Data center energy consumption and carbon emission prediction method and system based on ARIMA

Abstract

The invention relates to the technical field of electrical digital data processing, in particular to a data center energy consumption and carbon emission prediction method and system based on ARIMA, and the method comprises the steps: obtaining the equipment operation data of a data center, obtaining the abnormal characteristic value of the equipment operation through processing, and carrying out the risk early warning of the equipment of the data center; data center performance data and data processing task information are obtained, data center data processing task parameters and operation performance evaluation indexes are obtained through processing, and data center energy consumption evaluation indexes are obtained through comprehensive analysis; according to the data center energy consumption and carbon emission prediction method and system based on the ARIMA, the equipment operation abnormity, the operation performance and the data processing amount of the data center are comprehensively analyzed, the data center energy consumption and carbon emission prediction method and system based on the ARIMA are provided, and the data center energy consumption and carbon emission prediction method and system based on the ARIMA have the advantages that the data center energy consumption and carbon emission prediction efficiency is improved. The method can help to predict the energy consumption trend of the data center, and provides a basis for the updating of the ARIMA model.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical digital data processing, and in particular to a method and system for predicting energy consumption and carbon emissions of a data center based on ARIMA. Background Art

[0002] As an important infrastructure in the information age, data centers support the development of high-tech technologies such as cloud computing, big data, and artificial intelligence. With the rapid growth of the digital economy, the scale and number of data centers are also expanding, and their energy consumption and carbon emissions are increasingly receiving attention. Accurately predicting the energy consumption and carbon emission levels of power supply equipment is an important basis for data centers to optimize energy efficiency and formulate carbon emission reduction strategies.

[0003] For example, the invention patent with publication number: CN114860704A is a method and device for energy consumption monitoring and carbon emission accounting for data centers, which solves the problem that the existing technology cannot accurately recognize the power loss and invalid carbon emissions of each facility equipment in the facility room of the data center. The key points of its technical solution are: pre-processing the collected real-time energy consumption data to obtain energy consumption reports of each target device at equal time granularity; calculating the total carbon emissions of each target device according to the energy consumption report, and calculating the carbon emission efficiency of the IT equipment in the target device according to the total carbon emissions; generating a carbon emission report for each target device according to the total carbon emissions and the carbon emission efficiency, and inputting the energy consumption report and the carbon emission report into the deep neural network model for prediction to obtain prediction results.

[0004] For example, the invention patent with publication number: CN118819271A is a low-carbon control method, system and storage medium for data centers, which belongs to the field of data center technology, and specifically includes: determining the possibility of energy consumption control in different time periods and the energy consumption control time period according to historical processing conditions, determining that the time period can be energy-controlled by determining the energy consumption control possibility of different future adjacent time periods and the time interval with the time period, determining the minimum number of running servers in the time period according to the data processing fluctuation probability and energy consumption control possibility of the time period, determining the temperature control range of the matching data center according to the processing conditions of information data of different servers, and using the minimum number of running and temperature control range as constraints, and determining the energy consumption control method of the time period with the lowest carbon emission as the goal, thereby realizing low-carbon control of the data center.

[0005] However, in the process of implementing the technical solution of the invention in the embodiments of the present application, the present applicant discovered that the above technology has at least the following technical problems: since the data center requires a large number of servers and storage devices, these devices will generate a large amount of heat during operation, causing damage to the equipment, resulting in incomplete carbon emission data of the data center, thereby affecting the accuracy of the model prediction. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention provides a method and system for predicting energy consumption and carbon emissions of data centers based on ARIMA, which can effectively solve the problems involved in the above-mentioned background technology.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: The first aspect of the present invention provides a data center energy consumption and carbon emission prediction method based on ARIMA, including: obtaining the equipment operation data of the data center, obtaining the abnormal operation characteristic value of the equipment through processing, and issuing risk warnings for the data center equipment according to the abnormal operation characteristic value of the equipment.

[0008] The data center performance data and data processing task information are obtained, and the data center data processing task parameters and the data center operation performance evaluation index are obtained through processing. According to the abnormal operation characteristic values ​​of the equipment, the data center data processing task parameters and the data center operation performance evaluation index, a comprehensive analysis is performed to obtain the data center energy consumption evaluation index.

[0009] According to the data center energy consumption evaluation indicators, it is judged whether the ARIMA model can directly predict the data center energy consumption and carbon emissions, and the ARIMA model is fed back in real time based on the judgment results.

[0010] As a further method, the abnormal equipment operation characteristic value is obtained through processing, and the specific process is: the equipment operation data of the data center includes equipment operation environment data and equipment operation status data.

[0011] According to the equipment operating environment data, the operating environment impact characteristic value is processed.

[0012] The device operation status data includes: processor utilization, storage device utilization, cooling system pressure and network bandwidth utilization.

[0013] The critical standard value of processor utilization, the critical standard value of storage device utilization, the reference standard value of cooling system pressure, the allowable deviation value of cooling system pressure and the critical standard value of network bandwidth utilization are extracted from the data center information library.

[0014] According to the equipment operation status data and the operating environment impact characteristic values, comprehensive processing is performed to obtain the equipment operation abnormality characteristic values, which are used to quantitatively evaluate the degree of abnormal operation risk of data center equipment and provide a basis for equipment risk warning.

[0015] As a further method, the processing obtains the operating environment impact characteristic value, and the specific process is: the equipment operating environment data includes: the average temperature, average humidity and average noise level of the data center operating area and the power supply voltage.

[0016] The regional temperature standard value, regional humidity standard value, regional temperature allowable deviation value, regional humidity allowable deviation value, regional noise level critical value, rated power supply voltage and allowable deviation power supply voltage are extracted from the data center information database.

[0017] The operating environment impact characteristic value is processed according to the equipment operating environment data. The operating environment impact characteristic value is used to quantitatively evaluate the interference degree of the equipment operating environment on the equipment operation, and provide a basis for evaluating the abnormal operation of the equipment.

[0018] As a further method, the risk warning of data center equipment is carried out according to the abnormal operation characteristic value of the equipment. The specific process is: extracting the first threshold and the second threshold of the abnormal operation characteristic of the equipment from the data center information library, comparing the abnormal operation characteristic value of the equipment with the first threshold and the second threshold of the abnormal operation characteristic of the equipment; if the abnormal operation characteristic value of the equipment is higher than or equal to the first threshold of the abnormal operation characteristic of the equipment, the equipment operation status is evaluated as abnormal, the early warning platform issues a red alarm, and notifies the staff to perform abnormal analysis and equipment maintenance; if the abnormal operation characteristic value of the equipment is lower than the first threshold of the abnormal operation characteristic of the equipment and higher than the second threshold of the abnormal operation characteristic of the equipment, the early warning platform issues an orange alarm and automatically switches the power supply; if the abnormal operation characteristic value of the equipment is lower than or equal to the second threshold of the abnormal operation characteristic of the equipment, the equipment operation status is evaluated as normal, and the early warning platform outputs the result and displays a green indicator light.

[0019] As a further method, the processing obtains data center data processing task parameters and a data center operation performance evaluation index, including: the data processing task information, including: the amount of tasks processed by the data center within a preset monitoring period, the number of user visits and the number of user operations of the data center within the preset monitoring period.

[0020] The critical task volume, critical user access volume and critical user operation times are extracted from the data center information database.

