Power consumption prediction method, device and equipment for device cluster

Abstract

The present application provides a method, apparatus and device for predicting the power consumption of a device cluster. The method includes: obtaining an initial power consumption prediction model corresponding to each device in the device cluster; inputting the load parameter data of each device into the corresponding initial power consumption prediction model to obtain the predicted power consumption data corresponding to each device; obtaining the predicted total power consumption data of the device cluster according to the predicted power consumption data corresponding to each device; taking the error between the predicted total power consumption data and the actual total power consumption data of the device cluster as feedback, adjusting each power consumption prediction model until the error meets a preset condition, and obtaining the target power consumption prediction model corresponding to each device, so as to predict the power consumption of the device cluster through the target power consumption prediction model. The technical method of the present application can improve the efficiency of predicting the power consumption of the device cluster.

Description

Technical Field

[0001] This application belongs to the technical field of artificial intelligence, and particularly relates to a method for predicting power consumption of a device cluster, a device for predicting power consumption of a device cluster, and an electronic device. Background Art

[0002] Power consumption prediction is an important basis for the planning and operation optimization of device clusters such as data centers. For devices such as servers whose power consumption varies greatly with the load, a relatively complex model is required to predict power consumption. Currently, a simple linear model is usually used to predict power consumption. For example, the total power consumption includes a basic power consumption part and a part that varies linearly with the load. This simple model is not sufficient to accurately reflect the actual situation. In addition, for model solving, the power consumption needs to be independently counted for each device, and various dynamic loads need to be artificially created to provide enough input data for the model. Therefore, the efficiency of the existing power consumption prediction for device clusters is not high.

[0003] It should be noted that the information disclosed in the above background art section is only used to enhance the understanding of the background of this application, and thus may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0004] The purpose of this application is to provide a method, device, and device for predicting power consumption of a device cluster, which at least to a certain extent overcome the technical problems such as low efficiency of power consumption prediction of device clusters in related technologies.

[0005] Other features and advantages of this application will become apparent through the following detailed description, or will be learned in part through the practice of this application.

[0006] According to one aspect of the embodiments of this application, a method for predicting power consumption of a device cluster is provided, including: obtaining initial power consumption prediction models respectively corresponding to each device in the device cluster; inputting the load parameter data of each device into the corresponding initial power consumption prediction model to obtain prediction power consumption data respectively corresponding to each device; obtaining the predicted total power consumption data of the device cluster according to the prediction power consumption data respectively corresponding to each device; using the error between the predicted total power consumption data and the actual total power consumption data of the device cluster as feedback to adjust each power consumption prediction model until the error meets a preset condition, and obtaining the target power consumption prediction models respectively corresponding to each device, so as to predict the power consumption of the device cluster through the target power consumption prediction models.

[0007] In some embodiments, obtaining the initial power consumption prediction models respectively corresponding to the devices in the device cluster includes: obtaining the historical load parameter data of each device in the device cluster, and the historical power consumption data corresponding to the historical load parameter data; using the historical power consumption data and the historical load parameter data as training samples to train the original prediction model corresponding to the device, so as to obtain the initial power consumption prediction models respectively corresponding to the devices.

[0008] In some embodiments, obtaining the historical load parameter data of each device in the device cluster, and the historical power consumption data corresponding to the historical load parameter data includes: obtaining the static configuration parameters and dynamic load parameters respectively corresponding to each device in the device cluster during multiple historical time periods, and obtaining the historical power consumption data respectively corresponding to each device during each historical time period; before using the historical power consumption data and the historical load parameter data as training samples to train the original prediction model corresponding to the device, the method further includes: establishing the original prediction model corresponding to the device according to the static configuration parameters, dynamic load parameters, and historical power consumption data.

[0009] In some embodiments, obtaining the static configuration parameters and dynamic load parameters respectively corresponding to each device includes: obtaining the static configuration parameters respectively corresponding to each device; obtaining the dynamic load parameters respectively corresponding to each device at preset time intervals.

