Machine room energy-saving control method and device

Abstract

The invention discloses a machine room energy-saving control method and device. The method comprises the steps that a target state data set of a target area in a target machine room in a target period is acquired, a target causal reasoning algorithm matched with the target state data set is determined, and an energy-saving control model is constructed according to the algorithm; analyzing the state data by using the model to obtain an energy-saving control instruction; and regulating and controlling the target equipment according to the instruction. The technical problem of energy waste caused by the fact that complex causal effects among multiple variables cannot be accurately recognized and responded in a traditional machine room control method is solved.

Description

Technical Field

[0001] This application relates to the field of energy-saving control technology, and more specifically, to a method and device for energy-saving control of a computer room. Background Technology

[0002] In current data center operations, energy consumption has become a significant cost, especially since temperature control equipment, as a key component in maintaining a stable data center environment, accounts for a considerable proportion of total data center energy consumption. Traditional energy-saving control strategies often rely on fixed thresholds or simple feedback-based temperature control mechanisms. While these methods achieve basic temperature and humidity regulation to some extent, they often struggle to accurately capture the complex causal relationships between environmental parameters. This is particularly true when faced with the varying loads of IT equipment within the data center, fluctuations in external ambient temperature, and the internal structural layout of the data center. Traditional methods are prone to problems such as inadequate energy consumption control and low temperature control efficiency.

[0003] Specifically, existing technologies for evaluating energy-saving effects often employ statistical correlation analysis, such as regression models. While this can reveal the correlation between certain parameters, it struggles to distinguish between true causal relationships and the interference of confounding factors. For instance, when outdoor temperatures rise, the energy consumption of temperature control equipment naturally increases, but this does not necessarily mean that the air conditioning system itself has room for optimization. Furthermore, because the correlation between energy-saving measures (such as temperature adjustment by temperature control equipment and equipment load scheduling) and energy consumption data exhibits significant time lags and nonlinear characteristics, traditional methods often fail to accurately capture dynamic causal effects when dealing with such problems, resulting in poor response speed and adaptability of control strategies.

[0004] There is currently no effective solution to the above problems. Summary of the Invention

[0005] This application provides a data center energy-saving control method and apparatus to at least solve the technical problem of energy waste caused by the inability to accurately identify and respond to complex causal effects among multiple variables in traditional data center control methods.

[0006] According to one aspect of the embodiments of this application, a data center energy-saving control method is provided, comprising: acquiring a target state data set of a target area in a target data center within a target period, wherein the target state data set includes at least one of the following types of data: environmental state data and equipment state data; determining a target causal reasoning algorithm that matches the target state data set, and constructing a target energy-saving control model based on the target causal reasoning algorithm and the target state data set, wherein the target causal reasoning algorithm is used to analyze the mutual influence relationship between different types of state data; analyzing the target state data set and preset data center state data using the target energy-saving control model to obtain energy-saving control instructions; and regulating the target equipment in the target area according to the energy-saving control instructions.

[0007] Optionally, the target status data set of the target area in the target equipment room within the target period is obtained, including: obtaining status data reported by multiple sensors in the target area of ​​the target equipment room at multiple acquisition times within the target period, and performing data preprocessing on all status data, wherein the data preprocessing includes at least one of the following: data cleaning, normalization processing, and timestamp alignment; dividing the preprocessed status data according to data type to obtain multiple status time series corresponding to multiple types of status data, and forming the multiple status time series into a target status data set.

[0008] Optionally, determining a target causal inference algorithm that matches the target state data set includes: determining the mutual information between different types of state data in the target state data set, and determining the mutual information delay between different types of state data based on the mutual information; determining the data stationarity result of each type of state data in the target state data set using statistical testing methods, wherein the statistical testing methods include: the Enhanced Dick-Fuller Test (ADF) method; obtaining the amount of computing resources pre-allocated for the data center energy-saving control task; and determining a target causal inference algorithm that matches the target state data set based on target parameters, wherein the target parameters include at least one of the following: mutual information delay, data stationarity result, and computing resources.

[0009] Optionally, the target causal inference algorithm matching the target state data set is determined based on the target parameters, including: determining the maximum mutual information delay among the mutual information delays between different types of state data; using the Dynamic Continuous Time Bayesian Network (DCTBN) algorithm as the target causal inference algorithm when the computational resources are not less than a preset resource threshold and the maximum mutual information delay is not less than a preset delay threshold; and using the Structured Equation Modeling (SEM) algorithm as the target causal inference algorithm when the computational resources are less than a preset resource threshold or the maximum mutual information delay is less than a preset delay threshold.

[0010] Optionally, a target causal inference algorithm matching the target state data set is determined based on the target parameters, including: when the amount of computing resources is not less than a preset resource threshold and the data stationarity result indicates that the data is non-stationary, the DCTBN algorithm is used as the target causal inference algorithm; when the amount of computing resources is less than the preset resource threshold or the data stationarity result indicates that the data is stationary, the SEM algorithm is used as the target causal inference algorithm.

[0011] Optionally, a target energy-saving control model is constructed based on the target causal reasoning algorithm and the target state data set, including: acquiring multiple historical state data sets that match the target causal reasoning algorithm; using the target causal reasoning algorithm to analyze the state data in each historical state data set and the target state data set, determining each target state data that has a causal relationship with the preset computer room state data, and determining the influence relationship between each target state data and the preset computer room state data; and constructing the target energy-saving control model based on the influence relationship.

