Intelligent temperature control method and system for data center room, electronic device and storage medium

Abstract

The application relates to the technical field of network management, and provides a data center machine room intelligent temperature control method and system, electronic equipment and a storage medium, the method uses a multistage PID controller to perform parallel adjustment on current values of each target energy consumption index, can automatically adjust and optimize the temperature and humidity of the data center machine room based on an iterative closed-loop feedback mechanism, realizes full-process full-automatic intelligent operation and engineering application of an intelligent temperature control energy-saving algorithm for a complex scene, solves the problems of single adjustment measure, low efficiency, inaccurate temperature control and long time consumption of a single PID controller, has strong adaptability to the complexity and nonlinearity of the data center machine room system, can process complex temperature and humidity control problems, and has good energy-saving effect. Moreover, the parallel adjustment method can improve the response speed and control precision of the intelligent temperature control system, so that the intelligent temperature control system can better adapt to the working environment and load changes of the data center cabinet.

Description

Technical Field

[0001] This invention relates to the field of network management technology, and in particular to a method, system, electronic device and storage medium for intelligent temperature control in data center computer rooms. Background Technology

[0002] Currently, energy consumption in data center infrastructure has become a key factor restricting the high-quality development of computing networks. Data center development faces numerous challenges, including enormous energy consumption, increasingly stringent energy efficiency management requirements, and a lack of systematic energy-saving technologies, making it difficult to meet the energy-saving requirements of various industries. Applying intelligent temperature control algorithms to data centers can more precisely control cooling systems, reduce energy consumption and maintenance costs, and achieve energy conservation and efficiency improvements.

[0003] Existing intelligent temperature control algorithms primarily use machine learning or deep learning-based approaches to automatically optimize the parameters of proportional-integral-differential (PID) control algorithms. In this case, a PID controller is needed to adjust each environmental indicator in the data center room, thereby achieving intelligent temperature control.

[0004] However, the PID controller used is not very adaptable to the complexity and nonlinearity of data center computer room systems, and it is difficult to handle the complex control problems of multiple environmental indicators, thus resulting in limited energy-saving effects. Summary of the Invention

[0005] This invention provides a method, system, electronic device, and storage medium for intelligent temperature control in data center computer rooms, in order to overcome the deficiencies existing in the prior art.

[0006] This invention provides a method for intelligent temperature control in a data center computer room, comprising:

[0007] Obtain the current values ​​of each target energy consumption indicator of the data center computer room, and perform anomaly detection on the current values ​​of each target energy consumption indicator to determine the abnormal information of the data center computer room;

[0008] Based on the aforementioned anomaly information, the target values ​​for each target energy consumption indicator are determined.

[0009] Based on a multi-level PID controller, the target values ​​of each target energy consumption index are applied, and the current values ​​of each target energy consumption index are adjusted in parallel to control the temperature and humidity of the data center computer room.

[0010] According to a data center intelligent temperature control method provided by the present invention, the step of detecting anomalies in the current values ​​of the target energy consumption indicators to determine the abnormal information of the data center includes:

[0011] Obtain the current values ​​of each target computing power indicator of the data center computer room;

[0012] Anomaly detection is performed on the current values ​​of each target energy consumption indicator and each target computing power indicator to determine the anomaly information.

[0013] According to a data center intelligent temperature control method provided by the present invention, the step of performing anomaly detection on the current values ​​of the target energy consumption indicators and the current values ​​of the target computing power indicators to determine the anomaly information includes:

[0014] Collect first data of each initial energy consumption index, second data of each initial computing power index, and third data of the temperature and humidity index of the data center computer room;

[0015] Based on the first data, the second data, and the third data, a correlation analysis method is used to determine the target energy consumption index and the target computing power index by comparing the initial energy consumption index and the initial computing power index with the temperature and humidity index.

[0016] According to the present invention, a smart temperature control method for a data center computer room is provided, wherein each target energy consumption index corresponds one-to-one with a multi-level PID controller, and the multi-level PID controllers operate in parallel; the method involves adjusting the current values ​​of each target energy consumption index in parallel based on the multi-level PID controllers and applying the target values ​​of each target energy consumption index to control the temperature and humidity of the data center computer room, including:

[0017] For any target energy consumption index, based on the PID controller corresponding to the target energy consumption index, the current value of the target energy consumption index is adjusted by applying the target value of the target energy consumption index, thereby controlling the temperature and humidity of the data center computer room.

