of a pre-trained multi-dimensional temperature influence model for the computer room, and solving the target control temperature of the air conditioner corresponding to the real-time data and the target temperature through the multi-dimensional temperature influence model for the computer room, it further includes:

[0010] According to the real-time data and the target control temperature, estimate the temperature of the target computer room after implementing the temperature adjustment through a pre-trained computer room temperature prediction model as the predicted temperature;

[0011] After adjusting the temperature of the air conditioner based on the target control temperature, it further includes:

[0012] Obtain the measured temperature of the target computer room;

[0013] When the absolute value of the difference between the predicted temperature and the measured temperature satisfies a preset tuning threshold, perform tuning training on the computer room multi-dimensional temperature influence model.

[0014] Optionally, the target computer room is divided into several three-dimensional sub-spaces, and the computer room multi-dimensional temperature influence model includes: sub-models corresponding to each preset influence factor, a feature generation sub-model, and a feature encoding network.

[0015] Taking the real-time data and the target temperature as the input of a pre-trained computer room multi-dimensional temperature influence model, and solving the target control temperature of the air conditioner corresponding to the real-time data and the target temperature through the computer room multi-dimensional temperature influence model includes:

[0016] Calculate the gain of the value of the preset influence factor of each three-dimensional sub-space through the sub-model respectively to obtain the single-factor space temperature influence value of each three-dimensional sub-space;

[0017] Through the feature generation sub-model, fuse the single-factor space temperature influence values to obtain the cumulative space temperature influence value of the preset influence factor on each three-dimensional sub-space; and

[0018] Perform space dimensionality reduction processing on the cumulative space temperature influence value to obtain an air conditioner temperature gain matrix;

[0019] Encode and map the air conditioner temperature gain matrix through the feature encoding network to obtain the target control temperature of the air conditioner corresponding to the real-time data and the target temperature.

[0020] Optionally, the sub-model includes: a first sub-model expressing the influence of the air conditioner operation parameter factors on the computer room temperature, and the air conditioner operation parameter factors include one or more of the following data: air conditioner temperature, fan rotation speed, wind direction angle, air conditioner operation duration, air conditioner position, and thermal conductivity of air.

[0021] The calculating the gain of the value of the preset influence factor of each three-dimensional sub-space through the sub-model respectively to obtain the single-factor space temperature influence value of each three-dimensional sub-space includes:

[0022] Taking the values of the air conditioner operation parameter factors and the positions of the three-dimensional subspaces as the first input data, performing a preset operation on the first input data through the first sub-model to obtain the single-factor space temperature influence values of the air conditioner operation parameter factors on the three-dimensional subspaces.

[0023] Optionally, the sub-model includes: a second sub-model expressing the influence of the IT equipment load factor on the computer room temperature, and the IT equipment load factor includes one or more of the following data: CPU usage rate of the server, memory usage rate, and IO consumption rate, position of the server, temperature of the server.

[0024] The step of obtaining the single-factor space temperature influence values of the three-dimensional subspaces by performing gain calculations on the values of the preset influence factors of the three-dimensional subspaces through the sub-model includes:

[0025] Taking the values of the IT equipment load factor in the target computer room, the positions of the three-dimensional subspaces, and the thermal conductivity of the air as the second input data, performing a preset operation on the second input data through the second sub-model to obtain the single-factor space temperature influence values of the IT equipment load factor on the three-dimensional subspaces in the target computer room.

[0026] Optionally, the sub-model includes: a third sub-model expressing the influence of the environmental factor on the computer room temperature, and the environmental factor includes one or more of the following data: outdoor temperature and air humidity.

[0027] The step of obtaining the single-factor space temperature influence values of the three-dimensional subspaces by performing gain calculations on the values of the preset influence factors of the three-dimensional subspaces through the sub-model includes:

[0028] Taking the values of the environmental factor as the third input data, performing a preset operation on the third input data through the third sub-model to obtain the single-factor space temperature influence values of the environmental factor on the three-dimensional subspaces in the target computer room.

[0029] Optionally, the step of performing spatial dimensionality reduction processing on the cumulative space temperature influence value to obtain the air conditioner temperature gain matrix includes:

[0030] Determine the projection plane of the three-dimensional subspace, where the projection plane is: a plane parallel to the side surface of the three-dimensional subspace in the horizontal direction, or a plane parallel to the side surface of the three-dimensional subspace in a vertical direction;

[0031] Project the three-dimensional subspace onto the projection plane;

[0032] Taking the mean value of the cumulative space temperature influence values corresponding to the three-dimensional subspaces projected onto the same area of the projection plane as the cumulative space temperature influence value corresponding to the corresponding area of the projection plane;

[0033] According to the cumulative space temperature influence values corresponding to each area in the projection plane, an air-conditioning temperature gain matrix is obtained.

[0034] In a second aspect, an embodiment of the present application discloses a temperature regulation device for a computer room air conditioner, including:

[0035] A computer room target temperature and influence data acquisition module, configured to, in response to regulating the temperature of the air conditioner in the target computer room, acquire the target temperature of the target computer room and real-time data of preset influence factors, where the preset influence factors include: air-conditioning operation parameter factors, IT equipment load factors, and environmental factors;

[0036] An air-conditioning temperature solving module, configured to use the real-time data and the target temperature as inputs to a pre-trained computer room multi-dimensional temperature influence model, and solve the target control temperature of the air conditioner corresponding to the real-time data and the target temperature through the computer room multi-dimensional temperature influence model;

[0037] An air-conditioning temperature regulation module, configured to regulate the temperature of the air conditioner based on the target control temperature.

[0038] Optionally, the device further includes:

[0039] A computer room temperature prediction module, configured to, according to the real-time data and the target control temperature, estimate the computer room temperature after implementing the temperature regulation in the target computer room through a pre-trained computer room temperature prediction model as the predicted temperature;

[0040] A computer room measured temperature acquisition module, configured to acquire the measured temperature of the target computer room after regulating the temperature of the air conditioner based on the target control temperature;

[0041] A model tuning processing module, configured to perform tuning training on the computer room multi-dimensional temperature influence model when the absolute value of the difference between the predicted temperature and the measured temperature satisfies a preset tuning threshold.

[0042] Optionally, the target computer room is divided into several three-dimensional subspaces, and the computer room multi-dimensional temperature influence model includes: sub-models corresponding to each of the preset influence factors, a feature generation sub-model, and a feature encoding network,

[0043] The air-conditioning temperature solving module is further configured to:

[0044] The gain calculation is respectively performed on the values of the preset influence factors of each of the three-dimensional subspaces through the sub-model to obtain the single-factor space temperature influence value of each three-dimensional subspace;

[0045] Through the feature generation sub-model, the single-factor space temperature influence values are fused to obtain the cumulative space temperature influence value of the preset influence factor on each three-dimensional subspace; and,

[0046] Perform spatial dimensionality reduction processing on the cumulative space temperature influence value to obtain the air-conditioning temperature gain matrix;

[0047] Perform encoding mapping on the air-conditioning temperature gain matrix through the feature encoding network to obtain the target control temperature of the air-conditioning corresponding to the real-time data and the target temperature.

[0048] Optionally, the sub-model includes: a first sub-model expressing the influence of the air-conditioning operation parameter factor on the computer room temperature, and the air-conditioning operation parameter factor includes one or more of the following data: air-conditioning temperature, fan rotation speed, wind direction angle, air-conditioning operation duration, air-conditioning position, and the thermal conductivity of air. The gain calculation is respectively performed on the values of the preset influence factors of each of the three-dimensional subspaces through the sub-model to obtain the single-factor space temperature influence value of each three-dimensional subspace, including:

[0049] Taking the value of the air-conditioning operation parameter factor and the position of each three-dimensional subspace as the first input data, and performing a preset operation on the first input data through the first sub-model to obtain the single-factor space temperature influence value of the air-conditioning operation parameter factor on each three-dimensional subspace.

[0050] Optionally, the sub-model includes: a second sub-model expressing the influence of the IT equipment load factor on the computer room temperature, and the IT equipment load factor includes one or more of the following data: CPU usage rate of the server, memory usage rate, and IO consumption rate, the position of the server, and the temperature of the server. The gain calculation is respectively performed on the values of the preset influence factors of each of the three-dimensional subspaces through the sub-model to obtain the single-factor space temperature influence value of each three-dimensional subspace, including:

[0051] Taking the value of the IT equipment load factor in the target computer room, the position of each three-dimensional subspace, and the thermal conductivity of air as the second input data, and performing a preset operation on the second input data through the second sub-model to obtain the single-factor space temperature influence value of the IT equipment load factor on each three-dimensional subspace in the target computer room.

[0052] Optionally, the sub-model includes: a third sub-model for expressing the influence of the environmental factors on the temperature of the computer room. The environmental factors include one or more of the following data: outdoor temperature and air humidity. The gain calculation is respectively performed on the values of the preset influence factors of each three-dimensional subspace through the sub-model to obtain the single-factor space temperature influence value of each three-dimensional subspace, including:

[0053] Using the value of the environmental factor as the third input data, performing a preset operation on the third input data through the third sub-model to obtain the single-factor space temperature influence value of the environmental factor on each three-dimensional subspace in the target computer room.

[0054] Optionally, the spatial dimensionality reduction processing is performed on the cumulative space temperature influence value to obtain an air-conditioning temperature gain matrix, including:

[0055] Determine the projection plane of the three-dimensional subspace, where the projection plane is: a plane parallel to the side surface of the three-dimensional subspace in the horizontal direction, or a plane parallel to the side surface of the three-dimensional subspace in a vertical direction;

[0056] Project the three-dimensional subspace onto the projection plane;

[0057] Taking the mean value of the cumulative space temperature influence values corresponding to the three-dimensional subspaces projected onto the same area on the projection plane as the cumulative space temperature influence value corresponding to the corresponding area of the projection plane;

[0058] Obtain the air-conditioning temperature gain matrix according to the cumulative space temperature influence values corresponding to each area in the projection plane.

[0059] In a third aspect, an embodiment of the present application also discloses an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the computer room air-conditioning temperature regulation method described in the embodiments of the present application is implemented.

[0060] In a fourth aspect, an embodiment of the present application discloses a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the steps of the computer room air-conditioning temperature regulation method disclosed in the embodiments of the present application are implemented.

