Low-carbon regulation method, system and device of data center computing load and storage medium

Abstract

The application relates to the technical field of low-carbon regulation of data centers, and finally achieves the goal of carbon neutralization in the industry, and provides a low-carbon regulation method, system and device for the computing load of a data center and a storage medium, and solves the problems of low accuracy and effectiveness of low-carbon regulation. The application collects real-time computing power of each enterprise and real-time carbon price data of a carbon trading market; a digital twin model is constructed based on the two types of data, which can map the power consumption, cooling consumption and carbon emission of the data center in real time; relying on the digital twin model, the change of carbon emission under a preset computing power scheduling strategy is simulated to generate virtual environment feature data; a double-layer game model is designed: the upper unit receives basic data and virtual environment feature data to determine the final carbon quota allocation ratio; the lower unit dynamically adjusts the power supply voltage and frequency of the power consumption equipment, realizes low-carbon regulation of the computing load of the data center, and improves the accuracy and effectiveness of low-carbon regulation of the computing load of the data center.

Description

Technical Field

[0001] This application relates to the field of low-carbon control technology for data centers, and in particular to a method, system, device and storage medium for low-carbon control of data center computing load. Background Technology

[0002] With the rapid development of the digital economy, data centers, as the core infrastructure supporting computing power, exhibit dynamic allocation of computing loads across multiple enterprises and significant peak fluctuations. Simultaneously, according to relevant policies, data centers must achieve coordinated control of computing load and carbon emissions while ensuring the computing power needs of each enterprise. In the current low-carbon control scenario for data center computing load, the core technological requirements are reflected in three aspects: First, real-time acquisition of enterprise computing power data and carbon price data from the carbon trading market is necessary to ensure the timeliness of control measures; second, differentiated allocation of carbon quotas must be implemented based on the differences in computing power usage among different enterprises to avoid imbalances in computing power supply caused by single-mode control; and third, a correlation mechanism must be established between computing load, equipment energy consumption, and carbon emissions to ensure that adjustment measures directly contribute to carbon neutrality goals, balancing the accuracy of control with the quality of computing power services.

[0003] Currently, the industry's targeted solution for addressing the aforementioned technical needs is a "static carbon quota control scheme based on fixed energy consumption standards." This scheme first sets unit computing power energy consumption standards for various types of electrical equipment based on historical data center operating data; then, it evenly distributes the total carbon quota to each enterprise according to their contracted computing power scale, forming a fixed carbon quota; during the control process, it monitors the actual energy consumption of the corresponding equipment in each enterprise. If the energy consumption exceeds the energy consumption limit corresponding to the fixed carbon quota, the power supply of the corresponding equipment in that enterprise is reduced to control carbon emissions, while ensuring the normal use of computing power for enterprises that have not exceeded their quotas.

[0004] The existing scheme has significant flaws: First, carbon quota allocation lacks dynamic adaptability, allocating quotas solely based on contracted computing power rather than real-time computing power demand. When enterprise computing power demand surges temporarily, fixed quotas force equipment power to be reduced, affecting the stability of computing power services. Conversely, when demand decreases, quotas become idle, reducing carbon resource utilization efficiency. Second, it does not link to real-time carbon price data, and regulatory decisions rely solely on fixed energy consumption standards, failing to adjust regulatory strategies according to carbon price fluctuations. When carbon prices are low, it fails to fully utilize low-cost carbon resources to guarantee computing power; when carbon prices are high, it fails to strengthen regulatory efforts to reduce carbon costs, resulting in insufficient economic efficiency in regulation. Third, it does not establish a dynamic mapping relationship between computing power, energy consumption, and carbon emissions. It controls energy consumption solely through adjusting the power of a single device, ignoring the synergistic impact of data center cooling and electricity consumption. For example, reducing equipment power may lead to an imbalance between cooling supply and demand, increasing overall carbon emissions and failing to achieve true carbon neutrality and precise regulation. Summary of the Invention

[0005] The purpose of this application is to provide a low-carbon control method, system, device and storage medium for data center computing load, so as to solve the problems of accuracy and effectiveness of low-carbon control of data center computing load caused by the lack of dynamic adaptation of carbon quota allocation, insufficient economic efficiency and neglect of cooling and power coordination in the prior art.

