Control method and device of temperature control equipment, equipment and readable storage medium

Abstract

The embodiment of the application provides a kind of temperature control equipment control method, device, equipment and readable storage medium, comprising: the environmental temperature of data center is collected in real time, and the working state of temperature control equipment is controlled based on temperature control model.Therein, temperature control model is obtained by training as follows: the operating state data of temperature control equipment is obtained, and the operating state data is grouped according to first preset feature;For any operating state data group, remove the operating state data that does not meet the training requirements of operating state data;Remove the operating state data group that does not meet the grouping training requirements;Any remaining operating state data group is divided into training sample set and test sample set;After training sample set is aggregated, it is input to temperature control model for training, and after training is completed, test sample set is aggregated and input to temperature control model for testing;Determine that model evaluation index meets training target to end training, otherwise, training is carried out again.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automation, and in particular to a control method and device of a temperature control equipment, an equipment and a readable storage medium. BACKGROUND

[0002] With the rapid development of big data, the amount of data processing is growing, and the number of data centers is also increasing rapidly. A data center is usually composed of multiple cabinets, multiple servers, uninterruptible power supply (UPS), temperature control equipment, temperature and humidity sensors, etc. The environment of the data center is constantly affected by various factors such as heat generated by the server, environmental changes, etc., and the temperature control equipment (including cooling pumps, refrigeration pumps, cooling towers, etc.) is used to adjust the temperature of the data center, so as to ensure the constant temperature of the data center. The air conditioner consumes a large amount of electric energy in the use process, and the efficiency of electric energy conversion into cold energy is different under different working power, so how to control the temperature control equipment and reduce the total power consumption of the temperature control equipment is a problem to be solved. SUMMARY

[0003] The embodiments of the present application provide a control method, device, equipment and readable storage medium of a temperature control equipment, so as to reduce the total power consumption of the temperature control equipment of the data center.

[0004] The embodiments of the present application provide a control method of a temperature control equipment of a data center, comprising:

[0005] Real-time acquisition of the environmental temperature of the data center, and control of the working state of the temperature control equipment based on a temperature control model;

[0006] The temperature control model is obtained by training in the following manner:

[0007] Obtaining the running state data of the temperature control equipment, grouping the running state data according to the first preset characteristics of the running state data, and obtaining a plurality of running state data groups;

[0008] For any running state data group, removing the running state data in the running state data group that does not meet the training requirements of the running state data, and the remaining running state data constitutes a first running state data group;

[0009] Removing all first running state data groups that do not meet the grouping training requirements from the first running state data groups, and the remaining first running state data groups are used as second running state data groups;

[0010] For any second running state data group, the running state data in the second running state data group is divided into a training sample set and a test sample set;

[0011] aggregate the obtained training sample set, and aggregate the obtained test sample set;

[0012] input the aggregated training sample set to the temperature control model for training, and input the aggregated test sample set to the temperature control model for test calculation of a model evaluation index after the training ends, and if the model evaluation index meets a training target, end the training, otherwise retrain the temperature control model until the model evaluation index meets the training target.

[0013] Optionally, the running state data includes a working frequency and a power of the temperature control device.

[0014] The first preset feature includes a working frequency of the temperature control device.

[0015] Optionally, the removing of the running state data in the running state data group that does not meet the running state data training requirement includes:

[0016] sorting all the running state data in the running state data group according to power size;

[0017] removing a first number of running state data from a start position to a large size in the sorting and / or a second number of running state data from an end position to a small size;

[0018] The first number is a number of a preset first proportion of all the running state data in the running state data group, and the second number is a number of a preset second proportion of all the running state data in the running state data group.

[0019] Optionally, the removing of the first running state data group that does not meet the grouping training requirement in all the first running state data groups includes:

[0020] removing the first running state data group with a number of running state data less than a preset threshold in all the first running state data groups.

[0021] Optionally, the dividing of the running state data in the second running state data group into the training sample set and the test sample set includes:

[0022] determining a training sample proportion of the running state data in the training sample set to a number of the running state data in the second running state data group according to the number of the running state data in the second running state data group;

[0023] dividing the running state data in the second running state data group into the training sample set and the test sample set according to the training sample proportion.

