Data center carbon emission prediction method, device and equipment and computer storage medium

Abstract

The embodiment of the present application relates to the technical field of computer data processing, and discloses a data center carbon emission prediction method, which comprises the following steps: determining the predicted power consumption of the target computing power corresponding to the target computing requirement according to the computing power configuration information of the data center; the target computing power is at least one of a plurality of selectable computing powers configured in the data center; determining the energy consumption of the data center under a preset energy type when the target computing power is running according to the predicted power consumption; and determining the predicted carbon emission of the data center when the target computing power is running according to the energy consumption. In the above manner, the accuracy of data center carbon emission prediction is improved.

Description

Technical Field

[0001] This invention relates to the field of computer data processing technology, specifically to a method, apparatus, device, and computer storage medium for predicting carbon emissions from a data center. Background Technology

[0002] Data centers and the computing power they support are the cornerstone of the digital economy and have been growing rapidly in recent years. At the same time, due to the high energy consumption of data centers, accurately predicting their carbon emissions has become a crucial issue to be addressed in carbon reduction efforts.

[0003] The inventors of this application discovered during the implementation of the embodiments of this method that different industries employ entirely different carbon emission accounting methods due to variations in their business processes and technological flows. Existing carbon emission calculations are highly targeted, often specific to a particular industry, such as carbon emission prediction methods for the ceramics industry based on electricity data, or carbon emission accounting methods for the retail industry. Existing carbon emission calculation methods for non-data center industries are not applicable to the data center industry, resulting in low accuracy in carbon emission predictions for data centers. Summary of the Invention

[0004] In view of the above problems, embodiments of the present invention provide a method, apparatus, device and computer storage medium for predicting carbon emissions from data centers, in order to solve the problem of low accuracy in predicting carbon emissions from data centers in the prior art.

[0005] According to one aspect of the present invention, a method for predicting carbon emissions from a data center is provided, the method comprising:

[0006] The predicted power consumption of the target computing power corresponding to the target computing demand is determined based on the computing power configuration information of the data center; the target computing power is at least one of a variety of optional computing powers configured within the data center.

[0007] The energy consumption of the data center under a preset energy type is determined based on the predicted power consumption.

[0008] The predicted carbon emissions of the data center when the target computing power is running are determined based on the energy consumption.

[0009] In one optional approach, the computing power configuration information includes computing performance information of the multiple optional computing powers; the method further includes:

[0010] The target computing power is determined from the multiple selectable computing powers based on the computing performance information;

[0011] The predicted computing power efficiency is determined based on the historical energy consumption data of the target computing power.

[0012] The target computational requirement is converted based on the predicted computational efficiency to obtain the predicted power consumption.

[0013] In one alternative approach, the energy type is multiple; the method further includes:

[0014] The predicted power consumption is calculated based on the proportion of the target computing power to the power consumption of the data center, the power conversion coefficients corresponding to various energy types, and the energy availability status, to obtain the energy consumption amount.

[0015] In an alternative approach, the method further includes:

[0016] Based on the power consumption ratio and the power conversion coefficient, the predicted power consumption is converted to obtain the predicted energy consumption of the data center for the energy type when the target computing power is running.

[0017] The predicted energy consumption during operation is calculated based on the energy activation status and preset statistical time information to obtain the energy consumption.

[0018] In an alternative approach, the method further includes:

[0019] The actual power consumption of the target computing power is determined based on the first power utilization rate of the target computing power and the predicted power consumption.

[0020] The overall power consumption of the data center is determined based on the second power utilization rate of the data center and the total power consumption of the data center during operation.

[0021] The power consumption ratio of the target computing power is determined based on the ratio of the actual power consumption to the total power consumption.

[0022] In one alternative approach, the energy type is multiple; the method further includes:

[0023] For each of the aforementioned energy types, determine the corresponding carbon emission coefficient for that energy type;

[0024] The carbon emission coefficient and the energy consumption are used to determine the carbon emission of a single energy source corresponding to the energy type.

[0025] The predicted carbon emissions are determined based on the sum of the carbon emissions of each of the energy types.

[0026] In an alternative approach, the method further includes:

[0027] Determine the amount of heat recovered from the data center;

[0028] The predicted carbon emissions are calibrated based on the recovered carbon emissions corresponding to the recovered heat.

[0029] According to another aspect of the present invention, a data center carbon emission prediction device is provided, comprising:

[0030] The first determining module is used to determine the predicted power consumption of the target computing power corresponding to the target computing demand based on the computing power configuration information of the data center; the target computing power is at least one of a variety of optional computing powers configured in the data center;

[0031] The second determining module is used to determine the energy consumption of the data center under a preset energy type when the target computing power is running, based on the predicted power consumption.

[0032] The third determining module is used to determine the predicted carbon emissions of the data center when the target computing power is running, based on the energy consumption.

[0033] According to another aspect of the present invention, a data center carbon emission prediction device is provided, comprising: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other through the communication bus;

[0034] The memory is used to store at least one executable instruction that causes the processor to perform operations as described in any of the embodiments of the data center carbon emission prediction method.

[0035] According to another aspect of the present invention, a computer-readable storage medium is provided, the storage medium storing at least one executable instruction that causes a data center carbon emission prediction device to perform the operation of any of the data center carbon emission prediction method embodiments described above.

[0036] This invention, through its embodiments, determines the predicted power consumption of a target computing power corresponding to a target computing requirement based on the computing power configuration information of a data center. The target computing power is at least one of several selectable computing powers configured within the data center. Based on the predicted power consumption, the energy consumption of the data center under a preset energy type during the operation of the target computing power is determined. Based on the energy consumption, the predicted carbon emissions of the data center during the operation of the target computing power are determined. This enables the determination of the estimated carbon emissions of the data center to meet the target computing requirements, solving the problem that existing carbon emission calculation methods are not applicable to data centers and can only calculate carbon emissions for clearly configured computing equipment and already built data centers, but cannot predict carbon emissions to meet unmet computing requirements or planned data centers. This improves the practicality and accuracy of data center carbon emission prediction.