[0061] The method for adjusting the temperature of the computer room air conditioner disclosed in the embodiments of the present application obtains real-time data of the target temperature of the target computer room and preset influencing factors by responding to the temperature adjustment of the air conditioner in the target computer room, where the preset influencing factors include: air conditioner operation parameter factors, IT equipment load factors, and environmental factors; using the real-time data and the target temperature as the input of a pre-trained multi-dimensional temperature influence model for the computer room, and solving the target control temperature of the air conditioner corresponding to the real-time data and the target temperature through the multi-dimensional temperature influence model for the computer room; adjusting the temperature of the air conditioner based on the target control temperature, realizing the automatic control of the temperature of the computer room air conditioner by combining multi-dimensional factors affecting the temperature of the computer room, improving the real-time performance and temperature adjustment accuracy of the temperature adjustment of the computer room air conditioner, helping to reduce the energy consumption of the computer room air conditioner, and achieving a dynamic balance between the temperature of the computer room and the energy consumption of the air conditioner.

[0062] The above description is only an overview of the technical solution of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the description. And in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the following specifically gives the specific embodiments of the present application. Brief Description of the Drawings

[0063] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0064] Figure 1 is one of the flowcharts of the method for adjusting the temperature of the computer room air conditioner disclosed in the embodiments of the present application;

[0065] Figure 2 is the second flowchart of the method for adjusting the temperature of the computer room air conditioner disclosed in the embodiments of the present application;

[0066] Figure 3 is the third flowchart of the method for adjusting the temperature of the computer room air conditioner disclosed in the embodiments of the present application;

[0067] Figure 4 is a schematic diagram of the three-dimensional space division of the computer room in the method for adjusting the temperature of the computer room air conditioner disclosed in the embodiments of the present application;

[0068] Figure 5 is a schematic diagram of the space dimensionality reduction effect in the method for adjusting the temperature of the computer room air conditioner disclosed in the embodiments of the present application;

[0069] Figure 6It is a schematic diagram of the temperature interpolation principle in the method for adjusting the temperature of a computer room air conditioner disclosed in the embodiments of the present application;

[0070] Figure 7 It is one of the schematic diagrams of the structure of the device for adjusting the temperature of a computer room air conditioner disclosed in the embodiments of the present application;

[0071] Figure 8 It is the second schematic diagram of the structure of the device for adjusting the temperature of a computer room air conditioner disclosed in the embodiments of the present application;

[0072] Figure 9 Schematically shows a block diagram of an electronic device for executing the method according to the present application; and

[0073] Figure 10 Schematically shows a storage unit for holding or carrying program code for implementing the method according to the present application. Specific embodiments

[0074] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts shall fall within the scope of protection of the present application.

[0075] Next, specific examples will be given for the specific implementation manners of the method for adjusting the temperature of a computer room air conditioner disclosed in the embodiments of the present application.

[0076] As Figure 1 shown, a method for adjusting the temperature of a computer room air conditioner disclosed in the embodiments of the present application includes: step 120, step 130, and step 140.

[0077] Step 120, in response to adjusting the temperature of the air conditioner in the target computer room, obtain the target temperature of the target computer room and the real-time data of the preset influencing factors, where the preset influencing factors include: air conditioner operation parameter factors, IT equipment load factors, and environmental factors.

[0078] In the method for adjusting the temperature of a computer room air conditioner described in the embodiments of the present application, temperature adjustment is performed separately for each computer room. Among them, the target computer room is a computer room where air conditioner temperature monitoring is performed and air conditioner temperature adjustment is required.

[0079] At some positions in the target computer room, temperature acquisition devices (such as temperature sensors) are deployed to collect the real-time temperature of the target computer room. Usually, multiple temperature acquisition devices are deployed in the target computer room to collect the real-time temperatures of different positions in the target computer room.

[0080] The deployment location of the air conditioner in the computer room, the number of air conditioners, the operating temperature of the air conditioner, the fan speed, the air outlet direction, the placement location, quantity, and operating status of the IT equipment, as well as the temperature inside and outside the computer room and other factors will all affect the temperature at a certain location in the computer room. In the embodiments of the present application, multi-dimensional data that affects the temperature regulation effect of the air conditioner on the computer room temperature is determined as influencing factors for determining the target temperature of the air conditioner. For example, the operating parameters of the air conditioner in the computer room, the environmental conditions where the computer room is located, the IT equipment information operating in the computer room, etc., and a preset influence factor is defined according to the above multi-dimensional data.

[0081] Among them, the air conditioner operating parameter factor included in the preset influence factor includes: the relevant parameters of the air conditioner in the computer room; the IT equipment load factor included in the preset influence factor includes: the relevant parameters of the servers in the computer room; the environmental factor included in the preset influence factor includes: the outdoor temperature of the computer room, the indoor and / or outdoor air humidity.

[0082] In practical applications, the temperature of the air conditioner in the target computer room can be adjusted regularly. For example, the air conditioner temperature regulation method in the computer room is executed every 6 hours.

[0083] In other embodiments of the present application, the temperature regulation of the air conditioner in the target computer room can also be triggered according to the temperature monitoring result of the target computer room. For example, when the computer room temperature collected by the temperature acquisition device in the computer room exceeds the preset temperature range, the temperature of the air conditioner in the target computer room is adjusted.

[0084] For example, the air conditioner temperature regulation conditions can be set in advance. Among them, the air conditioner temperature regulation conditions include: the real-time temperature is less than the preset temperature threshold, or the absolute value of the difference between the real-time temperature and the preset temperature threshold is less than the preset temperature difference threshold. The preset air conditioner temperature regulation conditions are determined according to the computer room maintenance requirements.

[0085] After obtaining the real-time temperature of the target computer room, the real-time temperature is matched with the preset air conditioner temperature regulation conditions. If the real-time conditions meet the preset air conditioner temperature regulation conditions, the operating temperatures of each air conditioner in the target computer room are adjusted to achieve the purpose of energy conservation.

[0086] When implementing the energy-saving strategy and adjusting the operating temperature of the air conditioner, it is first necessary to determine how many degrees the temperature in the target computer room needs to be adjusted to, that is, to obtain the target temperature of the target computer room. In the embodiments of the present application, the target temperature of the target computer room can be determined according to the energy-saving requirements.

[0087] When implementing an energy-saving strategy and adjusting the operating temperature of the air conditioner, it is also necessary to obtain the real-time data of the above-mentioned preset influencing factors. Among them, the value of the air conditioner operation parameter factor can be obtained through the air conditioner setting system; the IT equipment load factor can be obtained through the IT equipment operation management system of the target computer room; the environmental factor can be obtained by combining temperature and humidity acquisition devices such as sensors set in the target computer room; the measured temperature of the first computer room can be obtained through the temperature acquisition device deployed in the target computer room.

[0088] Step 130, using the real-time data and the target temperature as the input of the pre-trained computer room multi-dimensional temperature influence model, and solving the target control temperature of the air conditioner corresponding to the real-time data and the target temperature through the computer room multi-dimensional temperature influence model.

[0089] When implementing an energy-saving strategy and adjusting the operating temperature of the air conditioner, it is first necessary to estimate to what temperature the operating temperature of each air conditioner will be adjusted, that is, to determine the target control temperature.

[0090] In the embodiments of the present application, the air conditioner temperature that satisfies the computer room multi-dimensional temperature influence model can be obtained by executing the computer room multi-dimensional temperature influence model.

[0091] Such as Figure 2 As shown, in some embodiments of the present application, before responding to the temperature adjustment of the air conditioner in the target computer room and obtaining the target temperature of the target computer room and the real-time data of the preset influencing factors, it further includes: Step 110.

[0092] Step 110, training a computer room multi-dimensional temperature influence model according to the historical temperature of the target computer room and the values of the preset influencing factors corresponding to the historical temperature.

[0093] In the embodiments of the present application, first, according to the relationship between the preset influencing factors and the computer room temperature, a computer room multi-dimensional temperature influence model expressing the mapping relationship between the values of the preset influencing factors and the computer room temperature is established. The computer room multi-dimensional temperature influence model includes: sub-models corresponding to each preset influencing factor, and each sub-model is used to calculate the gain of the value of the corresponding preset influencing factor to the computer room temperature. The gains of all preset influencing factors to the computer room temperature constitute an air conditioner temperature gain matrix. The air conditioner temperature gain matrix expresses the influence of the values of each preset influencing factor on the computer room temperature. From the above analysis, it can be seen that by training the computer room multi-dimensional temperature influence model based on the preset influencing factors of the target computer room and the historical data of the computer room temperature, the training of the air conditioner temperature gain matrix can be realized.

[0094] Taking the computer room multi-dimensional temperature influence model being expressed as the following expression as an example:

[0095] ΔTP * T0 * F(f(a), f(b), f(c)) = Troom; where a, b, and c represent the preset influencing factors, f(a), f(b), and f(c) represent the sub-models corresponding to the preset influencing factors, ΔTP represents the air-conditioning temperature gain matrix, the value of ΔTP is calculated according to the outputs of the sub-models, Troom represents the computer room temperature, T0 is the conversion matrix, which is a constant. It can be seen from the above formula that by substituting the values of several groups of preset influencing factors and the computer room temperature into the above formula, the mapping relationship between a certain preset influencing factor and the computer room temperature can be learned.

[0096] In the embodiments of the present application, the matrix element values of the air-conditioning temperature gain matrix ΔTP are calculated according to a preset method through the model parameters of the computer room multi-dimensional temperature influence model and the values of the above-mentioned preset influencing factors. Based on the above analysis, the process of training the computer room multi-dimensional temperature influence model is the process of learning the air-conditioning temperature gain matrix ΔTP.

[0097] Correspondingly, when implementing the energy-saving strategy, the computer room temperature Troom is a known quantity, that is, the target temperature. The value of the IT equipment load factor b is from the IT equipment load factor in the real-time data, the value of the environmental factor c is from the environmental factor in the real-time data, and the value of the air-conditioning operation parameter factor is from the air-conditioning operation parameter factor in the real-time data. Except for the air-conditioning temperature, other values remain unchanged. That is, in the air-conditioning operation parameter factor a, the air-conditioning temperature needs to be solved.

