Machine room energy saving rate evaluation method, device, equipment and computer program product

Abstract

This disclosure provides a method, apparatus, equipment, and computer program product for assessing the energy efficiency of data centers, relating to the field of computer and internet technology. The method includes: acquiring data center status profiles for multiple data centers, the profiles including at least one of data center cooling system capacity information, data center environmental information, and internal heat load information; the multiple data centers include data centers with energy consumption data and data centers without energy consumption data; based on the data center status profiles, performing unsupervised clustering on the data centers with energy consumption data to obtain multiple data center clusters; based on the data center status profiles, classifying the data centers without energy consumption data into clusters; and constructing an energy efficiency prediction model for each data center cluster, so as to predict the energy efficiency of the data centers within each cluster using the prediction model corresponding to each cluster. The embodiments of this disclosure can improve the accuracy of energy efficiency prediction without increasing hardware modification costs.

Description

Technical Field

[0001] This disclosure relates to the field of computer and Internet technology, and in particular to a method, apparatus, electronic device, computer-readable storage medium and computer program product for assessing the energy efficiency of a computer room. Background Technology

[0002] This section is intended to provide background or context for the embodiments of this disclosure as set forth in the claims. The description herein is not intended to be a prior art simply because it is included in this section.

[0003] Calculating the energy efficiency of data center cooling systems is a core method for quantitatively evaluating the effectiveness of energy-saving renovations, and it can objectively verify the actual implementation effect of energy-saving solutions. Ideally, by configuring smart meters for each computer room and cooling equipment, collecting energy consumption data before and after the renovation, and correcting for load, environmental, and other parameters, the energy efficiency can be accurately calculated.

[0004] However, due to limitations in construction investment and renovation conditions, most data centers lack comprehensive metering facilities for individual components. Some buildings are only equipped with a main electricity meter, making it impossible to accurately calculate the energy consumption of the cooling system in each computer room, which makes traditional calculation methods difficult to apply.

[0005] Therefore, how to reliably assess the energy efficiency of the cooling system in each computer room under conditions of inadequate energy metering facilities and incomplete statistical data has become a pressing technical problem in the field of data center energy-saving operation and maintenance. Summary of the Invention

[0006] The purpose of this disclosure is to provide a method, apparatus, electronic device, computer-readable storage medium, and computer program product for assessing data center energy efficiency, which can improve the accuracy of energy efficiency prediction.

[0007] Other features and advantages of this disclosure will become apparent from the following detailed description, or may be learned in part from practice of this disclosure.

[0008] This disclosure provides a method for assessing the energy efficiency of a data center, comprising: acquiring a data center status profile of multiple data centers, wherein the data center status profile includes at least one of data center cooling system capacity information, data center environmental information, and data center internal heat load information; the multiple data centers include data centers with energy consumption data and data centers without energy consumption data; based on the data center status profile, performing unsupervised clustering on the data centers with energy consumption data to obtain multiple data center clusters; based on the data center status profile, performing cluster classification processing on the data centers without energy consumption data to assign the data centers without energy consumption data to each of the data center clusters; and constructing an energy efficiency prediction model for each data center cluster to predict the energy efficiency of the data centers within the cluster using the prediction model corresponding to each cluster.

[0009] In some embodiments, the data center status profile includes dynamic fluctuation features; wherein, the method further includes: collecting observation data of the dynamic fluctuation features within a preset time period; dividing multiple continuous intervals based on the distribution characteristics of the observation data; constructing interval hesitant fuzzy features of the dynamic fluctuation features based on the multiple continuous intervals, so as to perform clustering processing on the data center based on the interval hesitant fuzzy features of the dynamic fluctuation features.

[0010] In some embodiments, the distribution characteristics of the observed data are used to divide multiple continuous intervals, including: determining the minimum, lower quartile, average, upper quartile, and maximum values ​​of the dynamic fluctuation characteristics based on the observed data; and constructing the multiple continuous intervals based on the minimum, lower quartile, average, upper quartile, and maximum values ​​of the dynamic fluctuation characteristics.

[0011] In some embodiments, based on the minimum, lower quartile, average, upper quartile, and maximum values ​​of the dynamic fluctuation feature, the plurality of continuous intervals are constructed, including: constructing a first interval representation based on the minimum and the lower quartile; constructing a second interval representation based on the lower quartile and the average; constructing a third interval representation based on the average and the upper quartile; and constructing a fourth interval representation based on the upper quartile and the maximum. Furthermore, based on the plurality of continuous intervals, the interval hesitation and ambiguity feature of the dynamic fluctuation feature is constructed, including: determining the interval hesitation and ambiguity feature of the dynamic fluctuation feature based on the first interval representation, the second interval representation, the third interval representation, and the fourth interval representation.

[0012] In some embodiments, the data center status profile includes dynamic fluctuation features, which are represented by interval hesitant fuzzy features. The process of unsupervised clustering of data centers with energy consumption data based on the data center status profile to obtain multiple data center clusters includes: determining the feature distance between each data center with energy consumption data based on the data center status profile; wherein the interval hesitant fuzzy features in the data center status profile are used to calculate the feature distance using interval hesitant fuzzy distance; and performing unsupervised clustering of the data centers with energy consumption data based on the feature distance between each data center with energy consumption data.

[0013] In some embodiments, the data center with energy consumption data includes a first data center and a second data center, and the dynamic fluctuation feature includes multiple interval representations. The dynamic fluctuation feature of the first data center is a first feature, and the dynamic fluctuation feature of the second data center is a second feature. Determining the feature distance between data centers with energy consumption data based on the data center status profile includes: sorting the interval representations in the first feature according to a first order; sorting the interval representations in the second feature according to the first order; calculating the feature difference between interval representations at the same position for the sorted first feature and second feature; and calculating the feature difference between interval representations at the same position for the sorted first feature and second feature to determine the feature distance between the first feature and the second feature.

[0014] In some embodiments, based on the data center status profile, the data centers without energy consumption data are classified into cluster categories to assign them to the respective data center clusters. This includes: determining the number of the plurality of data center clusters as a first value; performing supervised clustering with the first value as the number of cluster categories, wherein the cluster center of each category corresponds to a node within a data center cluster, and the cluster centers of different categories belong to different data center clusters.

[0015] In some embodiments, the heat load information inside the computer room is determined by the following methods: determining the set temperature of the computer room air conditioner, the difference between the return air temperature of the computer room air conditioner and the set temperature, and the difference between the set temperature of the computer room and the original mode set temperature; and determining the heat load information inside the computer room based on the set temperature of the computer room air conditioner, the difference between the return air temperature of the computer room air conditioner and the set temperature, and the difference between the set temperature of the computer room and the original mode set temperature.

[0016] This disclosure provides a data center energy efficiency assessment device, including: a profile acquisition module, an unsupervised clustering module, a cluster partitioning module, and a model building module.

[0017] The image acquisition module is used to acquire image profiles of multiple data centers, which include at least one of the following: data center cooling system capacity information, data center environmental information, and data center internal heat load information. The multiple data centers include data centers with energy consumption data and data centers without energy consumption data. The unsupervised clustering module can be used to perform unsupervised clustering on the data centers with energy consumption data based on the data center image profiles, resulting in multiple data center clusters. The cluster partitioning module can be used to perform cluster classification processing on the data centers without energy consumption data based on the data center image profiles, so as to assign the data centers without energy consumption data to the respective data center clusters. The model building module can be used to build an energy-saving rate prediction model for each data center cluster, so as to use the prediction model corresponding to each cluster to predict the energy-saving rate of the data centers within the cluster.

[0018] This disclosure provides an electronic device comprising: a memory and a processor; the memory for storing computer program instructions; and the processor for calling the computer program instructions stored in the memory to implement the data center energy efficiency assessment method described above.

[0019] This disclosure provides a computer-readable storage medium storing computer program instructions to implement the data center energy efficiency assessment method as described in any of the preceding embodiments.

[0020] This disclosure provides a computer program product or computer program that includes computer program instructions stored in a computer-readable storage medium. The computer program instructions are read from the computer-readable storage medium, and the processor executes the computer program instructions to implement the aforementioned method for evaluating the energy efficiency of a computer room.

[0021] The data center energy efficiency assessment method, apparatus, electronic device, computer-readable storage medium, and computer program product provided in this disclosure first perform unsupervised clustering on data centers with energy consumption data based on a data center status profile containing information on cooling capacity, environment, and heat load. Data centers with energy consumption data that have similar power consumption patterns and operating characteristics are grouped into the same cluster, achieving refined classification of data center groups. Then, based on the profile features, data centers without energy consumption data are assigned to the corresponding clusters, effectively solving the problem of difficult classification and assessment of data centers in scenarios where energy consumption data is missing. Finally, a dedicated energy efficiency prediction model is constructed for each cluster, which can achieve more accurate energy efficiency prediction based on the common patterns of similar data centers, significantly improving the applicability and accuracy of data center energy efficiency assessment under conditions of imperfect energy consumption metering.