[0010] In some embodiments, obtaining the initial power consumption prediction models respectively corresponding to the devices in the device cluster includes: classifying each device in the device cluster to obtain multiple device categories; establishing the initial power consumption prediction models corresponding to the device categories according to the device categories.

[0011] In some embodiments, establishing the initial power consumption prediction models corresponding to the device categories according to the device categories includes: determining the initial model parameters corresponding to the device categories based on the device categories; generating the initial power consumption prediction models respectively corresponding to the device categories according to the initial model parameters.

[0012] In some embodiments, using the error between the predicted total power consumption data and the actual total power consumption data of the device cluster as feedback to adjust each power consumption prediction model until the error meets the preset condition to obtain the target power consumption prediction models respectively corresponding to the devices includes: obtaining the actual total power consumption data of the device cluster in real time, and obtaining the predicted total power consumption data in real time; testing the initial power consumption prediction model according to the actually obtained total power consumption data and the predicted total power consumption data in real time; obtaining the target power consumption prediction models respectively corresponding to the devices when the test accuracy rate reaches the preset threshold.

[0013] In some embodiments, testing an initial power consumption prediction model based on the actually obtained real total power consumption data and the real total power consumption data includes: establishing a loss function according to the difference between the real total power consumption data and the predicted total power consumption data; performing gradient descent iteration on the loss function to obtain the minimized loss function and the model parameters of the initial power consumption prediction model.

[0014] According to one aspect of the embodiments of the present application, there is provided a power consumption prediction device for a device cluster, including:

[0015] A first acquisition unit, configured to acquire the initial power consumption prediction models respectively corresponding to each device in the device cluster;

[0016] A training unit, configured to input the load parameter data of each device into the corresponding initial power consumption prediction model respectively to obtain the predicted power consumption data respectively corresponding to each device;

[0017] A second acquisition unit, configured to obtain the predicted total power consumption data of the device cluster according to the predicted power consumption data respectively corresponding to each device;

[0018] A prediction unit, configured to adjust each power consumption prediction model with the error between the predicted total power consumption data and the actual total power consumption data of the device cluster as feedback until the error meets a preset condition, to obtain the target power consumption prediction models respectively corresponding to each device, so as to predict the power consumption of the device cluster through the target power consumption prediction models.

[0019] According to one aspect of the embodiments of the present application, there is provided a computer-readable medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the power consumption prediction method for a device cluster as in the above technical solution.

[0020] According to one aspect of the embodiments of the present application, there is provided an electronic device, which includes: a processor; and a memory, configured to store the executable instructions of the processor; wherein, the processor is configured to execute the power consumption prediction method for a device cluster as in the above technical solution by executing the executable instructions.

[0021] According to one aspect of the embodiments of the present application, there is provided a computer program product or a computer program, which includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the power consumption prediction method for a device cluster as in the above technical solution.

[0022] In the technical solution provided by the embodiments of the present application, first, the predicted total power consumption data is determined based on the predicted power consumption data corresponding to each device, and then the actual total power consumption data is obtained during the operation of the device cluster. Thus, taking the error between the predicted total power consumption data and the actual total power consumption data of the device cluster as feedback, each power consumption prediction model is adjusted until the error meets the preset conditions, and then the target power consumption prediction models corresponding to each device are obtained to predict the power consumption of the device cluster. In the present application, the convenience and accuracy of power consumption prediction of the device cluster are improved through the convenience and accuracy of data acquisition.

[0023] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The accompanying drawings herein are incorporated into the specification and constitute a part of this specification, showing embodiments consistent with the present application and used together with the specification to explain the principles of the present application. Obviously, the drawings in the following description are only some embodiments of the present application, and those of ordinary skill in the art can obtain other drawings based on these drawings without creative efforts.

[0025] Figure 1 Schematically shows an exemplary system architecture block diagram applying the technical solution of the present application.

[0026] Figure 2 Is a flowchart of a method for predicting power consumption of a device cluster according to an embodiment of the present application.

[0027] Figure 3 Is a flowchart of a method for predicting power consumption of a device cluster provided by another embodiment of the present application.