[0006] To address the aforementioned technical problems, in a first aspect, this application provides a low-carbon control method for data center computing load, comprising:

[0007] Real-time computing power data from various enterprises is collected, and real-time carbon price data from the carbon trading market is obtained. The real-time computing power data and the real-time carbon price data together serve as the basic data for low-carbon regulation of the data center's computing load.

[0008] Based on the real-time computing power data and the real-time carbon price data, a digital twin model is constructed, which maps the power consumption, cooling consumption and carbon emissions of the data center in real time.

[0009] Based on the digital twin model, carbon emission changes under a preset computing power scheduling strategy are simulated to generate virtual environment characteristic data.

[0010] A two-layer game model is designed. The upper-layer unit of the two-layer game model receives the basic data and the virtual environment feature data, and determines the final carbon quota allocation ratio according to the computing power needs of each enterprise. The lower-layer unit dynamically adjusts the power supply voltage and power frequency of the electrical equipment in the data center according to the final carbon quota allocation ratio, so as to achieve low-carbon control of the computing load of the data center.

[0011] Optionally, the two-layer game model is designed such that the upper-layer unit receives the basic data and the virtual environment feature data, determines the final carbon quota allocation ratio based on the computing power needs of each enterprise, and the lower-layer unit dynamically adjusts the power supply voltage and frequency of the electrical equipment in the data center according to the final carbon quota allocation ratio, so as to achieve low-carbon control of the computing load of the data center, including:

[0012] The upper unit of the two-layer game model receives the basic data and the virtual environment feature data, divides the computing power usage range of different enterprises according to the computing power demand declared by each enterprise, and generates the initial carbon quota allocation ratio according to the proportion of the computing power usage range.

[0013] The upper-level unit compares the initial carbon quota allocation ratio with the carbon emission change trend in the virtual environment feature data, and adjusts the initial carbon quota allocation ratio of each enterprise according to the comparison results to form the final carbon quota allocation ratio.

[0014] The lower-level unit determines the power supply adjustment threshold of the electrical equipment based on the final carbon quota allocation ratio;

[0015] Based on the power supply adjustment threshold, the power supply voltage of the electrical equipment is adjusted to a preset threshold range, and the power consumption frequency is adjusted according to the adjusted power supply voltage.

[0016] Based on the adjusted power supply voltage and power consumption frequency, the computing load of the data center is controlled in a low-carbon manner.

[0017] Secondly, this application provides a low-carbon control system for data center computing load, comprising:

[0018] The data acquisition module is used to collect real-time computing power data from various enterprises and obtain real-time carbon price data from the carbon trading market. The real-time computing power data and the real-time carbon price data together serve as the basic data for low-carbon regulation of the data center's computing load.

[0019] A construction module is used to construct a digital twin model based on the real-time computing power data and the real-time carbon price data. The digital twin model maps the power consumption, cooling consumption and carbon emissions of the data center in real time.

[0020] The generation module is used to simulate changes in carbon emissions under a preset computing power scheduling strategy based on the digital twin model, so as to generate virtual environment feature data.

[0021] An adjustment module is used to design a two-layer game model. The upper-layer unit of the two-layer game model receives the basic data and the virtual environment feature data, and determines the final carbon quota allocation ratio according to the computing power needs of each enterprise. The lower-layer unit dynamically adjusts the power supply voltage and power frequency of the electrical equipment in the data center according to the final carbon quota allocation ratio, so as to achieve low-carbon control of the computing load of the data center.

[0022] Thirdly, this application provides an electronic device, comprising:

[0023] Memory, used to store computer programs;

[0024] A processor for executing the computer program to implement the steps of the low-carbon control method for data center computing load as described in the first aspect above.

[0025] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps of the low-carbon control method for data center computing load as described in the first aspect above.