[0022] It should be understood that the above general description and the following detailed description are merely exemplary and do not limit this disclosure. Attached Figure Description

[0023] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure. It is obvious that the drawings described below are merely some embodiments of this disclosure, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0024] Figure 1 A schematic diagram of a scenario that can be applied to the data center energy efficiency assessment method or data center energy efficiency assessment device in the embodiments of this disclosure is shown.

[0025] Figure 2 This is a comparative schematic diagram illustrating an energy efficiency assessment method according to an exemplary embodiment.

[0026] Figure 3 This is a flowchart illustrating a method for evaluating the energy efficiency of a computer room according to an exemplary embodiment.

[0027] Figure 4 This is a flowchart illustrating a feature processing method according to an exemplary embodiment.

[0028] Figure 5 This is a flowchart illustrating a method for determining a continuous interval according to an exemplary embodiment.

[0029] Figure 6 This is a flowchart illustrating a method for constructing interval hesitant fuzzy features according to an exemplary embodiment.

[0030] Figure 7 This is a flowchart illustrating a method for determining characteristic distances between computer rooms according to an exemplary embodiment.

[0031] Figure 8 This is a flowchart illustrating a cluster classification method for a zero-energy data center according to an exemplary embodiment.

[0032] Figure 9 This is a flowchart illustrating an energy efficiency assessment method according to an exemplary embodiment.

[0033] Figure 10 This is a block diagram illustrating a data center energy efficiency assessment device according to an exemplary embodiment.

[0034] Figure 11 A schematic diagram of the structure of an electronic device suitable for implementing embodiments of the present disclosure is shown. Detailed Implementation

[0035] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the embodiments set forth herein; rather, they are provided so that this disclosure will be thorough and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted.

[0036] Those skilled in the art will recognize that embodiments of this disclosure can be a system, apparatus, device, method, or computer program product. Therefore, this disclosure can be implemented in the following forms: entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0037] The features, structures, or characteristics described in this disclosure can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a thorough understanding of embodiments of this disclosure. However, those skilled in the art will recognize that the technical solutions of this disclosure can be practiced with one or more specific details omitted, or other methods, components, apparatuses, steps, etc., can be employed. In other instances, well-known methods, apparatuses, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of this disclosure.

[0038] In this disclosure, the terms "module" or "unit" refer to a computer program or part of a computer program that has a predetermined function and works with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.

[0039] The accompanying drawings are merely illustrative of this disclosure, and the same reference numerals in the drawings denote the same or similar parts, thus omitting repeated descriptions of them. Some block diagrams shown in the drawings do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0040] The flowchart shown in the accompanying drawings is merely illustrative and does not necessarily include all content and steps, nor does it require execution in the described order. For example, some steps may be broken down, while others may be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0041] In the description of this disclosure, unless otherwise stated, " / " means "or," for example, A / B can mean A or B. "And / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone. Furthermore, "at least one" means one or more, and "multiple" means two or more. The terms "first," "second," etc., do not limit the quantity or order of execution, and "first," "second," etc., do not necessarily imply differences; the terms "contains," "includes," and "has" are used to indicate an open-ended meaning of inclusion and refer to the existence of additional elements / components / etc. besides those listed.

[0042] This disclosure embodiment can be implemented by a terminal and / or a server. The terminal can obtain data from a computer device and display that data. The computer device can interact with the terminal, and can be a server hosting the application, or it can belong to the terminal (i.e., the terminal's backend), etc., without limitation.

[0043] The terminal can be a mobile phone, a laptop computer, or a playback device in a vehicle, etc., without limitation. The terminal can be considered a playback device in a vehicle, and it can display the target application. The terminal is only one example of the devices listed; the terminal in this disclosure is not limited to the listed devices. The target application in this disclosure can be any application capable of displaying multimedia information.

[0044] It is understood that the terminal mentioned in the embodiments of this disclosure can be a computer device, including but not limited to a terminal or a server. In other words, the computer device can be a server or a terminal, or a system composed of a server and a terminal. The terminal mentioned above can be an electronic device, including but not limited to mobile phones, tablets, desktop computers, laptops, handheld computers, in-vehicle devices, augmented reality / virtual reality (AR / VR) devices, head-mounted displays, smart TVs, wearable devices, smart speakers, digital cameras, webcams, and other mobile internet devices (MIDs) with network access capabilities, or terminals in scenarios such as trains, ships, and flights.

[0045] The servers mentioned above can be independent physical servers, server clusters or distributed systems composed of multiple physical servers, or cloud servers that provide basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, vehicle-road cooperation, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0046] Optionally, the data involved in the embodiments of this disclosure may be stored in a computer device or may be stored based on cloud storage technology, without limitation.

[0047] To better understand the above-mentioned objectives, features and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of the present disclosure can be combined with each other.

[0048] It should be noted that the information collection, gathering, updating, analysis, processing, use, transmission, and storage involved in the technical solution disclosed herein all comply with relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. Necessary measures are taken to prevent unauthorized access to information data and to maintain information security and network security.

[0049] The following section will first explain some of the terms used in the embodiments of this disclosure so that those skilled in the art can understand them.

[0050] K-means algorithm: A partition-based unsupervised machine learning method. Its core objective is to divide a given dataset into K non-overlapping clusters, such that data points within the same cluster are similar to each other, while data points in different clusters are significantly different. K is an integer greater than or equal to 1.

[0051] Density-Based Spatial Clustering of Applications with Noise (DBSCAN) is a density-based spatial clustering method for applications with noise. Its core idea is to divide data points into clusters based on local density, grouping sufficiently high-density regions into clusters while simultaneously identifying low-density noise points. This method avoids the reliance of algorithms like K-means on "predefined cluster numbers" and "convex cluster distribution."

[0052] Interval-hesitant fuzzy sets are a mathematical tool developed from fuzzy sets and hesitant fuzzy sets, better suited to complex and uncertain decision-making scenarios. They were further developed by Torra et al. within the theoretical framework of hesitant fuzzy sets. Their core solution addresses the problems of traditional fuzzy sets, which can only express membership degrees through a single numerical value, and ordinary hesitant fuzzy sets, which, while allowing multiple discrete membership degrees, struggle to characterize the interval uncertainty inherent in the membership degree itself. Essentially, they define the membership relationship between elements and sets as a set of non-empty closed intervals, rather than a single numerical value or a set of discrete numerical values.

[0053] The preceding text introduced some terms and concepts involved in the embodiments of this disclosure. The following text introduces the technical features involved in the embodiments of this disclosure.

[0054] Data center (computer room) cooling system energy efficiency calculation serves as a core quantitative basis for evaluating energy-saving effectiveness, playing a crucial role in objectively verifying the actual effectiveness of energy-saving cooling system solutions. Furthermore, the calculation results can be adjusted by incorporating variables such as IT load fluctuations and environmental parameters, providing vital decision-making support for identifying energy consumption loopholes and determining the cost recovery period for energy-saving retrofits. Ideally, every computer room and even every air conditioning unit would be equipped with a smart meter. Only energy consumption data before and after the retrofit needs to be collected, and adjusted for key variables such as computer room load and environment, to obtain the energy efficiency rate of the computer room cooling system. However, due to investment constraints, some data centers lack comprehensive power consumption statistics facilities to accurately track the power consumption of the computer room cooling system or each computer room; some even have only one smart meter per building. This makes accurate calculation of the energy efficiency of the computer room cooling system difficult. Currently, assessing the energy efficiency of computer room cooling systems in scenarios lacking basic energy consumption metering facilities or with incomplete energy consumption statistics has become a crucial aspect of data center energy-saving operation and maintenance management.

[0055] Existing methods for assessing energy efficiency under conditions of data scarcity mainly fall into three categories: The first is based on mathematical model calculations. This involves establishing a mathematical model for equipment energy consumption based on knowledge of refrigeration system load and thermodynamic principles to estimate energy-saving performance. A drawback of this method is that it typically assumes the refrigeration system operates under ideal conditions, and the mathematical model considers few external influencing factors. However, the actual operating environment of refrigeration systems is complex, making energy efficiency prediction modeling difficult. The second method is based on simulation modeling. This involves simulating the operation of refrigeration systems in buildings and data centers, and then using the simulation results combined with thermodynamic calculations to estimate equipment energy consumption and energy efficiency. This method requires a deep understanding of the refrigeration system's operating mechanism and suffers from high simulation modeling difficulty and cost. The third method is based on artificial intelligence prediction. This method uses machine learning and deep learning prediction models to construct a mapping relationship between data center characteristics and energy efficiency, achieving energy efficiency prediction and assessment. However, the data used for model training may not have a similar distribution pattern to the data from actual data centers. In fact, data centers with similar energy consumption patterns have similar energy efficiency; differences in operating characteristics and energy consumption patterns may prevent the prediction model from accurately assessing the energy efficiency of the refrigeration system.