[0024] Optionally, the step of dividing the operation state data in the second operation state data set into a training sample set and a test sample set comprises:

[0025] dividing the operation state data in the second operation state data set into a training sample set and a test sample set;

[0026] respectively judging whether the numerical distribution state of the second preset feature in the operation state data in the training sample set and the test sample set meets the sample division requirement;

[0027] If the numerical distribution state of the second preset feature in the operation state data in the training sample set and / or the test sample set does not meet the sample division requirement, returning to the step of dividing the operation state data in the second operation state data set into a training sample set and a test sample set.

[0028] Optionally, the model evaluation index comprises at least one of:

[0029] mean absolute error (MAE), mean square error (MSE), root mean square error (RMSE), and determination coefficient R 2 , and mean absolute percentage error (MAPE).

[0030] Based on the same inventive concept, the embodiment of the present application also provides a control device of a temperature control equipment of a data center, comprising:

[0031] a control module configured to collect the ambient temperature of the data center in real time and control the working state of the temperature control equipment based on a temperature control model;

[0032] a training module configured to obtain operation state data of the temperature control equipment, group the operation state data according to a first preset feature of the operation state data, and obtain a plurality of operation state data sets; for any operation state data set, remove operation state data in the operation state data set that does not meet the training requirement of the operation state data, and the remaining operation state data constitutes a first operation state data set; remove the first operation state data set that does not meet the grouping training requirement from all the first operation state data sets, and the remaining first operation state data set is used as a second operation state data set; for any second operation state data set, divide the operation state data in the second operation state data set into a training sample set and a test sample set; aggregate the obtained training sample set and the obtained test sample set; input the aggregated training sample set into the temperature control model for training, and after the training is completed, input the aggregated test sample set into the temperature control model for test to calculate a model evaluation index; if the model evaluation index meets the training target, the training is ended, otherwise, the temperature control model is retrained until the model evaluation index meets the training target.

[0033] Based on the same inventive concept, the embodiment of the present application further provides an electronic device, comprising: a processor and a memory for storing instructions executable by the processor.

[0034] The processor is configured to execute the instructions to implement the control method of the temperature control device of the data center.

[0035] Based on the same inventive concept, the embodiment of the present application further provides a readable storage medium, the readable storage medium stores a computer program, the computer program is used to implement the control method of the temperature control device of the data center.

[0036] The present application has the following advantages:

[0037] The temperature control device control method, device, equipment and readable storage medium provided by the embodiment of the present application can effectively reduce the workload of manually controlling the temperature control device of the data center by using the regression model to regulate the working state of the temperature control device of the data center, not only can save manpower, but also can reduce the power consumption of the temperature control device of the data center. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 The flow chart of the temperature control device control method provided by the embodiment of the present application is provided.

[0039] Figure 2 One of the flow charts of part of steps of the temperature control device control method provided by the embodiment of the present application is provided.

[0040] Figure 3 One of the flow charts of part of steps of the temperature control device control method provided by the embodiment of the present application is provided

[0041] Figure 4 One of the flow charts of part of steps of the temperature control device control method provided by the embodiment of the present application is provided.

[0042] Figure 5 The structural schematic diagram of the temperature control device provided by the embodiment of the present application is provided.

[0043] Figure 6 The structural schematic diagram of the electronic device provided by the embodiment of the present application is provided. DETAILED DESCRIPTION

[0044] In order to make the above objectives, characteristics and advantages of the present application more apparent, comprehensible and easier to understand, the present application will be further described below in conjunction with the accompanying drawings and embodiments. However, the example embodiments can be implemented in various forms, and should not be understood as being limited to the embodiments set forth herein; on the contrary, these embodiments are provided so as to make the present application more comprehensive and complete, and to fully convey the ideas of the example embodiments to those skilled in the art. The same reference signs in the drawings represent the same or similar structures, and thus repeated descriptions thereof will be omitted. The expressions of position and direction described in the present application are described with reference to the drawings, but can be changed as needed, and the changes made are included in the scope of protection of the present application. The drawings of the present application are only used to illustrate the relative positional relationship and do not represent the true proportions.