[0026] This application provides a method for low-carbon control of data center computing load, comprising: collecting real-time computing power data from various enterprises and simultaneously acquiring real-time carbon price data from the carbon trading market, with the real-time computing power data and real-time carbon price data serving as the basic data for low-carbon control of data center computing load; constructing a digital twin model based on the real-time computing power data and real-time carbon price data, with the digital twin model mapping the data center's power consumption, cooling consumption, and carbon emissions in real time; simulating carbon emission changes under a preset computing power scheduling strategy based on the digital twin model to generate virtual environment characteristic data; and designing a two-layer game model, where the upper-layer unit receives the basic data and virtual environment characteristic data, determines the final carbon quota allocation ratio based on the computing power needs of each enterprise, and the lower-layer unit dynamically adjusts the power supply voltage and frequency of electrical equipment in the data center according to the final carbon quota allocation ratio to achieve low-carbon control of data center computing load. This application has the following advantages: By collecting real-time computing power data from various enterprises and simultaneously acquiring real-time carbon price data from the carbon trading market, and using both the real-time computing power data and the real-time carbon price data as the foundational data for low-carbon control of data center computing load, it can provide core basis for subsequent low-carbon control based on real-time operating conditions and market carbon prices, avoiding control decisions that are divorced from actual operation and market environment; by constructing a digital twin model based on the real-time computing power data and the real-time carbon price data, and by mapping the data center's power consumption, cooling consumption, and carbon emissions in real time, it can establish a dynamic correlation between computing power, carbon price, and data center energy consumption and carbon emissions, making key data center operating indicators quantifiable. Real-time perception; by simulating carbon emission changes under a preset computing power scheduling strategy based on the digital twin model to generate virtual environment feature data, it is possible to predict carbon emission trends under different computing power scheduling scenarios in advance, providing predictive support for subsequent carbon quota allocation and equipment adjustment; by designing a two-layer game model, the upper unit of the two-layer game model receives the basic data and the virtual environment feature data and determines the final carbon quota allocation ratio according to the computing power needs of each enterprise, and the lower unit dynamically adjusts the power supply voltage and power frequency of the electrical equipment in the data center according to the final carbon quota allocation ratio, which can realize the coordinated linkage of carbon quota allocation and equipment parameter adjustment, and provide a layered execution path for low-carbon regulation.

[0027] Furthermore, in designing the two-layer game model, the upper-layer unit first receives basic data and virtual environment characteristic data. Based on the computing power demand declared by each enterprise, it divides the computing power usage intervals of different enterprises and generates an initial carbon quota allocation ratio according to the interval proportion. Then, it compares and adjusts the initial carbon quota allocation ratio with the carbon emission change trend in the virtual environment characteristic data to form the final carbon quota allocation ratio. The lower-layer unit first associates the final carbon quota allocation ratio with the corresponding power consumption equipment of each enterprise to clarify the carbon quota proportion of each equipment. It sorts the equipment from high to low proportion and sets the corresponding benchmark adjustment coefficient. Based on real-time computing power data, it obtains the current operating power of each equipment and calculates the preliminary voltage adjustment range and preliminary frequency adjustment range by combining the operating power with the benchmark adjustment coefficient. It limits the preliminary adjustment range by combining the carbon emission change trend in the virtual environment characteristic data to define the upper limit of voltage fluctuation, the lower limit of voltage fluctuation, and the frequency fluctuation range. It integrates these to form the power supply adjustment threshold of the power consumption equipment. Then, based on the power supply adjustment threshold, it adjusts the power supply voltage of the power consumption equipment to the preset threshold range and adjusts the power consumption frequency according to the adjusted power supply voltage, ultimately realizing the low-carbon control of the data center computing load. By dynamically generating and adjusting the carbon quota allocation ratio through upper-level units, carbon quotas can be matched with enterprises' real-time computing power needs and carbon emission trends, solving the adaptability problem of static carbon quota allocation. Through lower-level units, carbon quotas are associated with specific electrical equipment, and power supply adjustment thresholds are calculated and optimized in combination with computing power data. Then, the equipment voltage and frequency are precisely adjusted, establishing a precise correspondence between carbon quotas and equipment operating parameters. This avoids imbalances in cooling and power coordination caused by single adjustments, while improving the accuracy of equipment adjustments and ensuring the effectiveness of low-carbon regulation and the stability of computing power services.