[0056] In conclusion, energy efficiency assessment of data center cooling systems under conditions of data scarcity is now essential. Although some methods based on mathematical analysis, simulation modeling, and intelligent prediction have been proposed, they suffer from problems such as excessive implementation constraints, high costs, and inaccurate predictions. There is an urgent need for a low-cost, highly practical, and reliable method for assessing data center energy efficiency, capable of achieving accurate assessments even in scenarios lacking basic energy metering facilities or with incomplete energy consumption statistics.

[0057] To address the aforementioned issues, this application proposes a method for assessing the energy efficiency of data centers, which can achieve accurate assessment of data center energy efficiency in scenarios where basic energy consumption metering facilities are lacking or energy consumption statistics are incomplete.

[0058] The exemplary embodiments of this disclosure will now be described in detail with reference to the accompanying drawings.

[0059] Figure 1 A schematic diagram of a scenario that can be applied to the data center energy efficiency assessment method or data center energy efficiency assessment device in the embodiments of this disclosure is shown.

[0060] Please refer to Figure 1 The diagram illustrates an implementation environment provided by an exemplary embodiment of this disclosure.

[0061] like Figure 1As shown, system architecture 100 may include terminal devices 101, 102, and 103, a network 104, and a server 105. Network 104 serves as the medium for providing communication links between terminal devices 101, 102, and 103 and server 105. Network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0062] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Terminal devices 101, 102, and 103 can be various electronic devices with displays and web browsing capabilities, including but not limited to smartphones, tablets, laptops, desktop computers, wearable devices, virtual reality devices, smart home devices, etc.

[0063] Server 105 can be a server that provides various services, such as a backend management server that supports the devices operated by users using terminal devices 101, 102, and 103. The backend management server can analyze and process received requests and other data, and feed the processing results back to the terminal devices.

[0064] A server can be a standalone physical server, a server cluster or a distributed system consisting of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms. This disclosure does not impose any restrictions on this.

[0065] Server 105 may, for example, acquire data center status profiles for multiple data centers, including at least one of the following: data center cooling system capacity information, data center environmental information, and data center internal heat load information; the multiple data centers include data centers with energy consumption data and data centers without energy consumption data; server 105 may, for example, perform unsupervised clustering on the data center status profiles for data centers with energy consumption data to obtain multiple data center clusters; server 105 may, for example, perform cluster classification processing on the data center status profiles for data centers without energy consumption data to assign the data centers without energy consumption data to each data center cluster; server 105 may, for example, construct an energy saving rate prediction model for each data center cluster to predict the energy saving rate of the data centers within the cluster using the prediction model corresponding to each cluster.

[0066] It should be understood that Figure 1The number of terminal devices, networks, and servers shown is merely illustrative. Server 105 can be a single physical server or a combination of multiple servers. Depending on actual needs, it can have any number of terminal devices, networks, and servers.

[0067] Under the above system architecture, this disclosure provides a method for evaluating the energy efficiency of a data center, which can be executed by any electronic device with computing power.

[0068] Figure 2 This is a comparative schematic diagram illustrating an energy efficiency assessment method according to an exemplary embodiment.

[0069] The difference between the method proposed in this application and traditional methods is as follows: Figure 2 As shown. The main steps of the energy-saving rate assessment method proposed in this application may include: 1. Constructing a feature representation method for the operating status of a data center in a hesitant fuzzy interval, realizing a representation of the data center's operating rules that considers dynamic fluctuations and uncertainties. 2. In the context of interval hesitant fuzzy intervals, proposing a two-layer clustering method based on partitioning and density to autonomously mine the types of data center energy consumption patterns and achieve similarity matching of data center energy consumption patterns. 3. Based on the energy consumption pattern partitioning, constructing a Kernel Ridge Regression energy-saving rate prediction model for data centers under each type of energy consumption pattern, realizing the prediction and assessment of data center energy-saving rates by energy consumption pattern category.

[0070] The above methods can solve the following technical problems.

[0071] 1. This application addresses the problem of significant discrepancies between model training data and prediction data in traditional data-driven methods for assessing the energy efficiency of refrigeration systems, leading to severely distorted prediction results. Recognizing that only data centers with similar energy consumption patterns exhibit similar energy efficiency, this application first performs cluster-based energy consumption similarity matching for data centers, and then assesses the energy efficiency of each type of data center separately, significantly improving the accuracy of the prediction assessment.

[0072] 2. To address the problem that simply using average values, maximum and minimum values, etc., is insufficient to accurately characterize the operational state of computer room dynamic characteristics, a novel approach is adopted: interval hesitant fuzzy sets are used to evaluate the uncertain characteristics of computer rooms. This provides a more comprehensive and objective evaluation of the dynamic characteristics of computer rooms. Furthermore, interval hesitant fuzzy sets are innovatively introduced into DBSCAN and weighted K-means clustering, achieving more accurate computer room clustering.

[0073] The energy efficiency assessment method proposed in this application will be explained and illustrated below with reference to specific embodiments.

[0074] Figure 3This is a flowchart illustrating a method for evaluating the energy efficiency of a computer room according to an exemplary embodiment. The method provided in this disclosure can be executed by any electronic device with computing power, for example, the method can be executed by the above-described... Figure 1 The execution can be performed by a server or terminal device in the embodiments, or it can be performed by both a server and a terminal device. In the following embodiments, the server is used as the execution subject for illustration, but this disclosure is not limited to this.

[0075] Reference Figure 3 The data center energy efficiency assessment method provided in this disclosure may include the following steps.

[0076] Step S302: Obtain the status profiles of multiple computer rooms. The status profiles include at least one of the following: computer room cooling system capacity information, computer room environment information, and computer room internal heat load information. The multiple computer rooms include computer rooms with energy consumption data and computer rooms without energy consumption data.

[0077] In some embodiments, a data center status profile can be a set of features that describe the full-dimensional characteristics of a single data center's cooling operation, and can be used to characterize the data center's hardware configuration, environmental conditions, heat load fluctuations, energy-saving operation strategies, etc.

[0078] The information on the capacity of the computer room cooling system can be static characteristics that characterize the inherent configuration of the computer room cooling equipment and the cooling hardware carrying capacity. It can include the rated power of the air conditioner or the number of air conditioners in the computer room, etc. This application does not limit this.

[0079] Data center environment information can describe the characteristics of the basic physical space and external environmental conditions of the data center, and may include at least one of the following: data center area, number of racks, hot and cold aisle closure status, and outdoor temperature.

[0080] The heat load information inside the computer room can be data related to the heat generated continuously by the operation of equipment and supporting facilities in the computer room, which is used to reflect the scale of heat generation and heat fluctuation inside the computer room.

[0081] In some embodiments, the heat load information inside the computer room can be described by at least one of multiple indicators such as the air conditioning set temperature, the temperature difference between the return air and the set temperature, and the temperature difference between the current and the original mode set temperature.

[0082] In some embodiments, the internal heat load information of the computer room can be determined by the following methods: determining the set temperature of the computer room air conditioner, the difference between the return air temperature of the computer room air conditioner and the set temperature, and the difference between the set temperature of the computer room and the original mode set temperature; and determining the internal heat load information of the computer room based on the set temperature of the computer room air conditioner, the difference between the return air temperature of the computer room air conditioner and the set temperature, and the difference between the set temperature of the computer room and the original mode set temperature.

[0083] In some embodiments, a data center status profile can be constructed using the following methods.

[0084] In some embodiments, three types of information are extracted: the capacity of the data center cooling system, the data center environment, and the internal heat load of the data center, in order to construct a data center status profile.

[0085] In some embodiments, the cooling system capacity characteristics may include: rated power of the air conditioner and the number of air conditioners in the computer room.

[0086] In some embodiments, the data center environment may include: data center area, number of server racks, hot and cold aisle enclosure, and outdoor temperature of the data center.