[0028] Figure 4 Schematically shows a structural block diagram of a device cluster power consumption prediction device provided by an embodiment of the present application.

[0029] Figure 5 Schematically shows a computer system structural block diagram of an electronic device suitable for implementing the embodiments of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0030] Now, the exemplary embodiments will be described more fully with reference to the accompanying drawings. However, the exemplary embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that the present application will be more comprehensive and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art.

[0031] In addition, the described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a thorough understanding of the embodiments of the present application. However, those skilled in the art will realize that the technical solutions of the present application can be practiced without one or more of the specific details, or other methods, components, devices, steps, etc. may be adopted. In other cases, well-known methods, devices, implementations, or operations are not shown or described in detail to avoid obscuring aspects of the present application.

[0032] The block diagrams shown in the drawings are only functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0033] The flowcharts shown in the drawings are only illustrative and do not necessarily include all the content and operations / steps, nor are they necessarily executed in the described order. For example, some operations / steps can be decomposed, while some operations / steps can be combined or partially combined, so the actual execution order may change according to the actual situation.

[0034] Figure 1 An exemplary system architecture block diagram applying the technical solution of the present application is schematically shown.

[0035] As Figure 1 shown, the system architecture 100 may include a terminal device 110, a network 120, and a server 130. The terminal device 110 may include various electronic devices such as a smart phone, a tablet computer, a laptop computer, a desktop computer, etc. The server 130 may be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services. The network 120 may be a communication medium of various connection types capable of providing a communication link between the terminal device 110 and the server 130, for example, it may be a wired communication link or a wireless communication link.

[0036] According to implementation requirements, the system architecture in the embodiments of the present application can have any number of terminal devices, networks, and servers. For example, the server 130 can be a server group composed of multiple server devices. The power consumption prediction of the device cluster in the present application can be specifically executed by the terminal device 110. Specifically, the terminal device 110 obtains the initial power consumption prediction models corresponding to each device in the device cluster. The terminal device 110 inputs the load parameter data of each device into the corresponding initial power consumption prediction model respectively to obtain the predicted power consumption data corresponding to each device. The terminal device 110 obtains the predicted total power consumption data of the device cluster based on the predicted power consumption data corresponding to each device, and uses the error between the predicted total power consumption data and the actual total power consumption data of the device cluster as feedback to adjust each power consumption prediction model until the error meets the preset conditions, and obtains the target power consumption prediction models corresponding to each device respectively, so as to predict the power consumption of the device cluster through the target power consumption prediction models. The power consumption prediction of the device cluster in the present application can be specifically executed by the server 130, or can be jointly implemented by the terminal device 110 and the server 130. The present application does not make special limitations on this.

[0037] The following makes a detailed description of the power consumption prediction of the device cluster provided by the present application in combination with specific implementation manners.

[0038] Figure 2 is a flowchart of a method for predicting the power consumption of a device cluster according to an embodiment of the present application. The method for predicting the power consumption of the device cluster can be specifically executed by a test terminal, or can be executed by a server, or can be jointly executed by a test terminal and a server. As Figure 2 shown, the method for predicting the power consumption of the device cluster can at least include the following steps S210 to step S240.

[0039] Step S210: Obtain the initial power consumption prediction models corresponding to each device in the device cluster.

[0040] Specifically, in the operation of a device cluster, such as a data center, the power consumption and load are constantly changing, providing rich data for characterizing the power consumption model. Thus, it is convenient to establish the initial power consumption prediction models corresponding to each device according to the operating power consumption and load data of each device in the device cluster, and the reliability of the data acquisition channels is high. Therefore, the accuracy of the established initial power consumption prediction models is relatively high.

[0041] Step S220: Input the load parameter data of each device into the corresponding initial power consumption prediction model respectively to obtain the predicted power consumption data corresponding to each device.

[0042] The load parameter data includes device configurations. By way of illustration, it includes the operating frequency of the server CPU, the number of CPUs, the input / output throughput, etc. In one embodiment, the load parameter data may include static configuration parameters and dynamic load parameters. Among them, the dynamic load parameter is the parameter data that is changing, and the static configuration parameter is the parameter data with a relatively small change range within a continuous time period, thereby improving the accuracy of the description of the parameter data.