[0028] These or other aspects of this application will become more apparent in the following description of the embodiments. Attached Figure Description

[0029] To more clearly illustrate the technical solutions of the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1 A flowchart illustrating a low-carbon control method for data center computing load provided in an embodiment of this application;

[0031] Figure 2 A schematic diagram illustrating a specific implementation of a low-carbon control method for data center computing load provided in this application embodiment;

[0032] Figure 3This is a schematic diagram of a low-carbon control system for data center computing load provided in an embodiment of this application. Detailed Implementation

[0033] To address the issues of lack of dynamic adaptation, insufficient economic efficiency, and neglect of cooling and power synergy in existing carbon quota allocation technologies, this application provides a low-carbon control method for data center computing load. This method employs the following design concept: First, it collects real-time computing power data from each enterprise and simultaneously acquires real-time price data from the carbon trading market; these two types of data serve as the basis for control. Then, using these two types of basic data, it builds a model that corresponds in real-time to the actual operation of the data center and reflects its true state. This model can display the data center's power consumption, cooling capacity, and carbon emissions in real time. Next, it uses this model to simulate how carbon emissions will change under different computing power arrangements, obtaining data that can reflect carbon emission trends in advance. Finally, it designs a two-layer control model. The upper layer, based on the aforementioned basic data and simulated carbon emission trend data, combined with the enterprise's computing power needs, determines the final proportion of carbon quotas allocated to each enterprise. The lower layer then flexibly adjusts the power supply voltage and frequency of electrical equipment in the data center based on this quota proportion. This approach allows carbon quotas to be adjusted in accordance with real-time business needs and carbon emission trends, solving the problem of fixed quotas; it also enables the optimization of control strategies by referencing real-time carbon prices, improving economic efficiency; and it establishes a dynamic correlation between computing power, energy consumption, and carbon emissions, avoiding problems caused by simply adjusting equipment power, ultimately achieving precise low-carbon control of data center computing load.

[0034] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0035] The core of this application is to provide a low-carbon control method for data center computing load, and a flowchart of one specific implementation is shown below. Figure 1 As shown, the method includes:

[0036] S11. Collect real-time computing power data from various enterprises, and simultaneously obtain real-time carbon price data from the carbon trading market. The real-time computing power data and real-time carbon price data together serve as the basic data for low-carbon regulation of data center computing load.

[0037] Among them, real-time computing power data is the relevant information generated by each enterprise when using the computing power of the data center at the moment, including the amount of computing power used by the enterprise when processing various tasks, the task processing speed, etc.; real-time carbon price data is the relevant information of the trading price of each unit of carbon emission rights in the carbon trading market at the current point in time; basic data is a data set formed by integrating the collected real-time computing power data and the acquired real-time carbon price data. This data set is used to provide a basis for subsequent low-carbon regulation of data center computing load. The final generated basic data will serve as the core reference for subsequent model building and regulation strategy formulation.

[0038] In this embodiment, firstly, a data acquisition tool is used to connect to the computing task processing systems of various enterprises within the data center to collect real-time computing power usage data for each enterprise. For example, if a data center has enterprises A, B, and C, enterprise A requires 200 units of computing power per hour to process order generation tasks, enterprise B requires 150 units of computing power per hour to process file storage tasks, and enterprise C requires 100 units of computing power per hour to process data statistics tasks. The hourly computing power usage of these enterprises is recorded in detail. Secondly, by accessing the official information release interface of the carbon trading market, the current carbon price data of the market is obtained in real-time within the same hour as the above-mentioned enterprise computing power data is collected. For example, if the trading price of each unit of carbon emission rights in the carbon trading market is 10 yuan, this price data is recorded. Finally, the recorded real-time computing power data and real-time carbon price data are organized into the same data document and used as the basic data for low-carbon control of the data center's computing load.

[0039] S12. Based on real-time computing power data and real-time carbon price data, construct a digital twin model. The digital twin model maps the power consumption, cooling consumption and carbon emissions of the data center in real time.