[0087] In some embodiments, the internal heat load of a computer room can refer to the temperature state inside the computer room caused by the heat load of the equipment. In some embodiments, the above state is not easily obtained directly, so this embodiment constructs three new features to describe the internal heat load of the computer room: the set temperature of the computer room air conditioner, the difference between the return air temperature and the set temperature of the computer room air conditioner, and the difference between the set temperature and the original mode temperature. The set temperature of the computer room air conditioner represents the desired ambient temperature of the computer room. The difference between the return air temperature and the set temperature represents the cooling capacity of the computer room air conditioner. The smaller the interpolation value, the lower the temperature of the heat source in the computer room and the better the cooling capacity of the computer room air conditioner. The difference between the set temperature and the original mode temperature represents the set temperature of the air conditioner under the energy-saving strategy adopted by the computer room.

[0088] In some embodiments, the features in the data center status profile are either fixed real-valued features or continuous, fluctuating variables.

[0089] For example, the characteristic attribute of the heat load inside a computer room can be a continuous and fluctuating variable. Traditional methods of calculating averages or extreme values ​​cannot objectively and comprehensively measure the heat load attribute of a computer room.

[0090] Interval hesitant fuzzy sets are an important tool in the field of fuzzy mathematics for handling uncertain decision-making and fuzzy information representation. They allow the use of multiple interval values ​​to express the "membership degree" of an element belonging to a certain set.

[0091] Interval hesitant fuzzy sets have the advantage of being able to express multiple assessment opinions with uncertainty. This embodiment uses the interval hesitant fuzzy set method to express the fluctuating changes (dynamic fluctuation characteristics) of attributes such as the heat load of the computer room. Specifically, for each dynamic time-series characteristic (dynamic fluctuation characteristic), the following can be selected for each characteristic: minimum value, lower quartile, average value, upper quartile, and maximum value. Then, interval representations are constructed: [minimum, lower quartile], [lower quartile, average value], [average value, upper quartile], [upper quartile, maximum value]. This forms an interval hesitant fuzzy feature. .

[0092] In some embodiments, the dynamic fluctuation features in the data center status profile can be represented by interval hesitant fuzzy features using the method described above.

[0093] Step S304: Based on the data center status profile, perform unsupervised clustering on data centers with energy consumption data to obtain multiple data center clusters.

[0094] In some embodiments, the data center status profile may include dynamic fluctuation features, which are represented by interval hesitant fuzzy features. Specifically, based on the data center status profile, unsupervised clustering is performed on data centers with energy consumption data to obtain multiple data center clusters. This includes: determining the feature distance between each data center with energy consumption data based on the data center status profile; wherein the interval hesitant fuzzy features in the data center status profile are used to calculate the feature distance using interval hesitant fuzzy distance; and performing unsupervised clustering on the data centers with energy consumption data based on the feature distance between each data center with energy consumption data.

[0095] In some embodiments, unsupervised clustering can be performed on data centers with energy consumption data based on features in the data center status profile.

[0096] In some embodiments, DBSCAN clustering can be used to perform density clustering on data centers with energy consumption data, thereby achieving unsupervised clustering.

[0097] In some embodiments, the energy consumption mode types of data centers with power consumption (data centers with energy consumption) can be determined first, and then, based on the determined energy consumption mode types, cluster similarity matching can be performed between data centers with no energy consumption and data centers with energy consumption.

[0098] In this embodiment, the energy consumption mode type for computer rooms with power is determined using interval hesitant fuzzy DBSCAN clustering: 1. Calculate the sample distance matrix: For the data Any two samples (such as data from a computer room) and , build Distance matrix .in, A status profile of data center i. This represents the status profile of data center j, where i and j are both positive integers. It's worth noting that real-valued features are calculated using real-valued distance, while interval-hesitant fuzzy features are calculated using interval-hesitant fuzzy distance.

[0099] 2. Using traditional DBSCAN clustering logic, density-based clustering is performed to obtain a set of clusters. Where C represents a data center cluster.

[0100] Step S306: Based on the data center status profile, perform cluster classification processing on data centers without energy consumption data, so as to classify the data centers without energy consumption data into each data center cluster.

[0101] In some embodiments, similarity matching can be performed on data centers without power consumption based on the existing energy consumption patterns of data centers with power consumption data (data centers with energy consumption data).

[0102] In some embodiments, clustering can be performed using the interval hesitant fuzzy K-Meams algorithm (based on the interval hesitant fuzzy K-Means algorithm), and the objective function for clustering is: The set number of cluster categories, K, is equal to the number of clusters corresponding to the data centers with energy consumption data. The cluster center point of each category... That is, it must be Data points in the data. For the sample The Euclidean distance to the cluster center. For real-valued features, the real-valued Euclidean distance measure is used to calculate the feature distance; for interval hesitant fuzzy features, the interval hesitant fuzzy distance measure is used to calculate the feature distance.

[0103] In some embodiments, after clustering is completed, the Davies-Bouldin Index can be used to measure the inter-cluster splitting performance of the clustering, i.e.: .

[0104] in, It is the number of clusters. Indicates the first The average distance from all sample points in a class to the cluster center measures the density of the cluster. Indicates the first Class and First The distance between cluster centers measures the degree of separation between clusters. i is a positive integer, and j is a positive integer.

[0105] Step S308: Construct an energy-saving rate prediction model for each data center cluster, so as to use the prediction model corresponding to each cluster to predict the energy-saving rate of the data centers within the cluster.

[0106] In some embodiments, the data centers within each cluster exhibit more similar operating states and energy consumption patterns, meaning their energy-saving performance patterns are more similar. Therefore, a prediction model is constructed for each clustered data center to achieve more accurate predictions of data center energy efficiency.

[0107] The Kernel Ridge Regression Prediction Model combines ridge regression with kernel methods, fundamentally addressing the problem that traditional linear models cannot handle nonlinear data. It maps low-dimensional nonlinear data to a high-dimensional feature space using a kernel function, constructing a linear ridge regression model in this high-dimensional space. This indirectly enables nonlinear prediction of low-dimensional data while retaining the overfitting advantage of ridge regression. The data used to train the prediction model consists of operational data for data centers with power consumption in each class: including the average difference between the set temperature and the original mode temperature, the number of air conditioners in the data center, and the energy saving rate of the data center. The prediction samples are denoted as... The model prediction results are as follows: .

[0108] That is, by combining the kernel function values ​​of the new sample with all training samples, and the dual coefficients... Calculate the prediction results, , Represents the kernel matrix. It is the identity matrix. y represents the kernel coefficients, and y represents the training sample labels.

[0109] In some embodiments, an energy-saving rate prediction model can be constructed for each data center cluster using a Kernel ridge regression prediction model, so as to use the prediction model corresponding to each cluster to predict the energy-saving rate of the data centers within the cluster.

[0110] The technical solution proposed in the above embodiments constructs a data center status profile that includes cooling capacity, environmental and heat load information, uses interval hesitant fuzzy sets to accurately express dynamic fluctuation characteristics, classifies data centers with and without energy consumption data into groups according to their operating characteristics, and then establishes a Kernel Ridge regression prediction model for each type of data center. This effectively solves the problems of difficulty in quantifying dynamic heat load in data centers and inability to evaluate data centers without energy consumption data, and significantly improves the accuracy and generalization applicability of data center energy saving rate prediction.

[0111] Figure 4 This is a flowchart illustrating a feature processing method according to an exemplary embodiment.

[0112] In some embodiments, the data center status profile may include dynamic fluctuation features.

[0113] In some embodiments, the indicators in the data center status profile that exhibit dynamic fluctuation characteristics may include at least one of the following: air conditioning set temperature, air conditioning return air temperature, and temperature difference between the current mode and the original mode. This application does not impose any restrictions on this.

[0114] refer to Figure 4 The above feature processing method may include the following steps.

[0115] Step S402: Collect observation data on dynamic fluctuation characteristics within a preset time period.

[0116] In some embodiments, time-series operational data can be collected within a preset time period, including dynamic fluctuation characteristics such as the set temperature of the computer room air conditioner, the temperature difference between the air conditioner return air and the set temperature, and the temperature difference between the current mode and the original mode set temperature.

[0117] In some embodiments, the collected data can be preprocessed by denoising, removing outliers, and filling in missing values ​​to obtain a complete and reliable dynamic feature observation dataset.

[0118] Step S404: Divide the observation data into multiple continuous intervals based on the distribution characteristics of the observation data.

[0119] In some embodiments, the distribution characteristics of the collected dynamic fluctuation characteristics observation data, such as the minimum value, lower quartile, average value, upper quartile and maximum value, can be statistically analyzed. Using these as dividing points, multiple continuous and non-overlapping numerical intervals can be successively divided to form a multi-segment interval structure (i.e. multiple continuous intervals) covering the overall data distribution.