[0037] The above description is merely an overview of the technical solutions of the embodiments of the present invention. In order to better understand the technical means of the embodiments of the present invention and to implement them in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the embodiments of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0038] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0039] Figure 1 A flowchart illustrating the data center carbon emission prediction method provided in an embodiment of the present invention is shown.

[0040] Figure 2 A schematic diagram of the structure of the data center carbon emission prediction device provided in an embodiment of the present invention is shown;

[0041] Figure 3 A schematic diagram of the structure of a data center carbon emission prediction device provided in an embodiment of the present invention is shown. Detailed Implementation

[0042] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0043] Before describing the embodiments of the present invention, the prior art and its problems will be further explained:

[0044] Data centers and their supporting computing facilities are the cornerstone of the digital economy and have been growing rapidly in recent years. At the same time, the high energy consumption of data centers cannot be ignored, making carbon emission prediction a crucial aspect of planning and construction. However, different industries employ vastly different carbon emission accounting methods due to variations in their business processes and technological workflows. Existing carbon emission calculation technologies are highly targeted, often specific to particular industries, such as carbon emission prediction methods for the ceramics industry based on electricity data, a carbon emission accounting method and device for the service industry, a method for predicting carbon emissions for city-scale industries based on uncertain sets, and a carbon emission accounting method for the retail industry. Existing carbon emission calculation methods for non-data center industries are not applicable to the data center industry.

[0045] Data centers are physical spaces that support the normal operation of IT equipment, typically taking the form of buildings. However, data center buildings differ significantly from residential and commercial buildings in terms of architectural style, energy composition, and energy consumption. Therefore, existing methods for calculating carbon emissions from building facilities are not applicable to data centers. For example, methods for predicting carbon emissions in the building industry generally divide the entire building into different spatial areas and separately analyze the energy consumption in each area, including heating, cooking activities, and daily activities, involving heating energy consumption, gas consumption, and electrical energy consumption from appliances, before summarizing the results to arrive at carbon emissions. This approach is not suitable for predicting carbon emissions from data centers.

[0046] Currently, there are few methods for calculating carbon emissions from data centers. Existing technologies for managing carbon emissions from data centers, including methods for evaluating carbon emission efficiency, generally use two indicators to calculate carbon emissions: the electricity consumption of IT equipment and the overall electricity consumption of the data center. However, existing methods do not comprehensively consider the energy structure of data centers, only taking into account the electricity consumed without considering the energy reduction achieved through renewable energy sources and waste heat recovery. This leads to inaccurate and incomplete calculations. Furthermore, existing methods for calculating carbon emissions from data centers typically calculate and summarize the carbon emissions of one or more of the four sources: electricity, heat, natural gas, and diesel, with each source including both input and output. These existing technologies are applicable to carbon emission calculations of existing data centers, representing backend statistics and calculations that reflect current or historical carbon emissions. They cannot predict data center carbon emissions and do not account for water resources or solar energy utilization.

[0047] Therefore, the existing technology has at least the following problems:

[0048] 1. Regarding IT equipment energy consumption prediction based on target computing power: Currently, there is no method in the industry to predict the power consumption of IT equipment supporting computing power based on computing power scale. The commonly used method is to calculate the corresponding total power consumption based on the composition of IT equipment and the power consumption of individual devices. This method is only applicable to systems with a clear IT equipment configuration plan, which has been refined to the type and quantity of equipment. Currently, the widely accepted computing power and computing efficiency models in the industry are all calculated based on theoretical computing power and theoretical power consumption, which deviates significantly from the actual operating conditions of data center resource pools. Therefore, given a clear target computing power, it is quite difficult to measure the energy consumption of IT equipment supporting this computing power, and existing methods are not mature enough.

[0049] 2. Regarding carbon emission calculation methods for data centers: Currently, there are few methods for calculating carbon emissions from data centers, and the proposed methods are incomplete. They only calculate the power consumption of IT equipment and the total power consumption of the data center, without addressing new technologies and new energy sources such as renewable energy utilization and waste heat recovery. Therefore, they cannot accurately determine the carbon emissions of data centers; or they are only applicable to the calculation of carbon emissions from data centers that have been built and are in operation, and cannot predict the carbon emissions of data centers under construction or in the planning stage. This proposal will combine various energy applications in data centers to construct a data center carbon emission prediction method based on computing power. This method can be widely applied to the estimation of carbon emissions during the planning, design, and construction phases of data centers.

[0050] 3. There is currently no method to predict data center carbon emissions based on computing power demand.

[0051] Figure 1 A flowchart of a data center carbon emission prediction method provided by an embodiment of the present invention is shown. This method is executed by a computer processing device. The computer processing device may include a mobile phone, laptop computer, etc. Figure 1 As shown, the method includes the following steps:

[0052] Step 10: Determine the predicted power consumption of the target computing power corresponding to the target computing requirements based on the computing power configuration information of the data center; the target computing power is at least one of the various optional computing powers configured in the data center.

[0053] Specifically, the preferred data center types are those other than supercomputing data centers and data centers primarily using traditional storage (such as disk arrays, tape libraries, and NAS). Computing power configuration information includes the computing performance information of at least one optional computing power configured within the data center. This computing performance information may include the number of such optional computing power units configured, their computing efficiency, and power consumption per unit time. Computing efficiency characterizes the computing power provided per unit of power consumed, and the unit may be GFLOPS / W. Target computing requirements refer to the computing needs to be met. These computing needs refer to the information required to complete the computing task, such as the amount of computing power required for the task.

[0054] Based on the data center's computing power configuration information, the target computing demand is broken down into various optional computing powers. The target computing power is then determined based on the computing demand undertaken by each optional computing power. Specifically, all optional computing powers with a computing demand greater than zero are selected as target computing powers. Optional computing powers can be categorized by processing capability level, such as based on the number of CPUs in the computing equipment. The computing power corresponding to general-purpose services with 1-2 CPUs is identified as the first optional computing power representing low-end computing power, denoted as CP. l--cyCorrespondingly, the computing power corresponding to general-purpose servers with 3-4 CPUs is determined as the second optional computing power representing mid-range computing power, denoted as CP. m-cy The computing power corresponding to general-purpose servers with 4 or more CPUs is identified as the third optional computing power to represent high-end computing power, denoted as CP. h-cy .