[0098] Based on the above principle analysis, when the air-conditioning temperature gain matrix is a determined value, all values in the real-time data except the air-conditioning temperature are determined values, and the target temperature is a determined value, using the real-time data and the target temperature as the input of the pre-trained computer room multi-dimensional temperature influence model, according to the operation relationship expressed by the computer room multi-dimensional temperature influence model, the air-conditioning temperature can be solved and used as the target control temperature.

[0099] Step 140, adjust the temperature of the air-conditioning based on the target control temperature.

[0100] After solving the target control temperature of the air-conditioning, the temperature of the air-conditioning can be adjusted according to the target control temperature to adjust the temperature of the target computer room to achieve the energy-saving effect.

[0101] In some embodiments of the present application, as Figure 3 shown, after using the real-time data and the target temperature as the input of the pre-trained computer room multi-dimensional temperature influence model and solving the target control temperature of the air-conditioning corresponding to the real-time data and the target temperature through the computer room multi-dimensional temperature influence model, it further includes: Step 135.

[0102] Step 135: According to the real-time data and the target control temperature, estimate the temperature of the target computer room after implementing the temperature adjustment through a pre-trained computer room temperature prediction model, and use it as the predicted temperature.

[0103] After adjusting the temperature of the air conditioner based on the target control temperature, as the air conditioner temperature is reset, the air conditioner operates at the reset temperature, and the temperature of the computer room will change. Theoretically, if the target control temperature of the air conditioner obtained by solving the foregoing steps is accurate, the temperature of the target computer room will gradually approach the target temperature until the target temperature is reached.

[0104] To verify whether the obtained target control temperature of the air conditioner is accurate, that is, to verify whether the energy-saving strategy is effective. In the embodiments of the present application, a computer room temperature prediction model that predicts the computer room temperature according to the air conditioner temperature is pre-trained, and the temperature that the target computer room will reach after adjusting the air conditioner temperature of the target computer room to the target control temperature is predicted and used as the computer room predicted temperature. Then, by comparing the computer room predicted temperature with the actual temperature that the target computer room reaches after adjusting the air conditioner temperature to the target control temperature, it is evaluated whether the energy-saving strategy is effective.

[0105] Among them, the computer room temperature prediction model is a neural network model trained based on the historical data of the target computer room. The historical data used to train the computer room temperature prediction model includes: the actual temperature of the computer room and the values of the foregoing preset influencing factors when the target computer room is at the actual temperature of the computer room.

[0106] The process of training the computer room temperature prediction model is the process of establishing the mapping relationship between the values of the preset influencing factors of the target computer room and the computer room temperature. In this way, for the trained computer room temperature prediction model, given the values of the preset influencing factors, the corresponding computer room temperature can be predicted.

[0107] The specific implementation manner of training the computer room temperature prediction model can refer to the prior art, and will not be elaborated in the embodiments of the present application.

[0108] Correspondingly, after adjusting the temperature of the air conditioner based on the target control temperature, it further includes: Step 150 and Step 160.

[0109] Step 150: Obtain the measured temperature of the target computer room.

[0110] After implementing the energy-saving strategy for the target computer room, wait for the air conditioner to operate at the target control temperature for a period of time, and then the real-time temperature of the target computer room can be collected to obtain the measured temperature. In some embodiments of the present application, the real-time temperature of the target computer room can be collected by a temperature collection device deployed in the target computer room, and the measured temperature can be obtained according to the real-time temperature.

[0111] Step 160, when the absolute value of the difference between the predicted temperature and the measured temperature meets a preset tuning threshold, perform tuning training on the multi-dimensional temperature impact model of the computer room.

[0112] In the embodiments of the present application, the energy-saving strategy can be secondarily tuned through the computer room temperature adaptation model. Optionally, the computer room temperature adaptation model can be expressed as: β(s) - P 预测 = α, where β(s) is the measured temperature of the target computer room obtained after implementing the energy-saving strategy, P 预测 is the predicted temperature, and α is a correction parameter. α is the preset tuning threshold.

[0113] The tuning threshold α is used for influencing factor correction. The optimal value of α is infinitely close to 0. If this tuning threshold is met, it indicates that the air-conditioning control temperature predicted by the multi-dimensional temperature impact model of the computer room based on the preset influencing factors is the most accurate.

[0114] In some embodiments of the present application, the correction parameter (i.e., the tuning threshold) can be preset according to the energy-saving requirements of the computer room. For example, the tuning threshold can be set to 1. When the absolute value of the difference between the predicted temperature and the measured temperature of the target computer room is greater than or equal to the preset tuning threshold (e.g., greater than or equal to 1), it can be considered that when implementing the energy-saving strategy, the solved target control temperature is not accurate enough, and the air-conditioning temperature gain matrix used when solving the target control temperature needs to be optimized and trained. That is, by updating the training samples, perform optimization training on the multi-dimensional temperature impact model of the computer room, so as to achieve the purpose of optimizing the air-conditioning temperature gain matrix. For example, update the real-time data of the preset influencing factors currently collected and the current computer room temperature to the training samples for training the multi-dimensional temperature impact model of the computer room, and iteratively train the multi-dimensional temperature impact model of the computer room.

[0115] When the absolute value of the difference between the predicted temperature and the measured temperature of the target computer room is less than the preset tuning threshold, it can be considered that the implemented energy-saving strategy is effective.

[0116] The method for regulating the temperature of a computer room air conditioner disclosed in the embodiments of the present application obtains real-time data of the target temperature of the target computer room and preset influencing factors in response to regulating the temperature of the air conditioner in the target computer room, where the preset influencing factors include: air conditioner operation parameter factors, IT equipment load factors, and environmental factors; uses the real-time data and the target temperature as inputs to a pre-trained multi-dimensional temperature influence model of the computer room, and solves the target control temperature of the air conditioner corresponding to the real-time data and the target temperature through the multi-dimensional temperature influence model of the computer room; and regulates the temperature of the air conditioner based on the target control temperature, realizing automatic control of the temperature of the computer room air conditioner by combining multi-dimensional factors affecting the temperature of the computer room, improving the real-time performance and temperature regulation accuracy of the temperature regulation of the computer room air conditioner, helping to reduce the energy consumption of the computer room air conditioner, and achieving a dynamic balance between the temperature of the computer room and the energy consumption of the air conditioner.

[0117] Further, the method for regulating the temperature of a computer room air conditioner disclosed in the embodiments of the present application, after obtaining real-time data of the target temperature of the target computer room and preset influencing factors in response to regulating the temperature of the air conditioner in the target computer room, where the preset influencing factors include: air conditioner operation parameter factors, IT equipment load factors, and environmental factors, and before regulating the temperature of the air conditioner based on the target control temperature, estimates the temperature of the target computer room after implementing the temperature regulation as a predicted temperature according to the real-time data and the target control temperature through a pre-trained computer room temperature prediction model, and after regulating the temperature of the air conditioner based on the target control temperature, obtains the measured temperature of the target computer room after implementing the temperature regulation, and compares the predicted temperature with the measured temperature; and when the absolute value of the difference between the predicted temperature and the measured temperature satisfies a preset tuning threshold, performs tuning training on the multi-dimensional temperature influence model of the computer room, thereby further improving the accuracy of the temperature regulation of the computer room air conditioner and improving the energy-saving effect.

[0118] To make it easier to understand the temperature regulation of the computer room air conditioner disclosed in the embodiments of the present application, the following further gives examples of the specific implementation manners of each step of the method for regulating the temperature of the computer room air conditioner.

[0119] As mentioned above, the preset influencing factors include: air conditioner operation parameter factors, IT equipment load factors, and environmental factors. Correspondingly, the multi-dimensional temperature influence model of the computer room includes: sub-models corresponding to each of the preset influencing factors, a feature generation sub-model, and a feature encoding network. Among them, the sub-models corresponding to each of the preset influencing factors include: a first sub-model corresponding to the air conditioner operation parameter factors, a second sub-model corresponding to the IT equipment load factors, and a third sub-model corresponding to the environmental factors.

[0120] In the embodiments of the present application, corresponding sub-models are constructed according to the influence of various types of influencing factors on the computer room temperature. Thus, through each sub-model, corresponding operations are performed on the values of the corresponding type of influencing factor, and the influence value of the value of each preset influencing factor on the computer room temperature is obtained.

[0121] In some embodiments of the present application, the target computer room is divided into several three-dimensional sub-spaces.

[0122] To improve the accuracy of temperature control, the overall three-dimensional space of the target computer room is divided into multiple local three-dimensional sub-spaces. As Figure 4 shown, using the existing data acquisition system in the computer room (such as temperature acquisition equipment, equipment management system) through the grid method to obtain multi-parameter indicators such as the computer room temperature measured by sensors in the target computer room, air-conditioning operation parameters (such as supply and return air temperatures, air-conditioning location, fan speed, air outlet direction, etc.), computer room IT load (such as server power consumption, temperature, memory, deployment location, etc.), environment (such as outdoor temperature, air humidity, etc.). Then, through each sub-model, the influence value of the value of the corresponding type of influencing factor on the temperature of each sub-space in the computer room is calculated.

[0123] Next, in combination with the training process of the computer room multi-dimensional temperature influence model, the execution principle of the computer room multi-dimensional temperature influence model, the execution principle of each sub-model, and the training method of the air-conditioning temperature gain matrix are elaborated.

[0124] In the aforementioned step 110, training the computer room multi-dimensional temperature influence model according to the historical temperature of the target computer room and the value of the preset influencing factor corresponding to the historical temperature includes: for each specified historical moment, respectively performing the following sub-steps 1101, sub-step 1102, and sub-step 1103 to obtain the air-conditioning temperature gain matrix corresponding to the specified historical moment; then, performing sub-step 1104.

[0125] Among them, the specified historical moment can be every 5 minutes within the past month as a historical moment.

[0126] When training the computer room multi-dimensional temperature influence model, the values of the preset influencing factors and the computer room temperature of the target computer room at each historical moment can be used as a set of historical data respectively, obtaining several sets of historical data. Then, the air-conditioning temperature gain matrix is calculated for each set of historical data respectively. In this way, for each historical moment, an air-conditioning temperature gain matrix can be obtained.

[0127] The specific implementation schemes of each sub-step are elaborated below.