[0018] According to the present invention, a method for intelligent temperature control in a data center computer room includes adjusting the current value of any target energy consumption index based on a PID controller corresponding to any target energy consumption index, using the target value of the target energy consumption index, to control the temperature and humidity of the data center computer room. Specifically, this method includes:

[0019] Based on the PID controller corresponding to any of the target energy consumption indicators, the target value of any of the target energy consumption indicators is applied to adjust the current value of any of the target energy consumption indicators, thereby generating a control signal;

[0020] Based on the control signal, the cooling system of the data center is controlled to control the temperature and humidity of the data center.

[0021] According to a data center intelligent temperature control method provided by the present invention, the method further includes, after performing anomaly detection on the current values ​​of the target energy consumption indicators to determine the anomaly information of the data center, the method further includes:

[0022] The abnormal information is sent to the operation and maintenance terminal.

[0023] The present invention also provides an intelligent temperature control system for data center computer rooms, comprising:

[0024] An anomaly information determination module is used to obtain the current values ​​of each target energy consumption indicator of the data center computer room, and to perform anomaly detection on the current values ​​of each target energy consumption indicator to determine the anomaly information of the data center computer room.

[0025] The target value determination module is used to determine the target values ​​of each target energy consumption index based on the abnormal information.

[0026] The parallel adjustment module is used to adjust the current values ​​of each target energy consumption index in parallel based on the target values ​​of each target energy consumption index using a multi-level PID controller, thereby controlling the temperature and humidity of the data center computer room.

[0027] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the intelligent temperature control method for data center computer rooms as described above.

[0028] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the intelligent temperature control method for data center computer rooms as described above.

[0029] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the intelligent temperature control method for data center computer rooms as described above.

[0030] This invention provides a data center intelligent temperature control method, system, electronic equipment, and storage medium. The method uses a multi-level PID controller to adjust the current values ​​of various target energy consumption indicators in parallel. Based on an iterative closed-loop feedback mechanism, it can automatically adjust and optimize the temperature and humidity of the data center, achieving fully automatic intelligent operation throughout the entire process and the engineering application of intelligent temperature control energy-saving algorithms in complex scenarios. It solves the problems of single PID controllers having limited adjustment measures, low efficiency, inaccurate temperature control, and long processing times. It exhibits strong adaptability to the complexity and nonlinearity of data center systems, can handle complex temperature and humidity control problems, and has good energy-saving effects. Furthermore, this parallel adjustment method can improve the response speed and control accuracy of the intelligent temperature control system, enabling it to better adapt to the working environment and load changes of data center racks. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, those skilled in the art can obtain other drawings based on the drawings described below without creative effort.

[0032] Figure 1 This is one of the flowcharts illustrating the intelligent temperature control method for data center computer rooms provided by the present invention;

[0033] Figure 2 This is the second flowchart of the intelligent temperature control method for data center computer rooms provided by the present invention;

[0034] Figure 3 This is a schematic diagram of the intelligent temperature control system for data center computer rooms provided by the present invention;

[0035] Figure 4 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0037] The terms "first" and "second" in the specification and claims of this invention may explicitly or implicitly include one or more of those features. In the description of the invention, unless otherwise stated, "a plurality of" means two or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0038] Existing intelligent temperature control algorithms primarily use machine learning or deep learning-based approaches to automatically fine-tune the parameters of PID control algorithms. For example, the fireworks algorithm from the field of neural algorithms is used to optimize the parameters of the PID control algorithm in real time, but this approach cannot eliminate the influence of abnormal data on parameter tuning. Another example is the generation of automatic temperature control models based on decision tree algorithms, but this requires a large amount of high-quality training data and a specific training strategy.

[0039] Existing solutions all employ a single PID controller. However, for complex controlled systems, the PID controller struggles to achieve ideal control performance and cannot guarantee the robustness of the controlled system due to the difficulty in accurately determining its model and parameters. Furthermore, the PID controller is a linear controller, capable of effectively controlling linear systems. If the system exhibits nonlinear characteristics, the PID controller cannot accurately compensate and adjust, leading to a decline in control performance.

[0040] Furthermore, PID controllers adjust the system based on the error between the desired setpoint and the actual output. In multidimensional systems, this error can be calculated along multiple dimensions, but a single PID controller can only handle the error in one dimension. Therefore, for multidimensional systems, a single PID controller cannot effectively handle errors across all dimensions.

[0041] The parameters (proportional, integral, and derivative) of a PID controller are tuned for a single output target. In a multidimensional system, each output target may require different PID parameters to achieve optimal control performance. However, a single PID controller can only use a fixed set of parameters, which may not meet the performance requirements of all output targets. Moreover, in multidimensional systems, building an accurate system model can be very difficult. Therefore, a single PID controller cannot fully realize its potential because it cannot account for all possible system behaviors and dynamics.