[0012] Optionally, the types of preset data room status data include at least one of the following: data room temperature, total energy consumption of data room equipment; the types of target status data include at least one of the following: ambient temperature, operating status parameters of each target device in the target area, the types of target devices include at least one of the following: IT equipment, temperature control equipment, and the operating status parameters include at least one of the following: operating power, heat dissipation, cooling capacity, and air volume.

[0013] Optionally, the target energy-saving control model is used to analyze the target state data set and the preset computer room state data to obtain energy-saving control instructions, including: substituting the preset computer room temperature and the operating state parameters of IT equipment into the target energy-saving control model to solve for the target operating state parameters of each temperature control device when the total energy consumption of the computer room equipment is minimized; determining the current operating state parameters of each temperature control device from the target state data set; and generating energy-saving control instructions for the corresponding temperature control device based on the difference between the current operating state parameters of each temperature control device and the corresponding target operating state parameters.

[0014] According to another aspect of the embodiments of this application, a data center energy-saving control device is also provided, comprising: an acquisition module, configured to acquire a target state data set of a target area in a target data center within a target period, wherein the target state data set includes at least one of the following types of data: environmental state data and equipment state data; a selection module, configured to determine a target causal reasoning algorithm that matches the target state data set, and construct a target energy-saving control model based on the target causal reasoning algorithm and the target state data set, wherein the target causal reasoning algorithm is used to analyze the mutual influence relationship between different types of state data; an analysis module, configured to analyze the target state data set and preset data center state data using the target energy-saving control model to obtain energy-saving control instructions; and a control module, configured to control the target equipment in the target area according to the energy-saving control instructions.

[0015] According to another aspect of the embodiments of this application, a computer program product is also provided, the computer program product comprising: a computer program, wherein the computer program, when executed by a processor, implements the above-described data center energy-saving control method.

[0016] According to another aspect of the embodiments of this application, an electronic device is also provided, the electronic device including: a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the above-described data center energy-saving control method through the computer program.

[0017] In this embodiment, an intelligent algorithm screening mechanism is used to rationally select a causal inference algorithm suitable for the current computing resources, reducing the burden on edge computing and ensuring the efficient operation of the system. Through causal inference technology, the impact of different environmental parameters on the energy consumption of the data center is accurately analyzed, avoiding ineffective regulation and significantly improving energy-saving effect. By identifying key variables and their influence paths through causal inference technology, the modeling of multivariate causal relationships is realized, thereby generating accurate energy-saving control commands, realizing closed-loop optimization of energy-saving strategies, reducing the energy consumption of data center equipment, and thus solving the technical problem of energy waste caused by the inability to accurately identify and respond to complex causal effects between multiple variables in traditional data center control methods. Attached Figure Description

[0018] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0019] Figure 1 This is a flowchart illustrating an optional energy-saving control method for a computer room according to an embodiment of this application;

[0020] Figure 2 This is a schematic diagram of an optional sensor deployment according to an embodiment of this application;

[0021] Figure 3 This is a schematic diagram of an optional energy-saving control device for a computer room according to an embodiment of this application;

[0022] Figure 4 This is a schematic diagram of the structure of an optional electronic device according to an embodiment of this application. Detailed Implementation

[0023] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0024] It should be noted that the terms "first," "second," etc., used in the specification, claims, and drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0025] To better understand the embodiments of this application, the following is a translation and explanation of some nouns or terms that appear in the description of the embodiments of this application:

[0026] The Augmented Dickey-Fuller Test (ADF) is a commonly used statistical test primarily used to determine whether a time series has a unit root, i.e., whether it is non-stationary. In time series analysis, the presence of non-stationary series can lead to spurious correlations or regression results; therefore, determining the stationarity of the series is crucial. The ADF test examines the presence of one or more unit roots by differencing the original series and then performing regression analysis on the differrated series. This method takes into account lag effects in the series, hence the name "enhanced" Dickey-Fuller test. If the ADF test result indicates that the series is stationary, it can be used for further analysis, such as modeling or prediction; otherwise, it may be necessary to differulate or perform other transformations to make the series stationary.

[0027] The Dynamic Continuous Time Bayesian Network (DCTBN) algorithm is an extended Bayesian network model used to describe complex dynamic relationships between variables in time-varying environments. In traditional Bayesian networks, nodes represent random variables, edges represent conditional dependencies between variables, and the network structure reflects the state transition probabilities of a discrete-event system. However, DCTBN can handle not only discrete states but also continuous-time variations, meaning it can capture the dynamic characteristics of variables changing continuously over time with greater precision. This algorithm is particularly suitable for handling causal relationships with different time scales and time delay effects. For example, in intelligent building management, DCTBN can be used to analyze the dynamic interactions between different sensor data and how these interactions evolve over time, thereby making more accurate predictions and control decisions.

[0028] Structural Equation Modeling (SEM) algorithms: SEM is a multivariate statistical analysis technique used to build, estimate, and test complex causal relationships between multiple variables, and is widely used, particularly in sociology, psychology, and economics. Compared to traditional regression analysis, SEM can handle more complex models, such as those involving latent variables (unobserved variables) and manifest variables (directly observed variables), as well as direct and indirect influences between variables. SEM typically uses path modeling to graphically represent causal relationships between variables, estimates model parameters using methods such as maximum likelihood estimation, and evaluates the model's validity and reasonableness using various indicators (such as fit indices). This modeling framework allows researchers to start from theory, construct complex causal networks, and then use data to verify and adjust these theoretical assumptions, thereby gaining a deeper understanding of phenomena.