[0055] Alternatively, computing power can be categorized based on the type of processing tasks performed by the computing devices. AI training server-type computing power can be identified as a fourth optional computing power representing intelligent training computing power, denoted as CP. at-cy AI inference server-type computing power is identified as the fifth optional computing power to represent intelligent inference computing power, denoted as CP. ai-cy The computing power of storage node-type servers is identified as the sixth optional computing power representing the computing power of storage nodes, denoted as CP. s-cy .

[0056] Optionally, step 10 further includes: step 101: determining the target computing power from the multiple selectable computing powers based on the computing performance information.

[0057] Specifically, the target computing power is selected and combined on demand from the available computing power to meet the target computing needs.

[0058] Specifically, based on the characteristics of the computing power and efficiency of different servers in a data center, the target computing requirements for a data center (unit: GFLOPS, Floating-point Operations Per Second, a quantity used to characterize and measure computer computing power) are decomposed into the aforementioned low-end computing power CP. l Mid-range computing power CP m High-end computing power CP h Intelligent training computing power CP at Intelligent computing power CP ai Computing power CP of storage nodes s The six types, meaning the target computing power is the sum of the six types of computing power, are represented as follows:

[0059] Target computing power CP = CP l +CP m +CP h +CP at +CP ai +CP s ;

[0060] Among them: CP l This represents the computing power requirements for general-purpose servers with 1-2 CPUs, measured in GFLOPS; CP. m This represents the computing power requirements for general-purpose servers with 3-4 CPUs, measured in GFLOPS; CP. hThis refers to the computing power requirements of general-purpose servers with 4 or more CPUs, measured in GFLOPS; CP. at Computing power requirements for AI training servers, in units of GFLOPS; CP ai : Computing power requirements for AI inference servers, unit: GFLOPS; CP S : This refers to the computing power requirements of storage node servers, measured in GFLOPS.

[0061] Step 102: Determine the predicted computing power efficiency of the target computing power based on the historical energy consumption data of the target computing power.

[0062] Specifically, historical energy consumption data includes energy consumption information during computational operations performed by the target computing power within a historical time interval. Based on the computational load information and corresponding energy consumption information of historical computational operations, the historical actual computing power efficiency of the target computing power is obtained. Then, based on this historical actual computing power efficiency, a prediction is made to obtain the target computing power. In calculating historical computing power efficiency, existing data center resource pools with similar business operations and computing power compositions can be selected based on the target computing power composition and the business requirements it supports. The low-end computing power CP1-cy and mid-end computing power CP1-cy of the sampled resource pools are obtained through network management platforms, environmental monitoring systems, and intelligent PDUs, respectively. m-cy High-end computing power CP h-cy Intelligent training computing power CP at-cy Intelligent computing power CP ai-cy Storage node computing power CP s-cy And their respective power consumption values ​​(unit: kWh), of which the cumulative power consumption should be no less than one month.

[0063] Specifically, based on the acquired historical energy consumption data, the actual computing power SACE value (unit: GFLOPS / W) for each type of computing power (including only server equipment) is calculated using the following formula:

[0064] SACEoCP l =CP l-cy / (CP l-cy Power consumption / CP l-cy (Statistical duration) / 1000

[0065] SACEoCP m =CP m-cy / (CP m-cy Power consumption / CP m-cy (Statistical duration) / 1000

[0066] SACEoCP h =CP h-cy / (CP h-cy Power consumption / CP h-cy (Statistical duration) / 1000

[0067] SACEoCP at =CP at-cy / (CP at-cy Power consumption / CP at-cy (Statistical duration) / 1000

[0068] SACEoCP ai =CP ai-cy / (CP ai-cy Power consumption / CP ai-cy (Statistical duration) / 1000

[0069] SACEoCP S =CP S-cy / (CP S-cy Power consumption / CP S-cy (Statistical duration) / 1000

[0070] in:

[0071] SACEoCP l The actual computing power of the sampled resource pool for low-end computing power (the computing power corresponding to general-purpose servers with 1-2 CPUs) is expressed in GFLOPS / W.

[0072] CP l-cy The computing power of the low-end computing power sampling source pool is expressed in GFLOPS.

[0073] CP l-cy The power consumption value is the total power consumption value within the statistical period of the low-end computing power sampling resource pool, in kWh;

[0074] CP l-cy The statistical duration refers to the duration of resource pool usage for low-end computing power, in hours (h).

[0075] SACEoCP m The actual computing power of the sampled resource pool for mid-range computing power (the computing power corresponding to general-purpose servers with 3-4 CPUs) is expressed in GFLOPS / W.

[0076] CP m-cy The computing power of the mid-range computing power sampling source pool is expressed in GFLOPS.

[0077] CP m-cy The power consumption value is the total power consumption value within the statistical period of the mid-range computing power sampling resource pool, in kWh;

[0078] CP m-cy The statistical duration refers to the duration of resource pool usage for mid-range computing power, in hours (h).

[0079] SACEoCP hThe actual computing power of the sampling resource pool for high-end computing power (the computing power corresponding to general-purpose servers with 4 or more CPUs), in units of GFLOPS / W;

[0080] CP h-cy The computing power of the high-end computing power sampling source pool, in GFLOPS;

[0081] CP h-cy The power consumption value is the total power consumption value within the statistical period of the high-end computing power sampling resource pool, in kWh;

[0082] CP h-cy The statistical duration refers to the duration during which high-end computing power utilizes the resource pool, in hours (h).

[0083] SACEoCP at The actual computing power of the AI ​​training server-type computing power sampling resource pool is expressed in GFLOPS / W.

[0084] CP at-cy The computing power of the AI ​​training server-type computing power sampling source pool, in units of GFLOPS;

[0085] CP at-cy The power consumption value is the total power consumption value of the AI ​​training server-type computing power sampling resource pool within the statistical duration, in kWh;

[0086] CP at-cy The statistical duration refers to the duration of AI training server computing power using resource pools, in hours (h).

[0087] SACEoCP ai The actual computing power of the AI ​​inference server-type computing power sampling resource pool is expressed in GFLOPS / W.