[0045] It should be noted that specific details are set forth in the following description in order to provide a thorough understanding of the present application. However, the present application can be practiced in a variety of ways other than those described herein, and skilled artisans can make similar substitutions without departing from the scope of the present application. Therefore, the present application is not limited to the specific embodiments disclosed below. The subsequent description of the specification is a preferred embodiment for implementing the present application, and is intended to illustrate the general principles of the present application, rather than to limit the scope of the present application. The scope of protection of the present application is defined by the appended claims.

[0046] The control method, device, equipment and readable storage medium of the temperature control device provided by the embodiments of the present application will be described in detail below in conjunction with the accompanying drawings.

[0047] The embodiments of the present application provide a control method of a temperature control device of a data center, comprising:

[0048] The environmental temperature of the data center is collected in real time, and the working state of the temperature control device is controlled based on a temperature control model.

[0049] In the specific implementation process, the temperature control model can be trained by using a regression model such as an xgboost model, a random forest (RF) model, a support vector machine (SVM) model, and a neural network model.

[0050] As shown in the formula (1), the temperature control model is trained by the following method: Figure 1

[0051] S110, obtaining running state data of the temperature control device, grouping the running state data according to a first preset feature of the running state data, and obtaining a plurality of running state data groups.

[0052] ​In the implementation process, the temperature control device can include but is not limited to a cooling pump, a cooling tower, a refrigeration pump and the like. In the process of obtaining the operation state data of the temperature control device, the operation state data can be obtained from the preset data source according to the settings of the algorithm development engineer.

[0053] In S120, for any of the operation state data sets, the operation state data that does not meet the operation state data training requirement in the operation state data set is removed, and the remaining operation state data constitutes a first operation state data set.

[0054] In S130, the first operation state data set that does not meet the grouping training requirement in all the first operation state data sets is removed, and the remaining first operation state data set is used as a second operation state data set.

[0055] In S140, for any of the second operation state data sets, the operation state data in the second operation state data set is divided into a training sample set and a test sample set.

[0056] In S150, the obtained training sample set is aggregated, and the obtained test sample set is aggregated.

[0057] In S160, the aggregated training sample set is input into a temperature control model for training, and after the training is completed, the aggregated test sample set is input into the temperature control model for test calculation of a model evaluation index.

[0058] In S170, it is judged whether the model evaluation index meets a training target.

[0059] If the result of S170 is no, the step S160 is returned; if the result of S170 is yes, the training is ended.

[0060] The application can effectively reduce the workload of manually controlling the temperature control device of the data center by using a regression model to regulate the working state of the temperature control device of the data center, not only can save manpower, but also can reduce the power consumption of the temperature control device of the data center through experiments. Of course, the technical solution of the application can be used in other similar scenarios in addition to the temperature control device of the data center.

[0061] Optionally, the model evaluation index includes at least one of the following:

[0062] ① Mean Absolute Error (MAE)

[0063]

[0064] The index is used to evaluate the closeness of the predicted results and the real data set, and the smaller the value is, the better the fitting effect is.

[0065] ②Mean Square Error (MSE)

[0066]

[0067] The index calculates the mean of the sum of squares of the error of the corresponding sample points of the fitted data and the original data. The smaller the value, the better the fitting effect.

[0068] ③Root Mean Square Error (RMSE)

[0069]

[0070] Used to indicate how large the error will be in the model prediction. The smaller the value, the better the fitting effect.

[0071] ④Coefficient of determination R 2 (R squared, Coefficient of determination)

[0072]

[0073] The index explains the variance score of the regression model. The value ranges from 0 to 1. The larger the value, the closer to 1, indicating that the independent variable can better explain the variance of the dependent variable.

[0074] ⑤Mean Absolute Percentage Error (MAPE)

[0075]

[0076] The value can accurately reflect the size of the actual prediction error. The smaller the value, the better the fitting effect of the prediction model, and the better the accuracy.