[0021] According to the data processing task information, comprehensive processing is performed to obtain the data center data processing task parameters, which are used to quantitatively evaluate the busyness of the data center data processing and provide a basis for analyzing the energy consumption of the data center.

[0022] As a further method, the processing obtains data center data processing task parameters and a data center operation performance evaluation index, and also includes: the data center performance data, including: data throughput of the data center, response time of the data center, processor utilization, network bandwidth utilization and data processing speed of the data center.

[0023] The data throughput critical value, response time critical value, processor utilization reference standard value, processor utilization allowable deviation value, network bandwidth utilization reference standard value, network bandwidth utilization allowable deviation value and critical data processing speed are extracted from the data center information library.

[0024] According to the data center performance data, a data center operation performance evaluation index is obtained by processing. The data center operation performance evaluation index is used to quantitatively evaluate the operation performance of the data center and provide a basis for analyzing the data processing energy efficiency of the data center.

[0025] As a further method, the comprehensive analysis obtains data center energy consumption evaluation indicators. The specific analysis process is: based on the abnormal characteristic values ​​of equipment operation, data center data processing task parameters and data center operation performance evaluation index, comprehensive analysis is performed to obtain data center energy consumption evaluation indicators. The data center energy consumption evaluation indicators are used to quantitatively evaluate the energy consumption of the data center and provide a basis for predicting the energy consumption and carbon emissions of the data center.

[0026] As a further method, the data center energy consumption evaluation index is used to determine whether the ARIMA model can directly predict the data center energy consumption carbon emissions. The specific process is: extracting the data center energy consumption reference standard index range from the data center information library, comparing the data center energy consumption evaluation index with the data center energy consumption reference standard index range, if the data center energy consumption evaluation index is within the data center energy consumption reference standard index range, then it is determined that the ARIMA model can directly predict the data center energy consumption carbon emissions; if the data center energy consumption evaluation index is higher than or equal to the maximum value of the data center energy consumption reference standard index range, or the data center energy consumption evaluation index is lower than or equal to the minimum value of the data center energy consumption reference standard index range, then it is determined that the ARIMA model cannot directly predict the data center energy consumption carbon emissions, and the ARIMA model is updated in real time according to the judgment result.

[0027] As a further method, the ARIMA model is fed back in real time according to the judgment result, and the specific feedback process is: obtaining the energy consumption prediction value of the ARIMA model.

[0028] If the data center energy consumption evaluation index is higher than or equal to the maximum value of the data center energy consumption reference standard index range, the first correction value of the data center energy consumption and the maximum threshold of the data center energy consumption evaluation index are obtained by matching the abnormal characteristic value of equipment operation, and the ARIMA model energy consumption prediction value is added to the first correction value of the data center energy consumption to obtain the final energy consumption prediction value. The final energy consumption prediction value is input into the ARIMA model to obtain the final energy consumption carbon emission prediction value.

[0029] If the data center energy consumption evaluation index is lower than or equal to the minimum value of the data center energy consumption reference standard index range, the second correction value of the data center energy consumption and the minimum threshold of the data center energy consumption evaluation index are obtained according to the matching of the data center operation performance evaluation index. The ARIMA model energy consumption prediction value is subtracted from the second correction value of the data center energy consumption to obtain the final energy consumption prediction value. The final energy consumption prediction value is input into the ARIMA model to obtain the final energy consumption carbon emission prediction value.

[0030] A second aspect of the present invention provides an ARIMA-based data center energy consumption and carbon emission prediction system, comprising: a data center equipment analysis module, a data center energy consumption analysis module and a comprehensive feedback module.

[0031] The data center equipment analysis module is used to obtain the equipment operation data of the data center, obtain the abnormal equipment operation characteristic value after processing, and issue risk warnings to the data center equipment based on the abnormal equipment operation characteristic value.

[0032] The data center energy consumption analysis module is used to obtain data center performance data and data processing task information, and process the data center data processing task parameters and the data center operation performance evaluation index. According to the abnormal operation characteristic values ​​of the equipment, the data center data processing task parameters and the data center operation performance evaluation index, a comprehensive analysis is performed to obtain the data center energy consumption evaluation index.

[0033] The comprehensive feedback module is used to determine whether the ARIMA model can directly predict the energy consumption and carbon emissions of the data center based on the data center energy consumption evaluation indicators, and to provide real-time feedback to the ARIMA model based on the judgment results.

[0034] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects: (1) The present invention provides a data center energy consumption and carbon emission prediction method and system based on ARIMA, which comprehensively analyzes the equipment operation anomalies, data processing volume and operation performance of the data center, can help predict the energy consumption trend of the data center, and provide a basis for updating the ARIMA model. At the same time, by introducing the first correction value or the second correction value to provide feedback to the ARIMA model, the model can more accurately predict the energy consumption of the data center.

[0035] (2) The present invention can detect indicators that deviate from the normal operating range at an early stage by comprehensively monitoring and analyzing the data center processor utilization, storage device utilization, cooling system pressure, network bandwidth utilization and operating environment, thereby early warning of potential equipment operation anomalies. It can also help identify performance bottlenecks, thereby guiding the optimal allocation of data center resources. It can also help predict resource requirements and provide data support for data center expansion or upgrades.

[0036] (3) The present invention can set performance benchmarks and goals for a data center by comprehensively analyzing the data processing speed, data throughput, response time, processor utilization, and network bandwidth utilization of the data center. This helps determine whether the data center meets current business needs and provides guidance for future expansion and upgrades. It also helps identify imbalances in resource allocation and potential bottlenecks. By optimizing resource allocation, the overall efficiency and performance of the data center can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0038] Figure 1 It is a schematic diagram of the method flow of the present invention.

[0039] Figure 2 It is a schematic diagram of system module connection of the present invention. DETAILED DESCRIPTION

[0040] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0041] Reference Figure 1 As shown, the first aspect of the present invention provides a data center energy consumption and carbon emission prediction method based on ARIMA, including: acquiring equipment operation data of the data center, obtaining abnormal equipment operation characteristic values ​​through processing, and performing risk warning for data center equipment according to the abnormal equipment operation characteristic values.

[0042] It should be explained that data center energy consumption carbon emissions refer to the carbon dioxide emissions generated by energy consumption during the operation of data centers. Data centers are the core part of information technology infrastructure. They support key functions such as cloud services, online computing, data storage and transmission. Since data centers require a large number of servers and storage devices, these devices will generate a lot of heat when running, so they need to be constantly cooled, which leads to higher energy consumption.

[0043] It needs to be explained that the ARIMA model, full name Autoregressive Integrated Moving Average Model, is a statistical model widely used in time series data analysis and prediction. It consists of three parts: autoregressive model (AR), difference operation (I) and moving average model (MA). The basic idea of ​​the ARIMA model is to capture the trend and periodicity of the time series by analyzing historical data, so as to predict future values.

[0044] Specifically, the abnormal characteristic value of equipment operation is obtained through processing, and the specific process is: the equipment operation data of the data center includes equipment operation environment data and equipment operation status data.

[0045] According to the equipment operating environment data, the operating environment impact characteristic value is processed.

[0046] The device operation status data includes: processor utilization, storage device utilization, cooling system pressure and network bandwidth utilization.