[0029] Mutual information: Mutual information is an important concept in information theory, used to measure the degree of interdependence between two random variables. Unlike the correlation coefficient, mutual information can capture all types of statistical dependencies between variables, including not only linear ones but also nonlinear ones. It is based on the concept of entropy, representing how observing one variable can reduce the amount of information we have about the uncertainty of the other variable.

[0030] Example 1

[0031] According to an embodiment of this application, a data center energy-saving control method is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0032] Figure 1 This is a flowchart illustrating a data center energy-saving control method according to an embodiment of this application, as shown below. Figure 1 As shown, the method includes the following steps:

[0033] Step S102: Obtain the target status data set of the target area in the target computer room within the target period, wherein the target status data set includes at least one of the following types of data: environmental status data and equipment status data;

[0034] Step S104: Determine the target causal reasoning algorithm that matches the target state data set, and construct the target energy-saving control model based on the target causal reasoning algorithm and the target state data set. The target causal reasoning algorithm is used to analyze the mutual influence relationship between different types of state data.

[0035] Step S106: Analyze the target state data set and preset computer room state data using the target energy-saving control model to obtain energy-saving control instructions;

[0036] Step S108: Adjust the target equipment in the target area according to the energy-saving control command.

[0037] The following section explains each step of the data center energy-saving control method in conjunction with the specific implementation process.

[0038] As an optional implementation method, obtaining the target status data set of the target area in the target data center within the target period can be achieved in the following ways:

[0039] First, acquire the status data reported by multiple sensors within the target area of ​​the target equipment room at multiple acquisition times within the target period.

[0040] Figure 2 A schematic diagram of an optional sensor deployment is shown. We consider a specific data center (DC) compartment within a large data center, specifically a separate temperature-controlled area within the target server room, equipped with an advanced sensor network and air conditioning system. Within the target area of ​​the target server room, sensors 1-6 are carefully deployed at key locations such as rack gaps and air conditioning vents 1-6. The control panel, located at the core, primarily provides an interface for displaying received real-time data, ensuring that relevant personnel can promptly access comprehensive monitoring information and understand changes in environmental conditions and other relevant data.

[0041] Within a specified target period, such as every 15 minutes, the sensor network reports status data in the computer room at multiple collection times, including but not limited to the following categories: IT equipment load data, air conditioning system parameter data, environmental parameter data, and power distribution data.

[0042] The data includes: IT equipment load data: monitoring CPU utilization, memory usage, and network traffic through sensors, reflecting the operating status and energy consumption requirements of IT equipment; Air conditioning system parameter data: collecting information such as air conditioning supply air temperature, compressor frequency, and refrigerant flow, which directly determine the energy consumption and cooling efficiency of the air conditioning system; Environmental parameter data: sensors continuously monitor the temperature and humidity distribution of the data center and the inlet and outlet temperatures of the server racks to assess the thermodynamic state of the data center; and Power distribution data: collecting PUE (Power Usage Effectiveness) and branch circuit power consumption information, which is crucial for comprehensively evaluating energy-saving effects.

[0043] Secondly, all state data undergoes data preprocessing, which includes at least one of the following: data cleaning, normalization, and timestamp alignment.

[0044] Specifically, all state data is transmitted to the controller for processing via wired or wireless networks. In the data preprocessing stage, the system first performs data cleaning to remove outliers and noise, ensuring data quality. Subsequently, the data is timestamped to ensure consistency across time series data from different sensors, which is fundamental for time series analysis. Finally, normalization is performed to transform data of different dimensions and ranges to a uniform scale, facilitating subsequent modeling and analysis. The order of these preprocessing operations can be selected as needed.

[0045] Finally, the preprocessed state data is divided according to data type to obtain multiple state time series corresponding to multiple state data types, and the multiple state time series are combined into a target state data set.

[0046] The preprocessed state data is further categorized by data type, such as separating IT equipment load data, air conditioning system parameter data, environmental parameter data, and power distribution data, with each type of data forming a corresponding state time series. For example, each parameter, such as CPU utilization and supply air temperature, has its own time series. These state time series are then integrated to form the target state data set, providing a complete and high-quality data foundation for subsequent causal reasoning analysis.

[0047] By meticulously preprocessing multiple sensor data points within the target area of ​​the target computer room, the quality and applicability of the data can be effectively improved, laying a solid foundation for subsequent causal reasoning analysis and ultimately enabling more precise energy-saving control of the computer room system.

[0048] Since the characteristics of the data are closely related to the selection of the causal inference algorithm, as an optional implementation method, after obtaining the target state data set, the target causal inference algorithm matching the target state set can be determined in the following way:

[0049] First, determine the mutual information between different types of state data in the target state data set, and then determine the mutual information delay between different types of state data based on the mutual information.

[0050] By calculating the maximum mutual information delay between environmental parameters, such as the time delay relationship between regional temperature and air conditioning energy consumption, we can gain insight into the dynamic response patterns between data. If a significant delay is detected, that is, the maximum mutual information delay exceeds a preset delay threshold (e.g., 5 minutes), it indicates that the causal relationship between variables has a significant time lag.

[0051] Secondly, statistical testing methods are used to determine the stationarity of each type of state data in the target state dataset. These statistical testing methods include the ADF method.