[0128] Sub-step 1101: Calculate the gain of the preset influence factor values of each three-dimensional subspace at the specified historical moment through the sub-model, and obtain the single-factor space temperature influence value of each three-dimensional subspace corresponding to the specified historical moment.

[0129] During the process of training the multi-dimensional temperature influence model of the computer room, after each sub-model operates on the corresponding preset influence factor value, the influence value of each type of preset influence factor on each three-dimensional subspace in the target computer room is obtained. In the embodiments of the present application, the influence value of the value of a type of preset influence factor on the temperature of a certain three-dimensional subspace in the computer room is denoted as the "single-factor space temperature influence value".

[0130] The following respectively elaborates on the specific implementation manners of obtaining the influence value of each type of preset influence factor on each three-dimensional subspace in the target computer room after each sub-model operates on the corresponding preset influence factor value.

[0131] I. The first sub-model

[0132] The sub-model includes: the first sub-model expressing the influence of the air-conditioning operation parameter factor on the computer room temperature.

[0133] By operating on the historical data of the air-conditioning operation parameter factor through the first sub-model, the single-factor space temperature influence value of the air-conditioning operation parameter factor on each three-dimensional subspace in the target computer room can be obtained.

[0134] In some embodiments of the present application, the air-conditioning operation parameter factor includes one or more of the following data: air-conditioning temperature, fan rotation speed, wind direction angle, air-conditioning operation duration, air-conditioning position, and the thermal conductivity of air. The step of calculating the gain of the preset influence factor values of each three-dimensional subspace at the specified historical moment through the sub-model to obtain the single-factor space temperature influence value of each three-dimensional subspace corresponding to the specified historical moment includes: using the value of the air-conditioning operation parameter factor and the positions of each three-dimensional subspace at the specified historical moment as the first input data, and performing a preset operation on the first input data through the first sub-model to obtain the single-factor space temperature influence value of the air-conditioning operation parameter factor on each three-dimensional subspace corresponding to the specified historical moment.

[0135] Based on the heat conduction model, a model expressing the consumption of air-conditioning temperature propagation can be constructed:

[0136]

[0137] Among them, p represents the three-dimensional subspace, i represents the air-conditioning identifier in the target computer room, λ is the thermal conductivity of air, T i represents the air-conditioning temperature of the i-th air conditioner, vi represents the fan rotation speed of the i-th air conditioner, t i represents the operation duration of the i-th air conditioner, O i represents the wind direction angle of the i-th air conditioner, S i,p represents the distance between the i-th air conditioner and the three-dimensional subspace P, O ip ` represents the angle of the three-dimensional subspace P relative to the i-th air conditioner, K1 is a model parameter, and Δt1(p) represents the single-factor space temperature influence value of the i-th air conditioner in the target computer room on the three-dimensional subspace p.

[0138] Among them, the angle O ip ` of the three-dimensional subspace P relative to the i-th air conditioner can be calculated by the following formula:

[0139]

[0140] The distance S i,p between the i-th air conditioner and the three-dimensional subspace P can be calculated by the following formula:

[0141] In the above formula, (x i , y i , z i ) represents the spatial position coordinates of the i-th air conditioner in the target computer room, and (x p , y p , z p ) represents the spatial position coordinates of the three-dimensional subspace P in the target computer room.

[0142] Then, the first sub-model expressing the single-factor space temperature influence value of the air conditioner operation parameters of n air conditioners on the three-dimensional subspace P can be expressed by the following formula:

[0143] Among them, n is the total number of air conditioners in the target computer room, and A1(p) represents the single-factor space temperature influence value of n air conditioners in the target computer room on the three-dimensional subspace p.

[0144] II. The second sub-model

[0145] The sub-model includes: a second sub-model expressing the influence of the IT equipment load factor on the computer room temperature.

[0146] By operating on the historical data of the IT equipment load factor through the second sub-model, the single-factor space temperature influence value of the IT equipment load factor on each three-dimensional subspace in the target computer room can be obtained.

[0147] In some embodiments of the present application, the IT device load factor includes one or more of the following data: the CPU usage rate of the server, the memory usage rate, the IO consumption rate, the location of the server, and the temperature of the server. By performing gain calculations on the values of the preset influence factors of each of the three-dimensional subspaces at the specified historical moment through the sub-model, the single-factor space temperature influence value of each three-dimensional subspace corresponding to the specified historical moment is obtained, including: using the value of the IT device load factor in the target computer room at the specified historical moment, the location of each three-dimensional subspace, and the thermal conductivity of the air as the second input data, and performing a preset operation on the second input data through the second sub-model to obtain the single-factor space temperature influence value of the IT device load factor on each three-dimensional subspace in the target computer room corresponding to the specified historical moment.

[0148] During the operation of IT devices (such as servers) in the computer room, heat is continuously released, thus affecting the computer room temperature. Based on the heat conduction model, a model expressing the propagation and consumption of IT load heat sources can be constructed:

[0149]

[0150] where p represents the three-dimensional subspace, j represents the server identifier in the target computer room, λ is the thermal conductivity of the air, and T j represents the temperature of the jth server predicted based on the CPU usage rate, memory usage rate, and IO consumption rate of the jth server, and S j,p represents the distance between the jth server and the three-dimensional subspace P, K2 is a model parameter, and Δt2(p) represents the single-factor space temperature influence value of the jth server in the target computer room on the three-dimensional subspace p.

[0151] Then, the second sub-model expressing the single-factor space temperature influence value of m IT load devices on the three-dimensional subspace P can be represented by the following formula:

[0152] where m is the total number of servers in the target computer room, and A2(p) represents the single-factor space temperature influence value of m servers in the target computer room on the three-dimensional subspace p.

[0153] where the calculation method of the distance S j,p between the jth server and the three-dimensional subspace P is described in the full text and will not be elaborated here.

[0154] where, when predicting the temperature of the jth server based on the CPU usage rate, memory usage rate, and IO consumption rate of the jth server, the temperature of the server can be predicted through the following server temperature prediction model pre-trained:

[0155] Tj = W1V cpu_j + W2V cache_j + W3V io_j + T base_j ;

[0156] Wherein, V cpu_j represents the CPU usage rate of the j-th server, V cache_j represents the memory usage rate of the j-th server, V io_j represents the IO consumption rate of the j-th server, W1, W2, and W3 represent the weights of the corresponding items, which are constants obtained through training, and T base_j is the temperature of the j-th server measured by the sensor, T j is the temperature of the server predicted based on the CPU usage rate, memory usage rate, and IO consumption rate of the j-th server.

[0157] The server temperature is closely related to its service load. The key evaluation indicators of the server load are the CPU usage rate, memory usage rate, and IO consumption. The actual temperature of the server is recorded through temperature sensing, and the temperature information of the server can be recorded at each time period during the operation of the server. The server temperature prediction model can be obtained by collecting the CPU usage rate, memory usage rate, IO consumption rate, and actual temperature of each server in the target computer room, and then training a regression model.

[0158] III. The third sub-model

[0159] The sub-model includes: a third sub-model expressing the influence of the environmental factor on the temperature of the computer room.

[0160] By performing operations on the corresponding historical data through the sub-models corresponding to each of the preset influence factors, the single-factor space temperature influence values of the corresponding preset influence factors on each three-dimensional subspace in the target computer room can be obtained.

[0161] In some embodiments of the present application, the environmental factor includes one or more of the following data: outdoor temperature and air humidity. By respectively performing gain calculations on the values of the preset influence factors of each three-dimensional subspace at the specified historical moment through the sub-model, the single-factor space temperature influence values of each three-dimensional subspace corresponding to the specified historical moment are obtained, including: using the values of the environmental factor at the specified historical moment as the third input data, and performing a preset operation on the third input data through the third sub-model to obtain the single-factor space temperature influence values of the environmental factor on each three-dimensional subspace in the target computer room corresponding to the specified historical moment.

[0162] The outdoor temperature and air humidity of the computer room have a certain impact on the temperature inside the computer room. Since the increase in the outdoor temperature value has a positive correlation with the three-dimensional space temperature inside the computer room, and the humidity in the computer room remains relatively stable, with humidity being a factor with a relatively small impact, the third sub-model can be expressed by the following formula:

[0163] A3(T E ,H E )=K3T E +ε(H E );where T E represents the outdoor temperature of the target computer room, H E represents the outdoor humidity of the target computer room, K3 is a model parameter, and A3(T E ,H E ) represents the influence value of the environmental factors of the target computer room on the single-factor space temperature of the three-dimensional subspace.

[0164] Sub-step 1102: For each of the specified historical moments, fuse the single-factor space temperature influence values to obtain the cumulative space temperature influence value of the preset influence factor on each three-dimensional subspace.

[0165] In the embodiments of the present application, various factors affecting the temperature of the subspaces in the computer room are fully considered, and the single-factor space temperature influence values calculated by the foregoing sub-models are fused to obtain the cumulative space temperature influence value of all preset influence factors on each three-dimensional subspace. For example, the single-factor space temperature influence values calculated by the above first sub-model, second sub-model, and third sub-model can be summed respectively, and then the obtained sum is used as the cumulative space temperature influence value of all preset influence factors on the three-dimensional subspace p. For example, the cumulative space temperature influence value ΔTP of all preset influence factors on the three-dimensional subspace p can be calculated by the following formula:

[0166] ΔTP=A1(p)+A2(p)+A3(T E ,H E ).

[0167] According to the distribution position relationship of the three-dimensional subspaces, taking the cumulative space temperature influence value corresponding to each subspace as a matrix element value, a three-dimensional matrix composed of the cumulative space temperature influence values of all preset influence factors on all three-dimensional subspaces in the target computer room can be obtained.

[0168] Sub-step 1103: Perform spatial dimensionality reduction processing on the cumulative space temperature influence value to obtain the air-conditioning temperature gain matrix corresponding to the specified historical moment.

[0169] According to the spatial distribution of each three-dimensional subspace, perform spatial dimensionality reduction on the cumulative spatial temperature influence value to obtain the influence value matrix of the preset influence factor on the two-dimensional space of the target computer room, denoted as the "air-conditioning temperature gain matrix".