[0047] It should be added that performance monitoring software installed in the data center, such as Nagios, Zabbix, SolarWinds, Datadog, etc., can be used to query these tools to obtain detailed processor usage, including processor utilization. Storage system management software can be used to monitor storage utilization, I / O performance, etc. in real time. Data center infrastructure management software can provide detailed monitoring data of the cooling system, including cooling system pressure. Network traffic and bandwidth utilization can be viewed through the network device management interface.

[0048] It should be added that the cooling system pressure refers to the pressure of the refrigerant gas in the exhaust pipe at the compressor outlet, which is also approximately equal to the pressure of the refrigerant gas condensing into liquid in the condenser. By detecting the exhaust pressure, the condensation temperature of the refrigeration system can be obtained, thereby understanding the operating status of the refrigeration system.

[0049] The critical standard value of processor utilization, the critical standard value of storage device utilization, the reference standard value of cooling system pressure, the allowable deviation value of cooling system pressure and the critical standard value of network bandwidth utilization are extracted from the data center information library.

[0050] According to the equipment operation status data and the operating environment impact characteristic values, comprehensive processing is performed to obtain the equipment operation abnormality characteristic values, which are used to quantitatively evaluate the degree of abnormal operation risk of data center equipment and provide a basis for equipment risk warning.

[0051] In a specific embodiment, the numerical expression of the abnormal operation characteristic value of the equipment is: ; In the formula, represents the abnormal characteristic value of equipment operation, e represents the natural constant, It represents the utilization rate of the nth processor in the data center at the tth time monitoring point, where n represents the number of each processor in the data center. , m represents the total number of processors in the data center, t represents the number of each time monitoring point, , s represents the total number of time monitoring points, represents the utilization rate of the rth storage device in the data center at the tth time monitoring point, r represents the number of each storage device in the data center, , g represents the total number of storage devices in the data center, represents the pressure of the data center cooling system at the monitoring point at time t, represents the utilization rate of the data center network bandwidth at the tth time monitoring point, Indicates the critical value of processor utilization. Indicates the critical value of storage device utilization. Indicates the reference standard value of the cooling system pressure. Indicates the allowable deviation of the cooling system pressure. Indicates the critical value of network bandwidth utilization. Indicates the operating environment affecting the characteristic value, Indicates the device operation abnormality characteristic factor corresponding to the preset processor utilization. Indicates the device operation abnormality characteristic factor corresponding to the preset storage device utilization rate. Indicates the equipment operation abnormality characteristic factor corresponding to the preset cooling system pressure, Indicates the device operation abnormality characteristic factor corresponding to the preset network bandwidth utilization. Indicates the preset operating environment impact characteristic factors.

[0052] It needs to be explained that when the processor utilization is higher, the storage device utilization is higher, the cooling system pressure is greater than the reference standard value, the network bandwidth utilization is higher, and the operating environment impact characteristic value is greater, the corresponding equipment operation abnormality characteristic value will be greater, indicating that the risk of abnormal operation of data center equipment is higher.

[0053] It should be explained that in this embodiment Indicates the device operation abnormality characteristic factor corresponding to the preset processor utilization. Indicates the device operation abnormality characteristic factor corresponding to the preset storage device utilization rate. Indicates the equipment operation abnormality characteristic factor corresponding to the preset cooling system pressure, Indicates the device operation abnormality characteristic factor corresponding to the preset network bandwidth utilization. The preset operating environment impact characteristic factors respectively represent the numerical values ​​of the influence degree of the data center processor utilization, data center storage device utilization, data center cooling system pressure, data center network bandwidth utilization and the unit value of the operating environment impact characteristic value on the abnormal operation of the equipment. When used, they can be directly obtained from the data center information library, and their corresponding relationship can be a pre-set mapping relationship. For example, the data center processor utilization, data center storage device utilization, data center cooling system pressure, data center network bandwidth utilization and operating environment impact characteristic value are respectively mapped with the equipment operation abnormality characteristic factors corresponding to the data center processor utilization, data center storage device utilization, data center cooling system pressure, data center network bandwidth utilization and operating environment impact characteristic value preset in the data center information library to form a mapping set, and the real-time data center processor utilization, data center storage device utilization, data center cooling system pressure, data center network bandwidth utilization and operating environment impact characteristic value are input into the mapping set to obtain the corresponding equipment operation abnormality characteristic factor, wherein the mapping relationship can be one-to-one or many-to-one. The above characteristic factors are all extracted from the data center information library, and the value range is between 0 and 1.

[0054] It is important to explain that high processor utilization usually means an increase in data processing demand, which may result in an increase in read and write operations to storage devices, thereby increasing storage device utilization. Increased processor utilization causes the processor to generate more heat, which increases the pressure on the data center cooling system, because the cooling system needs to remove heat more effectively to keep the equipment within a safe operating temperature range. If the cooling system cannot keep up with the heat generation, it may cause the data center temperature to rise, which in turn affects the performance and life of the equipment. Storage devices also generate heat when they are frequently read and written, so high storage device utilization may cause increased pressure on the cooling system. The efficiency of the cooling system directly affects the reliability and performance of the storage device, because overheating may cause data corruption or device failure. High network bandwidth utilization means that the network traffic in the data center is large, which usually requires the processor to process more network packets, thereby increasing processor utilization. High network bandwidth utilization means that there is a large amount of data transmitted on the network, which may involve read or write operations on storage devices, thereby increasing storage device utilization. The various systems in the data center are interrelated, and changes in one system may have an impact on other systems. Effective data center management requires monitoring and optimizing the relationship between these resources to ensure stable operation and efficient performance of the data center. For example, load balancing, resource allocation strategies, and proper cooling measures can reduce the impact of a system overload on the performance of the entire data center.

[0055] It should be explained that this embodiment can detect indicators that deviate from the normal operating range early by comprehensively monitoring and analyzing the processor utilization, storage device utilization, cooling system pressure, network bandwidth utilization and operating environment of the data center, thereby warning of potential equipment operation abnormalities. For example, a sudden increase in processor utilization may indicate an impending performance bottleneck or resource competition. When an abnormality occurs in a device, it can help the operation and maintenance team diagnose the cause of the failure. For example, if the storage device utilization rate increases abnormally and is accompanied by an increase in network bandwidth utilization, it may indicate a large amount of data transmission demand, which may be an abnormal backup operation or data leakage. It can also help identify performance bottlenecks, thereby guiding the optimal allocation of data center resources. For example, if it is found that the network bandwidth utilization rate is often close to the peak value, it may be necessary to increase the bandwidth or optimize the network architecture. It also helps to predict resource requirements and provide data support for data center expansion or upgrade. For example, if the storage device utilization rate continues to be high, it may be necessary to increase the storage capacity. Timely identification of abnormal changes in network bandwidth utilization can respond quickly and reduce security risks. At the same time, long-term trend analysis of the above parameters can help predict future performance requirements and potential problems, thereby achieving more proactive operation and maintenance management.

[0056] Further, the operating environment impact characteristic value is obtained by processing, and the specific process is: the equipment operating environment data includes: the average temperature, average humidity and average noise level and power supply voltage of the data center operating area.

[0057] It should be added that the power supply voltage refers to the power supply voltage of the power grid where the data center is located.

[0058] It should be added that several environmental monitoring points are deployed in the data center operation area, and the temperature, humidity and noise levels of each environmental monitoring point are collected in real time through temperature sensors, humidity sensors and sound level meters, and the average value is taken to obtain the average temperature, average humidity and average noise level of the data center operation area. The power supply voltage can be monitored in real time through the power management system of the data center.