[0052] Specifically, when conducting the data stationarity test, we use the ADF method to analyze each type of state data in the target state dataset, aiming to verify whether the statistical characteristics of the time series remain unchanged.

[0053] Next, obtain the amount of computing resources pre-allocated for processing the data center energy-saving control task;

[0054] We can assess the amount of computing resources pre-allocated to data center energy-saving control tasks. When the CPU utilization of edge computing devices exceeds 80%, it indicates resource scarcity. In this case, the intelligent filtering mechanism will force the selection of the SEM algorithm. The SEM algorithm, with its lightweight characteristics, can perform efficient data analysis in resource-constrained environments. Conversely, if computing resources are sufficient, we can try to retain the algorithm selected based on the analysis results of the first two steps, namely DCTBN or SEM.

[0055] Finally, a target causal inference algorithm matching the target state data set is determined based on the target parameters, wherein the target parameters include at least one of the following: mutual information delay, data stationarity result, and computational resource quantity.

[0056] Specifically, after obtaining the above target parameter information, the target causal reasoning algorithm that matches the target state data set can be determined in the following ways.

[0057] For example, the target causal inference algorithm can be determined based on mutual information latency and computational resource availability. The specific determination process is as follows: determine the maximum mutual information latency among the mutual information latency of different types of state data; if the computational resource availability is not less than a preset resource availability threshold and the maximum mutual information latency is not less than a preset latency threshold, the DCTBN algorithm is selected as the target causal inference algorithm; if the computational resource availability is less than the preset resource availability threshold or the maximum mutual information latency is less than the preset latency threshold, the SEM algorithm is selected as the target causal inference algorithm.

[0058] It should be noted that when the maximum mutual information delay exceeds a preset delay threshold, the DCTBN algorithm is the preferred choice due to its unique delay handling capabilities. DCTBN can effectively handle non-instantaneous causal effects, accurately capturing causal relationships between variables even in non-stationary time-series data. If the maximum mutual information delay is less than the preset delay threshold, we tend to use the SEM algorithm.

[0059] In addition, this application embodiment also provides another optional method for determining the target causal inference algorithm, which is mainly determined by the data stationarity result and the amount of computing resources. The specific determination process is as follows: the target causal inference algorithm that matches the target state data set is determined based on the target parameters. This can be achieved in the following ways: when the amount of computing resources is not less than a preset resource threshold and the data stationarity result indicates that the data is non-stationary, the DCTBN algorithm is used as the target causal inference algorithm; when the amount of computing resources is less than the preset resource threshold or the data stationarity result indicates that the data is stationary, the SEM algorithm is used as the target causal inference algorithm.

[0060] It should be noted that if all time series data for all variables show stationarity (i.e., the ADF test value is less than 0.05), we tend to choose the SEM algorithm. SEM can quickly build causal relationship models when dealing with stationary data, while providing high model accuracy. However, if the ADF test results indicate that the data is non-stationary, the DCTBN algorithm becomes the optimal choice again because it effectively handles non-stationary time series data, maintaining model stability and predictive accuracy even when data attributes change over time.

[0061] By introducing intelligent screening of causal inference algorithms, key variables affecting energy consumption and their causal paths can be identified more accurately, thereby enabling the development of more effective energy-saving measures. When determining the target causal inference algorithm, not only time delay characteristics or data stationarity results are considered, but also the size of computing resources. The algorithm selection is adjusted based on the real-time status of computing resources, avoiding the use of high-energy-consuming DCTBNs when resources are scarce, thus saving computing costs. Furthermore, the algorithm selection logic can adapt to changes in data characteristics; whether the data is stationary or non-stationary, the strategy can be quickly adjusted, improving the system's flexibility and response speed. The algorithm screening is dynamically adjusted based on real-time acquired data, allowing the system to continuously optimize its selection during operation, ensuring continuous improvement and best practices in energy-saving control strategies. Therefore, the entire dynamic screening process ensures that the system can dynamically adjust and select the most suitable causal inference algorithm under different data characteristics and computing resource constraints to achieve optimal energy-saving control in the data center.

[0062] After obtaining the target causal reasoning algorithm, we can use it to construct a causal network between parameters, identify core variables, and establish a causal relationship network, thereby laying the foundation for building a target energy-saving control model.

[0063] As an optional implementation method, the target energy-saving control model can be constructed based on the target causal reasoning algorithm and the target state data set, which can be achieved in the following way:

[0064] First, obtain multiple sets of historical state data that match the target causal reasoning algorithm.

[0065] Specifically, we obtain data that matches the target causal inference algorithm from multiple existing historical state datasets. The historical state datasets contain various state data during the past operation of the data center, such as data center temperature, total energy consumption of data center equipment (preset data center state data), and ambient temperature, operating status parameters of IT equipment and temperature control equipment (such as air conditioners) (target state data).

[0066] Secondly, the target causal reasoning algorithm is used to analyze the state data in each historical state data set and the target state data set to determine each target state data that has a causal relationship with the preset computer room state data, and to determine the influence relationship between each target state data and the preset computer room state data.

[0067] Subsequently, we employed selected causal reasoning algorithms (DCTBN or SEM) to conduct in-depth analysis of the state data in the historical and target state datasets. The aim was to identify target state data that exhibited causal relationships with the preset data center state data. For example, we might discover a positive causal relationship between data center temperature and air conditioning set temperature, and a negative causal relationship between ambient temperature and total energy consumption of data center equipment. Based on these findings, we were able to identify the main factors influencing data center temperature and energy consumption, such as the correlation between air conditioning load and ambient temperature, and the relationship between IT equipment heat dissipation and total energy consumption of data center equipment, and quantify the degree of influence of these factors.