[0170] In some embodiments of the present application, performing spatial dimensionality reduction on the cumulative spatial temperature influence value to obtain the air-conditioning temperature gain matrix corresponding to the specified historical moment includes: determining the projection plane of the three-dimensional subspace, where the projection plane is: a plane parallel to the side surface of the three-dimensional subspace in the horizontal direction, or a plane parallel to the side surface of the three-dimensional subspace in a vertical direction; projecting the three-dimensional subspace onto the projection plane; taking the mean value of the cumulative spatial temperature influence values corresponding to the three-dimensional subspaces projected onto the same area of the projection plane as the cumulative spatial temperature influence value corresponding to the corresponding area of the projection plane; and obtaining the air-conditioning temperature gain matrix corresponding to the specified historical moment according to the cumulative spatial temperature influence values corresponding to each area in the projection plane.

[0171] Taking Figure 4 the schematic diagram of the computer room subspace division shown as an example, a vertical plane determined by the X-axis and the Y-axis can be selected as the projection plane, or a horizontal plane determined by the X-axis and the Z-axis can be selected as the projection plane, or a vertical plane determined by the Y-axis and the Z-axis can be selected as the projection plane, or a plane parallel to any of the above projection planes can be selected as the projection plane.

[0172] If each of the three-dimensional subspaces is projected onto this projection plane, it can be seen that multiple subspaces located in the direction perpendicular to this projection plane will be projected onto the same grid in this projection plane. Therefore, according to the projection principle, the three-dimensional subspaces projected onto the same area of the projection plane can be determined. Then, the mean value of the cumulative spatial temperature influence values corresponding to the three-dimensional subspaces in the same area is taken as the cumulative spatial temperature influence value corresponding to the corresponding area of the projection plane.

[0173] For further example, assume that the length, width, and height of the target computer room are 3 meters, 3 meters, and 3 meters respectively, and the unit of space division in the three dimensions is 1 meter. According to Figure 4 the spatial coordinate system shown, taking Figure 4 the plane determined by the X-axis and the Z-axis in Figure 5As shown in the figure, for each grid area, the average value of the cumulative space temperature influence values of the three sub-spaces in the machine room space above the grid area can be used as the cumulative space temperature influence value corresponding to the grid area. For example, the average value of the space temperature influence values ΔTP1, ΔTP2, and ΔTP3 projected onto grid area A is used as the cumulative space temperature influence value corresponding to grid area A.

[0174] By performing the above-mentioned spatial dimensionality reduction process, a two-dimensional matrix of the cumulative space temperature influence values corresponding to each grid area of the corresponding projection plane will be obtained. This two-dimensional matrix can be used as the air-conditioning temperature gain matrix corresponding to the target machine room.

[0175] From the above dimensionality reduction method, it can be concluded that the air-conditioning temperature gain matrix expresses the temperature influence of the preset influence factors on each sub-space in the machine room.

[0176] The above sub-step 1102 and sub-step 1103 are executed by the feature generation sub-model.

[0177] Sub-step 1104: According to the air-conditioning temperature gain matrix corresponding to each of the specified historical moments and the historical temperature of the target machine room, train the sub-models corresponding to each of the preset influence factors and the feature encoding network.

[0178] According to the foregoing solution, for the values of the preset influence factors corresponding to different historical moments, the air-conditioning temperature gain matrix corresponding to this historical moment can be calculated. By inputting the sequence of the air-conditioning temperature gain matrix and the historical temperature of the target machine room into the feature encoding network, and performing feature encoding and mapping by the convolutional layer of the neural network model, the sub-models corresponding to each of the preset influence factors and the feature encoding network can be trained.

[0179] By continuously optimizing the model parameters of the foregoing sub-models and generating different air-conditioning temperature gain matrices based on the same historical data, iterative training of the multi-dimensional temperature influence model of the machine room is realized.

[0180] In the embodiments of the present application, based on the SENet (Squeeze-and-Excitation Networks) network, a feature encoding network for the multi-dimensional temperature influence model of the machine room is constructed.

[0181] During the process of training the multi-dimensional temperature influence model of the machine room, a cycle of 5 minutes can be used to select the air-conditioning parameters, server parameters, outdoor environment, and actual historical temperature of the machine room for one month. Through the foregoing sub-models, several air-conditioning temperature gain matrices are calculated to obtain several sequences of air-conditioning temperature gain matrices, and an original feature map is constructed; then, feature compression is performed through the Squeeze link of SENet.

[0182] The SENet network, through spatial dimensionality reduction and three convolutional trainings, completes the spatio-temporal feature learning of each parameter of the computer room environment temperature. Among them, the SEblock is expressed as:

[0183]

[0184] Among them, * represents the convolution operation, represents the convolution kernels of three convolution operations, C represents the time series, and U c represents the feature vector.

[0185] After that, feature dimensionality reduction is achieved through a fully connected neural network. The dimensionality reduction process can be expressed as:

[0186]

[0187] Among them, H and W respectively represent the number of grids in two directions in the projection plane, that is, H and W respectively correspond to the number of rows and columns of the air-conditioning temperature gain matrix, and F sq represents the convolution calculation process, and Z c is the average pooling result of the total area.

[0188] In order to limit the complexity of the model and improve the prediction of the conventional computer room air-conditioning temperature, in the embodiments of the present application, two fully connected layers are used to parameterize the gating mechanism. A connection layer is used to achieve dimensionality reduction, which is convenient for calculation and reduces the model complexity. After that, a connection layer is used to implement the dimensionality increase layer, and then it is transferred to the channel dimension of the output feature map.

[0189] In the Excitation link, in order to fully capture the channel dependence relationship, that is, to find the front-back relationship between different air-conditioning temperature gain matrices, a simple gating mechanism with Sigmoid activation is used.

[0190] This function must meet two criteria: First, it must be flexible in operation and be able to learn the non-linear relationship between channels; Second, it must learn non-exclusive relationships. In some embodiments of the present application, the following transformation form is adopted:

[0191] s = F ex (z, W) = σ(g(z, W)) = σ(W2δ(W1z));

[0192] In the above formula, δ represents the Relu activation function, σ represents the Sigmoid activation function, where W1 and W2 are model parameters.

[0193] After obtaining s, the final output of the SE_Block algorithm can be obtained through the following formula:

[0194] Q c = F scale (uc , s c ) = s c u c ;

[0195] Among them, Q c = [Q1, Q2,...], u c ∈R H×W , F scale is the product on the channel, and Q c is the target control temperature of the air conditioner corresponding to the target temperature of the computer room environment.

[0196] In the aforementioned step 130, when using the real-time data and the target temperature as the input of the pre-trained multi-dimensional temperature impact model of the computer room, and solving the target control temperature of the air conditioner corresponding to the real-time data and the target temperature through the multi-dimensional temperature impact model of the computer room, by inputting the obtained real-time data of the computer room environment and the target temperature of the computer room into the pre-trained multi-dimensional temperature impact model of the computer room, the Q c output by the multi-dimensional temperature impact model of the computer room is the target control temperature of the air conditioner.

[0197] As mentioned above, the target computer room is divided into several three-dimensional sub-spaces. The multi-dimensional temperature impact model of the computer room includes: sub-models corresponding to each of the preset impact factors, a feature generation sub-model, and a feature encoding network. Using the real-time data and the target temperature as the input of the pre-trained multi-dimensional temperature impact model of the computer room, and solving the target control temperature of the air conditioner corresponding to the real-time data and the target temperature through the multi-dimensional temperature impact model of the computer room includes: respectively performing gain calculations on the values of the preset impact factors of each of the three-dimensional sub-spaces through the sub-model to obtain the single-factor space temperature impact values of each three-dimensional sub-space; through the feature generation sub-model, fusing the single-factor space temperature impact values to obtain the cumulative space temperature impact values of the preset impact factors on each three-dimensional sub-space; and performing spatial dimensionality reduction processing on the cumulative space temperature impact values to obtain an air conditioner temperature gain matrix; and performing encoding mapping on the air conditioner temperature gain matrix through the feature encoding network to obtain the target control temperature of the air conditioner corresponding to the real-time data and the target temperature.

[0198] For the specific implementation of respectively performing gain calculations on the values of the preset impact factors of each of the three-dimensional sub-spaces through the sub-model to obtain the single-factor space temperature impact values of each three-dimensional sub-space, refer to the relevant descriptions in the model training stage, which will not be elaborated here.

[0199] Generate a sub-model through the above features, fuse the single-factor space temperature influence values, obtain the cumulative space temperature influence values of the preset influence factors on each three-dimensional subspace, and perform dimensionality reduction processing on the cumulative space temperature influence values to obtain the specific implementation manner of the air-conditioning temperature gain matrix. For details, refer to the relevant descriptions in the model training stage and will not be elaborated here.

[0200] Input the obtained air-conditioning temperature gain matrix into the feature encoding network, and the multi-dimensional temperature influence model of the computer room will correspondingly output a temperature sequence, which is the target control temperature of the air conditioner corresponding to the real-time data and the target temperature.

[0201] In some embodiments of the present application, the sub-model includes: a first sub-model expressing the influence of the air-conditioning operation parameter factors on the computer room temperature. The air-conditioning operation parameter factors include one or more of the following data: air-conditioning temperature, fan rotation speed, wind direction angle, air-conditioning operation duration, air-conditioning location, and the thermal conductivity of air. By calculating the gain of the values of the preset influence factors of each three-dimensional subspace through the sub-model respectively, the single-factor space temperature influence values of each three-dimensional subspace are obtained, including: taking the values of the air-conditioning operation parameter factors and the positions of each three-dimensional subspace as the first input data, and performing a preset operation on the first input data through the first sub-model to obtain the single-factor space temperature influence values of the air-conditioning operation parameter factors on each three-dimensional subspace.

[0202] In some embodiments of the present application, the sub-model includes: a second sub-model expressing the influence of the IT equipment load factor on the computer room temperature. The IT equipment load factor includes one or more of the following data: CPU usage rate, memory usage rate of the server, and IO consumption rate, the location of the server, and the temperature of the server. By calculating the gain of the values of the preset influence factors of each three-dimensional subspace through the sub-model respectively, the single-factor space temperature influence values of each three-dimensional subspace are obtained, including: taking the values of the IT equipment load factor in the target computer room, the positions of each three-dimensional subspace, and the thermal conductivity of air as the second input data, and performing a preset operation on the second input data through the second sub-model to obtain the single-factor space temperature influence values of the IT equipment load factor on each three-dimensional subspace in the target computer room.