[0077] Where i is the serial number of the test sample, n is the number of test samples, x i is the input value of the temperature control model in the test sample, f(x i ) is the predicted value output by the temperature control model according to x i , y i is the true value that the temperature control model should output according to x i , and is the average value of y i .

[0078] Further, the operating state data includes the working frequency and power of the temperature control device.

[0079] The first preset feature includes a working frequency of the temperature control device.

[0080] Further, as Figure 2 shown, in the step S120, the operation state data that does not meet the operation state data training requirement in the operation state data set is removed, specifically including:

[0081] S121, all operation state data in the operation state data set is sorted according to the power size.

[0082] S122, the first number of operation state data from the start position to the largest in the sorting and / or the second number of operation state data from the end position to the smallest is removed.

[0083] The first number is a number that occupies a preset first proportion of all operation state data in the operation state data set, and the second number is a number that occupies a preset second proportion of all operation state data in the operation state data set.

[0084] In the specific implementation process, the algorithm development engineer can set the preset first proportion and the preset second proportion according to experience. In this way, by removing the data with large and small power in the operation state data set, the abnormal data pollution of the training sample caused by the power shock of the temperature control device at the same working frequency can be avoided.

[0085] Further, in the step S130, the first operation state data set that does not meet the grouping training requirement in all first operation state data sets is removed, including:

[0086] Removing the first operation state data set with the number of operation state data less than the preset threshold in all first operation state data sets.

[0087] In the specific implementation process, the algorithm development engineer can set the preset threshold according to experience, so as to remove the operation state data set with too little data and avoid the overfitting phenomenon of the temperature control model obtained by training.

[0088] Further, as Figure 3 shown, in the step S140, the operation state data in the second operation state data set is divided into a training sample set and a test sample set, including:

[0089] S141, according to the number of operation state data of the second operation state data set, the training sample proportion of the operation state data in the training sample set to the number of operation state data of the second operation state data set is determined.

[0090] S142, according to the training sample ratio, the operating state data in the second operating state data set is divided into a training sample set and a test sample set.

[0091] In the implementation process, the algorithm development engineer can set the training sample ratio according to experience. For example, the training sample ratio can be relatively smaller when the number of operating state data in the second operating state data set is larger, so as to further increase the number of operating state data in the test sample set and facilitate the subsequent judgment of the training effect of the model.

[0092] In the process of manually controlling the temperature control equipment of the data center, the operation and maintenance personnel of the temperature control equipment of the data center may not implement automatic control on the equipment for safety protection, which will cause the operating state data of the temperature control equipment to be mostly segmented frequency data, and the numerical distribution of the operating state data is uneven. When constructing the training sample and the test sample, if no special processing is performed, the operating state data of the training sample and the test sample will be unbalanced, which will affect the training effect of the temperature control model. However, if the data is directly shuffled and the training sample set and the test sample set are randomly divided, the training effect of the temperature control model is also not good. To solve this problem, further, as shown in Figure 4 In the step S140, the operating state data in the second operating state data set is divided into a training sample set and a test sample set, including:

[0093] S143, the operating state data in the second operating state data set is divided into a training sample set and a test sample set.

[0094] In the implementation process, the implementation mode of the step S143 can be implemented in the mode of the steps S141 and S142 as described above, or can be implemented in other existing modes, which will not be described here.

[0095] S144, respectively judging whether the numerical distribution state of the second preset feature in the operating state data in the training sample set and the test sample set meets the sample division requirement.

[0096] In the implementation process, the sample division requirement can be set to the operating state data of a specified category in the training sample set and the test sample set meeting the uniform distribution and other numerical distribution conditions.

[0097] If the numerical distribution state of the second preset feature in the operating state data of the training sample set and / or the test sample set does not meet the sample division requirement, the step S141 is returned; if the numerical distribution state of the second preset feature in the operating state data of the training sample set and the test sample set meets the sample division requirement, the division process is completed, and the subsequent step S150 is executed.

[0098] In this way, by adjusting the training sample set and the test sample set according to the numerical values in the division construction process, the numerical imbalance problem of the training sample and the test sample is solved, so that the temperature control model with a relatively continuous output target value can be trained in the case that the original running state data of the temperature control device is relatively discrete.