[0042] In summary, existing intelligent temperature control algorithms using PID controllers are not well-suited to the complexity and nonlinearity of data center systems, making it difficult to handle complex control problems involving multiple environmental indicators, thus resulting in limited energy-saving effects. Therefore, this invention provides an intelligent temperature control method for data center computer rooms.

[0043] Figure 1This is a flowchart illustrating an intelligent temperature control method for a data center computer room provided in an embodiment of the present invention, as shown below. Figure 1 As shown, the method includes:

[0044] S1, obtain the current values ​​of each target energy consumption indicator of the data center computer room, and perform anomaly detection on the current values ​​of each target energy consumption indicator to determine the abnormal information of the data center computer room;

[0045] S2, Based on the abnormal information, determine the target values ​​of each target energy consumption index;

[0046] S3, based on a multi-level PID controller, applies the target values ​​of each target energy consumption index to adjust the current values ​​of each target energy consumption index in parallel, and performs temperature and humidity control on the data center computer room.

[0047] Specifically, the intelligent temperature control method for data center computer rooms provided in this embodiment of the invention is executed by an intelligent temperature control system for data center computer rooms. This system can be configured in a computer, which can be a local computer or a cloud computer. The local computer can be a computer, tablet, etc., and no specific limitation is made here.

[0048] First, execute step S1. The target energy consumption indicators of the data center can be either initial energy consumption indicators that the intelligent temperature control system of the data center can read in real time from the sensors and then centrally collect through the data center infrastructure management platform (DCIM), such as the energy usage of the data center and the energy usage of the cooling system, or they can be obtained by screening the initial energy consumption indicators. No specific limitation is made here.

[0049] Initial energy consumption indicators can specifically include air conditioning operation indicators, daily electricity consumption indicators, and equipment power indicators. Air conditioning operation indicators are shown in Table 1, and can include operating status, return air temperature, supply air temperature, outdoor temperature, operating current, and compressor status.

[0050] Table 1 Air Conditioning Operation Indicators

[0051]

[0052] The daily electricity consumption indicators are shown in Table 2, which may include total daily electricity consumption, daily IT electricity consumption, daily refrigeration equipment electricity consumption, daily power supply and distribution system electricity consumption, and daily lighting and other electricity consumption.

[0053] Table 2 Daily Electricity Consumption Indicators

[0054]

[0055] The equipment power indicators are shown in Table 3, which may include total daily power consumption, daily IT power consumption, daily cooling equipment power consumption, daily power supply and distribution system power consumption, and daily lighting and other power consumption.

[0056] Table 3 Equipment Power Indicators

[0057]

[0058] After obtaining the current values ​​of each target energy consumption indicator in the data center, an anomaly detection model can be used to detect anomalies in the current values ​​of each target energy consumption indicator, thus identifying anomalies in the data center. Here, anomalies in the data center can include anomalies in temperature and humidity indicators. As shown in Table 4, temperature and humidity indicators can include rack temperature / humidity, equipment temperature / humidity, and aisle temperature / humidity within the data center. Rack temperature / humidity can include the temperature / relative humidity of the racks within the data center; equipment temperature / humidity can include the temperature / relative humidity of each piece of equipment within the data center; and aisle temperature / humidity can include the temperature / relative humidity of the aisles within the data center.

[0059] Table 4 Temperature and Humidity Indicators

[0060]

[0061] In this embodiment of the invention, each target energy consumption indicator can correspond to an anomaly detection model. For any target energy consumption indicator, the corresponding anomaly detection model can be constructed based on the CBLOF (Cluster-based Local OutlierFactor) algorithm, and can be trained using the historical values ​​of any target energy consumption indicator through unsupervised learning.

[0062] By inputting the current value of each target energy consumption indicator into the corresponding anomaly detection model, the model can determine whether the current value of the target energy consumption indicator is abnormal.

[0063] After performing anomaly detection on the current values ​​of all target energy consumption indicators, the abnormal information of the data center can be determined, that is, the abnormal equipment or temperature and humidity indicators in the data center can be located.

[0064] The CBLOF algorithm is a clustering algorithm for anomaly detection. It is based on the LOF (Local Outlier Factor) algorithm. It identifies outliers by dividing a dataset consisting of historical values ​​of any target energy consumption index into multiple clusters and calculating the local outlier factor for each data point.

[0065] The main steps of the CBLOF algorithm include:

[0066] 1) Data partitioning: Divide the dataset into multiple clusters. This step can use common clustering algorithms, such as K-means or DBSCAN, to partition the dataset.