[0203] In some embodiments of the present application, the sub-model includes: a third sub-model for expressing the influence of the environmental factors on the temperature of the computer room, where the environmental factors include one or more of the following data: outdoor temperature and air humidity. By performing gain calculations on the values of the preset influence factors of each of the three-dimensional sub-spaces through the sub-model, the single-factor space temperature influence values of each three-dimensional sub-space are obtained, including: using the values of the environmental factors as the third input data, and performing a preset operation on the third input data through the third sub-model to obtain the single-factor space temperature influence values of the environmental factors on each three-dimensional sub-space in the target computer room.

[0204] In some embodiments of the present application, the spatial dimensionality reduction process of the cumulative space temperature influence value to obtain the air-conditioning temperature gain matrix includes: determining the projection plane of the three-dimensional sub-space, where the projection plane is: a plane parallel to the side surface of the three-dimensional sub-space in the horizontal direction, or a plane parallel to the side surface of the three-dimensional sub-space in a vertical direction; projecting the three-dimensional sub-space onto the projection plane; taking the mean value of the cumulative space temperature influence values corresponding to the three-dimensional sub-spaces projected onto the same area on the projection plane as the cumulative space temperature influence value corresponding to the corresponding area of the projection plane; and obtaining the air-conditioning temperature gain matrix according to the cumulative space temperature influence values corresponding to each area in the projection plane.

[0205] In the model application stage, the data processing processes of each sub-model and the feature encoding network in the model are the same as those in the model training stage. For the specific implementation manners of the foregoing steps, refer to the relevant descriptions in the previous model training stage, which will not be elaborated here.

[0206] In the embodiments of the present application, in order to improve the energy-saving effect and the accuracy of air-conditioning temperature control, the temperature of the computer room is adjusted by dividing the computer room space into several three-dimensional sub-spaces. Therefore, the size of the foregoing air-conditioning temperature gain matrix is equivalent to the number of projection grids of the three-dimensional sub-spaces of the target computer room on a projection plane. Taking the target computer room including H×W×C three-dimensional sub-spaces as an example, where H can represent the number of three-dimensional sub-spaces in the length direction of the computer room, and W can represent the number of three-dimensional sub-spaces in the width direction of the computer room, the size of the air-conditioning temperature gain matrix can be H×W. Correspondingly, the target temperature of the computer room is a temperature matrix of H×W.

[0207] After implementing the energy-saving strategy (for example, setting the target temperature of the computer room to 28 degrees, that is, the value of the matrix element in the temperature matrix is 28), it is necessary to measure the real-time temperature corresponding to each of the foregoing projection grids in the computer room for comparison with the predicted temperature.

[0208] As described above, in order to reduce the hardware cost and deployment complexity, temperature acquisition devices are usually deployed at designated positions in the target computer room, rather than in each three-dimensional subspace. Therefore, when it is necessary to obtain the temperature of each three-dimensional subspace in the target computer room, it is necessary to perform interpolation processing on the real-time temperature of the designated monitoring position in the target computer room obtained by the temperature acquisition device to obtain the real-time temperature of each three-dimensional subspace as the measured temperature.

[0209] For example, obtaining the measured temperature of the target computer room includes: obtaining the real-time temperature of the designated computer room position in the target computer room after the temperature adjustment, which is collected by the temperature acquisition device deployed in the target computer room; performing temperature interpolation in the designated space dimension according to the real-time temperature to obtain the measured temperature in the designated space dimension in the target computer room. Wherein, the designated space dimension is consistent with the space dimension corresponding to the air-conditioning temperature gain matrix.

[0210] For example, the real-time temperature collected by the temperature acquisition device can be used as the base temperature of the projection grid of the three-dimensional subspace corresponding to the temperature acquisition device in the aforementioned projection plane, so that the temperatures of multiple projection grids in the projection plane can be obtained. Taking Figure 6 the projection grid of the projection plane shown (such as the plane determined by the space coordinate axes X and Z) as an example, where T base_11 , T base_12 , T base_21 and T base_22 are the temperatures in the projection grids corresponding to the projection plane coordinates (X1, Z1), (X1, Z2), (X2, Z1) and (X2, Z2) collected by the temperature acquisition device respectively. Then, for the projection grid in the projection plane corresponding to the three-dimensional subspace projected to the projection plane without temperature records for the projection plane coordinate P(X, Z), linear interpolation can be performed according to the distance between the projection grid to be interpolated and the projection grids with known temperatures in each dimension. Specifically, taking the method of obtaining the base temperature of the projection grid where the projection plane coordinate P(X, Z) is located by interpolation as an example:

[0211] First, perform linear interpolation in the X direction through the following formula to obtain the temperature R1 of the point (X, Z1) and the temperature R2 of the point (X, Z2):

[0212]

[0213]

[0214] ]]First, perform linear interpolation in the Z direction through the following formula to obtain the base temperature T base_P of the projection grid where the projection plane coordinate P(X, Z) is located:

[0215] [[ID=--33]]

[0216] According to the above method, temperature interpolation calculation is performed on the projection grids of the subspace without temperature acquisition equipment. After obtaining the base temperatures of all projection grids, the base temperatures of all projection grids constitute the measured temperature matrix of the target computer room.

[0217] Traditional energy-saving methods require a large number of temperature sensors for temperature acquisition, resulting in high investment costs. In this application, by establishing a multi-dimensional gain temperature residual model for the computer room and using grid interpolation method to fill the temperature of the local three-dimensional space of the computer room, the number of installed temperature sensors can be reduced, effectively reducing the deployment cost of the computer room.

[0218] When the target computer room is divided into H×W×C three-dimensional subspaces, the computer room temperature prediction model adopted in the foregoing step 135 can be constructed based on SENet (Squeeze-and-Excitation Networks), and the historical data of each preset influencing factor in the target computer room and the historical temperatures of each subspace of the target computer room are used as inputs to train the computer room temperature prediction model to learn the prediction ability of the temperatures of each subspace of the computer room corresponding to different air-conditioning temperature settings.

[0219] For the training process and implementation principle of the computer room temperature prediction model, refer to the training process and principle of the SENet network model in the prior art, which will not be elaborated in this embodiment of the application.

[0220] When performing temperature prediction, the real-time values of each preset influencing factor of the collected target computer room and the target control temperature of the air conditioner obtained in the foregoing step are input into the computer room temperature prediction model, and the computer room temperature prediction model will output the predicted temperatures of each subspace of the target computer room.

[0221] Furthermore, according to the calculation needs, the predicted temperatures corresponding to each grid on the foregoing projection plane can be calculated based on the predicted temperatures of each subspace.

[0222] When the target computer room is divided into H×W×C three-dimensional subspaces and the target temperature of the predicted computer room is set as a temperature matrix of H×W, the preset tuning threshold α can be set as a matrix of H×W, such as:

[0223]

[0224] In this way, in the foregoing step 160, the matrix element values corresponding to each subspace in the target computer room are used to judge the predicted temperature and the measured temperature corresponding to the subspace, and determine whether to perform tuning training on the multi-dimensional temperature influence model of the computer room.

[0225] Those skilled in the art should understand that the above specific embodiments are merely one or more feasible embodiments, not the only feasible embodiments, and should not be construed as a limitation to this application.

[0226] Correspondingly, the embodiment of this application also discloses a temperature regulation device for a computer room air conditioner, as Figure 7 shown, the device includes:

[0227] A computer room target temperature and influence data acquisition module 720, configured to obtain real-time data of the target temperature of the target computer room and a preset influence factor in response to temperature regulation of the air conditioner in the target computer room, wherein the preset influence factor includes: an air conditioner operation parameter factor, an IT device load factor, and an environmental factor;

[0228] An air conditioner temperature solving module 730, configured to use the real-time data and the target temperature as inputs of a pre-trained multi-dimensional temperature influence model for a computer room, and solve the target control temperature of the air conditioner corresponding to the real-time data and the target temperature through the multi-dimensional temperature influence model for the computer room;

[0229] An air conditioner temperature regulation module 740, configured to regulate the temperature of the air conditioner based on the target control temperature.

[0230] In some embodiments of this application, as Figure 8 shown, the device further includes:

[0231] A computer room temperature prediction module 735, configured to estimate the computer room temperature after implementing the temperature regulation of the target computer room as a predicted temperature according to the real-time data and the target control temperature through a pre-trained computer room temperature prediction model;

[0232] A computer room measured temperature acquisition module 750, configured to obtain the measured temperature of the target computer room after regulating the temperature of the air conditioner based on the target control temperature;

[0233] A model tuning processing module 760, configured to perform tuning training on the multi-dimensional temperature influence model for the computer room when the absolute value of the difference between the predicted temperature and the measured temperature satisfies a preset tuning threshold.

[0234] Optionally, the target computer room is divided into several three-dimensional sub-spaces, the multi-dimensional temperature influence model for the computer room includes: sub-models corresponding to each of the preset influence factors, a feature generation sub-model, and a feature encoding network, and the air conditioner temperature solving module 730 is further configured to:

[0235] Perform gain calculation on the values of the preset influence factors in each of the three-dimensional sub-spaces through the sub-models to obtain the single-factor space temperature influence values of each three-dimensional sub-space;

[0236] Generate a sub-model based on the said features, fuse the single-factor space temperature influence values, and obtain the cumulative space temperature influence values of the preset influence factors on each three-dimensional subspace; and,

[0237] Perform spatial dimensionality reduction processing on the cumulative space temperature influence values to obtain an air-conditioning temperature gain matrix;

[0238] Perform encoding mapping on the air-conditioning temperature gain matrix through the said feature encoding network to obtain the target control temperature of the air-conditioning corresponding to the real-time data and the target temperature.

[0239] Optionally, the sub-model includes: a first sub-model expressing the influence of the air-conditioning operation parameter factors on the computer room temperature, and the air-conditioning operation parameter factors include one or more of the following data: air-conditioning temperature, fan rotation speed, wind direction angle, air-conditioning operation duration, air-conditioning location, and the thermal conductivity of air. By using the sub-model to perform gain calculations on the values of the preset influence factors of each three-dimensional subspace respectively, the single-factor space temperature influence values of each three-dimensional subspace are obtained, including:

[0240] Using the values of the air-conditioning operation parameter factors and the positions of each three-dimensional subspace as the first input data, perform a preset operation on the first input data through the first sub-model to obtain the single-factor space temperature influence values of the air-conditioning operation parameter factors on each three-dimensional subspace.