[0099] For a certain data center, through experiments, the evaluation index of the temperature control model of the cooling pump trained by using the prior art is: MAE = 0.2497364524, R 2 = 0.97544387. After training the temperature control model of the cooling pump by using all the above embodiments, MAE = 0.1609193671, R 2 = 0.994254946. The control effect of the temperature control model on the cooling pump is obviously improved.

[0100] Based on the same inventive concept, the embodiment of the present application also provides a control device of a temperature control equipment of a data center, as shown in Figure 5 , comprising:

[0101] a control module M1, configured to collect the environmental temperature of the data center in real time, and control the working state of the temperature control equipment based on the temperature control model;

[0102] a training module M2, configured to obtain the running state data of the temperature control equipment, group the running state data according to a first preset feature of the running state data, to obtain a plurality of running state data groups; for any one of the running state data groups, remove the running state data in the running state data group that does not meet the training requirements of the running state data, and the remaining running state data constitutes a first running state data group; remove the first running state data group that does not meet the grouping training requirements from all the first running state data groups, and the remaining first running state data group is used as a second running state data group; for any one of the second running state data groups, divide the running state data in the second running state data group into a training sample set and a test sample set; aggregate the obtained training sample set, and aggregate the obtained test sample set; input the aggregated training sample set into the temperature control model for training, and after the training is completed, input the aggregated test sample set into the temperature control model for test calculation of the model evaluation index; if the model evaluation index meets the training target, the training is ended, otherwise the temperature control model is retrained until the model evaluation index meets the training target.

[0103] Optionally, the running state data includes the working frequency and the power of the temperature control equipment.

[0104] The first preset feature includes the working frequency of the temperature control equipment.

[0105] Optionally, the removing the operation state data that does not meet the operation state data training requirement in the operation state data set comprises:

[0106] sequencing all the operation state data in the operation state data set according to power size;

[0107] removing the first number of operation state data from the start position to the end position in the sequence from small to large and / or the second number of operation state data from the end position to the start position in the sequence from large to small;

[0108] The first number is a number that accounts for a preset first proportion of all the operation state data in the operation state data set, and the second number is a number that accounts for a preset second proportion of all the operation state data in the operation state data set.

[0109] Optionally, the removing the first operation state data set that does not meet the grouping training requirement in all the first operation state data sets comprises:

[0110] removing the first operation state data set whose number of operation state data is less than a preset threshold value in all the first operation state data sets.

[0111] Optionally, the dividing the operation state data in the second operation state data set into a training sample set and a test sample set comprises:

[0112] determining a training sample proportion of the operation state data in the training sample set to the number of operation state data in the second operation state data set according to the number of operation state data in the second operation state data set;

[0113] dividing the operation state data in the second operation state data set into the training sample set and the test sample set according to the training sample proportion.

[0114] Optionally, the dividing the operation state data in the second operation state data set into a training sample set and a test sample set comprises:

[0115] dividing the operation state data in the second operation state data set into the training sample set and the test sample set;

[0116] respectively judging whether the value distribution state of the second preset feature in the operation state data in the training sample set and the test sample set meets the sample division requirement;

[0117] If the value distribution state of the second preset feature in the operation state data in the training sample set and / or the test sample set does not meet the sample division requirement, returning to the step of dividing the operation state data in the second operation state data set into the training sample set and the test sample set.

[0118] Optionally, the model evaluation metrics include at least one of the following:

[0119] Mean Absolute Error (MAE), Mean Square Error (MSE), Root Mean Square Error (RMSE), Coefficient of Determination (R²) 2 Mean Absolute Percentage Error (MAPE)

[0120] In the embodiments provided in this application, it should be understood that the device embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the functional modules in the various embodiments of this application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module. The integrated modules described above can be implemented in hardware or as software functional modules. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can be stored in a retrievable storage medium.

[0121] Since the specific methods by which the various modules of the electronic device perform their operations have been described in detail in the embodiments of the control method for the temperature control equipment in the data center, they will not be repeated here.