[0067] 2) Calculate the Local Outlier Factor (LOF): For each data point, calculate its Local Outlier Factor (LOF). LOF is a metric that measures the degree of a data point's anomaly relative to its neighborhood. It is calculated by comparing the density of a data point with that of its neighboring data points. Specifically, for each data point, the CBLOF algorithm calculates its average distance to its neighboring data points and compares it to the average distance of its neighboring data points. If the average distance of a data point is large, it indicates that the density difference between the data point and its neighboring data points is significant, and it may be an outlier.

[0068] 3) Calculate the CBLOWF value: For each data point, calculate its CBLOWF value. The CBLOWF value is obtained by comparing the data point's LOF value with the average LOF value of its cluster. Generally, data points with high local outlier factors are considered outliers because they have a large density difference from their neighboring data points.

[0069] In the CBLOF algorithm, appropriate thresholds can be set for different target energy consumption indicators based on the experience of temperature and humidity control in data center computer rooms to determine the number and severity of outliers. If a data point has a high LOF value and the average LOF value of its cluster is low, then the data point is considered an outlier.

[0070] 4) Outlier labeling: Outliers are labeled based on the calculated CBLOF values. A set threshold can be used to determine which data points are considered outliers.

[0071] The CBLOF algorithm can take into account the density differences between data points and their neighborhoods, thus more accurately identifying outliers.

[0072] Identifying anomalies in data center computer rooms can solve the problems of traditional temperature control methods that lack consideration for the spatiotemporal characteristics of data and overly simplistic data analysis, which leads to the inability to filter noise and abnormal data.

[0073] Then, step S2 is executed to determine the target values ​​for each target energy consumption indicator using the abnormal information from the data center. If the abnormal information includes abnormal temperature and humidity indicators, the target values ​​for each target energy consumption indicator corresponding to the abnormal information can be determined based on the pre-defined correspondence between the temperature and humidity indicators and each target energy consumption indicator.

[0074] Finally, step S3 is executed, utilizing a multi-level PID controller to apply the target values ​​of each target energy consumption indicator. Temperature and humidity control of the data center is achieved by adjusting the current values ​​of each target energy consumption indicator in parallel. Here, the multi-level PID controller corresponds one-to-one with each target energy consumption indicator. For each target energy consumption indicator, the corresponding PID controller adjusts the current value of that indicator to its target value.

[0075] Each PID controller has proportional, integral, and derivative parameters, which are used to adjust the control signal output by the PID controller based on the error and rate of change between the current value and the target energy consumption index. The specific calculation method for PID is as follows:

[0076]

[0077] Where u(x) is the control signal curve output by the PID controller, Kp is the proportional parameter, e(t) is the difference curve between the current value and the target value of the target energy consumption index, that is, the curve of the deviation between the current value and the target value of the target energy consumption index changing over time, and T is the integral time. d For the differential time.

[0078] First, calculate the input deviation, which is the difference between the current value and the target value of the target energy consumption index. This can be obtained through simple subtraction. Based on the proportional parameter Kp, multiply the input deviation by a constant to obtain the proportional adjustment. Integrate the input deviation to obtain the integral adjustment over the integral time, thus adjusting the input deviation. The larger the integral time T, the slower the adjustment rate. Differentially calculate the input deviation to obtain the derivative adjustment, allowing for proactive adjustments to the control signal output by the PID controller. The larger the derivative time constant Td, the smaller the overshoot.

[0079] The proportional control is proportional to the input deviation and is used to quickly respond to changes in the target value. The integral control is proportional to the cumulative value of the input deviation and is used to eliminate steady-state error. The derivative control is proportional to the rate of change of the input deviation and is used to suppress overshoot and oscillation.

[0080] The proportional, integral, and derivative control variables are weighted and summed to obtain the control signal output by the PID control algorithm. This control signal is then applied to the cooling system of the data center to achieve intelligent temperature and humidity control.

[0081] When all PID controllers adjust the current value of the corresponding target energy consumption index to the target value of the corresponding target index, the temperature and humidity of the data center computer room meet the requirements, and the temperature and humidity control process of the data center computer room is completed.

[0082] Subsequently, the actual values ​​of each target energy consumption index are measured after PID feedback. After anomaly detection to identify abnormalities, the results of the next round are fed back to the multi-stage PID controller. This feedback information is used to update the error and perform the next control calculation. The actual values ​​of each target energy consumption index will be continuously monitored to obtain real-time feedback from the multi-stage PID system, achieving closed-loop feedback control of the multi-stage PID controller.