[0043] Step S230: Obtain the predicted total power consumption data of the device cluster according to the predicted power consumption data corresponding to each device.

[0044] It is easy to understand that by adding the predicted power consumption data corresponding to each device, the predicted total power consumption data of the device cluster can be obtained.

[0045] Step S240: Using the error between the predicted total power consumption data and the actual total power consumption data of the device cluster as feedback, adjust each power consumption prediction model until the error meets the preset conditions, and obtain the target power consumption prediction model corresponding to each device, so as to predict the power consumption of the device cluster through the target power consumption prediction model.

[0046] By way of illustration, the actual total power consumption data of the device cluster can be obtained according to the power consumption.

[0047] Therefore, in this application, first, the predicted total power consumption data is determined through the predicted power consumption data corresponding to each device. Then, the actual total power consumption data is obtained during the operation of the device cluster. Thus, using the error between the predicted total power consumption data and the actual total power consumption data of the device cluster as feedback, each power consumption prediction model is adjusted until the error meets the preset conditions. Furthermore, the target power consumption prediction model corresponding to each device is obtained, so as to predict the power consumption of the device cluster through the target power consumption prediction model, thereby improving the convenience and accuracy of the power consumption prediction of the device cluster. And the training process does not require independent power consumption statistics for each device, and a large amount of real dynamic data can be used for model training, thereby ensuring the effectiveness of the model.

[0048] Figure 3 It is a flowchart of the power consumption prediction method for a device cluster provided by another embodiment of this application. As Figure 3 shown, in this embodiment, obtaining the initial power consumption prediction model corresponding to each device in the device cluster specifically may include the following steps:

[0049] Step S310: Obtain the historical load parameter data of each device in the device cluster, and the historical power consumption data corresponding to the historical load parameter data;

[0050] Step S320: Use the historical power consumption data and historical load parameter data as training samples to train the original prediction model corresponding to the device, and obtain the initial power consumption prediction models corresponding to each device respectively.

[0051] Specifically, the historical load parameter data can be obtained from the device's operation log. The historical power consumption data can be obtained by testing the device separately or through a chip installed in the device's power supply device. It is easy to understand that the historical load parameter data and historical power consumption data can be obtained conveniently and accurately.

[0052] It is possible to obtain the static configuration parameters and dynamic load parameters corresponding to each device in the device cluster during multiple historical periods, and obtain the historical power consumption data corresponding to each device during each historical period. Then, based on the static configuration parameters, dynamic load parameters, and historical power consumption data, establish the original prediction model corresponding to the device. The multiple historical periods can be adjacent periods or non-adjacent periods. Schematically, the same period of each day can be selected, for example, from 8 o'clock to 9 o'clock, so as to make the external influencing factors as consistent as possible. Different periods within the same season can also be selected to reduce the impact of external temperature changes on the device power consumption prediction model.

[0053] Among them, there can be multiple static configuration parameters, which are schematic, such as the number and frequency of CPUs. There can also be multiple dynamic load parameters, such as CPU utilization rate, IO throughput, etc.

[0054] It should be noted that the static configuration parameters and dynamic load parameters need to be distinguished according to different devices, and data with a relatively high correlation with power consumption in the corresponding devices should be included as much as possible. In addition, for parameters with uncertain high correlation, they are also used as neural network inputs, and weights are automatically assigned by the neural network during the training process with real data. The neural network for single-device modeling can be an ordinary multi-layer fully connected network, a more compact convolutional network, or a residual structure can be introduced, which is not limited here.

[0055] In one embodiment, obtaining the static configuration parameters and dynamic load parameters corresponding to each device respectively may specifically include the following steps: obtaining the static configuration parameters corresponding to each device respectively; obtaining the dynamic load parameters corresponding to each device respectively at preset time intervals. Thus, the reliability of system operation can be improved, and the change of dynamic load parameters can be obtained in a timely manner, thereby improving the accuracy of the initial power consumption prediction model training.