[0040] Among them, the digital twin model is a model that can correspond to the actual operating status of the data center in real time. It can reflect the power consumption, cooling consumption and carbon emissions of the data center during operation like a mirror. The power consumption is the total electricity consumed by all electrical equipment in the data center during operation, the cooling consumption is the total cooling consumed by the data center to maintain the normal operating temperature of the equipment, and the carbon emissions are the total amount of greenhouse gases such as carbon dioxide generated by the energy consumption during the operation of the data center.

[0041] S13. Based on the digital twin model, simulate the changes in carbon emissions under the preset computing power scheduling strategy to generate virtual environment characteristic data.

[0042] The virtual environment characteristic data is a dataset compiled by simulating the changes in data center carbon emissions under different computing power scheduling strategies. This dataset contains information on carbon emission fluctuations over different time periods under different strategies. The digital twin model is consistent with the concept in S12 and is a model that can map the data center's power consumption, cooling consumption, and carbon emissions in real time. The preset computing power scheduling strategy is a way to allocate computing power to each enterprise in different time periods based on the data center's operational needs and the characteristics of enterprise business. For example, different amounts of computing power are allocated to enterprises during peak and off-peak periods. The final virtual environment characteristic data will serve as an important reference for determining the carbon quota allocation ratio and help to formulate more reasonable control strategies.

[0043] S14. Design a two-layer game model. The upper-layer unit of the two-layer game model receives basic data and virtual environment characteristic data, and determines the final carbon quota allocation ratio according to the computing power needs of each enterprise. The lower-layer unit dynamically adjusts the power supply voltage and power frequency of the electrical equipment in the data center according to the final carbon quota allocation ratio, so as to achieve low-carbon control of the computing load of the data center, and thus achieve carbon neutrality control.

[0044] The two-layer game model is a control model that operates in two layers. The upper layer is mainly responsible for receiving relevant data and determining the final proportion of carbon allowances allocated to each enterprise. This proportion is the final carbon allowance allocation ratio. The lower layer is mainly responsible for adjusting the power supply voltage and power frequency of the electrical equipment in the data center according to the final carbon allowance allocation ratio. The power supply voltage is the voltage value used by the electrical equipment when it is running, and the power frequency is the frequency value of the current when the electrical equipment is running.