[0241] Optionally, the sub-model includes: a second sub-model expressing the influence of the IT equipment load factor on the computer room temperature, and the IT equipment load factor includes one or more of the following data: CPU usage rate of the server, memory usage rate, and IO consumption rate, location of the server, and temperature of the server. By using the sub-model to perform gain calculations on the values of the preset influence factors of each three-dimensional subspace respectively, the single-factor space temperature influence values of each three-dimensional subspace are obtained, including:

[0242] Using the values of the IT equipment load factor in the target computer room, the positions of each three-dimensional subspace, and the thermal conductivity of air as the second input data, perform a preset operation on the second input data through the second sub-model to obtain the single-factor space temperature influence values of the IT equipment load factor on each three-dimensional subspace in the target computer room.

[0243] Optionally, the sub-model includes: a third sub-model expressing the influence of the environmental factor on the computer room temperature, and the environmental factor includes one or more of the following data: outdoor temperature and air humidity. By using the sub-model to perform gain calculations on the values of the preset influence factors of each three-dimensional subspace respectively, the single-factor space temperature influence values of each three-dimensional subspace are obtained, including:

[0244] Using the value of the environmental factor as the third input data, perform a preset operation on the third input data through the third sub-model to obtain the single-factor space temperature influence value of the environmental factor on each three-dimensional subspace in the target computer room.

[0245] Optionally, the spatial dimensionality reduction process on the cumulative space temperature influence value to obtain the air-conditioning temperature gain matrix includes:

[0246] Determine the projection plane of the three-dimensional subspace, where the projection plane is: a plane parallel to the side surface of the three-dimensional subspace in the horizontal direction, or a plane parallel to the side surface of the three-dimensional subspace in a vertical direction;

[0247] Project the three-dimensional subspace onto the projection plane;

[0248] Take the mean value of the cumulative space temperature influence values corresponding to the three-dimensional subspaces projected onto the same area on the projection plane as the cumulative space temperature influence value corresponding to the corresponding area of the projection plane;

[0249] Obtain the air-conditioning temperature gain matrix according to the cumulative space temperature influence values corresponding to each area in the projection plane.

[0250] In some embodiments of the present application, the device further includes: a computer room multi-dimensional temperature influence model training module (not shown in the figure);

[0251] The computer room multi-dimensional temperature influence model training module is used to train a computer room multi-dimensional temperature influence model according to the historical temperature of the target computer room and the values of the preset influence factors corresponding to the historical temperature.

[0252] Optionally, training the computer room multi-dimensional temperature influence model according to the historical temperature of the target computer room and the values of the preset influence factors corresponding to the historical temperature includes:

[0253] For each specified historical moment, perform the following steps respectively to obtain the air-conditioning temperature gain matrix corresponding to the specified historical moment:

[0254] [[ID=3,1]]Perform gain calculation on the values of the preset influence factors of each three-dimensional subspace at the specified historical moment through the sub-model to obtain the single-factor space temperature influence value of each three-dimensional subspace corresponding to the specified historical moment;

[0255] Fuse the single-factor space temperature influence values to obtain the cumulative space temperature influence value of the preset influence factor on each three-dimensional subspace;

[0256] Perform spatial dimensionality reduction on the cumulative spatial temperature influence value to obtain the air-conditioning temperature gain matrix corresponding to the specified historical moment;

[0257] Train the sub-models corresponding to the preset influence factors in the multi-dimensional temperature influence model of the computer room, and the feature encoding network, according to the air-conditioning temperature gain matrix corresponding to each specified historical moment and the historical temperature of the target computer room.

[0258] Optionally, the sub-model includes: a first sub-model expressing the influence of the air-conditioning operation parameter factor on the computer room temperature, and the air-conditioning operation parameter factor includes one or more of the following data: air-conditioning temperature, fan rotation speed, wind direction angle, air-conditioning operation duration, air-conditioning position, and the thermal conductivity of air. By using the sub-model to perform gain calculation on the values of the preset influence factors of each three-dimensional subspace at the specified historical moment respectively, the single-factor spatial temperature influence value of each three-dimensional subspace corresponding to the specified historical moment is obtained, including:

[0259] Taking the value of the air-conditioning operation parameter factor at the specified historical moment and the positions of each three-dimensional subspace as the first input data, and performing a preset operation on the first input data through the first sub-model to obtain the single-factor spatial temperature influence value of the air-conditioning operation parameter factor on each three-dimensional subspace corresponding to the specified historical moment.

[0260] Optionally, the sub-model includes: a second sub-model expressing the influence of the IT equipment load factor on the computer room temperature, and the IT equipment load factor includes one or more of the following data: CPU usage rate, memory usage rate of the server, and IO consumption rate, position of the server, temperature of the server. By using the sub-model to perform gain calculation on the values of the preset influence factors of each three-dimensional subspace at the specified historical moment respectively, the single-factor spatial temperature influence value of each three-dimensional subspace corresponding to the specified historical moment is obtained, including:

[0261] Taking the value of the IT equipment load factor in the target computer room at the specified historical moment, the positions of each three-dimensional subspace, and the thermal conductivity of air as the second input data, and performing a preset operation on the second input data through the second sub-model to obtain the single-factor spatial temperature influence value of the IT equipment load factor on each three-dimensional subspace in the target computer room corresponding to the specified historical moment.

[0262] Optionally, the sub-model includes: a third sub-model for expressing the influence of the environmental factors on the computer room temperature, where the environmental factors include one or more of the following data: outdoor temperature and air humidity, and the gain calculation is respectively performed on the values of the preset influence factors of each of the three-dimensional sub-spaces at the specified historical moment through the sub-model to obtain the single-factor space temperature influence value of each three-dimensional sub-space corresponding to the specified historical moment, including:

[0263] Taking the values of the environmental factors at the specified historical moment as the third input data, and performing a preset operation on the third input data through the third sub-model to obtain the single-factor space temperature influence value of the environmental factors on each three-dimensional sub-space in the target computer room corresponding to the specified historical moment.

[0264] Optionally, the spatial dimensionality reduction processing is performed on the cumulative space temperature influence value to obtain the air-conditioning temperature gain matrix corresponding to the specified historical moment, including:

[0265] Determine the projection plane of the three-dimensional sub-space, where the projection plane is: a plane parallel to the side surface of the three-dimensional sub-space in the horizontal direction, or a plane parallel to the side surface of the three-dimensional sub-space in a vertical direction;

[0266] Project the three-dimensional sub-space onto the projection plane;

[0267] Taking the mean value of the cumulative space temperature influence values corresponding to the three-dimensional sub-spaces projected onto the same area on the projection plane as the cumulative space temperature influence value corresponding to the corresponding area of the projection plane;

[0268] According to the cumulative space temperature influence values corresponding to each area in the projection plane, obtain the air-conditioning temperature gain matrix corresponding to the specified historical moment.

[0269] The computer room air-conditioning temperature regulation device disclosed in the embodiments of the present application is used to implement the computer room air-conditioning temperature regulation method described in the embodiments of the present application. The specific implementation manners of the modules of the device will not be elaborated here, and reference may be made to the specific implementation manners of the corresponding steps in the method embodiments.

[0270] An air conditioner temperature regulation device for a computer room disclosed in an embodiment of the present application responds to the temperature regulation of the air conditioner in a target computer room, and obtains real-time data of the target temperature of the target computer room and preset influencing factors, where the preset influencing factors include: air conditioner operation parameter factors, IT equipment load factors, and environmental factors; using the real-time data and the target temperature as the input of a pre-trained multi-dimensional temperature influence model for the computer room, and solving the target control temperature of the air conditioner corresponding to the real-time data and the target temperature through the multi-dimensional temperature influence model for the computer room; based on the target control temperature, adjusting the temperature of the air conditioner, realizing the automatic control of the temperature of the air conditioner in the computer room by combining multi-dimensional factors affecting the temperature of the computer room, improving the real-time performance and temperature regulation accuracy of the temperature regulation of the air conditioner in the computer room, helping to reduce the energy consumption of the air conditioner in the computer room, and achieving a dynamic balance between the temperature of the computer room and the energy consumption of the air conditioner.

[0271] Further, the air conditioner temperature regulation device for a computer room disclosed in an embodiment of the present application, after responding to the temperature regulation of the air conditioner in a target computer room and obtaining the real-time data of the target temperature of the target computer room and the preset influencing factors, and before adjusting the temperature of the air conditioner based on the target control temperature, according to the real-time data and the target control temperature, estimates the temperature of the target computer room after implementing the temperature regulation through a pre-trained computer room temperature prediction model as the predicted temperature, and after adjusting the temperature of the air conditioner based on the target control temperature, obtains the measured temperature of the target computer room after implementing the temperature regulation, and compares the predicted temperature with the measured temperature; when the absolute value of the difference between the predicted temperature and the measured temperature meets a preset optimization threshold, the multi-dimensional temperature influence model for the computer room is optimized and trained, thereby further improving the accuracy of the temperature regulation of the air conditioner in the computer room and improving the energy-saving effect.

[0272] Each embodiment in this specification is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. The same or similar parts among the embodiments can be referred to each other. For the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method embodiment.

[0273] The above provides a detailed introduction to an air conditioner temperature regulation method and device for a computer room provided by the present application. Specific examples are used in this article to elaborate on the principle and implementation manner of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for those of ordinary skill in the art, according to the idea of the present application, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present application.

[0274] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative labor.

[0275] Each component embodiment of the present application can be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art should understand that a microprocessor or a digital signal processor (DSP) can be used in practice to implement some or all of the functions of some or all of the components in the electronic device according to the embodiments of the present application. The present application can also be implemented as a device or apparatus program (for example, a computer program and a computer program product) for executing some or all of the methods described herein. Such a program implementing the present application can be stored on a computer-readable medium or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.