[0068] Common types of preset data for computer room status include, but are not limited to, one of the following: computer room temperature, total energy consumption of computer room equipment; types of target status data include at least one of the following: ambient temperature, operating status parameters of each target device in the target area, the types of target devices include at least one of the following: IT equipment, temperature control equipment, and operating status parameters include at least one of the following: operating power, heat dissipation, cooling capacity, and air volume.

[0069] By setting the computer room temperature and total equipment energy consumption as preset state data, and the ambient temperature and equipment operating status parameters as target state data, key parameters for computer room energy-saving control are covered. Generally, preset state data is used to define the energy-saving control target, while target state data is used for real-time monitoring and analysis of the computer room environment to build a causal relationship model. The technical solution in this application embodiment can generate differentiated control commands by analyzing the causal relationship between preset state data and target state data, such as adjusting the air conditioning power of specific areas and shutting down redundant air conditioners, to achieve more efficient energy-saving control. Furthermore, we can also solve energy-saving control problems in complex environments by introducing more types of preset state data and target state data, such as humidity and airflow distribution.

[0070] After determining the causal relationships and the degree of influence, we further derived the influence relationship between each target state data and the preset computer room state data. For example, the dynamic relationship between the air conditioning set temperature and the current target area (DC compartment) temperature can be simplified using a differential equation model:

[0071]

[0072] Where k is the air conditioner's cooling / heating regulation rate, and Q... ext It is an external thermal interference term, T set It is the set temperature of the air conditioning equipment, T DC These are the actual temperatures inside the DC compartment in the computer room. These relationships clearly demonstrate the immediate impact of changes in the air conditioning set temperature on the DC compartment temperature, as well as the mathematical description of its long-term steady-state behavior.

[0073] For the parameter k, we can generally solve it in the following way:

[0074] For example, to set an ideal condition, which can be achieved by disabling thermal interference, let Q... ext =0, record the time t required for the DC compartment to cool from 30℃ to 26℃, then

[0075]

[0076] The calculation process of the k value more intuitively reflects the relationship between the amount of temperature change and the required time, thereby estimating the ability of the air conditioning system to lower or raise the temperature per unit time.

[0077] Determine the steady-state solution (at a stable temperature) of the above differential equation. at this time, It should be noted that in the presence of an external heat source, the cooling output of the air conditioning system needs to slightly exceed the demand under balanced conditions to ensure that the computer room temperature does not exceed the set temperature due to the influence of the external heat source.

[0078] Regarding the relationship between the total energy consumption of IT equipment and data center equipment, we can derive the first equation for this relationship using the principle of heat balance:

[0079]

[0080] Among them, T in Q represents the intake air temperature of the DC compartment cabinet. IT Q is the total heat generated by IT equipment. loss It is heat transfer in building structure, C p F is the specific heat capacity of air, and F is the air volume supplied by the air conditioning system. The first relationship above is determined based on the principle that the cooling capacity of the air conditioning system needs to offset the heat generated by the IT equipment and the heat leakage from the environment.

[0081] Furthermore, the heat channel capture efficiency η can be introduced to modify the above first relationship to obtain the corresponding second relationship:

[0082]

[0083] By introducing thermal channel capture efficiency, temperature can be controlled more precisely, avoiding the increased energy consumption caused by temperature stratification.

[0084] It should be noted that the above-listed influence relationships are merely examples and do not constitute specific limitations.

[0085] Finally, a target energy-saving control model is constructed based on a series of influence relationships obtained.

[0086] Finally, we utilize these influence relationships to construct a target energy-saving control model. The goal of model construction is to find an optimal control strategy that minimizes the total energy consumption of the computer room equipment while maintaining the computer room temperature within a suitable operating range. For example, by increasing the air conditioning set temperature T... set This can reduce |T DC -T set This reduces the compressor's operating time and energy consumption. For example, by adjusting the airflow F using a variable frequency drive, the air volume T can be maintained. in -T set The difference is optimized within the range of 2℃ to 4℃ to achieve energy-saving effects under dynamic air volume adjustment.

[0087] As an optional implementation method, the target energy-saving control model is used to analyze the target state data set and the preset computer room state data to obtain energy-saving control instructions. This can be achieved in the following way: Substitute the preset computer room temperature and the operating state parameters of the IT equipment into the target energy-saving control model to solve for the target operating state parameters of each temperature control device that minimizes the total energy consumption of the computer room equipment; determine the current operating state parameters of each temperature control device from the target state data set; and generate the corresponding energy-saving control instructions for the temperature control device based on the difference between the current operating state parameters of each temperature control device and the corresponding target operating state parameters.

[0088] Suppose we have established a target energy-saving control model based on causal reasoning. This model identifies and quantifies the causal paths between key variables by comprehensively analyzing the complex relationships between computer room temperature, total equipment energy consumption, ambient temperature, and operating status parameters of temperature control equipment.

[0089] Let's take a specific scenario as an example, assuming the current temperature in the computer room is T. DC =27℃, the total heat dissipation Q of IT equipment IT It has reached the preset high level, while the external ambient temperature T env Relatively low.