[0045] In this embodiment, the upper-level unit of the two-layer game model is first designed to receive the basic data of S11 and the virtual environment feature data of S13. Then, based on the computing power requirements of each enterprise, the upper-level unit first calculates the proportion of each enterprise's computing power to the total computing power. The calculation process is that the total computing power equals 220 + 140 + 110, resulting in 470 units. Enterprise A's computing power proportion is 220 ÷ 470 ≈ 44%, Enterprise B's computing power proportion is 140 ÷ 470 ≈ 29.8%, and Enterprise C's computing power proportion is 110 ÷ 470 ≈ 23.4%. The initial carbon quota allocation ratio is generated according to this ratio. Then, the upper-level unit compares the initial carbon quota allocation ratio with the virtual environment feature data. The system compares carbon emission changes in environmental characteristic data. For example, it finds that allocating carbon allowances according to the initial ratio will cause data center carbon emissions to exceed the limit during peak business periods. Therefore, the initial ratio is adjusted, reducing the ratio for company A to 42%, increasing the ratio for company B to 30%, and increasing the ratio for company C to 28%, forming the final carbon allowance allocation ratio. Secondly, a two-layer game model is designed with the lower-level unit receiving the final carbon allowance allocation ratio determined by the upper-level unit. This ratio is then associated with the corresponding power consumption equipment of each company. For example, company A's 42% carbon allowance corresponds to server group 1, company B's 30% carbon allowance corresponds to server group 2, and company C's 28% carbon allowance corresponds to server group 3. 3. Define the carbon allowance percentage for each equipment group; then sort the equipment groups from highest to lowest carbon allowance percentage, with server group 1 > server group 2 > server group 3 (28%). Set a corresponding baseline adjustment factor for each sorted equipment group, for example, 1.2 for server group 1, 1.0 for server group 2, and 0.8 for server group 3. Then, use the real-time computing power data of S11 to obtain the current operating power of each equipment group, for example, 2000 watts for server group 1, 1500 watts for server group 2, and 1000 watts for server group 3. Multiply the operating power of each equipment group by the corresponding baseline adjustment factor to obtain the basic adjustment value. The calculation process is as follows: Server group 1 has a voltage limit of 2000 × 1.2 = 2400, server group 2 has a voltage limit of 1500 × 1.0 = 1500, and server group 3 has a voltage limit of 1000 × 0.8 = 800. Based on the basic adjustment values, the initial voltage adjustment range and the initial frequency adjustment range are obtained. Then, based on the carbon emission change trend in the virtual environment characteristic data, the initial voltage adjustment range and the initial frequency adjustment range are limited. For example, the upper limit of voltage for server group 1 is reduced to 225 volts and the upper limit of frequency is reduced to 50.8 Hz, the upper limit of voltage for server group 2 is reduced to 215 volts and the upper limit of frequency is reduced to 50.3 Hz, and the lower limit of voltage for server group 3 is increased to 195 volts and the lower limit of frequency is increased to 49 Hz.At 8 Hz, the upper and lower limits of voltage fluctuation and the frequency fluctuation range for each equipment group are defined, determining the power supply adjustment threshold for each equipment group. Finally, the lower-level unit adjusts the power supply voltage of each equipment group to the preset threshold range based on the power supply adjustment threshold. For example, server group 1 is adjusted to 220 volts, server group 2 to 210 volts, and server group 3 to 200 volts. Then, based on the adjusted power supply voltage, the power frequency is adjusted, for example, server group 1 to 50.5 Hz, server group 2 to 50.2 Hz, and server group 3 to 50 Hz. By controlling the energy consumption and carbon emissions of each equipment group through the adjusted power supply voltage and power frequency, low-carbon regulation of the data center's computing load is ultimately achieved.

[0046] Figure 2 This is a schematic diagram illustrating a specific implementation of a low-carbon control method for data center computing load provided in this application embodiment. Its process is similar to... Figure 1 The process is similar, so it will not be repeated here. Figure 2The provided method and process are illustrated in the following specific example: When a data center conducts low-carbon control of computing load, step S11 is executed first: Real-time computing power data of enterprises A, B, and C are automatically collected every 15 minutes through deployed computing power monitoring software. During one collection, enterprise A's computing power usage increased from 200 units to 220 units due to an increase in order processing volume, enterprise B's computing power usage decreased from 150 units to 140 units due to a decrease in file storage, and enterprise C's computing power remained stable at 110 units. At the same time, through data connection with the carbon trading platform, real-time carbon price data of 11 yuan per unit is obtained at the same time point, and these data are summarized into the control data folder as basic data. Next, execute step S12: Associate the real-time computing power data in the basic data with the equipment, determine that A corresponds to server group 1 with 3 servers, B corresponds to server group 2 with 2 servers, and C corresponds to server group 3 with 1 server. Tests show that server group 1 consumes 10 kWh of electricity for every 20 units of computing power increase. Therefore, when A has 220 units of computing power, the electricity consumption of server group 1 is calculated to be 110 kWh. Combined with the real-time carbon price of 11 yuan, it is determined that each kWh of electricity generates 0.45 units of carbon emissions and requires 4 units of cooling capacity for every 10 kWh of electricity. Thus, the cooling capacity of server group 1 is calculated to be 110 / 10×4=44 units, and the carbon emissions are 53.3 units. Similarly, the data for server group 2 and server group 3 are calculated and integrated into a digital twin model. Then execute step S13: Based on the enterprise's business characteristics (A: morning peak, B: afternoon peak, C: no peak), set two computing power scheduling strategies. Strategy 1 is A230 units, B160 units, and C120 units from 9:00 to 17:00, and A180 units, B130 units, and C90 units for other periods. Input strategy 1 into the digital twin model to simulate the following from 9:00 to 17:00, A230 units of computing power corresponds to server group 1 with 115 kWh of electricity ((230-180) / 20×10+90), 46 units of cooling, and 59.9 units of carbon emissions; B160 units corresponds to server group 2 with 80 kWh of electricity, 32 units of cooling, and 42.4 units of carbon emissions; and C120 units corresponds to server group 3 with 60 kWh of electricity, 24 units of cooling, and 31.8 units of carbon emissions. The total carbon emissions for this period are 134.1 units, and the total carbon emissions for other periods are 106.3 units. Similarly, simulate the data of strategy 2 and organize it into virtual environment characteristic data. Finally, step S14 is executed: a two-layer game model is designed. The upper-layer unit receives basic data and virtual environment characteristic data, calculates the initial carbon quota ratio, and adjusts it to a final ratio of A 42%, B 30%, and C 28% after comparing with the virtual data. The lower-layer unit associates the ratio with the equipment group, sets the benchmark adjustment coefficient according to the percentage, obtains the current power of the equipment group, calculates the basic adjustment value, determines the preliminary adjustment range, and obtains the power supply adjustment threshold after combining the carbon emission trend limit. Finally, the voltage of group A is adjusted to 220 volts and the frequency to 50.6 Hz, the voltage of group B to 210 volts and the frequency to 50.3 Hz, and the voltage of group C to 205 volts and the frequency to 50.2 Hz, thus achieving low-carbon regulation.