[0120] Step S406: Based on multiple continuous intervals, construct interval hesitant fuzzy features with dynamic fluctuation characteristics, so as to perform clustering processing on the computer room based on the interval hesitant fuzzy features with dynamic fluctuation characteristics.

[0121] In some embodiments, multiple consecutive intervals obtained from the division can be used as membership intervals and combined to form an interval hesitant fuzzy set, thereby constructing an interval hesitant fuzzy feature corresponding to the dynamic fluctuation characteristics, which can be used for subsequent clustering processing of the data center.

[0122] The technical solution provided in the above embodiments, by collecting and preprocessing time-series observation data with dynamic fluctuation characteristics, dividing continuous intervals according to data distribution, and then constructing interval hesitant fuzzy features, can accurately characterize the fluctuations and uncertainties of the dynamic characteristics of the data center, and improve the rationality and reliability of subsequent data center clustering.

[0123] Figure 5 This is a flowchart illustrating a method for determining a continuous interval according to an exemplary embodiment.

[0124] refer to Figure 5 Dividing multiple continuous intervals based on the distribution characteristics of observed data may include the following steps.

[0125] Step S502: Determine the minimum, lower quartile, average, upper quartile, and maximum values ​​of the dynamic fluctuation characteristics based on the observed data.

[0126] Step S504: Based on the minimum, lower quartile, average, upper quartile, and maximum values ​​of the dynamic fluctuation characteristics, construct multiple continuous intervals.

[0127] In some embodiments, the minimum, lower quartile, average, upper quartile, and maximum values ​​of the dynamic fluctuation characteristics can be calculated first based on the observed data. Then, using these five key statistics as dividing points, multiple continuous intervals that do not overlap and cover the entire data range can be sequentially connected to complete the interval construction based on the data distribution characteristics.

[0128] The technical solution provided in the above embodiments divides continuous intervals by using key distribution characteristics such as extreme values, quartiles, and means of statistical observation data. This objectively matches the actual distribution pattern of the data, accurately reflects the fluctuation range of dynamic characteristics, and provides a stable and reliable interval basis for the subsequent construction of interval hesitant and fuzzy features.

[0129] Figure 6 This is a flowchart illustrating a method for constructing interval hesitant fuzzy features according to an exemplary embodiment.

[0130] refer to Figure 6 The aforementioned interval hesitation and ambiguity features may include the following steps.

[0131] Step S602: Construct the first interval representation based on the minimum value and the lower quartile.

[0132] Step S604: Construct a second interval representation based on the lower quartiles and the mean.

[0133] Step S606: Construct a third interval representation based on the mean and the upper quartile.

[0134] Step S608: Construct the fourth interval representation based on the upper quartiles and the maximum value.

[0135] Step S610: Based on the first interval representation, the second interval representation, the third interval representation, and the fourth interval representation, determine the interval hesitant fuzzy features of the dynamic fluctuation features.

[0136] In some embodiments, the first to fourth interval representations can be constructed sequentially using the minimum value and the lower quartile, the lower quartile and the average value, the average value and the upper quartile, and the upper quartile and the maximum value as the upper and lower boundaries, respectively. These four intervals are then merged and integrated to finally determine the interval hesitant fuzzy feature corresponding to the dynamic fluctuation feature.

[0137] The technical solution provided in the above embodiments, by constructing four continuous intervals with data statistical features as boundaries and merging them, can completely and meticulously describe the distribution and uncertainty of dynamic fluctuation features, forming a standardized and unified interval hesitant fuzzy feature, providing an accurate and robust feature expression for subsequent data center clustering and distance calculation.

[0138] Figure 7This is a flowchart illustrating a method for determining characteristic distances between computer rooms according to an exemplary embodiment.

[0139] In some embodiments, the data center containing energy consumption data may include a first data center and a second data center, and the dynamic fluctuation characteristics may be represented by multiple intervals, wherein the dynamic fluctuation characteristics of the first data center are the first characteristics, and the dynamic fluctuation characteristics of the second data center are the second characteristics.

[0140] refer to Figure 7 Determining the characteristic distance between data centers with energy consumption data based on data center status profiles may include the following steps.

[0141] Step S702: Sort the interval representations in the first feature according to the first order.

[0142] Step S704: Sort the interval representations in the second feature according to the first order.

[0143] Step S706: For the sorted first feature and second feature, calculate the feature difference between the interval representations corresponding to the same position.

[0144] Step S708: For the sorted first feature and second feature, calculate the feature difference between the interval representations at the same position to determine the feature distance between the first feature and the second feature.

[0145] In some embodiments, the intervals in the hesitant fuzzy set corresponding to the dynamic fluctuation characteristics of the two computer rooms can be sorted in the same order (e.g., ascending) first, then the feature differences of intervals at the same position can be calculated one by one, and finally the feature differences of intervals at the same position can be combined to obtain the feature distance between the two computer rooms for the dynamic characteristic. Specific implementation methods can be found in the following embodiments, which will not be repeated here.

[0146] The technical solution provided in the above embodiments can accurately quantify the similarity of dynamic fluctuation features between two computer rooms by sorting the interval hesitant and fuzzy features of the two computer rooms according to a unified rule and calculating the interval differences bit by bit. This results in a stable and reliable feature distance, providing a reasonable similarity measurement basis for subsequent computer room clustering.

[0147] Figure 8 This is a flowchart illustrating a cluster classification method for a zero-energy data center according to an exemplary embodiment.

[0148] refer to Figure 8 Based on the data center status profile, data centers without energy consumption data are classified into cluster categories to assign them to different data center clusters. This process may include the following steps.

[0149] Step S802: Determine the number of multiple data center clusters as the first value.

[0150] Step S804: Perform supervised clustering with the first value as the number of clustering categories, where the cluster center of each category corresponds to a node in a data center cluster, and the cluster centers of different categories belong to different data center clusters.

[0151] In some embodiments, the total number of cluster categories can be determined first based on the number of pre-defined clusters of data centers with energy consumption, and then clustering can be performed using this number as the number of categories. Each cluster center is set as the data point of the data center within the corresponding original cluster, thereby assigning data centers without energy consumption to the matching data center clusters.

[0152] The technical solution provided in the above embodiments uses the number of existing clusters as the number of cluster categories and uses the corresponding data center nodes within the cluster as cluster centers for division. This can efficiently and accurately classify data centers without energy consumption into suitable cluster categories, realize unified clustering of data centers with and without energy consumption, and improve the integrity and applicability of data center clustering.

[0153] In some embodiments, to address the problem that some data centers (computer rooms) lack energy consumption metering facilities such as electricity meters, making it impossible to accurately measure the energy efficiency of the cooling system, a cooling system energy consumption assessment method combining interval hesitant fuzzy clustering and Kernel ridge regression prediction is proposed.

[0154] Below, this application will explain and illustrate the above-mentioned method for assessing the energy consumption (energy saving rate assessment) of a refrigeration system with reference to specific embodiments.

[0155] 1. Scheme design concept.

[0156] The basic principle of this embodiment is that computer rooms with similar operating and configuration states have similar energy consumption patterns, and their energy-saving effects are more similar. Using information from computer rooms with similar energy-saving effects for energy-saving rate prediction yields more reliable results. Therefore, this embodiment proposes a method for evaluating the energy-saving rate of computer room cooling systems using interval hesitant fuzzy clustering and Kernel ridge regression prediction. This method includes computer room similarity attribute matching using interval hesitant fuzzy clustering and energy-saving rate prediction and evaluation based on Kernel ridge regression. The former uses interval hesitant fuzzy numbers to express the dynamic uncertainty characteristics of the computer room and performs interval hesitant fuzzy clustering to autonomously divide computer room clusters with similar energy consumption patterns and energy-saving spaces. Based on this, data on computer rooms with available power are extracted for each energy consumption pattern to form training datasets for different energy consumption patterns. A Kernel ridge regression prediction model is constructed for each energy consumption pattern, and the energy-saving strategies for each computer room are input to achieve energy-saving rate evaluation for computer rooms under different energy consumption pattern types. The method flow steps are as follows: Figure 9 As shown. Figure 9This is a flowchart illustrating an energy efficiency assessment method according to an exemplary embodiment.

[0157] Matching of similarity attributes of computer rooms based on interval hesitant fuzzy clustering.

[0158] (1) Construct a data center status profile.