[0090] We will preset the computer room temperature T DC and the operating status parameters of IT equipment (such as CPU utilization, heat dissipation Q) IT Substitute these parameters into the target energy-saving control model. The model calculates the target operating state parameters of the temperature control equipment (i.e., the air conditioning system) that minimize the total energy consumption of the computer room equipment, such as the target supply air temperature T. set_target Target air volume F target and target air conditioning power P target wait.

[0091] To solve for this optimal state, the model may employ various mathematical methods, including but not limited to linear programming, dynamic programming, or genetic algorithms, to find the temperature control equipment state parameters that minimize energy consumption without affecting the stability of the computer room's temperature and humidity. Taking a simplified model as an example, if T... set_target The temperature is 22℃, while the current set temperature T set The temperature was 18°C, which indicates that by increasing the supply air temperature to be closer to room temperature, the compressor's operating time can be reduced, thus achieving energy savings on the air side.

[0092] Next, we determine the current operating status parameters of each temperature control device from the target status dataset. For example, the current supply air temperature T set The temperature is 18℃, and the air supply volume F = 0.5m³. 3 The data includes information such as the air conditioner's operating power ( / s) and the fact that the air conditioner's operating power is 70% of its rated power. Obtaining this information through real-time monitoring and data analysis is a prerequisite for executing energy-saving control commands.

[0093] Finally, based on the difference between the current operating parameters of the temperature control equipment and the target operating parameters obtained from the model solution, we generate specific energy-saving control commands. Taking the supply air temperature as an example, if the current temperature is set to (T... set =18℃, while the model suggests a target temperature of (T) set_target If the temperature is 22℃, then the energy-saving control command may include "gradually increasing the air conditioning supply air temperature from 18℃ to 22℃". Furthermore, for the supply air volume F, if the model predicts the optimal supply air volume to be 0.7m³ / h... 3 / s, while the current value is 0.5m. 3 If / s, the instruction may include "adjust the air volume to the optimal air volume via the inverter" to optimize the cooling efficiency of the air conditioning system.

[0094] Furthermore, considering the reasonableness of the predicted adjustment values ​​given by the model, we need to continuously monitor the data center status after executing the control commands to evaluate their effectiveness. If, after execution, the data center temperature stabilizes, energy consumption decreases, and the performance of IT equipment remains unaffected, this demonstrates the effectiveness of the energy-saving control model and the successful implementation of the energy-saving strategy. Based on real-time feedback data, the energy-saving model will iteratively optimize to update its parameters. For example, adjusting the air conditioning modulation rate k value to adapt to new environmental conditions and changing equipment load states, this closed-loop feedback mechanism ensures continuous optimization of energy-saving control, maintaining high energy efficiency even when internal and external conditions of the data center change.

[0095] Through the above steps, combined with an intelligent algorithm screening mechanism, a causal inference algorithm suitable for the current computing resources is rationally selected, reducing the burden on edge computing and ensuring the efficient operation of the system. Causal inference technology accurately analyzes the impact of different environmental parameters on data center energy consumption, avoiding ineffective regulation and significantly improving energy-saving effects. By identifying key variables and their influence paths through causal inference technology, modeling multivariate causal relationships is achieved, thereby generating precise energy-saving control commands, realizing closed-loop optimization of energy-saving strategies, reducing the energy consumption of data center equipment, and thus solving the technical problem of energy waste caused by the inability to accurately identify and respond to complex causal effects between multiple variables in traditional data center control methods.

[0096] Example 2

[0097] According to an embodiment of this application, a data center energy-saving control device for implementing the data center energy-saving control method in Embodiment 1 is also provided, such as... Figure 3 As shown, the energy-saving control device for the computer room includes at least: an acquisition module 31, a selection module 32, an analysis module 33, and a control module 34, wherein:

[0098] The acquisition module 31 can acquire a set of target status data of the target area in the target computer room within the target period, wherein the target status data set includes at least one of the following types of data: environmental status data and equipment status data;

[0099] Selecting module 32 allows you to determine the target causal reasoning algorithm that matches the target state data set, and to construct a target energy-saving control model based on the target causal reasoning algorithm and the target state data set. The target causal reasoning algorithm can analyze the mutual influence between different types of state data.

[0100] Analysis module 33 can use the target energy-saving control model to analyze the target state data set and the preset computer room state data to obtain energy-saving control instructions;

[0101] The control module 34 can control the target equipment in the target area according to the energy-saving control command.

[0102] The following section explains the functions of each module of the data center energy-saving control device in conjunction with the specific implementation process.

[0103] As an optional implementation, the acquisition module acquires a set of target status data for the target area in the target room within the target period. This can be achieved by: acquiring status data reported by multiple sensors within the target area of ​​the target room at multiple acquisition times within the target period, and performing data preprocessing on all status data. The data preprocessing includes at least one of the following: data cleaning, normalization, and timestamp alignment. The preprocessed status data is then divided according to data type to obtain multiple status time series corresponding to multiple types of status data, and these multiple status time series are combined to form a target status data set.

[0104] As an optional implementation, the selection module determines the target causal inference algorithm that matches the target state data set, which can be achieved by: determining the mutual information between different types of state data in the target state data set, and determining the mutual information delay between different types of state data based on the mutual information; determining the data stationarity result of each type of state data in the target state data set using statistical testing methods, wherein the statistical testing methods include: the ADF method; obtaining the amount of computing resources pre-allocated for the data center energy-saving control task; and determining the target causal inference algorithm that matches the target state data set based on target parameters, wherein the target parameters include at least one of the following: mutual information delay, data stationarity result, and computing resources.