[0276] For example, Figure 9 An electronic device that can implement the method according to the present application is shown. The electronic device can be a PC, a mobile terminal, a personal digital assistant, a tablet computer, etc. Traditionally, the electronic device includes a processor 910, a memory 920, and program code 930 stored on the memory 920 and executable on the processor 910. When the processor 910 executes the program code 930, the method described in the above embodiment is implemented. The memory 920 can be a computer program product or a computer-readable medium. The memory 920 can be an electronic memory such as a flash memory, an EEPROM (electrically erasable programmable read-only memory), an EPROM, a hard disk, or a ROM. The memory 920 has a storage space 9201 for the program code 930 of a computer program for executing any method step in the above method. For example, the storage space 9201 for the program code 930 can include respective computer programs for implementing various steps in the above method. The program code 930 is computer-readable code. These computer programs can be read from or written into one or more computer program products. These computer program products include program code carriers such as hard disks, compact discs (CDs), memory cards, or floppy disks. The computer program includes computer-readable code, and when the computer-readable code runs on the electronic device, it causes the electronic device to execute the method according to the above embodiment.

[0277] The embodiment of the present application also discloses a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the steps of the computer room air conditioner temperature regulation method described in the embodiment of the present application are implemented.

[0278] Such a computer program product may be a computer-readable storage medium, and the computer-readable storage medium may have storage segments, storage spaces, etc. arranged similarly to the memory 920 in the Figure 9 shown electronic device. The program code may be stored in the computer-readable storage medium in a compressed form in an appropriate manner, for example. The computer-readable storage medium is generally a portable or fixed storage unit as described in reference Figure 10 above. Generally, the storage unit includes computer-readable code 930'. The computer-readable code 930' is the code read by the processor. When these codes are executed by the processor, the respective steps in the method described above are implemented.

[0279] As used herein, the terms "one embodiment", "embodiment", or "one or more embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present application. In addition, note that the examples of the phrase "in one embodiment" herein do not necessarily all refer to the same embodiment.

[0280] In the specification provided herein, a large number of specific details are set forth. However, it can be understood that the embodiments of the present application may be practiced without these specific details. In some instances, well-known methods, structures, and technologies have not been shown in detail so as not to obscure the understanding of this specification.

[0281] In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in the claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present application may be implemented by means of hardware including several different elements and by means of a suitably programmed computer. In a unit claim listing several devices, several of these devices may be embodied by the same item of hardware. The use of the words first, second, and third, etc. does not denote any order. These words may be interpreted as names.

[0282] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements 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 the present application.

Claims

1. A method for adjusting the temperature of an air conditioner in a computer room, characterized in that, Including: In response to temperature adjustment of the air conditioner in the target computer room, obtaining the target temperature of the target computer room and real-time data of preset influence factors, where the preset influence factors include: air conditioner operation parameter factors, IT equipment load factors, and environmental factors; Using the real-time data and the target temperature as inputs to a pre-trained computer room multi-dimensional temperature influence model, and solving the target control temperature of the air conditioner corresponding to the real-time data and the target temperature through the computer room multi-dimensional temperature influence model; Adjusting the temperature of the air conditioner based on the target control temperature; The target computer room is divided into several three-dimensional sub-spaces, and the computer room multi-dimensional temperature influence model includes: sub-models corresponding to each of the preset influence factors, a feature generation sub-model, and a feature encoding network; The step of using the real-time data and the target temperature as inputs to a pre-trained computer room multi-dimensional temperature influence model, and solving the target control temperature of the air conditioner corresponding to the real-time data and the target temperature through the computer room multi-dimensional temperature influence model includes: Performing gain calculation on the values of the preset influence factors in each of the three-dimensional sub-spaces through the sub-model to obtain the single-factor space temperature influence values of each three-dimensional sub-space; Through the feature generation sub-model, fusing the single-factor space temperature influence values to obtain the cumulative space temperature influence values of the preset influence factors on each three-dimensional sub-space; and Performing space dimensionality reduction processing on the cumulative space temperature influence values to obtain an air conditioner temperature gain matrix; Performing encoding mapping on the air conditioner temperature gain matrix through the feature encoding network to obtain the target control temperature of the air conditioner corresponding to the real-time data and the target temperature.

2. The method according to claim 1, wherein After the step of using the real-time data and the target temperature as inputs to a pre-trained computer room multi-dimensional temperature influence model, and solving the target control temperature of the air conditioner corresponding to the real-time data and the target temperature through the computer room multi-dimensional temperature influence model, it further includes: According to the real-time data and the target control temperature, estimating the computer room temperature after implementing the temperature adjustment in the target computer room through a pre-trained computer room temperature prediction model as the predicted temperature; After the step of adjusting the temperature of the air conditioner based on the target control temperature, it further includes: Obtaining the measured temperature of the target computer room; When the absolute value of the difference between the predicted temperature and the measured temperature satisfies a preset tuning threshold, performing tuning training on the computer room multi-dimensional temperature influence model.

3. The method according to claim 1, wherein The sub-model includes: a first sub-model expressing the influence of the air conditioner operation parameter factors on the computer room temperature, and the air conditioner operation parameter factors include one or more of the following data: air conditioner temperature, fan rotation speed, wind direction angle, air conditioner operation duration, air conditioner position, and thermal conductivity of air, The step of performing gain calculation on the values of the preset influence factors in each of the three-dimensional sub-spaces through the sub-model to obtain the single-factor space temperature influence values of each three-dimensional sub-space includes: Taking the values of the air conditioner operation parameter factors and the positions of the three-dimensional subspaces as the first input data, performing a preset operation on the first input data through the first sub-model, and obtaining the single-factor space temperature influence values of the air conditioner operation parameter factors on the three-dimensional subspaces.

4. The method according to claim 1, characterized in that The sub-model includes: a second sub-model expressing the influence of the IT equipment load factor on the computer room temperature, where the IT equipment load factor includes one or more of the following data: CPU usage rate of the server, memory usage rate, and IO consumption rate, the position of the server, and the temperature of the server. The obtaining of the single-factor space temperature influence values of the three-dimensional subspaces by respectively performing gain calculations on the values of the preset influence factors of the three-dimensional subspaces through the sub-model includes: Taking the values of the IT equipment load factor in the target computer room, the positions of the three-dimensional subspaces, and the thermal conductivity of the air as the second input data, performing a preset operation on the second input data through the second sub-model, and obtaining the single-factor space temperature influence values of the IT equipment load factor on the three-dimensional subspaces in the target computer room.

5. The method according to claim 1, characterized in that, The sub-model includes: a third sub-model expressing the influence of the environmental factor on the computer room temperature, where the environmental factor includes one or more of the following data: outdoor temperature and air humidity. The obtaining of the single-factor space temperature influence values of the three-dimensional subspaces by respectively performing gain calculations on the values of the preset influence factors of the three-dimensional subspaces through the sub-model includes: Taking the values of the environmental factor as the third input data, performing a preset operation on the third input data through the third sub-model, and obtaining the single-factor space temperature influence values of the environmental factor on the three-dimensional subspaces in the target computer room.

6. The method according to claim 1, characterized in that, The performing of space dimensionality reduction processing on the cumulative space temperature influence value to obtain the air conditioner temperature gain matrix includes: Determining the projection plane of the three-dimensional subspace, where the projection plane is: a plane parallel to the side surface of the three-dimensional subspace in the horizontal direction, or a plane parallel to the side surface of the three-dimensional subspace in a vertical direction; Projecting the three-dimensional subspace onto the projection plane; Taking the mean value of the cumulative space temperature influence values corresponding to the three-dimensional subspaces projected onto the same area on the projection plane as the cumulative space temperature influence value corresponding to the corresponding area of the projection plane; Obtaining the air conditioner temperature gain matrix according to the cumulative space temperature influence values corresponding to the areas in the projection plane.

7. An air conditioner temperature adjustment device for a computer room, characterized in that, Including: A computer room target temperature and influence data acquisition module, configured to, in response to adjusting the temperature of the air conditioner in the target computer room, acquire the target temperature of the target computer room and the real-time data of the preset influence factors, where the preset influence factors include: air conditioner operation parameter factors, IT equipment load factors, and environmental factors; An air conditioner temperature solving module, configured to use the real-time data and the target temperature as the input of a pre-trained computer room multi-dimensional temperature influence model, and solve the target control temperature of the air conditioner corresponding to the real-time data and the target temperature through the computer room multi-dimensional temperature influence model. An air conditioner temperature adjustment module for adjusting the temperature of the air conditioner based on the target control temperature; The target computer room is divided into several three-dimensional sub-spaces. The multi-dimensional temperature influence model of the computer room includes: sub-models corresponding to each of the preset influence factors, a feature generation sub-model, and a feature encoding network; The air conditioner temperature solution module is further configured to: Perform gain calculation on the values of the preset influence factors of each of the three-dimensional sub-spaces through the sub-model to obtain the single-factor space temperature influence values of each three-dimensional sub-space; Fuse the single-factor space temperature influence values through the feature generation sub-model to obtain the cumulative space temperature influence values of the preset influence factors on each three-dimensional sub-space; and Perform spatial dimensionality reduction processing on the cumulative space temperature influence values to obtain an air conditioner temperature gain matrix; Perform encoding mapping on the air conditioner temperature gain matrix through the feature encoding network to obtain the target control temperature of the air conditioner corresponding to the real-time data and the target temperature.

8. The device according to claim 7, characterized in that, It further includes: A computer room temperature prediction module for predicting, according to the real-time data and the target control temperature, the computer room temperature of the target computer room after implementing the temperature adjustment through a pre-trained computer room temperature prediction model as the predicted temperature; A computer room measured temperature acquisition module for acquiring the measured temperature of the target computer room after adjusting the temperature of the air conditioner based on the target control temperature; A model optimization processing module for optimizing and training the multi-dimensional temperature influence model of the computer room when the absolute value of the difference between the predicted temperature and the measured temperature satisfies a preset optimization threshold.

9. An electronic device, comprising a memory, a processor, and program code stored on the memory and executable on the processor, characterized in that, When the processor executes the program code, it implements the computer room air conditioner temperature adjustment method according to any one of claims 1 to 6.

10. A computer-readable storage medium having program code stored thereon, characterized in that, When the program code is executed by the processor, it implements the steps of the computer room air conditioner temperature adjustment method according to any one of claims 1 to 6.

Images