[0159] This paper extracts three types of information—cooling system capacity, server room environment, and internal heat load—to construct a server room status profile. Cooling system capacity characteristics include: rated power of air conditioners and the number of air conditioners in the server room. The server room environment includes: server room area, number of server racks, hot and cold aisle enclosure, and outdoor temperature. Internal heat load mainly refers to the temperature state inside the server room caused by the heat load of IT equipment. However, these states are not easily obtained directly. Therefore, this application constructs three new features: server room air conditioner set temperature, the difference between server room air conditioner return air temperature and set temperature, and the difference between server room set temperature and original mode temperature. The server room air conditioner set temperature represents the desired ambient temperature of the server room. The difference between the server room air conditioner return air temperature and set temperature represents the cooling capacity of the server room air conditioner; a smaller interpolation value indicates a lower heat source temperature and better cooling capacity. The difference between the server room set temperature and the original mode temperature represents the air conditioner set temperature under the energy-saving strategy adopted by the server room.

[0160] (2) Interval hesitant fuzzy feature representation.

[0161] The characteristic attribute of the heat load inside a computer room is a continuous and fluctuating quantity. Traditional methods of calculating averages or extreme values ​​cannot objectively and comprehensively measure the heat load attribute of a computer room. Interval hesitant fuzzy sets are important tools in the field of fuzzy mathematics for handling uncertain decision-making and fuzzy information representation. They allow the use of multiple interval values ​​to express the "membership degree" of an element belonging to a certain set.

[0162] Interval hesitant fuzzy sets offer the advantage of expressing multiple assessment opinions with uncertainty. This embodiment employs the interval hesitant fuzzy set method to express the heat load attributes of the computer room. Specifically, for each dynamic time-series feature, the following values ​​are selected for each feature: minimum, lower quartile, average, upper quartile, and maximum. Then, interval representations are constructed: [minimum, lower quartile], [lower quartile, average], [average, upper quartile], [upper quartile, maximum]. This forms an interval hesitant fuzzy feature. .

[0163] (3) Construction of distance measure for interval hesitant fuzzy evaluation.

[0164] For two interval hesitant fuzzy numbers, the distance between them needs to be calculated. If the two intervals have different lengths, length alignment is required first. For the shorter interval, the longest interval is used for length padding. For two intervals of the same length, the evaluation is denoted as... and First, align them in ascending order. Then, use the interval hesitant fuzzy Euclidean distance to calculate the distance between two interval hesitant fuzzy attribute features: .

[0165] (4) Classification of data centers with energy consumption.

[0166] Based on the determination of the interval hesitant fuzzy distance measure, it is first necessary to determine the energy consumption mode types of data centers with power. Then, based on the determined energy consumption mode categories, cluster similarity matching is performed between data centers with and without power. For determining the energy consumption mode types of data centers with power, this embodiment innovatively adopts interval hesitant fuzzy DBSCAN clustering, as detailed in the following steps.

[0167] 1) Calculate the sample distance matrix: For the data Any two samples and , build Distance matrix It is worth noting that real-valued features are calculated using real-valued distance, while interval-hesitant fuzzy features are calculated using interval-hesitant fuzzy distance.

[0168] 2) Using the traditional DBSCAN clustering logic, density-based clustering is performed to obtain a set of clusters. .

[0169] (5) Matching similarity of data center energy consumption patterns.

[0170] Based on the existing energy consumption patterns of data centers with power consumption data, similarity matching is performed on data centers without power consumption data. An interval-hesitant fuzzy K-Meams algorithm is proposed for clustering, with the objective function being: .

[0171] The number of clusters, K, is set to be equal to the number of clusters in DBSCAN. The cluster center point of each cluster... That is, it must be Data points in the data. This represents the Euclidean distance between the sample and the cluster center. For real-valued features, the real-valued Euclidean distance measure is used to calculate the feature distance; for interval-hesitant fuzzy features, the interval-hesitant fuzzy distance measure is used to calculate the feature distance.

[0172] After clustering is completed, the Davies-Bouldin Index can be used to measure the inter-cluster splitting performance, i.e.: .in, It is the number of clusters. Indicates the first The average distance from all sample points in a class to the cluster center measures the density of the cluster. Indicates the first Class and First The distance between cluster centers measures the degree of separation between clusters.

[0173] (5) Kernel Ridge regression prediction of data center energy consumption.

[0174] The data centers within each cluster exhibit more similar operating states and energy consumption patterns, implying greater similarity in their energy-saving performance. Therefore, a predictive model is constructed for each clustered data center to achieve more accurate predictions of data center energy efficiency.

[0175] The Kernel Ridge Regression Prediction Model combines ridge regression with kernel methods, fundamentally addressing the problem that traditional linear models cannot handle nonlinear data. It maps low-dimensional nonlinear data to a high-dimensional feature space using a kernel function, constructing a linear ridge regression model in this high-dimensional space. This indirectly enables nonlinear prediction of low-dimensional data while retaining the overfitting advantage of ridge regression. The data used to train the prediction model consists of operational data for data centers with power consumption in each class: including the average difference between the set temperature and the original mode temperature, the number of air conditioners in the data center, and the energy saving rate of the data center. The prediction samples are denoted as... The model prediction results are as follows: .

[0176] That is, by combining the kernel function values ​​of the new sample with all training samples, and the dual coefficients... Calculate the prediction results, , Represents the kernel matrix. It is the identity matrix. y represents the kernel coefficients, and y represents the training sample labels.

[0177] This application also proposes a method for evaluating the energy efficiency of a data center cooling system in scenarios where energy consumption data is lacking. Specifically, it may include the following process.

[0178] Phase 1: Matching similarity of data center energy consumption patterns.

[0179] Step 1: Obtain static and dynamic characteristic data of the computer room. Static characteristics include: rated power of air conditioners, number of air conditioners in the computer room, area of ​​the computer room, number of racks in the computer room, and whether the hot and cold aisles are closed. Dynamic characteristics include: outdoor temperature of the computer room, original mode temperature of the computer room air conditioners, set temperature of the computer room air conditioners, and return air temperature of the computer room.

[0180] Step 2: Normalize the static and dynamic characteristic data of the computer room.

[0181] Step 3: For the normalized dynamic data, extract the minimum, lower quartile, mean, upper quartile, and maximum. Then construct four interval numbers: [minimum, lower quartile], [lower quartile, mean], [mean, upper quartile], and [upper quartile, maximum], and form the interval hesitant fuzzy number evaluation feature.

[0182] Step 4: Perform similarity matching of data center energy consumption patterns.

[0183] Step 4-1: Extract the data from the computer room containing power data, and calculate the distance between any two computer room data points. Real number features are calculated using the real number distance measure, while interval hesitant fuzzy number features are calculated using the interval hesitant fuzzy distance measure. Step 4-2: Based on the data center distance measurement, perform DBSCAN clustering on the data with power consumption to determine how many categories these data centers with power consumption patterns can be divided into. Step 4-3: Using all data including data from data centers with and without power, perform multiple weighted K-meams clustering algorithms, where the number of clusters is equal to the number of clusters in DBSCAN, and the cluster center of each cluster is selected from the data points of the data centers with power. Step 4-4: The Davies-Bouldin Index is used to measure the inter-cluster segmentation performance of weighted K-Meams clustering. The clustering result with the lowest index is selected as the data center energy consumption similarity matching result. The final clustering results are shown in the second to last column of Table 1.

[0184] Phase Two: Energy Saving Rate Prediction and Assessment for Kernel Ridge Return Data Center.

[0185] Step 1: Divide the data center data into multiple types based on the clustering results; Step 2: Select each type of computer room in turn, and extract the data of the computer rooms with electricity in that type of data as the training set. The features are the average difference between the set temperature of the computer room and the original mode temperature and the number of air conditioners in the computer room. The label is the energy saving rate of the computer room. Step 3: Train a Kernel Ridge regression prediction model using each class of training dataset; Step 4: Perform energy saving rate prediction and evaluation for the computer room under each type of energy consumption mode.

[0186] The above method proposes an energy-saving rate assessment approach for data center cooling systems in scenarios with missing energy consumption data. First, it proposes a data center energy consumption pattern similarity matching method based on interval hesitant fuzzy clustering, enabling autonomous discovery of data center energy consumption pattern types and similarity matching. Second, based on different data center energy consumption patterns, a Kernel Ridge Regression energy-saving rate prediction method is constructed to accurately predict the energy-saving rate under different data center energy consumption patterns.

[0187] Compared to existing technologies, this embodiment addresses the reality that only data centers with the same energy consumption mode have similar energy-saving rates. It solves the problem that traditional data-driven energy-saving rate prediction and evaluation of cooling systems often suffers from significant discrepancies between model training data and real data, leading to severely distorted prediction results. This method collects static attributes that evaluate the cooling capacity of the data center, dynamic data reflecting the heat load of the data center, and energy-saving strategy data. First, it performs cluster-based energy consumption similarity matching. Then, based on the differences in energy consumption categories, it evaluates the energy-saving rate of the data center under different energy consumption modes, significantly improving the accuracy of the prediction and evaluation.