[0047] By executing steps S11 to S14, this embodiment of the application collects real-time computing power data of enterprises and real-time carbon price data of the carbon trading market in step S11 and integrates them into basic data. This provides a basis for subsequent regulation that is in line with actual operation and market dynamics, avoiding regulation deviation from needs due to outdated data. Step S12 constructs a digital twin model based on the basic data, which maps electricity, cooling consumption and carbon emissions in real time. This allows staff to intuitively grasp the impact of computing power changes on energy consumption and carbon emissions, solving the problem of not being able to clearly perceive the internal operating status of the data center in the past. Step S13 simulates carbon emission changes under different computing power scheduling strategies through the digital twin model and generates virtual environment characteristic data. This allows for early prediction of carbon emission trends and provides a forward-looking reference for carbon quota allocation, avoiding situations where carbon emissions exceed the limit or computing power is insufficient after quota allocation. Step S14 designs a two-layer game model. The upper-layer unit dynamically adjusts the carbon quota allocation ratio, solving the problem of poor adaptability of traditional fixed quotas. The lower-layer unit accurately adjusts the equipment voltage and frequency based on the quota, avoiding cooling and power imbalance caused by adjusting power alone. This ensures the enterprise's computing power needs while effectively controlling carbon emissions. The coordinated efforts of each step ultimately achieve precision, flexibility, and foresight in the low-carbon regulation of data center computing load, ensuring that the regulation work can not only match the real-time needs of enterprises, but also align with carbon neutrality goals and market carbon price dynamics.

[0048] In one possible embodiment, S14, a two-layer game model is designed. The upper-layer unit of the two-layer game model receives basic data and virtual environment characteristic data, and determines the final carbon quota allocation ratio based on the computing power needs of each enterprise. The lower-layer unit dynamically adjusts the power supply voltage and power frequency of the electrical equipment in the data center according to the final carbon quota allocation ratio, so as to achieve low-carbon control of the data center's computing load, including:

[0049] Step 141: The upper unit of the two-layer game model receives the basic data and virtual environment feature data, divides the computing power usage range of different enterprises according to the computing power demand declared by each enterprise, and generates the initial carbon quota allocation ratio according to the proportion of the computing power usage range.