[0122] Based on the same inventive concept, embodiments of the present invention also provide an electronic device, such as... Figure 6 As shown, it includes: a processor 110 and a memory 120 for storing instructions executable by the processor 110;

[0123] The processor 110 is configured to execute the instructions to implement the control method of the temperature control device in the data center.

[0124] In specific implementations, the device may vary significantly due to differences in configuration or performance. It may include one or more processors 110, memory 120, and readable storage media 130. The memory 120 and / or readable storage media 130 may contain one or more application programs 131 or data 132. The memory 120 and / or readable storage media 130 may also contain one or more operating systems 133, such as Windows, Mac OS, Linux, iOS, Android, Unix, FreeBSD, etc. The memory 120 and readable storage media 130 may be temporary or persistent storage. The application program 131 may include one or more of the aforementioned modules (…). Figure 6Each module can include a series of instruction operations (not shown in FIG. 1). Further, the processor 110 can be configured to communicate with the readable storage medium 130, and execute the series of instruction operations in the readable storage medium 130 on the device. The device can also include one or more power supplies 131 (not shown in FIG. 1) ; one or more network interfaces 140, including a wired network interface 141 and / or a wireless network interface 142; and one or more input / output interfaces 143. Figure 6

[0125] Based on the same inventive concept, the embodiment of the present application further provides a readable storage medium, which stores a computer program used for implementing the control method of the temperature control device of the data center.

[0126] The control method, device, equipment and readable storage medium of the temperature control device provided by the embodiment of the present application can effectively reduce the workload of manually controlling the temperature control device of the data center by using the regression model to regulate the working state of the temperature control device of the data center, which not only can save manpower, but also can reduce the power consumption of the temperature control device of the data center.

[0127] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system or a computer program product. Therefore, the present application can be in the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Moreover, the present application can be in the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0128] The present application is described with reference to the flowcharts and / or block diagrams of the method, device (system) and computer program product according to the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing equipment to produce a machine, so that the instructions executed by the computer or other programmable data processing equipment produce a device for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 The device for implementing the functions specified in one or more flows and / or blocks. Figure 1 The device for implementing the functions specified in one or more flows and / or blocks.

[0129] ​These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the flow Figure 1 The flow or flows and / or blocks Figure 1 The flow or flows and / or blocks

[0130] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions that execute on the computer or other programmable apparatus provide steps for implementing the flow Figure 1 The flow or flows and / or blocks Figure 1 The flow or flows and / or blocks

[0131] Obviously, numerous modifications and variations of the present application are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.

Claims

1. A control method of a temperature control device of a data center, characterized by, The application relates to a temperature control method and device. Real-time acquisition of the ambient temperature of a data center and control of the working state of a temperature control device based on a temperature control model; The temperature control device comprises a cooling pump, a cooling tower and a refrigeration pump. The temperature control model is an xgboost model, a random forest model or a support vector machine model, and the temperature control model is obtained by training in the following manner: Obtaining running state data of the temperature control device, grouping the running state data according to a first preset feature of the running state data, and obtaining a plurality of running state data groups; For any running state data group, all the running state data in the running state data group are sorted according to power size; a first quantity of running state data from the start position to the end position in the sorting from small to large and / or a second quantity of running state data from the end position to the start position in the sorting from large to small are removed; and the remaining running state data constitutes a first running state data group; wherein the first quantity is a quantity that accounts for a preset first proportion of all the running state data in the running state data group, and the second quantity is a quantity that accounts for a preset second proportion of all the running state data in the running state data group; All first running state data groups with a quantity of running state data less than a preset threshold are removed, and the remaining first running state data groups are used as second running state data groups; For any second running state data group, the running state data in the second running state data group is divided into a training sample set and a test sample set; The obtained training sample set is aggregated, and the obtained test sample set is aggregated; The aggregated training sample set is input into the temperature control model for training, and after the training is completed, the aggregated test sample set is input into the temperature control model for test calculation of a model evaluation index; if the model evaluation index meets a training target, the training is ended; otherwise, the temperature control model is retrained until the model evaluation index meets the training target. The model evaluation index comprises at least one of the following: mean absolute error MAE, mean squared error MSE, root mean squared error RMSE, coefficient of determination , mean absolute percentage error MAPE; The division of the running state data in the second running state data group into the training sample set and the test sample set comprises: According to the quantity of running state data in the second running state data group, a training sample proportion of the running state data in the training sample set to the quantity of running state data in the second running state data group is determined; wherein the more the quantity of running state data in the second running state data group, the smaller the training sample proportion; According to the training sample proportion, the running state data in the second running state data group is divided into the training sample set and the test sample set; It is respectively judged whether the value distribution state of a second preset feature in the running state data in the training sample set and the test sample set meets sample division requirements; If the value distribution state of the second preset feature in the running state data in the training sample set and / or the test sample set does not meet the sample division requirements, the step of dividing the running state data in the second running state data group into the training sample set and the test sample set is returned. The operation state data includes a working frequency and a power of the temperature control device; and the first preset feature includes the working frequency of the temperature control device. The sample division requirement includes that operation state data of a specified category in the training sample set and the test sample set conforms to uniform distribution.