[0088] CP ai-cy The computing power of the AI ​​inference server-type computing power sampling source pool, in units of GFLOPS;

[0089] CP ai-cy The power consumption value is the total power consumption value of the AI ​​inference server-type computing power sampling resource pool within the statistical period, in kWh;

[0090] CP ai-cy The statistical duration refers to the duration of AI inference server computing power using the resource pool, in hours (h).

[0091] SACEoCP s The actual computing power of the storage node / server class computing power sample resource pool is expressed in GFLOPS / W.

[0092] CP s-cyThe computing power of the sampling source pool for storage node-type server-type computing power, in units of GFLOPS;

[0093] CP s-cy The power consumption value is the total power consumption of the storage node-type server-type computing power sampling resource pool within the statistical period, in kWh;

[0094] CP s-cy The statistical duration is the duration of the resource pool used by the computing power of storage nodes and servers, in hours.

[0095] When predicting computing performance based on historical actual computing performance, the following formula can be used to calculate the predicted computing performance ACE (unit: GFLOPS / W) for various computing powers according to the preset overall power consumption ratio of network devices and servers in the resource pool:

[0096] ACEoCP l =SACEoCP l / (1+P n )

[0097] ACEoCP m =SACEoCP m / (1+P n )

[0098] ACEoCP h =SACEoCP h / (1+P n )

[0099] ACEoCP at =SACEoCP at / (1+P n )

[0100] ACEoCP ai =SACEoCP ai / (1+P n )

[0101] ACEoCP s =SACEoCP s / (1+P n )

[0102] Where: P n The ratio of the total power consumption of the network devices and servers in the resource pool is generally set to 0.1 based on experience.

[0103] ACEoCP l Predicted computing performance for low-end computing power requirements within the target computing power, unit: GFLOPS / W; ACEoCP mPredicted computing power efficiency for mid-range computing power requirements within the target computing power, unit: GFLOPS / W; ACEoCP h Predicted computing performance of high-end computing power requirements within the target computing power, unit: GFLOPS / W; ACEoCP at Predicted computational efficiency of AT training server-type computing power requirements in the target computing power, unit: GFLOPS / W; ACEoCP ai Predicted computing power efficiency for AI inference server-type computing power requirements within the target computing power, unit: GFLOPS / W; ACEoCP s Predicted computing power efficiency of storage node and server computing power requirements in the target computing power, in units of GFLOPS / W.

[0104] Step 103: Convert the target computing requirements according to the predicted computing efficiency to obtain the predicted power consumption.

[0105] Specifically, combining the computing power decomposition and computing efficiency prediction methods in steps 101-102 above, the predicted power consumption P of the IT equipment carrying the target computing power can be calculated using the following formula. 算 As follows: P 算 (Unit: kW) =

[0106] (CP l / ACEoCP l +CP m / ACEoC m +CP h / ACEoCP h +CP at / ACEoCP at +CP ai / ACEoC P ai +CP s / ACEoCP S ) / 1000.

[0107] Step 20: Determine the energy consumption of the data center under the preset energy type when the target computing power is running based on the predicted power consumption.

[0108] In one embodiment of the present invention, the preset energy types include one or more of municipal electricity, fossil fuels, green energy, water resources, heat, and waste heat output. For each energy type, the data center's consumption of that energy type during the target computing power operation is calculated based on the energy conversion coefficient between that energy type and electricity, and the activation status of that energy type. Specifically, the activation status can be the activation time information for that energy type. For example, for fossil fuels, the activation status is when the data center generates electricity through its own generator set when the municipal power supply fails (this is not included if the data center where the computing power facility is located does not have a backup generator set). Therefore, the activation status of fossil fuels can be determined based on the situation of municipal power supply failure. For municipal electricity, it can be determined based on its year-round operation or the activation time within a historical statistical time period.

[0109] Furthermore, considering that the energy consumption of the target computing power accounts for a certain proportion of the total energy consumption of the data center, the energy consumption of the target computing power for various energy types during operation is also a corresponding proportion of the data center's consumption of various energy types. Therefore, the proportion of the target computing power's power consumption relative to the data center's power consumption can be determined first. Based on this proportion and the overall consumption of various energy types by the data center during operation, the consumption of various energy types by the target computing power during operation can be further determined.

[0110] Therefore, in one embodiment of the present invention, the energy type is multifaceted, including municipal electricity, fossil fuels, green energy, water resources, and heat.

[0111] Step 20 further includes: Step 201: Calculate the predicted power consumption based on the proportion of the target computing power to the power consumption of the data center, the power conversion coefficients corresponding to various energy types, and the energy activation status, to obtain the energy consumption.

[0112] Specifically, the overall power consumption of the data center during operation is determined based on the proportion of electricity consumption and the predicted power consumption. The power consumption of each energy type is determined based on the overall power consumption and the energy availability. Finally, the power consumption of each energy type is converted according to the power conversion factor to obtain the energy consumption of that energy type.

[0113] Therefore, in another embodiment of the present invention, step 201 further includes: step 2011: converting the predicted power consumption according to the power consumption ratio and the power conversion coefficient to obtain the predicted energy consumption of the data center for the energy type when the target computing power is running.

[0114] Specifically, the overall power consumption of the data center when the target computing power is running is first calculated based on the proportion of electricity consumption and the predicted power consumption. Then, the overall power consumption is converted according to the power conversion coefficient to obtain the predicted energy consumption of the data center for the energy type when the target computing power is running.

[0115] The process for determining the proportion of electricity consumption is as follows:

[0116] Step 2021: Determine the actual power consumption of the target computing power based on the first power utilization rate of the target computing power and the predicted power consumption.

[0117] Specifically, the first power utilization rate is used to characterize the expected efficiency of the target computing power in utilizing power, and can be an indicator value under ideal conditions. The actual power consumption of the target computing power is calculated as follows:

[0118] Actual power consumption = P 算 *PUE1;

[0119] Among them, P 算 : Total power of the target computing power IT equipment, calculated according to the computing power energy consumption model, unit: kW; PUE1: Expected energy utilization rate of the target computing power IT equipment, which is the expected target.

[0120] Step 2022: Determine the overall power consumption of the data center based on the second power utilization rate of the data center and the total power consumption of the data center during operation.