[0056] In one embodiment, in order to reduce the computational amount and improve the system stability, the devices in the device cluster can also be classified to obtain multiple device categories, and then based on the device categories, establish the initial power consumption prediction models corresponding to the device categories.

[0057] Thus, the initial power consumption prediction models corresponding to each device category can be obtained, and then the power consumption of the device cluster can be predicted through the initial power consumption prediction models of each device category. Among them, due to the similarity of the same device category, the model parameters and network structure of the same device category can be shared, thereby further reducing the system calculation amount and cost and improving the stability of the prediction system operation.

[0058] Specifically, according to the device category, establishing the initial power consumption prediction model corresponding to the device category may include the following steps: determining the initial model parameters corresponding to the device category based on the device category; generating the initial power consumption prediction models corresponding to the device category respectively according to the initial model parameters. Schematically, the device category may include servers, client machines, test machines, etc., so as to establish different initial power consumption prediction models according to the different hardware and software configurations of the devices in each device category.

[0059] In one embodiment, taking the error between the predicted total power consumption data and the actual total power consumption data of the device cluster as feedback, adjusting each power consumption prediction model until the error meets the preset condition to obtain the target power consumption prediction models corresponding to each device respectively, which may specifically include the following steps: obtaining the actual total power consumption data of the device cluster in real time, and obtaining the predicted total power consumption data in real time; testing the initial power consumption prediction model according to the actually obtained total power consumption data and the predicted total power consumption data in real time; obtaining the target power consumption prediction models corresponding to each device respectively when the test accuracy rate reaches the preset threshold.

[0060] Specifically, a loss function can be established according to the difference between the actual total power consumption data and the predicted total power consumption data; performing gradient descent iteration on the loss function to obtain the minimized loss function and the model parameters of the initial power consumption prediction model. Thus, the target power consumption prediction model can be reliably obtained. Specifically, it can be the square of the difference between the actual total power consumption data and the predicted total power consumption data. During training, the load parameter data of each device or each device category can be statistically counted at preset time intervals and input into the power consumption prediction models corresponding to each device, and at the same time, the actual total power consumption data within this period is collected, and then gradient descent iteration is performed to obtain the model parameters of each power consumption prediction model.

[0061] It should be noted that although the steps of the method in this application are described in a specific order in the drawings, this does not require or imply that these steps must be executed in this specific order, or that all the steps shown must be executed to achieve the desired result. Additionally or alternatively, some steps can be omitted, multiple steps can be combined into one step for execution, and / or one step can be decomposed into multiple steps for execution, etc.

[0062] The following introduces the device embodiments of the present application, which can be used to execute the power consumption prediction method for the device cluster in the above embodiments of the present application. Figure 4 The structural block diagram of the power consumption prediction device 400 for the device cluster provided by the embodiment of the present application is schematically shown. As Figure 4 shown, the power consumption prediction device for the device cluster includes:

[0063] A first acquisition unit 410, configured to acquire the initial power consumption prediction models respectively corresponding to each device in the device cluster;

[0064] A training unit 420, configured to input the load parameter data of each device into the corresponding initial power consumption prediction model respectively, and obtain the predicted power consumption data respectively corresponding to each device;

[0065] A second acquisition unit 430, configured to obtain the predicted total power consumption data of the device cluster according to the predicted power consumption data respectively corresponding to each device;

[0066] A prediction unit 440, configured to use the error between the predicted total power consumption data and the actual total power consumption data of the device cluster as feedback, adjust each power consumption prediction model until the error meets a preset condition, and obtain the target power consumption prediction models respectively corresponding to each device, so as to predict the power consumption of the device cluster through the target power consumption prediction models.

[0067] Wherein, in one embodiment, the first acquisition unit 410 is further configured to acquire the historical load parameter data of each device in the device cluster, and the historical power consumption data corresponding to the historical load parameter data; use the historical power consumption data and the historical load parameter data as training samples to train the original prediction model corresponding to the device, and obtain the initial power consumption prediction models respectively corresponding to each device.