2. A control device of a temperature control apparatus of a data center, characterized by, Comprise: The control module is configured to collect the ambient temperature of the data center in real time and control the working state of the temperature control device based on a temperature control model; the temperature control device includes a cooling pump, a cooling tower, and a refrigeration pump; and the temperature control model is an xgboost model, a random forest model, or a support vector machine model. The training module is configured to obtain operation state data of the temperature control device, group the operation state data according to a first preset feature of the operation state data, and obtain a plurality of operation state data groups; for any operation state data group, sort all operation state data in the operation state data group according to power size, remove a first number of operation state data from a start position to a large size and / or a second number of operation state data from an end position to a small size in the sorting, and form a first operation state data group with the remaining operation state data; the first number is a number of the operation state data in the operation state data group that accounts for a first preset proportion of all the operation state data, and the second number is a number of the operation state data in the operation state data group that accounts for a second preset proportion of all the operation state data; remove a first operation state data group with a number of operation state data less than a preset threshold from all the first operation state data groups, and use the remaining first operation state data groups as second operation state data groups; for any second operation state data group, divide the operation state data in the second operation state data group into a training sample set and a test sample set; aggregate the obtained training sample set and the obtained test sample set; input the aggregated training sample set into a temperature control model for training, and after the training is completed, input the aggregated test sample set into the temperature control model for test calculation of a model evaluation index; if the model evaluation index meets a training target, the training is ended, otherwise the temperature control model is retrained until the model evaluation index meets the training target; the model evaluation index includes at least one of the following: mean absolute error MAE, mean square error MSE, root mean square error RMSE, determination coefficient , mean absolute percentage error MAPE; the dividing the operating state data in the second operating state data set into the training sample set and the test sample set comprises: determining a training sample proportion of the operating state data in the training sample set in the operating state data quantity of the second operating state data set according to the operating state data quantity of the second operating state data set; wherein the more the operating state data quantity in the second operating state data set is, the smaller the training sample proportion is relatively; dividing the operating state data in the second operating state data set into the training sample set and the test sample set according to the training sample proportion; respectively judging whether the numerical value distribution state of the second preset feature in the operating state data in the training sample set and the test sample set meets the sample division requirement; if the numerical value distribution state of the second preset feature in the operating state data of the training sample set and / or the test sample set does not meet the sample division requirement, returning to the step of dividing the operating state data in the second operating state data set into the training sample set and the test sample set; wherein the operating state data comprises the working frequency and the power of the temperature control device; the first preset feature comprises the working frequency of the temperature control device; and the sample division requirement comprises that the operating state data of the specified category in the training sample set and the test sample set meets the uniform distribution.

3. An electronic device, comprising: Comprise: The processor and a memory for storing executable instructions of the processor; The processor is configured to execute the instructions to implement the control method of the temperature control device of the data center according to claim 1.

4. A readable storage medium, characterized by, The readable storage medium stores a computer program used to implement the control method of the temperature control device of the data center according to claim 1.

Images