[0121] Specifically, the second power utilization rate is used to characterize the expected efficiency of power utilization in a data center, and can be an indicator value under ideal conditions. The overall power consumption of a data center is calculated as follows:

[0122] Total power consumption = P 总 *PUE 总 ;

[0123] Among them, P 总 Power Usage Effectiveness (PUE) is the power consumption of all IT equipment in a data center, measured in kW. 总 Expected power utilization rate of data centers.

[0124] Step 2023: Determine the power consumption ratio of the target computing power based on the ratio of the actual power consumption to the total power consumption.

[0125] Considering that the target computing power in this embodiment of the invention is a part or all of the total computing power carried by the data center, the power consumption of the target computing power IT equipment in the overall power consumption of the data center based on computing power in this embodiment of the invention has already included the power consumption of the IT equipment with the target computing power in the total power consumption of all IT equipment in the data center. Therefore, the ratio of actual power consumption to total power consumption can be determined as the proportion of power consumption.

[0126] Step 2012: Calculate the predicted energy consumption during operation based on the energy activation status and preset statistical time information to obtain the energy consumption.

[0127] Specifically, the calculation process for energy consumption of various energy sources is as follows:

[0128] Energy consumption E1 from municipal power (annual calculation): Considering that the main energy source for the IT equipment with the target computing power is municipal electricity, the annual energy consumption E1 from municipal power is:

[0129] E1 = P 算 *PUE1*8760*α;

[0130] Where E1: Annual municipal power consumption of the target computing power IT equipment, unit: kW; P 算 : Total power of the target computing power IT equipment, calculated based on the computing power energy consumption model, unit: kW; PUE1: Expected energy utilization rate of the target computing power IT equipment, which is the expected target. α: Municipal power supply reliability data for the previous year, which can be obtained from the statistical data of the power supply department where the target computing power is deployed; The IT equipment of the target computing power is assumed to operate throughout the year, with an operating time of 8760 hours.

[0131] Energy consumption calculation for fossil fuels: When the municipal power supply fails, the data center generates electricity using its own backup generators (this is not included if the data center housing the computing facilities does not have backup generators). The backup generators operate in two modes: First, in the event of a mains power failure, the backup generators operate until the mains power is restored; second, to ensure the backup power equipment operates normally, the backup power equipment undergoes regular testing, typically monthly. The annual fossil fuel consumption E2 for the target computing power IT equipment under mains power failure and backup generator operation conditions is calculated as follows:

[0132] E2 = P 算 *PUE1*8760*(1-α)*M1 / 1000;

[0133] The annual fossil fuel consumption E3 of the target computing power IT equipment during generator set test operation is calculated as follows:

[0134] E3 = P 算 *PUE1 / (P 总 *PUE 总 )*12*N*K*M2 / 60 / 1000;

[0135] Therefore, the data center corresponding to the target computing power IT equipment consumes fossil fuels (E4).

[0136] E4 = E2 + E3;

[0137] Wherein, E4: fossil energy consumption of the data center corresponding to the target computing power IT equipment, in tons (t); E2: fossil energy consumption of the self-contained generator set, in tons; M1: fossil energy consumption per kilowatt-hour when the standby generator set is in power generation mode, in kg / kWh; E3: fossil energy consumption of the self-contained generator set during test operation, in tons; P 总 Power Usage Effectiveness (PUE) is the power consumption of all IT equipment in a data center, measured in kW. 总 : Expected power utilization rate of the data center; N: Number of backup generators in the data center; K: Monthly operating time of each backup power supply device, in minutes; M2: Fossil energy consumption per unit time during backup power supply test operation, in kg / h.

[0138] Regarding green energy consumption: Data center green energy mainly consists of two parts: green electricity traded through the market and self-built green energy power generation facilities. The data center commits to purchasing E5 worth of green electricity, and its self-built green energy generation is E6. Therefore, the total green energy consumed by the target computing power IT equipment is E7.

[0139] E7 = (E5 + E6) * P 算 *PUE1 / (P 总 *PUE 总 );

[0140] Among them, E5: the data center's commitment to purchase green electricity, in kWh; E6: the data center's expected annual power generation from its self-built green power generation facilities, in kWh; and E7: the annual green energy consumption of the target computing power IT equipment, in kWh.

[0141] Calculation of energy consumption for water resources: If the water resource utilization rate of the target computing power IT equipment is determined to be WUE1, then the water consumption of the target computing power IT equipment is E8.

[0142] E8 = P 算 *PUE1*8760*WUE1*10 -3 ;

[0143] Wherein, E8: the total water resources consumed by the target computing power IT equipment, in tons; WUE1: the water utilization efficiency parameter of the target computing power IT equipment, which is the ratio of the total water consumption of all water-using equipment in the data center to the total electrical energy consumed by all electronic information equipment, in liters per kWh.

[0144] Regarding the calculation of heat energy consumption: This embodiment of the invention predicts the heat consumption from two sources: heating and solar thermal conversion. Heat consumed by municipal heating, centralized heating, etc., increases the carbon emissions of data centers. Heat generated by solar thermal conversion, if used to supply domestic hot water or heating hot water, requires a corresponding reduction in carbon emissions. The calculation methods for the two heat sources are as follows:

[0145] Heating heat consumption E9:

[0146] E9 = F * A * t1 * 0.0036, in GJ;

[0147] Where E9 represents the heating heat consumption of the target computing power area, in tons (t); and F represents the heating load per unit area of ​​the target computing power area, in kW / m². 2 A: Heating area of ​​the target computing power region, in m² 2 t1: Heating time, in hours (h).

[0148] Solar thermal energy generates heat E 10 :

[0149] E 10 =P*S*t²*y*0.0036, in GJ;

[0150] Where P represents the average annual solar irradiance of the city where the data center is located, in kW / m². 2 S: Solar thermal area of ​​the region where the target computing power is located, in m². 2 t2: Solar energy usage time in the area where the target computing power is located, in hours; y: Photothermal conversion efficiency.