[0059] The regional temperature standard value, regional humidity standard value, regional temperature allowable deviation value, regional humidity allowable deviation value, regional noise level critical value, rated power supply voltage and allowable deviation power supply voltage are extracted from the data center information database.

[0060] The operating environment impact characteristic value is processed according to the equipment operating environment data. The operating environment impact characteristic value is used to quantitatively evaluate the interference degree of the equipment operating environment on the equipment operation, and provide a basis for evaluating the abnormal operation of the equipment.

[0061] In a specific embodiment, the numerical expression of the operating environment impact characteristic value is: ; In the formula, Indicates the operating environment affecting the characteristic value, Indicates the average temperature of the data center operating area. Indicates the average humidity in the data center operating area. Represents the average noise level in the area where the data center operates. Indicates the power supply voltage, Indicates the standard value of regional temperature, Indicates the standard value of regional humidity. Indicates the allowable deviation value of regional temperature. Indicates the allowable deviation value of regional humidity. represents the critical value of regional noise level, Indicates the rated power supply voltage, Indicates the allowable deviation power supply voltage, Indicates the operating environment influence characteristic factor corresponding to temperature, Indicates the operating environment impact characteristic factor corresponding to humidity, Indicates the operating environment impact characteristic factor corresponding to the noise level, Indicates the operating environment impact characteristic factor corresponding to the power supply voltage.

[0062] It needs to be explained that when the absolute difference between the regional temperature and the standard value is greater, the absolute difference between the regional humidity and the standard value is greater, the regional noise level is greater, and the absolute difference between the power supply voltage and the rated voltage is greater, the corresponding operating environment impact characteristic value will be greater, indicating that the degree of interference of the equipment operating environment on the equipment operation is greater.

[0063] It should be explained that in this embodiment Indicates the operating environment influence characteristic factor corresponding to temperature, Indicates the operating environment impact characteristic factor corresponding to humidity, Indicates the operating environment impact characteristic factor corresponding to the noise level, The operating environment impact characteristic factors corresponding to the power supply voltage respectively represent the numerical values ​​of the interference degree of the regional average temperature, regional average humidity, regional average noise level and power supply voltage unit value on the operation of the equipment. When used, they can be directly obtained from the data center information library, and their corresponding relationship can be a pre-set mapping relationship. For example, the regional average temperature, regional average humidity, regional average noise level and power supply voltage respectively form a mapping set with the operating environment impact characteristic factors corresponding to the regional average temperature, regional average humidity, regional average noise level and power supply voltage preset in the data center information library, and the real-time regional average temperature, regional average humidity, regional average noise level and power supply voltage are input into the mapping set to obtain the corresponding operating environment impact characteristic factors, wherein the mapping relationship can be one-to-one or many-to-one. The above-mentioned impact characteristic factors are all extracted from the data center information library, and the value range is between 0 and 1.

[0064] It should be explained that electronic equipment operating in a high temperature environment will accelerate aging, causing circuit board deformation, solder joint shedding, and electronic component performance degradation, which may cause equipment failure or system crash. At the same time, too low a temperature may cause the elasticity of certain materials to decrease, affecting the mechanical properties of the equipment and even causing circuit failure. High humidity can cause condensation of water vapor inside the equipment, causing circuit shorts, corroding metal parts, reducing insulation performance, and increasing the risk of equipment failure. Low humidity can increase the generation of static electricity, and static discharge may damage sensitive electronic components and affect the normal operation of the equipment. Abnormal noise levels may be a precursor to equipment failure, such as fan damage, hard drive failure, or wear and tear of other mechanical parts. Power supply voltage fluctuations can cause equipment to restart, data loss, or hardware damage, especially during power conversion and regulation. Excessive voltage can damage electronic components and shorten the life of the equipment. Low voltage may cause the device to fail to start or run unstably, affecting data processing capabilities.

[0065] It should be explained that an increase in temperature usually leads to an increase in the saturated water vapor pressure of the air, which reduces the relative humidity, and vice versa. Therefore, in a data center, the operation of the cooling system affects both temperature and humidity. During the cooling process, if the humidity is not properly controlled, overcooling and condensation may occur, which can damage the equipment. The cooling system of the data center will generate noise during operation. Generally, the higher the temperature, the more frequently the cooling system works, and the noise level may also increase. Conversely, if the temperature of the data center is effectively controlled, the noise level of the cooling system will also decrease accordingly. Humidity control equipment will also generate noise during operation. Adjustment of humidity may affect the operating frequency and noise level of these devices. Too high or too low humidity may require humidity control equipment to operate more frequently, which may increase the noise level. The stability of the power supply voltage directly affects the operation of the cooling system, humidity control equipment and other IT equipment. Voltage fluctuations may cause these devices to work improperly, which in turn affects the temperature, humidity and noise level. If the power supply voltage is unstable, it may cause the cooling system to be less efficient, which in turn affects the temperature control of the data center.

[0066] It should be explained that this embodiment can predict possible equipment failures by monitoring and analyzing the interference degree of regional average temperature, regional average humidity, regional average noise level and power supply voltage on equipment operation. For example, an abnormal increase in temperature may be a sign of poor heat dissipation of the equipment, and early intervention can avoid equipment failure. Humidity fluctuations may cause the circuit board to get damp, and regular analysis of humidity data helps to take preventive measures before the problem occurs. When the equipment is operating abnormally, the data of environmental parameters can help diagnose the cause of the problem. For example, if the power supply voltage fluctuation coincides with the equipment failure time, it may be caused by a power supply problem. Comprehensive analysis of environmental parameters helps to optimize the operating environment of the data center and reduce interference with the equipment. For example, adjusting the cooling system to maintain optimal temperature and humidity reduces the performance degradation of the equipment due to environmental factors. Analyzing environmental parameters can also help the data center use energy more efficiently. For example, by optimizing the operation of the cooling system, energy consumption can be reduced while maintaining a suitable temperature. At the same time, a stable power supply voltage can ensure that the equipment operates at the highest efficiency and reduce energy waste. By maintaining suitable environmental parameters, the service life of the equipment can be extended. Reducing fluctuations in temperature and humidity and avoiding equipment from operating in extreme environments can slow down equipment aging. Stable environmental parameters can also improve the overall reliability of the data center. When equipment is running in the best environment, the failure rate is reduced and the system stability is improved. Reducing equipment failures and downtime can save the cost of repairing and replacing equipment.

[0067] Specifically, risk warnings are issued to data center equipment based on abnormal equipment operation characteristic values. The specific process is as follows: extract the first threshold and the second threshold of the abnormal equipment operation characteristic from the data center information database, compare the abnormal equipment operation characteristic value with the first threshold and the second threshold of the abnormal equipment operation characteristic, if the abnormal equipment operation characteristic value is higher than or equal to the first threshold of the abnormal equipment operation characteristic, then the equipment operation status is evaluated as abnormal, the early warning platform issues a red alarm, and notifies the staff to conduct abnormal analysis and equipment maintenance, if the abnormal equipment operation characteristic value is lower than the first threshold of the abnormal equipment operation characteristic and higher than the second threshold of the abnormal equipment operation characteristic, the early warning platform issues an orange alarm and automatically switches the power supply, if the abnormal equipment operation characteristic value is lower than or equal to the second threshold of the abnormal equipment operation characteristic, then the equipment operation status is evaluated as normal, the early warning platform outputs the result and displays a green indicator light.