[0068] In one embodiment, the first acquisition unit 410 is further configured to acquire the static configuration parameters and dynamic load parameters respectively corresponding to each device in the device cluster in multiple historical time periods, and acquire the historical power consumption data respectively corresponding to each device in each historical time period; establish the original prediction model corresponding to the device according to the static configuration parameters, dynamic load parameters, and historical power consumption data.

[0069] In one embodiment, the first acquisition unit 410 is further configured to acquire the static configuration parameters respectively corresponding to each device; and acquire the dynamic load parameters respectively corresponding to each device at preset time intervals.

[0070] In one embodiment, the first acquisition unit 410 is further configured to classify each device in the device cluster to obtain multiple device categories; and establish the initial power consumption prediction models corresponding to the device categories according to the device categories.

[0071] In one embodiment, the first acquisition unit 410 is further configured to determine initial model parameters corresponding to the device category based on the device category; and generate initial power consumption prediction models corresponding to the device categories respectively according to the initial model parameters.

[0072] The specific details of the power consumption prediction device for the device cluster provided in the embodiments of the present application have been described in detail in the corresponding method embodiments, and will not be elaborated here.

[0073] Figure 5 Schematically shown is a block diagram of a computer system of an electronic device for implementing the embodiments of the present application.

[0074] It should be noted that Figure 5 The computer system 500 of the shown electronic device is only an example, and should not impose any limitation on the functions and usage scope of the embodiments of the present application.

[0075] As Figure 5 shown, the computer system 500 includes a central processing unit 501 (Central Processing Unit, CPU), which can perform various appropriate actions and processes according to a program stored in a read-only memory 502 (Read-Only Memory, ROM) or a program loaded from a storage section 508 into a random access memory 503 (Random Access Memory, RAM). In the random access memory 503, various programs and data required for system operation are also stored. The central processing unit 501, the read-only memory 502, and the random access memory 503 are connected to each other via a bus 504. An input / output interface 505 (Input / Output interface, i.e., I / O interface) is also connected to the bus 504.

[0076] The following components are connected to the input / output interface 505: an input section 506 including a keyboard, a mouse, etc.; an output section 507 including, for example, a cathode ray tube (Cathode Ray Tube, CRT), a liquid crystal display (Liquid Crystal Display, LCD), etc., and a speaker, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a local area network card, a modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the input / output interface 505 as needed. A removable medium 511, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 510 as needed, so that a computer program read from it can be installed into the storage section 508 as needed.

[0077] In particular, according to the embodiments of the present application, the processes described in each method flow chart can be implemented as computer software programs. For example, embodiments of the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains program codes for executing the methods shown in the flow charts. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 509, and / or installed from the removable medium 511. When the computer program is executed by the central processing unit 501, various functions defined in the system of the present application are executed.

[0078] It should be noted that the computer-readable medium shown in the embodiments of the present application can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium can include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, the computer-readable storage medium can be any tangible medium that contains or stores a program, and the program can be used by or in combination with an instruction execution system, apparatus, or device. In the present application, the computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program codes. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium can also be any computer-readable medium other than the computer-readable storage medium, and the computer-readable medium can send, propagate, or transmit a program for use by or in combination with an instruction execution system, apparatus, or device. The program codes contained on the computer-readable medium can be transmitted by any appropriate medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0079] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a part of code that contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram or flowchart, as well as combinations of blocks in the block diagram or flowchart, can be implemented by a dedicated hardware-based system that performs the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0080] It should be noted that although several modules or units of devices for action execution are mentioned in the above detailed description, such a division is not mandatory. In fact, according to the embodiments of the present application, the features and functions of two or more of the above-mentioned modules or units can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0081] From the description of the above embodiments, those skilled in the art can easily understand that the example embodiments described herein can be implemented by software, or can be implemented by a combination of software and necessary hardware. Therefore, the technical solutions according to the embodiments of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a touch terminal, or a network device, etc.) to execute the method according to the embodiments of the present application.

[0082] After considering the specification and practicing the invention disclosed herein, those skilled in the art will readily conceive of other embodiments of the present application. The present application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of the present application and include well-known common general knowledge or conventional technical means in the technical field not disclosed in the present application.