[0105] As an optional implementation, the selection module determines the target causal inference algorithm that matches the target state data set based on the target parameters. This can be achieved in the following ways: determining the maximum mutual information delay among the mutual information delays between different types of state data; using the DCTBN algorithm as the target causal inference algorithm when the computational resources are not less than a preset resource threshold and the maximum mutual information delay is not less than a preset delay threshold; and using the SEM algorithm as the target causal inference algorithm when the computational resources are less than a preset resource threshold or the maximum mutual information delay is less than a preset delay threshold.

[0106] As an optional implementation, the selection module determines the target causal inference algorithm that matches the target state data set based on the target parameters. This can be achieved in the following ways: when the amount of computing resources is not less than a preset resource threshold and the data stationarity result indicates that the data is not stationary, the DCTBN algorithm is used as the target causal inference algorithm; when the amount of computing resources is less than the preset resource threshold or the data stationarity result indicates that the data is stationary, the SEM algorithm is used as the target causal inference algorithm.

[0107] As an optional implementation, the selection module constructs a target energy-saving control model based on the target causal reasoning algorithm and the target state data set. This can be achieved in the following way: acquiring multiple historical state data sets that match the target causal reasoning algorithm; using the target causal reasoning algorithm to analyze the state data in each historical state data set and the target state data set, determining each target state data that has a causal relationship with the preset computer room state data, and determining the influence relationship between each target state data and the preset computer room state data; and constructing the target energy-saving control model based on the influence relationship.

[0108] As an optional implementation, the types of preset data room status include at least one of the following: data room temperature, total energy consumption of data room equipment; the types of target status data include at least one of the following: ambient temperature, operating status parameters of each target device in the target area, the types of target devices include at least one of the following: IT equipment, temperature control equipment, and the operating status parameters include at least one of the following: operating power, heat dissipation, cooling capacity, and air volume.

[0109] As an optional implementation, the analysis module uses a target energy-saving control model to analyze the target state data set and the preset computer room state data to obtain energy-saving control instructions. This can be achieved in the following way: Substitute the preset computer room temperature and the operating state parameters of the IT equipment into the target energy-saving control model to solve for the target operating state parameters of each temperature control device that minimizes the total energy consumption of the computer room equipment; determine the current operating state parameters of each temperature control device from the target state data set; and generate energy-saving control instructions for the corresponding temperature control device based on the difference between the current operating state parameters of each temperature control device and the corresponding target operating state parameters.

[0110] It should be noted that each module in the data center energy-saving control device in this application embodiment corresponds one-to-one with each implementation step of the data center energy-saving control method in embodiment 1. Since embodiment 1 has been described in detail, some details not shown in this embodiment can be referred to embodiment 1, and will not be elaborated further here.

[0111] Example 3

[0112] According to an embodiment of this application, a computer program product is also provided, which includes a computer program, wherein when the computer program is executed by a processor, it implements the data center energy-saving control method in embodiment 1.

[0113] According to an embodiment of this application, a non-volatile storage medium is also provided, which includes a stored computer program, wherein the device containing the non-volatile storage medium executes the data center energy-saving control method in Embodiment 1 by running the computer program.

[0114] According to an embodiment of this application, a processor is also provided for running a computer program, wherein the computer program executes the data center energy-saving control method in embodiment 1 during runtime.

[0115] According to an embodiment of this application, an electronic device is also provided, comprising: a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the data center energy-saving control method of Embodiment 1 through the computer program.

[0116] Specifically, the computer program executes the following steps during runtime: acquiring a target state data set of the target area in the target computer room within the target period, wherein the target state data set includes at least one of the following data types: environmental state data and equipment state data; determining a target causal inference algorithm that matches the target state data set, and constructing a target energy-saving control model based on the target causal inference algorithm and the target state data set, wherein the target causal inference algorithm is used to analyze the mutual influence relationship between different types of state data; analyzing the target state data set and the preset computer room state data using the target energy-saving control model to obtain energy-saving control instructions; and regulating the target equipment in the target area according to the energy-saving control instructions.

[0117] As an alternative implementation, the above-mentioned electronic device may exist in the form of a mobile terminal, a computer terminal, or a similar computing device. Figure 4 A hardware block diagram of an electronic device for implementing an energy-saving control method in a computer room is shown. Figure 4 As shown, the electronic device 40 may include one or more processors 402 (shown as 402a, 402b, ..., 402n in the figure) 402 (processor 402 may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 404 for storing data, and a transmission device 406 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 4 The structure shown is for illustrative purposes only and does not limit the structure of the electronic device described above. For example, electronic device 40 may also include... Figure 4 The more or fewer components shown, or having the same Figure 4 The different configurations shown.

[0118] It should be noted that the aforementioned one or more processors 402 and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element of the electronic device 40. As involved in the embodiments of this application, the data processing circuits serve as a processor control mechanism (e.g., selection of a variable resistor termination path connected to an interface).

[0119] The memory 404 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the data center energy-saving control method in this embodiment. The processor 402 executes various functional applications and data processing by running the software programs and modules stored in the memory 404, thereby implementing the aforementioned application vulnerability detection method. The memory 404 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 404 may further include memory remotely located relative to the processor 402, and these remote memories can be connected to the electronic device 40 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0120] The transmission device 406 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the electronic device 40. In one example, the transmission device 406 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 406 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.