[0068] The data center performance data and data processing task information are obtained, and the data center data processing task parameters and the data center operation performance evaluation index are obtained through processing. According to the abnormal operation characteristic values ​​of the equipment, the data center data processing task parameters and the data center operation performance evaluation index, a comprehensive analysis is performed to obtain the data center energy consumption evaluation index.

[0069] Specifically, the data center data processing task parameters are processed, and the specific process is: the data processing task information includes: the task volume processed by the data center within a preset monitoring period, the user access volume and the number of user operations of the data center within the preset monitoring period.

[0070] The critical task volume, critical user access volume and critical user operation times are extracted from the data center information database.

[0071] According to the data processing task information, comprehensive processing is performed to obtain the data center data processing task parameters, which are used to quantitatively evaluate the busyness of the data center data processing and provide a basis for analyzing the energy consumption of the data center.

[0072] In a specific embodiment, the numerical expression of the data center data processing task parameter is: ; In the formula, Indicates the data processing task parameters of the data center, Indicates the amount of tasks processed by the data center within the preset monitoring period. Indicates the number of user visits to the data center within the preset monitoring period. Indicates the number of user operations in the data center within the preset monitoring period. represents the critical task volume, Indicates the critical user access volume, Indicates the critical number of user operations. Indicates the influencing factor of the data center data processing task parameters corresponding to the preset processing task volume, Indicates the impact factor of the data center data processing task parameter corresponding to the preset user access volume, Indicates the impact factor of the data center data processing task parameter corresponding to the preset number of user operations.

[0073] It needs to be explained that when the data center processes a larger amount of tasks within the preset monitoring period, the larger the user visits, and the more user operations, the corresponding data center data processing task parameters will be larger, indicating that the data center data processes a larger amount of tasks within the preset monitoring period.

[0074] It should be explained that in this embodiment Indicates the influencing factor of the data center data processing task parameters corresponding to the preset processing task volume, Indicates the impact factor of the data center data processing task parameter corresponding to the preset user access volume, The data center data processing task parameter influencing factors corresponding to the preset number of user operations respectively represent the numerical values ​​of the amount of tasks, user visits and number of user operations processed by the data center data within the preset monitoring period on the data center data processing energy efficiency. When used, they can be directly obtained from the data center information library, and the corresponding relationship can be a pre-set mapping relationship. For example, the amount of tasks, user visits and number of user operations processed by the data center data within the preset monitoring period are respectively mapped with the data center data processing task parameter influencing factors corresponding to the amount of tasks, user visits and number of user operations processed by the data center data within the preset monitoring period preset in the data center information library to form a mapping set. The amount of tasks, user visits and number of user operations processed by the real-time data center data within the preset monitoring period are input into the mapping set to obtain the corresponding data center data processing task parameter influencing factors, wherein the mapping relationship can be one-to-one or many-to-one. The above-mentioned influencing factors are all extracted from the data center information library, and the value range is between 0 and 1.

[0075] It should be explained that the amount of tasks processed by the data center during the preset monitoring period directly determines the total demand for data processing. The larger the amount of tasks, the more data processing resources and time are required. The amount of tasks, user visits, and user operation times reflect the amount of data and requests that the data center needs to process.

[0076] Further, the data center operation performance evaluation index is obtained by processing, and the specific process is: the data center performance data includes: data throughput of the data center, response time of the data center, processor utilization, network bandwidth utilization and data processing speed of the data center.

[0077] It needs to be explained that data processing speed refers to the time required from data input to output processing results.

[0078] The data throughput critical value, response time critical value, processor utilization reference standard value, processor utilization allowable deviation value, network bandwidth utilization reference standard value, network bandwidth utilization allowable deviation value and critical data processing speed are extracted from the data center information library.

[0079] According to the data center performance data, a data center operation performance evaluation index is obtained by processing. The data center operation performance evaluation index is used to quantitatively evaluate the operation performance of the data center and provide a basis for analyzing the data processing energy efficiency of the data center.

[0080] In a specific embodiment, the numerical expression of the data center operation performance evaluation index is: ; In the formula, represents the data center operation performance evaluation index, represents the data throughput of the data center at time t, represents the response time of the data center at time t, represents the processor utilization of the data center at time t, represents the network bandwidth utilization of the data center at time t, represents the data processing speed of the data center at time t, represents the data throughput critical value, Indicates the critical value of response time, Indicates the processor utilization reference standard value. Indicates the allowable deviation value of processor utilization. Indicates the reference standard value of network bandwidth utilization. Indicates the allowable deviation value of network bandwidth utilization. represents the critical data processing speed, Indicates the impact factor of data center operation performance corresponding to the preset data processing speed, Indicates the impact factor of data center operation performance corresponding to the preset data throughput, Indicates the impact factor of data center operation performance corresponding to the preset response time, Indicates the impact factor of data center operation performance corresponding to the preset processor utilization, Indicates the data center operation performance impact factor corresponding to the preset network bandwidth utilization.

[0081] It should be noted that when the data throughput of the data center is greater, the data processing speed is faster, the response time is shorter, the absolute difference between the processor utilization and the reference standard value is smaller, and the absolute difference between the network bandwidth utilization and the reference standard value is smaller, the corresponding data center operation performance evaluation index is greater, indicating that the operation performance of the data center is better.

[0082] It should be explained that in this embodiment Indicates the impact factor of data center operation performance corresponding to the preset data processing speed, Indicates the impact factor of data center operation performance corresponding to the preset data throughput, Indicates the impact factor of data center operation performance corresponding to the preset response time, Indicates the impact factor of data center operation performance corresponding to the preset processor utilization, The data center operation performance influencing factors corresponding to the preset network bandwidth utilization rate are respectively the numerical values ​​of the degree of influence of the unit values ​​of the data center data processing speed, the data center data throughput, the data center response time, the data center processor utilization rate and the data center network bandwidth utilization rate on the data center operation performance. When used, they can be directly obtained from the data center information library, and their corresponding relationship can be a preset mapping relationship. For example, the data center data processing speed, the data center data throughput, the data center response time, the data center processor utilization rate and the data center network bandwidth utilization rate are respectively mapped with the data center operation performance influencing factors corresponding to the data center data processing speed, the data center data throughput, the data center response time, the data center processor utilization rate and the data center network bandwidth utilization rate preset in the data center information library to form a mapping set, and the real-time data center data processing speed, the data center data throughput, the data center response time, the data center processor utilization rate and the data center network bandwidth utilization rate are input into the mapping set to obtain the corresponding data center operation performance influencing factors, wherein the mapping relationship can be one-to-one or a many-to-one relationship. The above-mentioned influencing factors are all extracted from the data center information library, and the value range is between 0 and 1.

[0083] It should be explained that data throughput refers to the ability of a data center to process data per unit time, usually measured in the amount of data processed per second (such as bits / second, transactions / second). Response time refers to the average time required from sending a request to receiving a response. Processor utilization refers to the percentage of time the processor spends performing tasks in the total time. An increase in data throughput usually leads to an increase in processor utilization because more computing resources are needed to process data. Network bandwidth utilization refers to the ratio of the actual bandwidth used by the network to the total available bandwidth. As data throughput increases, network bandwidth utilization will also increase accordingly. When processor utilization is low, response time is usually short because the processor has enough resources to process requests quickly. As processor utilization increases, response time may become longer because the processor needs to queue up more tasks. Low network bandwidth utilization usually means less network congestion and shorter response time. If network bandwidth utilization is too high, especially when it is close to or reaches the limit, it will cause network congestion, which will increase the transmission delay of data packets and thus increase response time.