[0083] The intelligent temperature control method for data center computer rooms provided in this invention first acquires the current values ​​of various target energy consumption indicators of the data center computer room and performs anomaly detection on the current values ​​of each target energy consumption indicator to determine the abnormal information of the data center computer room; then, based on the abnormal information, it determines the target values ​​of each target energy consumption indicator; finally, based on a multi-level PID controller, it applies the target values ​​of each target energy consumption indicator to perform parallel adjustment of the current values ​​of each target energy consumption indicator, thereby controlling the temperature and humidity of the data center computer room. This method uses a multi-level PID controller to perform parallel adjustment of the current values ​​of each target energy consumption indicator. Based on an iterative closed-loop feedback mechanism, it can automatically adjust and optimize the temperature and humidity of the data center computer room, achieving fully automatic intelligent operation throughout the entire process and the engineering application of intelligent temperature control and energy-saving algorithms in complex scenarios. It solves the problems of single PID controller adjustment measures, low efficiency, inaccurate temperature control, and long processing time. It has strong adaptability to the complexity and nonlinearity of data center computer room systems and can handle complex temperature and humidity control problems, resulting in good energy-saving effects. Moreover, this parallel adjustment method can also improve the response speed and control accuracy of the intelligent temperature control system, enabling the intelligent temperature control system to better adapt to the working environment and load changes of the data center cabinets.

[0084] In data center server rooms, computing power data often changes before energy consumption data. That is, the physical environment changes only after the equipment's operating status changes. For example, when a server in the server room experiences excessive CPU utilization, its CPU heat output will not immediately increase, and the server rack will not detect any temperature abnormalities.

[0085] Therefore, PID controllers can predict temperature changes in advance based on computational data, rather than adjusting only after the temperature rises or falls. Traditional methods using energy consumption data for analysis imply a lag in the control process. If the temperature is too high, the PID controller may increase the cooling system's power, potentially leading to overcooling and excessively low temperatures. This adjustment process can cause the temperature-controlled machine to start and stop frequently, thus shortening its lifespan.

[0086] Furthermore, lag in the adjustment process can lead to energy waste. If the PID controller fails to adjust the temperature in a timely manner, the data center's cooling system may run for extended periods, consuming more energy. This not only increases the data center's operating costs but may also have negative environmental impacts.

[0087] Based on the above embodiments, the step of performing anomaly detection on the current values ​​of each target energy consumption indicator to determine the abnormal information of the data center computer room includes:

[0088] Obtain the current values ​​of each target computing power indicator of the data center computer room;

[0089] Anomaly detection is performed on the current values ​​of each target energy consumption indicator and each target computing power indicator to determine the anomaly information.

[0090] Specifically, when using the current values ​​of various target energy consumption indicators to determine anomalies in a data center, the current values ​​of each target computing power indicator can be obtained first. Then, an anomaly detection model is used to perform anomaly detection on the current values ​​of each target energy consumption indicator and each target computing power indicator, ultimately determining the anomaly in the data center. The target energy consumption indicators and target computing power indicators can be collectively referred to as computing power indicators, which comprehensively reflect the energy efficiency level of the data center.

[0091] In this embodiment of the invention, a corresponding anomaly detection model can be added for each target computing power indicator. For any target computing power indicator, the corresponding anomaly detection model can be constructed based on the CBLOF (Cluster-based Local OutlierFactor) algorithm, and can be trained by unsupervised learning using the historical values ​​of any target computing power indicator.

[0092] By inputting the current value of each target computing power indicator into the corresponding anomaly detection model, the model can determine whether the current value of the target computing power indicator is abnormal.

[0093] After performing anomaly detection on the current values ​​of all target energy consumption indicators and all target computing power indicators, the abnormal information of the data center can be determined.

[0094] The target computing power metrics for the data center can be either initial computing power metrics that can be collected by the data center's computing power monitoring tools, such as the computing performance, resource utilization, and resource load of data center servers, or metrics obtained by filtering the initial computing power metrics. No specific limitations are specified here. Computing power monitoring tools may include deployed cloud resource management systems or pre-probe programs.

[0095] The initial computing power metrics are shown in Table 5, which may include network bandwidth of the data center, server memory capacity, server memory utilization, server disk utilization, server hard disk capacity, server CPU utilization, server GPU utilization, virtual machine CPU utilization, and virtual machine memory utilization.

[0096] Table 5 Initial Computing Power Indicators

[0097]

[0098]

[0099] Furthermore, the abnormal information can also include abnormal information related to target computing power indicators. In this case, the target values ​​of each target energy consumption indicator corresponding to the abnormal information can be determined based on the pre-determined abnormal adjustment strategy corresponding to the target computing power indicator. For example, if the abnormal information is that the CPU utilization of server A is too high, the corresponding abnormal adjustment strategy could be to automatically reduce the target temperature of the cooling system to a specified value. Then, based on the pre-determined correspondence between the temperature indicator and each target energy consumption indicator, the target values ​​of each target energy consumption indicator corresponding to the abnormal information can be determined.