[0121] The display can be, for example, a touchscreen liquid crystal display (LCD), which allows the user to interact with the user interface of the electronic device 40.

[0122] The sequence numbers of the above embodiments are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0123] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0124] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between units or modules may be electrical or other forms.

[0125] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0126] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0127] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0128] The above are merely preferred embodiments of this application. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for energy-saving control of a computer room, characterized in that, include: Obtain a target status data set of the target area in the target computer room within the target period, wherein the target status data set includes at least one of the following data types: environmental status data and equipment status data; A target causal reasoning algorithm matching the target state data set is determined, and a target energy-saving control model is constructed based on the target causal reasoning algorithm and the target state data set. The target causal reasoning algorithm is used to analyze the mutual influence relationship between different types of state data. The target energy-saving control model is used to analyze the target state data set and the preset computer room state data to obtain energy-saving control instructions; The target equipment in the target area is regulated according to the energy-saving control command.

2. The method according to claim 1, characterized in that, Obtain the target status data set of the target area in the target data center within the target period, including: The status data reported by multiple sensors in the target area of ​​the target room at multiple collection times within the target period are obtained, and all status data are preprocessed, wherein the data preprocessing includes at least one of the following: data cleaning, normalization processing, and timestamp alignment. The preprocessed state data is divided according to data type to obtain multiple state time series corresponding to multiple state data types, and the multiple state time series are combined to form the target state data set.

3. The method according to claim 2, characterized in that, Determining a target causal inference algorithm that matches the target state data set includes: Determine the mutual information between different types of state data in the target state data set, and determine the mutual information delay between different types of state data based on the mutual information; The stationarity of each type of state data in the target state dataset is determined using statistical testing methods, including the Enhanced Dick-Fuller Test (ADF). Obtain the amount of computing resources pre-allocated for processing data center energy-saving control tasks; A target causal reasoning algorithm matching the target state data set is determined based on target parameters, wherein the target parameters include at least one of the following: mutual information delay, data stationarity result, and computational resource quantity.

4. The method according to claim 3, characterized in that, The target causal reasoning algorithm, which determines the target state data set based on the target parameters, includes: Determine the maximum mutual information delay among the mutual information delays between different types of state data; When the amount of computing resources is not less than a preset resource threshold and the maximum mutual information delay is not less than a preset delay threshold, the Dynamic Continuous Time Bayesian Network (DCTBN) algorithm is used as the target causal inference algorithm. If the amount of computing resources is less than a preset resource threshold, or the maximum mutual information delay is less than a preset delay threshold, the Structural Equation Modeling (SEM) algorithm will be used as the target causal inference algorithm.

5. The method according to claim 3, characterized in that, The target causal reasoning algorithm, which determines the target state data set based on the target parameters, includes: When the amount of computing resources is not less than a preset resource threshold and the data stationarity result indicates that the data is non-stationary, the DCTBN algorithm will be used as the target causal inference algorithm. When the amount of computing resources is less than a preset resource threshold, or when the data stability result indicates that the data is stable, the SEM algorithm will be used as the target causal inference algorithm.

6. The method according to claim 1, characterized in that, Based on the target causal reasoning algorithm and the target state data set, a target energy-saving control model is constructed, including: Obtain multiple sets of historical state data that match the target causal reasoning algorithm; The target causal reasoning algorithm is used to analyze the state data in each of the historical state data sets and the target state data sets to determine each target state data that has a causal relationship with the preset data room state data, and to determine the influence relationship between each target state data and the preset data room state data. The target energy-saving control model is constructed based on the aforementioned influence relationship.

7. The method according to claim 6, characterized in that, The types of preset data room status include at least one of the following: data room temperature, total energy consumption of data room equipment; The target status data includes at least one of the following types: ambient temperature, operating status parameters of each target device in the target area, the target device type includes at least one of the following types: IT equipment, temperature control equipment, and the operating status parameters include at least one of the following types: operating power, heat dissipation, cooling capacity, and air volume.

8. The method according to claim 7, characterized in that, The target energy-saving control model is used to analyze the target state data set and the preset computer room state data to obtain energy-saving control instructions, including: Substitute the preset computer room temperature and the operating status parameters of IT equipment into the target energy-saving control model to solve for the target operating status parameters of each temperature control device when the total energy consumption of the computer room equipment is minimized. Determine the current operating status parameters of each temperature control device from the target status data set; Based on the difference between the current operating status parameters of each temperature control device and the corresponding target operating status parameters, energy-saving control instructions are generated for the corresponding temperature control device.

9. A computer room energy-saving control device, characterized in that, include: The acquisition module is used to acquire a set of target status data of a target area in the target computer room within a target period, wherein the set of target status data includes at least one of the following types of data: environmental status data and equipment status data; The selection module is used to determine the target causal reasoning algorithm that matches the target state data set, and to construct the target energy-saving control model based on the target causal reasoning algorithm and the target state data set. The target causal reasoning algorithm is used to analyze the mutual influence relationship between different types of state data. The analysis module is used to analyze the target state data set and the preset computer room state data using the target energy-saving control model to obtain energy-saving control instructions; The control module is used to control the target equipment in the target area according to the energy-saving control command.

10. A computer program product, characterized in that, include: A computer program, wherein when executed by a processor, the computer program implements the data center energy-saving control method according to any one of claims 1 to 8.

11. An electronic device, characterized in that, include: A memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the data center energy-saving control method according to any one of claims 1 to 8 through the computer program.

Images