[0084] It should be explained that this embodiment can set performance benchmarks and goals for the data center by comprehensively analyzing the data processing speed, data throughput, response time, processor utilization, and network bandwidth utilization of the data center. This helps to determine whether the data center meets current business needs and provides guidance for future expansion and upgrades. It also helps to identify imbalances in resource allocation and potential bottlenecks. By optimizing resource allocation, the overall efficiency and performance of the data center can be improved. Understanding data throughput and resource utilization can help data center managers predict future resource requirements, thereby performing effective capacity planning and avoiding excess or insufficient resources. It can also reveal patterns that may lead to system failures, such as processor overload or network congestion. This helps to identify problems in advance and prevent them.

[0085] Specifically, a comprehensive analysis is performed to obtain a data center energy consumption evaluation index. The specific analysis process is as follows: based on the abnormal characteristic values ​​of equipment operation, data center data processing task parameters and data center operation performance evaluation index, a comprehensive analysis is performed to obtain a data center energy consumption evaluation index. The data center energy consumption evaluation index is used to quantitatively evaluate the energy consumption of a data center and provide a basis for predicting the energy consumption and carbon emissions of a data center.

[0086] In a specific embodiment, the numerical expression of the data center energy consumption evaluation index is: ; In the formula, represents the energy consumption evaluation index of the data center, represents the abnormal characteristic value of equipment operation, N represents the data processing task parameter of the data center, represents the data center operation performance evaluation index, Indicates the data center energy consumption correction factor corresponding to the preset equipment operation abnormality characteristic value, Indicates the data center energy consumption correction factor corresponding to the preset data center data processing task parameters, Indicates the data center energy consumption correction factor corresponding to the preset data center operation performance evaluation index.

[0087] It needs to be explained that when the abnormal characteristic value of equipment operation is larger and the data center data processing task parameter is smaller, the corresponding data center energy consumption evaluation index is larger, indicating that the energy consumption of the data center is higher.

[0088] It should be explained that in this embodiment Indicates the data center energy consumption correction factor corresponding to the preset equipment operation abnormality characteristic value, Indicates the data center energy consumption correction factor corresponding to the preset data center data processing task parameters, The data center energy consumption correction factors corresponding to the preset data center operation performance evaluation index respectively represent the numerical values ​​of the degree of influence of the abnormal operation characteristic values ​​of the data center equipment, the data center data processing task parameters and the unit values ​​of the data center operation performance evaluation index on the energy consumption of the data center. When used, they can be directly obtained from the data center information library, and the corresponding relationship can be a pre-set mapping relationship. For example, the abnormal operation characteristic values ​​of the data center equipment, the data center data processing task parameters and the data center operation performance evaluation index respectively form a mapping set with the data center energy consumption correction factors corresponding to the abnormal operation characteristic values ​​of the data center equipment, the data center data processing task parameters and the data center operation performance evaluation index preset in the data center information library. The real-time abnormal operation characteristic values ​​of the data center equipment, the data center data processing task parameters and the data center operation performance evaluation index are input into the mapping set to obtain the corresponding data center energy consumption correction factors, wherein the mapping relationship can be one-to-one or many-to-one. The above correction factors are all extracted from the data center information library, and the value range is between 0 and 1.

[0089] It needs to be explained that when data center equipment is abnormal, its performance may decline. This may lead to slower data processing speed, thereby reducing the overall data processing energy efficiency. Equipment abnormality may cause certain resources (such as CPU, memory, storage space, etc.) to be underutilized, which will reduce the resource utilization and energy efficiency of the data center. Equipment may require more electricity to maintain operation in an abnormal state, or cause more energy waste due to reduced efficiency, which will increase the overall energy consumption of the data center. Equipment abnormality may cause the heat dissipation system to be burdened. For example, an overheated server may require more cooling resources. This not only increases energy consumption, but may also further affect the stable operation of other equipment. Equipment abnormality may lead to more maintenance work and troubleshooting, and these activities may temporarily or permanently reduce the processing capacity of the data center while increasing operating costs.

[0090] It should be explained that this embodiment can predict the energy consumption trend of the data center by analyzing equipment operation anomalies, data center operation performance and data processing volume. Provide a basis for updating the ARIMA model. For example, if it is found that equipment anomalies occur frequently, it may indicate an increase in energy consumption in the future. Identifying the relationship between equipment anomalies and energy inefficiency can help the data center operation team take measures to optimize energy consumption. For example, by repairing or replacing faulty equipment, energy efficiency can be improved and energy consumption can be reduced. It can also help prevent potential equipment failures, thereby avoiding increased energy consumption caused by failures. At the same time, reducing energy consumption not only helps save costs, but also helps reduce the carbon footprint of the data center and promote environmental sustainability. By optimizing the operating status of equipment and data processing energy efficiency, the overall performance of the data center can be improved, and the reliability of services and user experience can be improved.

[0091] According to the data center energy consumption evaluation indicators, it is judged whether the ARIMA model can directly predict the data center energy consumption and carbon emissions, and the ARIMA model is fed back in real time based on the judgment results.

[0092] Specifically, according to the data center energy consumption evaluation index, it is judged whether the ARIMA model can directly predict the data center energy consumption carbon emissions. The specific process is: extract the data center energy consumption reference standard index range from the data center information database, compare the data center energy consumption evaluation index with the data center energy consumption reference standard index range, if the data center energy consumption evaluation index is within the data center energy consumption reference standard index range, then it is judged that the ARIMA model can directly predict the data center energy consumption carbon emissions; if the data center energy consumption evaluation index is higher than or equal to the maximum value of the data center energy consumption reference standard index range, or the data center energy consumption evaluation index is lower than or equal to the minimum value of the data center energy consumption reference standard index range, then it is judged that the ARIMA model cannot directly predict the data center energy consumption carbon emissions, and the ARIMA model is updated in real time according to the judgment result.

[0093] It should be added that the process of using the ARIMA model to predict the energy consumption and carbon emissions of data centers is to first collect the historical energy consumption and carbon emissions data of the data center and ensure that the data is arranged in chronological order, such as daily, weekly or monthly carbon emissions. Check the integrity and consistency of the data and handle missing values ​​or outliers. Then determine the order (p, d, q) of the ARIMA model through the analysis of the autocorrelation function (ACF) and the partial autocorrelation function (PACF). Among them, p is the autoregressive term, d is the difference order, and q is the number of moving average terms. Then use statistical methods to estimate the parameters of the ARIMA model, which describe the autoregressive and moving average parts of the time series. It is also necessary to perform diagnostic tests on the model to ensure the degree of fit of the model and the randomness of the residuals. This usually includes checking the autocorrelation, normality and heteroscedasticity of the residuals. Depending on the results of the diagnostic test, the order or parameters of the model may need to be adjusted to improve the fit of the model and the accuracy of the prediction. Once the model has been trained and verified, it can be used to predict future data center energy consumption and carbon emissions. This usually involves specifying the time range for the prediction and using the model to generate corresponding predictions. Finally, the accuracy of the prediction results is evaluated using appropriate evaluation metrics such as mean squared error, mean absolute error, etc. This helps to understand the performance of the model and make further adjustments and optimizations if necessary.