[0083] It should be understood that the present application is not limited to the exact structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the present application is only limited by the appended claims.

Claims

1. A method for predicting the power consumption of a device cluster, characterized in that, it includes: Obtain the historical load parameter data of each device in the device cluster, and the historical power consumption data corresponding to the historical load parameter data; use the historical power consumption data and the historical load parameter data as training samples to train the original prediction model corresponding to the device, and obtain the initial power consumption prediction models corresponding to each device respectively; Input the load parameter data of each device into the corresponding initial power consumption prediction model respectively, and obtain the predicted power consumption data corresponding to each device respectively; According to the predicted power consumption data corresponding to each device respectively, obtain the predicted total power consumption data of the device cluster; Obtain the actual total power consumption data of the device cluster in real time, and obtain the predicted total power consumption data in real time; Establish a loss function according to the difference between the actual total power consumption data and the predicted total power consumption data; perform gradient descent iteration on the loss function to obtain the minimized loss function and the model parameters of the initial power consumption prediction model; when the test accuracy reaches a preset threshold, obtain the target power consumption prediction models corresponding to each device respectively, so as to predict the power consumption of the device cluster through the target power consumption prediction model.

2. The method according to claim 1, characterized in that, the obtaining of the historical load parameter data of each device in the device cluster, and the historical power consumption data corresponding to the historical load parameter data includes: Obtain the static configuration parameters and dynamic load parameters corresponding to each device in the device cluster in multiple historical periods, and obtain the historical power consumption data corresponding to each device in each historical period; Before using the historical power consumption data and the historical load parameter data as training samples to train the original prediction model corresponding to the device, the method further includes: Establish the original prediction model corresponding to the device according to the static configuration parameters, the dynamic load parameters, and the historical power consumption data.

3. The method according to claim 2, characterized in that, the obtaining of the static configuration parameters and dynamic load parameters corresponding to each device in the device cluster in multiple historical periods includes: Obtain the static configuration parameters corresponding to each device respectively; At every preset time interval, obtain the dynamic load parameters corresponding to each device respectively.

4. The method according to claim 1, characterized in that, the obtaining of the initial power consumption prediction models corresponding to each device in the device cluster includes: Classify each device in the device cluster to obtain multiple device categories; Establish the initial power consumption prediction model corresponding to the device category according to the device category.

5. The method according to claim 4, characterized in that, the establishing of the initial power consumption prediction model corresponding to the device category according to the device category includes: Based on the device category, determine the initial model parameters corresponding to the device category; Generate the initial power consumption prediction models corresponding to the device category according to the initial model parameters.

6. A power consumption prediction device for a device cluster, characterized in that, it includes: A first acquisition unit, configured to acquire the historical load parameter data of each device in the device cluster, and the historical power consumption data corresponding to the historical load parameter data; use the historical power consumption data and the historical load parameter data as training samples to train the original prediction model corresponding to the device, and obtain the initial power consumption prediction models corresponding to each device respectively; A training unit, configured to input the load parameter data of each device into the corresponding initial power consumption prediction model respectively, and obtain the predicted power consumption data corresponding to each device respectively; A second acquisition unit, configured to obtain the predicted total power consumption data of the device cluster according to the predicted power consumption data corresponding to each device respectively; A prediction unit, configured to acquire the actual total power consumption data of the device cluster in real time, and acquire the predicted total power consumption data in real time; Establish a loss function according to the difference between the actual total power consumption data and the predicted total power consumption data; perform gradient descent iteration on the loss function to obtain the minimized loss function and the model parameters of the initial power consumption prediction model; when the test accuracy reaches a preset threshold, obtain the target power consumption prediction models corresponding to each device respectively, so as to predict the power consumption of the device cluster through the target power consumption prediction models.

7. An electronic device, characterized in that, it includes: A processor; and A memory, configured to store the executable instructions of the processor; wherein, the processor is configured to execute the power consumption prediction method of the device cluster according to any one of claims 1 to 5 by executing the executable instructions.

Images