[0151] Optionally, considering the abundant waste heat resources in data centers, for air-cooled air conditioning systems, heat from the system condenser or hot aisle of the server room can be recovered; for water-cooled air conditioning systems, heat from the water system can be recovered. Therefore, the predicted carbon emissions of the data center can be reduced based on the carbon emissions corresponding to the waste heat recovery energy. Specifically, the calculation of waste heat recovery is as follows:

[0152] E 11 =c*ρ*V*ΔT*t3*10 -9 ;

[0153] Among them, E 11 : Waste heat recovery in the target computing area: unit is GJ; c: specific heat capacity of air or water, unit is J / (kg·℃); ρ: density of air or water, unit is kg / m³ 3 V: Air or water flow rate in the target computing power area, in meters per second (m³). 3 / h; ΔT: heat exchange temperature difference, in °C; t3: waste heat recovery time in the target computing area, in h.

[0154] Step 30: Determine the predicted carbon emissions of the data center when the target computing power is running based on the energy consumption.

[0155] Specifically, energy consumption is converted based on the carbon emission coefficient corresponding to each energy type to obtain the carbon emission amount corresponding to that energy type. Finally, the carbon emission amount corresponding to the target computing power is used to obtain the predicted carbon emission amount of the data center when the target computing power is running. Furthermore, considering that in addition to consuming various types of energy, data centers may also have energy recycling and reuse, such as using green energy such as solar energy and recovering waste heat for power supply, the total energy consumption of all energy types corresponding to the target computing power can be further reduced based on the energy recycling and reuse of the data center, making the carbon emission prediction of the data center when the target computing power is running more accurate.

[0156] In another embodiment of the present invention, the energy type is multiple; step 30 further includes:

[0157] Step 301: For each of the energy types, determine the carbon emission coefficient corresponding to that energy type.

[0158] The carbon emission factor is used to characterize the amount of carbon emissions produced per unit of energy during combustion or use. Energy types can include municipal electricity, fossil fuels, green energy, water resources, and heat.

[0159] Step 302: Determine the carbon emission amount per energy source corresponding to the energy type based on the carbon emission coefficient and the energy consumption.

[0160] Specifically, the calculation of carbon emissions per energy source under each energy type is explained:

[0161] The carbon emissions from mains electricity during the operation of IT equipment with target computing power are calculated. After adopting green energy, the amount of municipal electricity will be reduced, and the carbon emissions from mains electricity need to be reduced by the green energy portion.

[0162] The carbon emissions (C1) from mains electricity after the reduction are calculated as follows:

[0163] C1 = (E1 - E7) * KC1;

[0164] Wherein, C1 is the amount of carbon emissions from municipal power generation reduced by green energy, in tons; E1 is the predicted energy consumption of municipal power sources, in kWh; E7 is the predicted energy consumption of green energy sources, in kWh; and KC1 is the carbon emission coefficient of municipal power sources at the location of the data center, in tCO2 / kWh.

[0165] The carbon emissions (C2) corresponding to the fossil fuel consumption of data centers during the operation of the target computing power IT equipment are calculated as follows:

[0166] C2 = E4 * KC2;

[0167] Wherein, C2: carbon emissions corresponding to fossil energy, in tons; E4: energy consumption corresponding to the fossil energy type obtained in the previous steps, predicting the annual consumption of fossil energy, in tons; KC2: carbon emission coefficient corresponding to fossil energy, in tons (tCO2 / t); where, when green energy power generation equipment is used as backup power, the C2 part is not included.

[0168] The carbon emissions (C4) generated by the water consumption of IT equipment operating at the target computing power are as follows:

[0169] C4 = E8 * KC3;

[0170] Wherein, C4: carbon emissions corresponding to water energy, in tons; E8: total predicted water resources consumed, in tons; KC3: carbon emission coefficient corresponding to water resource consumption, in tons (tCO2 / t).

[0171] Calculate the heat energy and carbon emissions consumed by IT equipment during operation to calculate the target computing power.

[0172] C5 = (E9 - E 10 )*KC4, unit is t;

[0173] Wherein, C5: carbon emissions corresponding to thermal energy, in tons (t); E9: heat consumed for heating in data centers, in gigabytes (GJ).

[0174] E 10 : Heat generated by solar thermal energy in data centers, in GJ; KC4: Carbon emission factor corresponding to heat energy consumption, in (tCO2 / GJ).

[0175] Step 303: Determine the predicted carbon emissions based on the sum of the carbon emissions of each single energy source corresponding to all the energy types.

[0176] Specifically, the sum of the carbon emissions of each energy type is determined as the predicted carbon emissions. That is, the predicted carbon emissions C = C1 + C2 + C3 + C4 + C5.

[0177] Optionally, step 303 may be followed by:

[0178] Step 3031: Determine the amount of heat recovery from the data center.

[0179] Specifically, heat recovery volume characterizes the amount of heat recycled and reused by the data center, which can include waste heat recovery from the data center and heat obtained through recovery methods, such as solar power generation. The carbon emissions from waste heat recovery corresponding to the target computing power are as follows:

[0180] C6 = E 11 *KC4;

[0181] Where C6: Carbon emissions corresponding to waste heat recovery, in tons (t); E 11 : Waste heat recovery from the data center, in GJ; KC4: Carbon emission factor corresponding to heat energy consumption, in (tCO2 / GJ).

[0182] Step 3032: Calibrate the predicted carbon emissions based on the recovered carbon emissions corresponding to the recovered heat amount.

[0183] Specifically, the recovered carbon emissions are subtracted from the predicted carbon emissions, and the difference is used as the calibrated predicted carbon emissions. That is, the calibrated predicted carbon emissions C. 校 =C1+C2+C3+C4+C5-C6.

[0184] This invention, through its embodiments, determines the predicted power consumption of a target computing power corresponding to a target computing requirement based on the computing power configuration information of a data center. The target computing power is at least one of several selectable computing powers configured within the data center. Based on the predicted power consumption, the energy consumption of the data center under a preset energy type during the operation of the target computing power is determined. Based on the energy consumption, the predicted carbon emissions of the data center during the operation of the target computing power are determined. This enables the determination of the estimated carbon emissions of the data center to meet the target computing requirements, solving the problem that existing carbon emission calculation methods are not applicable to data centers and can only calculate carbon emissions for clearly configured computing equipment and already built data centers, but cannot predict carbon emissions to meet unmet computing requirements or planned data centers. This improves the practicality and accuracy of data center carbon emission prediction.