[0094] Furthermore, the ARIMA model is fed back in real time according to the judgment result, and the specific feedback process is: obtaining the energy consumption prediction value of the ARIMA model.

[0095] If the data center energy consumption evaluation index is higher than or equal to the maximum value of the data center energy consumption reference standard index range, the first correction value of the data center energy consumption and the maximum threshold of the data center energy consumption evaluation index are obtained by matching the abnormal characteristic value of equipment operation, and the ARIMA model energy consumption prediction value is added to the first correction value of the data center energy consumption to obtain the final energy consumption prediction value. The final energy consumption prediction value is input into the ARIMA model to obtain the final energy consumption carbon emission prediction value.

[0096] If the data center energy consumption evaluation index is lower than or equal to the minimum value of the data center energy consumption reference standard index range, the second correction value of the data center energy consumption and the minimum threshold of the data center energy consumption evaluation index are obtained according to the matching of the data center operation performance evaluation index. The ARIMA model energy consumption prediction value is subtracted from the second correction value of the data center energy consumption to obtain the final energy consumption prediction value. The final energy consumption prediction value is input into the ARIMA model to obtain the final energy consumption carbon emission prediction value.

[0097] It should be added that if the data center energy consumption evaluation index is higher than the maximum value of the data center energy consumption reference standard index interval and higher than or equal to the maximum threshold of the data center energy consumption evaluation index, at the same time, if the data center energy consumption evaluation index is lower than the minimum value of the data center energy consumption reference standard index interval and lower than or equal to the minimum threshold of the data center energy consumption evaluation index, then the ARIMA model needs to be retrained.

[0098] It needs to be explained that the first correction value of data center energy consumption and the maximum threshold value of data center energy consumption evaluation index are obtained by matching the abnormal equipment operation characteristic value. The specific process is to extract the mapping sets of the abnormal equipment operation characteristic value, the first correction value of data center energy consumption and the maximum threshold value of data center energy consumption evaluation index from the data center information library, input the real-time abnormal equipment operation characteristic value, and obtain the first correction value of data center energy consumption and the maximum threshold value of data center energy consumption evaluation index corresponding to the abnormal equipment operation characteristic value according to the mapping set.

[0099] It needs to be explained that the second correction value of data center energy consumption and the minimum threshold value of data center energy consumption evaluation index are obtained according to the matching of the data center operation performance evaluation index. The specific process is to extract the mapping sets of the data center operation performance evaluation index, the second correction value of data center energy consumption and the minimum threshold value of data center energy consumption evaluation index from the data center information library, input the real-time data center operation performance evaluation index, and obtain the second correction value of data center energy consumption and the minimum threshold value of data center energy consumption evaluation index corresponding to the data center operation performance evaluation index according to the mapping set.

[0100] It should be explained that, in this embodiment, by introducing the first correction value or the second correction value, the ARIMA model can more accurately predict the energy consumption of the data center. When the energy consumption evaluation index of the data center exceeds the standard range, the ARIMA model is adjusted in time to achieve real-time monitoring and rapid response to energy consumption changes, so as to take measures to reduce energy consumption. According to the model prediction results, the data center can optimize resource allocation, such as adjusting server load, optimizing cooling system operation, etc., to reduce unnecessary energy consumption. By effectively managing energy consumption, the data center can reduce electricity costs, thereby reducing overall operating costs. When the model predicts potential energy consumption problems, preventive maintenance can be performed to avoid increased energy consumption caused by equipment failure.

[0101] In a specific embodiment, the data center information library is used to store relevant parameters in the process of evaluating the energy consumption of the data center, including the critical standard value of processor utilization, the critical standard value of storage device utilization, the critical standard value of cooling system pressure and the critical standard value of network bandwidth utilization, as well as the data extracted from the data center information library in the above embodiments. Data centers usually install various monitoring systems and sensors to collect real-time data, including temperature, humidity, power usage, server load, etc. These systems can automatically transmit data to the information library. The data center infrastructure management system software can integrate and manage data from various devices and systems in the data center. The data center infrastructure management system software usually has data storage and analysis functions and can be directly used to build an information library. The network management system can provide performance and utilization data of network equipment. Servers and storage devices usually generate log files to record their own operating status and performance data. These log files can be exported and imported into the information library regularly.

[0102] Reference Figure 2 As shown, the second aspect of the present invention provides an ARIMA-based data center energy consumption and carbon emission prediction system, including: a data center equipment analysis module, a data center energy consumption analysis module and a comprehensive feedback module.

[0103] The data center equipment analysis module is used to obtain the equipment operation data of the data center, obtain the abnormal equipment operation characteristic value after processing, and issue risk warnings to the data center equipment based on the abnormal equipment operation characteristic value.

[0104] The data center energy consumption analysis module is used to obtain data center performance data and data processing task information, and process the data center data processing task parameters and the data center operation performance evaluation index. According to the abnormal operation characteristic values ​​of the equipment, the data center data processing task parameters and the data center operation performance evaluation index, a comprehensive analysis is performed to obtain the data center energy consumption evaluation index.

[0105] The comprehensive feedback module is used to determine whether the ARIMA model can directly predict the energy consumption and carbon emissions of the data center based on the data center energy consumption evaluation indicators, and to provide real-time feedback to the ARIMA model based on the judgment results.

[0106] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should all fall within the protection scope of the present invention.

Claims

1. A data center energy consumption and carbon emission prediction method based on ARIMA, characterized in that: include: Obtain the equipment operation data of the data center, obtain the abnormal equipment operation characteristic value after processing, and issue risk warning for the data center equipment according to the abnormal equipment operation characteristic value; Obtain data center performance data and data processing task information, process to obtain data center data processing task parameters and data center operation performance evaluation index, and comprehensively analyze to obtain data center energy consumption evaluation index based on equipment operation abnormality characteristic values, data center data processing task parameters and data center operation performance evaluation index; Determine whether the ARIMA model can directly predict the energy consumption and carbon emissions of the data center based on the data center energy consumption evaluation indicators, and provide real-time feedback to the ARIMA model based on the judgment results; The numerical expression of the data center energy consumption evaluation index is: ； In the formula, represents the energy consumption evaluation index of the data center, represents the abnormal characteristic value of equipment operation, N represents the data processing task parameter of the data center, represents the data center operation performance evaluation index, Indicates the data center energy consumption correction factor corresponding to the preset equipment operation abnormality characteristic value, Indicates the data center energy consumption correction factor corresponding to the preset data center data processing task parameters, Indicates the data center energy consumption correction factor corresponding to the preset data center operation performance evaluation index.

2. According to claim 1, a data center energy consumption and carbon emission prediction method based on ARIMA is characterized by: The abnormal characteristic value of the equipment operation is obtained through processing, and the specific process is: The equipment operation data of the data center includes equipment operation environment data and equipment operation status data; According to the equipment operating environment data, the operating environment impact characteristic value is obtained by processing; The device operation status data includes: processor utilization, storage device utilization, cooling system pressure and network bandwidth utilization; Extracting from the data center information database a critical standard value of processor utilization, a critical standard value of storage device utilization, a reference standard value of cooling system pressure, an allowable deviation value of cooling system pressure, and a critical standard value of network bandwidth utilization; According to the equipment operation status data and the operating environment impact characteristic values, comprehensive processing is performed to obtain the equipment operation abnormality characteristic values, which are used to quantitatively evaluate the degree of abnormal operation risk of data center equipment and provide a basis for equipment risk warning.