[0100] In this embodiment of the invention, in addition to considering various target energy consumption indicators, target computing power indicators are introduced to fully explore the core indicators reflecting the system's temperature control and energy-saving efficiency. This assists in constructing a hybrid intelligent temperature control and energy-saving system for data centers based on proactive early warning of computing power data and real-time feedback of energy consumption data. This system can detect anomalies in advance and achieve more precise temperature and humidity control while minimizing energy consumption. Moreover, by using anomaly detection models corresponding to each indicator to determine whether the current value of each indicator is abnormal, unsupervised automatic early warning and screening of time-series anomalies can be achieved, enabling timely detection and location of abnormal equipment or indicators.

[0101] In existing technologies, for high-dimensional data, PID control algorithms can only perform dimensionality reduction, transforming the high-dimensional data into low-dimensional data for processing and analysis. For example, each device in a data center may have different temperatures and humidity levels. These parameters can reflect the operating status of the data center and the performance of the equipment, but traditional methods require filtering out the most important features. Processing high-dimensional data is challenging; without suitable filtering techniques to extract useful control signals, the impact of noise and outliers is difficult to mitigate.

[0102] Based on the above embodiments, the step of performing anomaly detection on the current values ​​of each target energy consumption index and each target computing power index to determine the anomaly information includes:

[0103] Collect first data of each initial energy consumption index, second data of each initial computing power index, and third data of the temperature and humidity index of the data center computer room;

[0104] Based on the first data, the second data, and the third data, a correlation analysis method is used to determine the target energy consumption index and the target computing power index by comparing the initial energy consumption index and the initial computing power index with the temperature and humidity index.

[0105] Specifically, in this embodiment of the invention, the initial energy consumption indicators are shown in Tables 1-3. The temperature and humidity indicators of the data center computer room are shown in Table 4.

[0106] After collecting the first data of each initial energy consumption index, the second data of each initial computing power index, and the third data of temperature and humidity index in the data center, the first, second, and third data can be preprocessed, such as data cleaning and missing value completion.

[0107] Subsequently, the first, second, and third data can be used to perform correlation analysis on the initial energy consumption indicators, initial computing power indicators, and temperature and humidity indicators. For example, principal component analysis (PCA) or linear discriminant analysis (LDA) can be used to achieve this.

[0108] Several initial energy consumption indicators with high correlation to temperature and humidity were selected as target energy consumption indicators, and several initial computing power indicators with high correlation to temperature and humidity were selected as target computing power indicators. Thus, the introduction of correlation analysis can reduce the dimensionality of the indicators, thereby reducing the computational load.

[0109] Based on the above embodiments, each target energy consumption index corresponds one-to-one with the multi-level PID controller, and the multi-level PID controllers operate in parallel; the step of applying the target values ​​of each target energy consumption index to the multi-level PID controllers and adjusting the current values ​​of each target energy consumption index in parallel to control the temperature and humidity of the data center server room includes:

[0110] For any target energy consumption index, based on the PID controller corresponding to the target energy consumption index, the current value of the target energy consumption index is adjusted by applying the target value of the target energy consumption index, thereby controlling the temperature and humidity of the data center computer room.

[0111] Specifically, in this embodiment of the invention, each PID controller works in parallel, and each PID controller adjusts the current value of a target energy consumption index individually. After the PID controllers work in parallel, temperature and humidity control of the data center computer room can be achieved.

[0112] Based on the above embodiments, the step of using a PID controller corresponding to any target energy consumption index to adjust the current value of any target energy consumption index, and controlling the temperature and humidity of the data center server room, specifically includes:

[0113] Based on the PID controller corresponding to any of the target energy consumption indicators, the target value of any of the target energy consumption indicators is applied to adjust the current value of any of the target energy consumption indicators, thereby generating a control signal;

[0114] Based on the control signal, the cooling system of the data center is controlled to control the temperature and humidity of the data center.

[0115] Specifically, when a single PID controller adjusts the current value of a target energy consumption indicator, it first generates a control signal based on the difference between the current value and the target value. This control signal is then input to the cooling system of the data center server room to control the system and achieve temperature and humidity control. Here, the cooling system can control the temperature and humidity of the data center server room through air conditioning or other means.

[0116] Based on the above embodiments, the step of performing anomaly detection on the current values ​​of each target energy consumption indicator to determine the anomaly information of the data center computer room further includes:

[0117] The abnormal information is sent to the operation and maintenance terminal.