[0188] The aforementioned method also proposes a clustering method based on interval hesitant fuzzy sets. This method innovatively uses interval hesitant modulo to evaluate data center characteristics, providing a multi-dimensional assessment of data fluctuations and uncertainties. It also introduces interval hesitant fuzzy distance measures into DBSCAN and K-means clustering, achieving more accurate clustering. Compared to existing technologies, this method addresses the similarity of data center characteristics, particularly the dynamic characteristics of data centers over time. Simply using average, maximum, or minimum values ​​is insufficient to accurately characterize the operational state of these characteristics. Interval hesitant fuzzy set theory uses multiple elements to evaluate data uncertainty. Therefore, this method innovatively extracts the minimum, lower quartile, average, upper quartile, and maximum values ​​of each time-series data point to form an interval hesitant fuzzy number evaluation. This provides a more comprehensive and objective evaluation of the dynamic characteristics of the data center. Furthermore, the innovative introduction of interval hesitant fuzzy set theory into DBSCAN and weighted K-means clustering achieves more accurate data center clustering.

[0189] The aforementioned method also proposes a two-level clustering method using DBSCAN and K-means in interval-hesitant fuzzy environments. DBSCAN first autonomously mines the energy consumption patterns of data center rooms with power. Then, K-means clustering is used to match the similarity of energy consumption patterns between data center rooms without power and those with power. Compared to existing technologies, this method focuses on the changing characteristics of data center energy consumption patterns. It does not use a method of manually classifying data center energy consumption patterns, but dynamically mines the types of data center energy consumption patterns based on the data center's operating rules and changes. This ensures the reliability of data center energy consumption pattern similarity judgment.

[0190] It should be particularly noted that the steps in each embodiment of the above-described data center energy efficiency assessment method can be interchanged, substituted, added to, or deleted from each other. Therefore, these reasonable permutations and combinations of the data center energy efficiency assessment method should also fall within the scope of protection of this disclosure, and the scope of protection of this disclosure should not be limited to the described embodiments.

[0191] It should be noted that the scope of protection of this application should include, but is not limited to, the specific implementation methods described in the embodiments. Any alternative solution that uses a different name but substantially performs the same function and achieves the same technical effect falls within the scope of protection defined by the claims of this application.

[0192] Based on the same inventive concept, this disclosure also provides a data center energy efficiency assessment device, as described in the following embodiments. Since the principle by which this device addresses the problem is similar to that of the method embodiments described above, the implementation of this device embodiment can refer to the implementation of the method embodiments described above, and repeated details will not be elaborated further.

[0193] Figure 10 This is a block diagram illustrating a data center energy efficiency assessment device according to an exemplary embodiment. (Refer to...) Figure 10 The data center energy efficiency assessment device 1000 provided in this embodiment may include: a profile acquisition module 1001, an unsupervised clustering module 1002, a cluster partitioning module 1003, and a model building module 1004.

[0194] The image acquisition module 1001 can be used to acquire the status images of multiple computer rooms. The status images of the computer rooms include at least one of the following: computer room cooling system capacity information, computer room environmental information, and computer room internal heat load information. The multiple computer rooms include computer rooms with energy consumption data and computer rooms without energy consumption data. The unsupervised clustering module 1002 can be used to perform unsupervised clustering on the computer rooms with energy consumption data based on the status images of the computer rooms to obtain multiple computer room clusters. The cluster partitioning module 1003 can be used to perform cluster classification processing on the computer rooms without energy consumption data based on the status images of the computer rooms to assign the computer rooms without energy consumption data to the respective computer room clusters. The model building module 1004 can be used to build an energy saving rate prediction model for each computer room cluster so as to use the prediction model corresponding to each cluster to predict the energy saving rate of the computer rooms within the cluster.

[0195] It should be noted that the aforementioned image acquisition module 1001, unsupervised clustering module 1002, cluster partitioning module 1003, and model building module 1004 correspond to S302 to S308 in the method embodiment. The examples and application scenarios implemented by these modules and their corresponding steps are the same, but they are not limited to the content disclosed in the above method embodiment. It should also be noted that these modules, as part of a device, can be executed in a computer system, such as a set of computer-executable instructions.

[0196] In some embodiments, the data center status profile includes dynamic fluctuation characteristics; wherein, the data center energy efficiency assessment device 1000 may include an observation data acquisition module, an interval division module, and a feature construction module.

[0197] The observation data acquisition module can be used to collect observation data of the dynamic fluctuation characteristics within a preset time period; the interval division module can be used to divide multiple continuous intervals based on the distribution characteristics of the observation data; the feature construction module can be used to construct interval hesitant fuzzy features of the dynamic fluctuation characteristics based on the multiple continuous intervals, so as to perform clustering processing on the computer room based on the interval hesitant fuzzy features of the dynamic fluctuation characteristics.

[0198] In some embodiments, the distribution characteristics of the observed data are used to divide multiple continuous intervals, including: determining the minimum, lower quartile, average, upper quartile, and maximum values ​​of the dynamic fluctuation characteristics based on the observed data; and constructing the multiple continuous intervals based on the minimum, lower quartile, average, upper quartile, and maximum values ​​of the dynamic fluctuation characteristics.

[0199] In some embodiments, based on the minimum, lower quartile, average, upper quartile, and maximum values ​​of the dynamic fluctuation feature, the plurality of continuous intervals are constructed, including: constructing a first interval representation based on the minimum and the lower quartile; constructing a second interval representation based on the lower quartile and the average; constructing a third interval representation based on the average and the upper quartile; and constructing a fourth interval representation based on the upper quartile and the maximum. Furthermore, based on the plurality of continuous intervals, the interval hesitation and ambiguity feature of the dynamic fluctuation feature is constructed, including: determining the interval hesitation and ambiguity feature of the dynamic fluctuation feature based on the first interval representation, the second interval representation, the third interval representation, and the fourth interval representation.

[0200] In some embodiments, the data center status profile includes dynamic fluctuation features, which are represented by interval hesitant fuzzy features. The process of unsupervised clustering of data centers with energy consumption data based on the data center status profile to obtain multiple data center clusters includes: determining the feature distance between each data center with energy consumption data based on the data center status profile; wherein the interval hesitant fuzzy features in the data center status profile are used to calculate the feature distance using interval hesitant fuzzy distance; and performing unsupervised clustering of the data centers with energy consumption data based on the feature distance between each data center with energy consumption data.

[0201] In some embodiments, the data center with energy consumption data includes a first data center and a second data center, and the dynamic fluctuation feature includes multiple interval representations. The dynamic fluctuation feature of the first data center is a first feature, and the dynamic fluctuation feature of the second data center is a second feature. Determining the feature distance between data centers with energy consumption data based on the data center status profile includes: sorting the interval representations in the first feature according to a first order; sorting the interval representations in the second feature according to the first order; calculating the feature difference between interval representations at the same position for the sorted first feature and second feature; and calculating the feature difference between interval representations at the same position for the sorted first feature and second feature to determine the feature distance between the first feature and the second feature.

[0202] In some embodiments, based on the data center status profile, the data centers without energy consumption data are classified into cluster categories to assign them to the respective data center clusters. This includes: determining the number of the plurality of data center clusters as a first value; performing supervised clustering with the first value as the number of cluster categories, wherein the cluster center of each category corresponds to a node within a data center cluster, and the cluster centers of different categories belong to different data center clusters.

[0203] In some embodiments, the heat load information inside the computer room is determined by the following methods: determining the set temperature of the computer room air conditioner, the difference between the return air temperature of the computer room air conditioner and the set temperature, and the difference between the set temperature of the computer room and the original mode set temperature; and determining the heat load information inside the computer room based on the set temperature of the computer room air conditioner, the difference between the return air temperature of the computer room air conditioner and the set temperature, and the difference between the set temperature of the computer room and the original mode set temperature.

[0204] Since the functions of the device 1000 have been described in detail in their respective method embodiments, they will not be repeated here.

[0205] The modules described in the embodiments of this disclosure can be implemented in software or hardware. The described modules can also be located in a processor. The names of these modules do not, in some cases, constitute a limitation on the module itself.

[0206] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a portion of a module or program segment containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer program instructions.

[0207] Furthermore, the above figures are merely illustrative of the processes included in the method according to exemplary embodiments of this disclosure and are not intended to be limiting. It is readily understood that the processes shown in the above figures do not indicate or limit the temporal order of these processes. Additionally, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0208] Figure 11 A schematic diagram of an electronic device suitable for implementing embodiments of the present disclosure is shown. It should be noted that... Figure 11 The illustrated electronic device 1100 is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments disclosed herein.