[0185] Figure 2 A schematic diagram of the structure of a data center carbon emission prediction device provided in an embodiment of the present invention is shown. Figure 2 As shown, the device 40 includes: a first determining module 401, a second determining module 402, and a third determining module 403, wherein the first determining module 401 is used to determine the predicted power consumption of the target computing power corresponding to the target computing demand based on the computing power configuration information of the data center; the target computing power is at least one of a variety of optional computing powers configured in the data center;

[0186] The second determining module 402 is used to determine the energy consumption of the data center under a preset energy type when the target computing power is running, based on the predicted power consumption.

[0187] The third determining module 403 is used to determine the predicted carbon emissions of the data center when the target computing power is running, based on the energy consumption.

[0188] The operation process of the data center carbon emission prediction device provided in this embodiment of the invention is largely the same as that of the aforementioned method embodiment, and will not be described again.

[0189] The data center carbon emission prediction device provided in this invention determines the predicted power consumption of the target computing power corresponding to the target computing demand based on the computing power configuration information of the data center. The target computing power is at least one of a variety of optional computing powers configured in the data center. Based on the predicted power consumption, the device determines the energy consumption of the data center under a preset energy type when the target computing power is running. Based on the energy consumption, the device determines the predicted carbon emissions of the data center when the target computing power is running. This device can determine the expected carbon emissions of the data center to meet the target computing demand, solving the problem that the carbon emission calculation methods in the prior art are not applicable to data centers and can only calculate carbon emissions for clearly configured computing equipment and already built data centers, but cannot predict the carbon emissions to meet unmet computing demands and planned data centers. This device can improve the practicality and accuracy of data center carbon emission prediction.

[0190] Figure 3 The diagram shows a schematic of the structure of a data center carbon emission prediction device provided in an embodiment of the present invention. The specific embodiments of the present invention do not limit the specific implementation of the data center carbon emission prediction device.

[0191] like Figure 3 As shown, the data center carbon emission prediction device may include: a processor 502, a communications interface 504, a memory 506, and a communications bus 508.

[0192] The processor 502, communication interface 504, and memory 506 communicate with each other via communication bus 508. Communication interface 504 is used to communicate with other network elements, such as clients or other servers. Processor 502 executes program 510, specifically performing the relevant steps described in the embodiment of the data center carbon emission prediction method.

[0193] Specifically, program 510 may include program code, which includes computer-executable instructions.

[0194] Processor 502 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention. The data center carbon emission prediction device includes one or more processors, which may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0195] Memory 506 is used to store program 510. Memory 506 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0196] Specifically, program 510 can be called by processor 502 to cause the data center carbon emission prediction device to perform the following operations:

[0197] The predicted power consumption of the target computing power corresponding to the target computing demand is determined based on the computing power configuration information of the data center; the target computing power is at least one of a variety of optional computing powers configured within the data center.

[0198] The energy consumption of the data center under a preset energy type is determined based on the predicted power consumption.

[0199] The predicted carbon emissions of the data center when the target computing power is running are determined based on the energy consumption.

[0200] The operation process of the data center carbon emission prediction device provided in this embodiment of the invention is largely the same as that of the aforementioned method embodiment, and will not be described again.

[0201] The data center carbon emission prediction device provided in this invention determines the predicted power consumption of the target computing power corresponding to the target computing demand based on the computing power configuration information of the data center. The target computing power is at least one of a variety of optional computing powers configured in the data center. Based on the predicted power consumption, the device determines the energy consumption of the data center under a preset energy type when the target computing power is running. Based on the energy consumption, the device determines the predicted carbon emissions of the data center when the target computing power is running. This device can determine the expected carbon emissions of the data center to meet the target computing demand, solving the problem that the carbon emission calculation methods in the prior art are not applicable to data centers and can only calculate carbon emissions for clearly configured computing equipment and already built data centers, but cannot predict the carbon emissions to meet unmet computing demands and planned data centers. This device can improve the practicality and accuracy of data center carbon emission prediction.

[0202] This invention provides a computer-readable storage medium storing at least one executable instruction that, when executed on a data center carbon emission prediction device, causes the data center carbon emission prediction device to perform the data center carbon emission prediction method in any of the above method embodiments.

[0203] Specifically, the executable instructions can be used to cause the data center carbon emission prediction equipment to perform the following operations:

[0204] The predicted power consumption of the target computing power corresponding to the target computing demand is determined based on the computing power configuration information of the data center; the target computing power is at least one of a variety of optional computing powers configured within the data center.

[0205] The energy consumption of the data center under a preset energy type is determined based on the predicted power consumption.

[0206] The predicted carbon emissions of the data center when the target computing power is running are determined based on the energy consumption.

[0207] The operation process of storing executable instructions on the computer storage medium provided in this embodiment of the invention is largely the same as that in the aforementioned method embodiments, and will not be described again.

[0208] The executable instructions stored in the computer storage medium provided in this invention determine the predicted power consumption of the target computing power corresponding to the target computing demand based on the computing power configuration information of the data center. The target computing power is at least one of a variety of optional computing powers configured in the data center. Based on the predicted power consumption, the energy consumption of the data center under a preset energy type when the target computing power is running is determined. Based on the energy consumption, the predicted carbon emissions of the data center when the target computing power is running are determined. This can determine the expected carbon emissions of the data center to meet the target computing demand, solving the problem that the carbon emission calculation methods in the prior art are not applicable to data centers, and can only calculate carbon emissions for clearly configured computing equipment and already built data centers, but cannot predict the carbon emissions to meet unmet computing demands and planned data centers. This can improve the practicality and accuracy of data center carbon emission prediction.

[0209] This invention provides a data center carbon emission prediction device for executing the above-described data center carbon emission prediction method.

[0210] This invention provides a computer program that can be invoked by a processor to cause a data center carbon emission prediction device to execute the data center carbon emission prediction method in any of the above method embodiments.