[0118] Specifically, after identifying anomalies in the data center, this information can be sent to the operations and maintenance (O&M) terminal. This O&M terminal refers to the device used by the O&M personnel, such as a smartphone, tablet, or laptop. The anomaly information can be sent via email, SMS, etc. Typically, the anomaly information is sent in the form of a message body, as shown in Table 6.

[0119] Table 6. Abnormal Information and Corresponding Message Bodies

[0120]

[0121] In this embodiment of the invention, when abnormal information is identified, it can be conveyed to the data center operation and maintenance personnel through various means such as email and SMS, enabling them to know about the problem in advance and intervene manually, thus realizing an abnormal early warning notification mechanism.

[0122] After abnormal information is sent to the data center maintenance personnel, if the monitoring system attempts to adjust automatically but fails to return to normal, or if the system adjustment strategy is determined to be incorrect, the data center maintenance personnel need to manually intervene to resolve the problem.

[0123] In summary, this invention provides a smart temperature control and energy-saving method for data center computer rooms that combines computing power data and AI anomaly detection technology, such as... Figure 2 As shown, the method includes:

[0124] Determine the computing power data of the data center, which includes the energy consumption data and computing power data of each terminal device in the data center.

[0125] The CBLOF algorithm is used to detect anomalies in computing power data and locate abnormal information. Here, an initial anomaly detection model is constructed using the CBLOF algorithm, and then trained using historical values ​​of the corresponding indicators to obtain an anomaly detection model, which is used to detect anomalies in the current values ​​of the corresponding indicators.

[0126] The multi-level PID controller adjusts the current values ​​of various target energy consumption indicators in the data center in parallel, and controls the temperature and humidity of the data center by sending control signals to the cooling system.

[0127] like Figure 3 As shown, based on the above embodiments, this embodiment of the invention provides an intelligent temperature control system for data center computer rooms, including:

[0128] The abnormal information determination module 31 is used to obtain the current values ​​of each target energy consumption index of the data center computer room, and to perform abnormal detection on the current values ​​of each target energy consumption index to determine the abnormal information of the data center computer room.

[0129] The target value determination module 32 is used to determine the target value of each target energy consumption index based on the abnormal information.

[0130] The parallel adjustment module 33 is used to adjust the current values ​​of each target energy consumption index in parallel based on the target values ​​of each target energy consumption index using a multi-level PID controller, and to control the temperature and humidity of the data center computer room.

[0131] Based on the above embodiments, the intelligent temperature control system for data center computer rooms provided in this embodiment of the invention, wherein the abnormal information determination module is specifically used for:

[0132] Obtain the current values ​​of each target computing power indicator of the data center computer room;

[0133] Anomaly detection is performed on the current values ​​of each target energy consumption indicator and each target computing power indicator to determine the anomaly information.

[0134] Based on the above embodiments, the intelligent temperature control system for data center computer rooms provided in this embodiment of the invention, wherein the abnormal information determination module is specifically used for:

[0135] Collect first data of each initial energy consumption index, second data of each initial computing power index, and third data of the temperature and humidity index of the data center computer room;

[0136] Based on the first data, the second data, and the third data, a correlation analysis method is used to determine the target energy consumption index and the target computing power index by comparing the initial energy consumption index and the initial computing power index with the temperature and humidity index.

[0137] Based on the above embodiments, the intelligent temperature control system for data center computer rooms provided in this embodiment of the invention, wherein the parallel adjustment module is specifically used for:

[0138] For any target energy consumption index, based on the PID controller corresponding to the target energy consumption index, the current value of the target energy consumption index is adjusted by applying the target value of the target energy consumption index, thereby controlling the temperature and humidity of the data center computer room.

[0139] Based on the above embodiments, the intelligent temperature control system for data center computer rooms provided in this embodiment of the invention, wherein the parallel adjustment module is specifically used for:

[0140] Based on the PID controller corresponding to any of the target energy consumption indicators, the target value of any of the target energy consumption indicators is applied to adjust the current value of any of the target energy consumption indicators, thereby generating a control signal;

[0141] Based on the control signal, the cooling system of the data center is controlled to control the temperature and humidity of the data center.

[0142] Based on the above embodiments, the intelligent temperature control system for data center computer rooms provided in this embodiment of the invention further includes an early warning module, used for:

[0143] The abnormal information is sent to the operation and maintenance terminal.

[0144] Specifically, the functions of each module in the intelligent temperature control system for data center computer rooms provided in this embodiment of the invention correspond one-to-one with the operation flow of each step in the above method-like embodiments, and the achieved effects are also the same. For details, please refer to the above embodiments, and this will not be repeated in this embodiment of the invention.