3. According to claim 2, a data center energy consumption and carbon emission prediction method based on ARIMA is characterized by: The processing obtains the operating environment impact characteristic value, and the specific process is: The equipment operating environment data includes: average temperature, average humidity, average noise level and power supply voltage of the data center operating area; Extract the regional temperature standard value, regional humidity standard value, regional temperature allowable deviation value, regional humidity allowable deviation value, regional noise level critical value, rated power supply voltage and allowable deviation power supply voltage from the data center information database; The operating environment impact characteristic value is processed according to the equipment operating environment data. The operating environment impact characteristic value is used to quantitatively evaluate the interference degree of the equipment operating environment on the equipment operation, and provide a basis for evaluating the abnormal operation of the equipment.

4. According to claim 2, a data center energy consumption and carbon emission prediction method based on ARIMA is characterized by: The specific process of performing risk warning on data center equipment according to abnormal equipment operation characteristic values ​​is as follows: The first threshold and the second threshold of the abnormal operation characteristic of the equipment are extracted from the data center information library, and the abnormal operation characteristic value of the equipment is compared with the first threshold and the second threshold of the abnormal operation characteristic of the equipment. If the abnormal operation characteristic value of the equipment is higher than or equal to the first threshold of the abnormal operation characteristic of the equipment, the operation status of the equipment is evaluated as abnormal, and the early warning platform issues a red alarm and notifies the staff to perform abnormal analysis and equipment maintenance. If the abnormal operation characteristic value of the equipment is lower than the first threshold of the abnormal operation characteristic of the equipment and higher than the second threshold of the abnormal operation characteristic of the equipment, the early warning platform issues an orange alarm and automatically switches the power supply. If the abnormal operation characteristic value of the equipment is lower than or equal to the second threshold of the abnormal operation characteristic of the equipment, the operation status of the equipment is evaluated as normal, and the early warning platform outputs the result and displays a green indicator light.

5. According to claim 1, a data center energy consumption and carbon emission prediction method based on ARIMA is characterized by: The processing obtains the data center data processing task parameters and the data center operation performance evaluation index, including: The data processing task information includes: the amount of tasks processed by the data center within a preset monitoring period, the number of user visits and the number of user operations of the data center within the preset monitoring period; Extract the critical task volume, critical user access volume and critical user operation times from the data center information database; According to the data processing task information, comprehensive processing is performed to obtain the data center data processing task parameters, which are used to quantitatively evaluate the busyness of the data center data processing and provide a basis for analyzing the energy consumption of the data center.

6. According to claim 1, a data center energy consumption and carbon emission prediction method based on ARIMA is characterized by: The processing obtains the data center data processing task parameters and the data center operation performance evaluation index, and also includes: The data center performance data includes: data throughput of the data center, response time of the data center, processor utilization, network bandwidth utilization and data processing speed of the data center; Extracting data throughput critical value, response time critical value, processor utilization reference standard value, processor utilization allowable deviation value, network bandwidth utilization reference standard value, network bandwidth utilization allowable deviation value and critical data processing speed from the data center information database; According to the data center performance data, a data center operation performance evaluation index is obtained by processing. The data center operation performance evaluation index is used to quantitatively evaluate the operation performance of the data center and provide a basis for analyzing the data processing energy efficiency of the data center.

7. The method for predicting energy consumption and carbon emissions of a data center based on ARIMA according to claim 6 is characterized by: The comprehensive analysis obtains the data center energy consumption evaluation index, and the specific analysis process is as follows: Based on the abnormal characteristic values ​​of equipment operation, data center data processing task parameters and data center operation performance evaluation index, a comprehensive analysis is conducted to obtain the data center energy consumption evaluation index, which is used to quantitatively evaluate the energy consumption of the data center and provide a basis for predicting the energy consumption and carbon emissions of the data center.

8. According to claim 7, a data center energy consumption and carbon emission prediction method based on ARIMA is characterized by: The specific process of judging whether the ARIMA model can directly predict the energy consumption carbon emission of the data center based on the data center energy consumption evaluation index is as follows: The data center energy consumption reference standard indicator range is extracted from the data center information database, and the data center energy consumption evaluation indicator is compared with the data center energy consumption reference standard indicator range. If the data center energy consumption evaluation indicator is within the data center energy consumption reference standard indicator range, it is judged that the ARIMA model can directly predict the data center energy consumption carbon emissions. If the data center energy consumption evaluation indicator is higher than or equal to the maximum value of the data center energy consumption reference standard indicator range, or the data center energy consumption evaluation indicator is lower than or equal to the minimum value of the data center energy consumption reference standard indicator range, it is judged that the ARIMA model cannot directly predict the data center energy consumption carbon emissions, and the ARIMA model is updated in real time according to the judgment result.

9. The method for predicting energy consumption and carbon emissions of a data center based on ARIMA according to claim 1, characterized in that: The ARIMA model is fed back in real time according to the judgment result, and the specific feedback process is as follows: Get the energy consumption forecast value of the ARIMA model; If the data center energy consumption evaluation index is higher than or equal to the maximum value of the data center energy consumption reference standard index interval, the first correction value of the data center energy consumption and the maximum threshold of the data center energy consumption evaluation index are obtained according to the matching of the abnormal characteristic value of equipment operation, and the ARIMA model energy consumption prediction value is added to the first correction value of the data center energy consumption to obtain the final energy consumption prediction value. The final energy consumption prediction value is input into the ARIMA model to obtain the final energy consumption carbon emission prediction value; If the data center energy consumption evaluation index is lower than or equal to the minimum value of the data center energy consumption reference standard index range, the second correction value of the data center energy consumption and the minimum threshold of the data center energy consumption evaluation index are obtained according to the matching of the data center operation performance evaluation index. The ARIMA model energy consumption prediction value is subtracted from the second correction value of the data center energy consumption to obtain the final energy consumption prediction value. The final energy consumption prediction value is input into the ARIMA model to obtain the final energy consumption carbon emission prediction value.

10. A data center energy consumption and carbon emission prediction system based on ARIMA, using a data center energy consumption and carbon emission prediction method based on ARIMA as claimed in any one of claims 1 to 9, characterized in that: include: Data center equipment analysis module, data center energy consumption analysis module and comprehensive feedback module; The data center equipment analysis module is used to obtain the equipment operation data of the data center, obtain the abnormal equipment operation characteristic value after processing, and issue risk warning for the data center equipment according to the abnormal equipment operation characteristic value; The data center energy consumption analysis module is used to obtain data center performance data and data processing task information, process the data center data processing task parameters and the data center operation performance evaluation index, and comprehensively analyze the data center energy consumption evaluation index based on the abnormal operation characteristic value of the equipment, the data center data processing task parameters and the data center operation performance evaluation index; The comprehensive feedback module is used to determine whether the ARIMA model can directly predict the energy consumption and carbon emissions of the data center based on the data center energy consumption evaluation indicators, and to provide real-time feedback to the ARIMA model based on the judgment results.

Images