[0211] This invention provides a computer program product, which includes a computer program stored on a computer-readable storage medium. The computer program includes program instructions that, when executed on a computer, cause the computer to perform the data center carbon emission prediction method in any of the above method embodiments.

[0212] The algorithms or displays provided herein are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used in conjunction with the teachings herein. The required structure for constructing such systems is apparent from the above description. Furthermore, the embodiments of the present invention are not directed to any particular programming language. It should be understood that the content of the invention described herein can be implemented using various programming languages, and the above description of specific languages ​​is for the purpose of disclosing the best mode of implementation of the invention.

[0213] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0214] Similarly, it should be understood that, in order to streamline the invention and aid in understanding one or more of the various aspects of the invention, features of the embodiments of the invention are sometimes grouped together in a single embodiment, figure, or description thereof in the above description of exemplary embodiments of the invention. However, this disclosure should not be construed as reflecting an intention that the claimed invention requires more features than are expressly recited in each claim.

[0215] Those skilled in the art will understand that modules in the device of the embodiments can be adaptively changed and placed in one or more devices different from that embodiment. Modules, units, or components in the embodiments can be combined into a single module, unit, or component, and can be divided into multiple sub-modules, sub-units, or sub-components. Except where at least some of such features and / or processes or units are mutually exclusive, any combination can be used to combine all features disclosed in this specification (including the accompanying claims, abstract, and drawings) and all processes or units of any method or device so disclosed. Unless expressly stated otherwise, each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) may be replaced by an alternative feature that serves the same, equivalent, or similar purpose.

[0216] It should be noted that the above embodiments are illustrative of the invention and not restrictive, and that those skilled in the art can devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses should not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third, etc., does not indicate any order. These words can be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be construed as limiting the order of execution.

Claims

1. A method for predicting carbon emissions from data centers, characterized in that, The method includes: The predicted power consumption of the target computing power corresponding to the target computing demand is determined based on the computing power configuration information of the data center; the target computing power is at least one of a variety of optional computing powers configured in the data center; the computing power configuration information includes the computing performance information of the various optional computing powers; wherein, the target computing power is determined from the various optional computing powers based on the computing performance information; the predicted computing efficiency of the target computing power is determined based on the historical energy consumption data of the target computing power; the target computing demand is converted based on the predicted computing efficiency to obtain the predicted power consumption; The energy consumption of the data center under a preset energy type is determined based on the predicted power consumption. The predicted carbon emissions of the data center during the operation of the target computing power are determined based on the energy consumption. The energy types are multiple; determining the energy consumption of the data center under preset energy types when the target computing power is running based on the predicted power consumption further includes: calculating the predicted power consumption based on the power consumption ratio of the target computing power relative to the data center, the power conversion coefficients corresponding to each of the energy types, and the energy activation status, to obtain the energy consumption; wherein, the predicted power consumption is converted based on the power consumption ratio and the power conversion coefficient to obtain the predicted energy consumption of the data center for the energy type when the target computing power is running; the predicted energy consumption is calculated based on the energy activation status and preset statistical time information to obtain the energy consumption.

2. The method according to claim 1, characterized in that, The step of calculating the predicted power consumption based on the proportion of the target computing power to the power consumption of the data center, the power conversion coefficients corresponding to various energy types, and the energy availability, to obtain the energy consumption, further includes: The actual power consumption of the target computing power is determined based on the first power utilization rate of the target computing power and the predicted power consumption. The overall power consumption of the data center is determined based on the second power utilization rate of the data center and the total power consumption of the data center during operation. The power consumption ratio of the target computing power is determined based on the ratio of the actual power consumption to the total power consumption.

3. The method according to claim 1, characterized in that, The step of determining the predicted carbon emissions of the data center when the target computing power is running based on the energy consumption further includes: For each of the aforementioned energy types, determine the corresponding carbon emission coefficient for that energy type; The carbon emission coefficient and the energy consumption are used to determine the carbon emission of a single energy source corresponding to the energy type. The predicted carbon emissions are determined based on the sum of the carbon emissions of each of the energy types.

4. The method according to claim 3, characterized in that, After determining the predicted carbon emissions based on the sum of the carbon emissions of each of the energy types, the method further includes: Determine the amount of heat recovered from the data center; The predicted carbon emissions are calibrated based on the recovered carbon emissions corresponding to the recovered heat.

5. A data center carbon emission prediction device, characterized in that, The device includes: The first determining module is used to determine the predicted power consumption of the target computing power corresponding to the target computing demand based on the computing power configuration information of the data center; the target computing power is at least one of a variety of optional computing powers configured in the data center; the computing power configuration information includes the computing performance information of the variety of optional computing powers; wherein, the target computing power is determined from the variety of optional computing powers based on the computing performance information; the predicted computing efficiency of the target computing power is determined based on the historical energy consumption data of the target computing power; and the target computing demand is converted based on the predicted computing efficiency to obtain the predicted power consumption. The second determining module is used to determine the energy consumption of the data center under a preset energy type when the target computing power is running, based on the predicted power consumption. The third determining module is used to determine the predicted carbon emissions of the data center when the target computing power is running, based on the energy consumption. The energy types are multiple; the second determining module is further configured to: calculate the predicted power consumption based on the power consumption ratio of the target computing power relative to the data center, the power conversion coefficients corresponding to each of the energy types, and the energy activation status, to obtain the energy consumption; wherein, the predicted power consumption is converted based on the power consumption ratio and the power conversion coefficient to obtain the predicted runtime energy consumption of the data center for the energy type when the target computing power is running; the predicted runtime energy consumption is calculated based on the energy activation status and preset statistical time information to obtain the energy consumption.

6. A data center carbon emission prediction device, characterized in that, include: The processor, memory, communication interface, and communication bus are provided, wherein the processor, memory, and communication interface communicate with each other via the communication bus. The memory is used to store at least one executable instruction that causes the processor to perform the operation of the data center carbon emission prediction method as described in any one of claims 1-4.

7. A computer-readable storage medium, characterized in that, The storage medium stores at least one executable instruction, which, when executed on the data center carbon emission prediction device, causes the data center carbon emission prediction device to perform the operation of the data center carbon emission prediction method as described in any one of claims 1-4.

Images