[0145] Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4 As shown, the electronic device may include a processor 410, a communications interface 420, a memory 430, and a communication bus 440, wherein the processor 410, the communications interface 420, and the memory 430 communicate with each other via the communication bus 440. The processor 410 can call logical instructions in the memory 430 to execute the intelligent temperature control method for data center computer rooms provided in the above embodiments.

[0146] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a 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 described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0147] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the intelligent temperature control method for data center computer rooms provided in the above embodiments.

[0148] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the intelligent temperature control method for data center computer rooms provided in the above embodiments.

[0149] The system embodiments described above are merely illustrative. 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 network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0150] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0151] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent temperature control in a data center computer room, characterized in that, include: Obtain the current values ​​of each target energy consumption indicator of the data center computer room, and perform anomaly detection on the current values ​​of each target energy consumption indicator to determine the abnormal information of the data center computer room; Based on the abnormal information and the pre-determined correspondence between temperature and humidity indicators and each target energy consumption indicator, the target value of each target energy consumption indicator is determined. Based on a multi-level PID controller, the target values ​​of each target energy consumption index are applied, and the current values ​​of each target energy consumption index are adjusted in parallel to control the temperature and humidity of the data center computer room. The step of detecting anomalies in the current values ​​of each target energy consumption indicator to determine the abnormal information of the data center computer room includes: Obtain the current values ​​of each target computing power index of the data center computer room; Anomaly detection is performed on the current values ​​of each target energy consumption indicator and each target computing power indicator to determine the anomaly information.

2. The intelligent temperature control method for data center computer rooms according to claim 1, characterized in that, The step of performing anomaly detection on the current values ​​of each target energy consumption indicator and each target computing power indicator to determine the anomaly information includes: Collect first data of each initial energy consumption index, second data of each initial computing power index, and third data of the temperature and humidity index of the data center computer room; Based on the first data, the second data, and the third data, a correlation analysis method is used to determine the target energy consumption index and the target computing power index by comparing the initial energy consumption index and the initial computing power index with the temperature and humidity index.

3. The intelligent temperature control method for data center computer rooms according to any one of claims 1-2, characterized in that, Each target energy consumption indicator corresponds one-to-one with a multi-level PID controller, and the multi-level PID controllers operate in parallel. The step of applying the target values ​​of each target energy consumption indicator to the multi-level PID controllers, and then adjusting the current values ​​of each target energy consumption indicator in parallel to control the temperature and humidity of the data center server room includes: For any target energy consumption index, based on the PID controller corresponding to the target energy consumption index, the current value of the target energy consumption index is adjusted by applying the target value of the target energy consumption index, thereby controlling the temperature and humidity of the data center computer room.

4. The intelligent temperature control method for data center computer rooms according to claim 3, characterized in that, The PID controller based on any target energy consumption index adjusts the current value of any target energy consumption index by applying the target value of the target energy consumption index, thereby controlling the temperature and humidity of the data center server room. Specifically, this includes: Based on the PID controller corresponding to any of the target energy consumption indicators, the target value of any of the target energy consumption indicators is applied to adjust the current value of any of the target energy consumption indicators, thereby generating a control signal; Based on the control signal, the cooling system of the data center is controlled to control the temperature and humidity of the data center.

5. The intelligent temperature control method for data center computer rooms according to any one of claims 1-2, characterized in that, The step of performing anomaly detection on the current values ​​of the target energy consumption indicators to determine the abnormal information of the data center computer room further includes: The abnormal information is sent to the operation and maintenance terminal.

6. A smart temperature control system for a data center computer room, characterized in that, include: An anomaly information determination module is used to obtain the current values ​​of each target energy consumption indicator of the data center computer room, and to perform anomaly detection on the current values ​​of each target energy consumption indicator to determine the anomaly information of the data center computer room. The target value determination module is used to determine the target value of each target energy consumption index based on the abnormal information and the pre-determined correspondence between temperature and humidity indexes and each target energy consumption index. The parallel adjustment module is used to adjust the current values ​​of each target energy consumption index in parallel based on the target values ​​of each target energy consumption index using a multi-level PID controller, and to control the temperature and humidity of the data center computer room. The anomaly information determination module is specifically used for: Obtain the current values ​​of each target computing power index of the data center computer room; Anomaly detection is performed on the current values ​​of each target energy consumption indicator and each target computing power indicator to determine the anomaly information.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the intelligent temperature control method for data center computer rooms as described in any one of claims 1-5.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the intelligent temperature control method for data center computer rooms as described in any one of claims 1-5.

9. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the intelligent temperature control method for data center computer rooms as described in any one of claims 1-5.

Images