[0209] like Figure 11 As shown, the electronic device 1100 includes a central processing unit (CPU) 1101, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1102 or a program loaded from a storage section 1108 into a random access memory (RAM) 1103. The RAM 1103 also stores various programs and data required for the operation of the electronic device 1100. The CPU 1101, ROM 1102, and RAM 1103 are interconnected via a bus 1104. An input / output (I / O) interface 1105 is also connected to the bus 1104.

[0210] The following components are connected to I / O interface 1105: an input section 1106 including a keyboard, mouse, etc.; an output section 1107 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 1108 including a hard disk, etc.; and a communication section 1109 including a network interface card such as a LAN card, modem, etc. The communication section 1109 performs communication processing via a network such as the Internet. A drive 1110 is also connected to I / O interface 1105 as needed. Removable media 1111, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 1110 as needed so that computer programs read from them can be installed into storage section 1108 as needed.

[0211] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable storage medium, the computer program containing computer program instructions for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 1109, and / or installed from removable medium 1111. When the computer program is executed by central processing unit (CPU) 1101, it performs the functions defined above in the system of this disclosure.

[0212] It should be noted that the computer-readable storage medium disclosed herein may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable computer program instructions. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable storage medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. Computer program instructions contained on a computer-readable storage medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0213] In another aspect, this disclosure also provides a computer-readable storage medium, which may be included in the device described in the above embodiments; or it may exist independently and not assembled into the device. The computer-readable storage medium carries one or more programs, which, when executed by the device, enable the device to perform the following functions: acquiring a data center status profile of multiple data centers, the data center status profile including at least one of data center cooling system capacity information, data center environmental information, and data center internal heat load information; the multiple data centers include data centers with energy consumption data and data centers without energy consumption data; based on the data center status profile, performing unsupervised clustering on the data centers with energy consumption data to obtain multiple data center clusters; based on the data center status profile, performing cluster classification processing on the data centers without energy consumption data to assign the data centers without energy consumption data to each of the data center clusters; and constructing an energy-saving rate prediction model for each data center cluster to predict the energy-saving rate of the data centers within the cluster using the prediction model corresponding to each cluster.

[0214] According to one aspect of this disclosure, a computer program product or computer program is provided, comprising computer program instructions stored in a computer-readable storage medium. The computer program instructions are read from the computer-readable storage medium, and a processor executes the computer program instructions to implement the methods provided in various optional implementations of the above embodiments.

[0215] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions of the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, or portable hard drive) and includes several computer program instructions to cause an electronic device (such as a server or terminal device) to execute the method according to the embodiments of this disclosure.

[0216] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

[0217] It should be understood that this disclosure is not limited to the detailed structures, drawing arrangements or implementations shown herein; rather, this disclosure is intended to cover various modifications and equivalent arrangements contained within the spirit and scope of the appended claims.

Claims

1. A method for evaluating the energy efficiency of a computer room, characterized in that, include: Obtain data center status profiles for multiple data centers, wherein the data center status profiles include at least one of the following: data center cooling system capacity information, data center environmental information, and data center internal heat load information; the multiple data centers include data centers with energy consumption data and data centers without energy consumption data. Based on the data center status profile, unsupervised clustering is performed on the data centers with energy consumption data to obtain multiple data center clusters. Based on the data center status profile, the data centers without energy consumption data are classified into cluster categories to assign them to the respective data center clusters. Energy saving rate prediction models are constructed for each data center cluster, so that the energy saving rate of the data centers within the cluster can be predicted using the prediction models corresponding to each cluster.

2. The method according to claim 1, characterized in that, The data center status profile includes dynamic fluctuation features; the method further includes: Collect observational data on the dynamic fluctuation characteristics within a preset time period; The observed data is divided into multiple continuous intervals based on its distribution characteristics. Based on the multiple continuous intervals, an interval hesitant fuzzy feature of the dynamic fluctuation characteristics is constructed so as to perform clustering processing on the computer room based on the interval hesitant fuzzy feature of the dynamic fluctuation characteristics.

3. The method according to claim 2, characterized in that, Based on the distribution characteristics of the observed data, multiple continuous intervals are divided, including: The minimum, lower quartile, average, upper quartile, and maximum values ​​of the dynamic fluctuation characteristics are determined based on the observed data. Based on the minimum, lower quartile, average, upper quartile, and maximum values ​​of the dynamic fluctuation characteristics, the multiple continuous intervals are constructed.

4. The method according to claim 3, characterized in that, Based on the minimum, lower quartile, average, upper quartile, and maximum values ​​of the dynamic fluctuation characteristics, the multiple continuous intervals are constructed, including: A first interval representation is constructed based on the minimum value and the lower quartile; A second interval representation is constructed based on the lower quartiles and the mean; A third interval representation is constructed based on the mean and the upper quartile; A fourth interval representation is constructed based on the upper quartiles and the maximum values; Among them, based on the multiple continuous intervals, the interval hesitant fuzzy features of the dynamic fluctuation characteristics are constructed, including: Based on the first interval representation, the second interval representation, the third interval representation, and the fourth interval representation, the interval hesitation fuzzy feature of the dynamic fluctuation feature is determined.

5. The method according to claim 1, characterized in that, The data center status profile includes dynamic fluctuation features, which are represented by interval hesitant fuzzy features. Based on the data center status profile, unsupervised clustering is performed on the data centers with energy consumption data to obtain multiple data center clusters, including: Based on the data center status profile, the feature distance between each data center with energy consumption data is determined; wherein the interval hesitant fuzzy feature in the data center status profile is calculated using the interval hesitant fuzzy distance. Based on the feature distance between each data center with energy consumption data, unsupervised clustering is performed on the data centers with energy consumption data.

6. The method according to claim 5, characterized in that, The data centers containing energy consumption data include a first data center and a second data center. The dynamic fluctuation characteristics include multiple interval representations. The dynamic fluctuation characteristics of the first data center are the first characteristics, and the dynamic fluctuation characteristics of the second data center are the second characteristics. Based on the data center status profile, the characteristic distance between data centers with energy consumption data is determined, including: The interval representations in the first feature are sorted according to a first order; The interval representations in the second feature are sorted according to the first order; For the sorted first feature and second feature, calculate the feature difference between the interval representations corresponding to the same position; For the sorted first feature and second feature, calculate the feature difference between the interval representations at the same position to determine the feature distance between the first feature and the second feature.

7. The method according to claim 1, characterized in that, Based on the data center status profile, the data centers without energy consumption data are classified into clusters to assign them to their respective clusters, including: The number of clusters of the multiple computer rooms is determined and used as the first value; Supervised clustering is performed using the first value as the number of cluster categories, where the cluster center of each category corresponds to a node in a data center cluster, and the cluster centers of different categories belong to different data center clusters.

8. The method according to claim 1, characterized in that, The internal heat load information of the computer room is determined using the following method: Determine the set temperature of the computer room air conditioner, the difference between the computer room air conditioner return air temperature and the set temperature, and the difference between the computer room set temperature and the original mode set temperature. The heat load information inside the computer room is determined based on the set temperature of the computer room air conditioner, the difference between the return air temperature of the computer room air conditioner and the set temperature, and the difference between the set temperature of the computer room and the original mode set temperature.

9. A device for evaluating the energy efficiency of a computer room, characterized in that, include: The profile acquisition module is used to acquire profiles of the status of multiple computer rooms. The profiles of the computer rooms include at least one of the following: computer room cooling system capacity information, computer room environment information, and computer room internal heat load information. The multiple computer rooms include computer rooms with energy consumption data and computer rooms without energy consumption data. An unsupervised clustering module is used to perform unsupervised clustering on the data centers with energy consumption data based on the data center status profile, and obtain multiple data center clusters. The cluster partitioning module is used to perform cluster category classification processing on the data center without energy consumption data based on the data center status profile, so as to classify the data center without energy consumption data into the respective data center clusters. The model building module is used to build an energy-saving rate prediction model for each data center cluster, so as to use the prediction model corresponding to each cluster to predict the energy-saving rate of the data centers within the cluster.

10. An electronic device, characterized in that, include: Memory and processor; The memory is used to store computer program instructions; the processor calls the computer program instructions stored in the memory to implement the data center energy efficiency assessment method as described in any one of claims 1-8.

11. A computer-readable storage medium storing computer program instructions thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the data center energy efficiency assessment method as described in any one of claims 1-8.

12. A computer program product comprising computer program instructions stored in a computer-readable storage medium, characterized in that, When the computer program instructions are executed by the processor, they implement the method according to any one of claims 1-8.

Images