[0050] The upper-level unit of the two-layer game model is responsible for receiving relevant data and initially determining the carbon quota allocation ratio. It is mainly used to process information related to carbon quota allocation. The basic data is a collection of real-time computing power data of various enterprises and real-time carbon price data of the carbon trading market, which is used to provide a basis for quota allocation. The virtual environment characteristic data is carbon emission change-related data obtained by simulating different computing power scheduling strategies through digital twin models, which is used to help judge the rationality of quota allocation. The computing power demand declared by enterprises is the application for computing power usage submitted by enterprises according to their own business needs, including the scale of computing power expected to be used by enterprises. The computing power usage range is the exclusive computing power usage range of each enterprise divided according to the computing power demand declared by enterprises. The initial carbon quota allocation ratio is the proportion of carbon quota that each enterprise should obtain, which is initially determined by the upper-level unit according to the proportion of each enterprise's computing power usage range to the total computing power usage range of all enterprises. This ratio is the basis for subsequent carbon quota adjustments, and the final generated initial carbon quota allocation ratio will serve as the initial basis for further adjustments by the upper-level unit.

[0051] Step 142: The upper-level unit compares the initial carbon quota allocation ratio with the carbon emission change trend in the virtual environment characteristic data, and adjusts the initial carbon quota allocation ratio of each enterprise according to the comparison results to form the final carbon quota allocation ratio.

[0052] Among them, the carbon emission change trend in the virtual environment characteristic data is the trend of data center carbon emissions over time or computing power under different computing power scheduling strategies recorded in the data, such as the increase of carbon emissions during peak business periods and the decrease during off-peak periods; the final carbon quota allocation ratio is the determined quota ratio obtained by the upper-level unit after comparing the initial carbon quota allocation ratio with the carbon emission change trend. This ratio will serve as the direct basis for the lower-level unit to adjust the parameters of power equipment, ensuring that subsequent equipment adjustments can meet the carbon neutrality target.

[0053] Step 143: The lower-level unit determines the power supply adjustment threshold for electrical equipment based on the final carbon quota allocation ratio.

[0054] The lower-level unit is the module in the two-layer game model responsible for adjusting the parameters of electrical equipment according to the carbon quota ratio. It mainly handles adjustment operations related to electrical equipment. Electrical equipment refers to the equipment in the data center that provides computing power support to enterprises, such as server groups composed of multiple servers. Each enterprise has its own dedicated electrical equipment. The benchmark adjustment coefficient is a coefficient set by the lower-level unit based on the carbon quota ratio corresponding to the electrical equipment to calculate the equipment adjustment parameters. The coefficient is usually different for different ratios. The operating power is the power value of the electrical equipment when it is currently working, which can be obtained from the basic data. The initial voltage adjustment range and the initial frequency adjustment range are the initial adjustable ranges of the equipment voltage and frequency calculated by the lower-level unit through the operating power and the benchmark adjustment coefficient. The power supply adjustment threshold is the final adjustable range of the power supply voltage and frequency of the electrical equipment formed by the lower-level unit after limiting the initial adjustment range. It includes the upper limit of voltage fluctuation, the lower limit of voltage fluctuation, and the frequency fluctuation range, which is used to regulate the subsequent adjustment of equipment parameters.

[0055] Step 144: Based on the power supply adjustment threshold, adjust the power supply voltage of the electrical equipment to the preset threshold range, and adjust the power consumption frequency according to the adjusted power supply voltage.

[0056] Among them, the power supply voltage is the voltage value required for the operation of electrical equipment. The preset threshold range is the range between the upper limit and the lower limit of voltage fluctuation in the power supply adjustment threshold. The power frequency is the frequency value of the current when the electrical equipment is running. Its adjustment needs to be determined based on the adjusted power supply voltage to ensure that the adjusted frequency and voltage match each other and ensure the stable operation of the equipment. Finally, the adjusted power supply voltage and power frequency will be used as the parameters for the operation of the equipment.

[0057] Step 145: Based on the adjusted power supply voltage and adjusted power consumption frequency, perform low-carbon regulation of the data center computing load.

[0058] Among them, data center computing load is the total amount of computing tasks undertaken by all electrical equipment in the data center during operation. Its size is related to the voltage and frequency of the equipment. When the voltage and frequency are matched, the computing load can better meet the needs of enterprises. Low-carbon regulation is a process of controlling the operating parameters of the equipment to balance the carbon emissions of the data center with the emission reduction measures, and ultimately achieves the carbon emission target requirements. The adjusted power supply voltage and power frequency are the key parameters for achieving this regulation.

[0059] This application provides the following specific example: First, step 141 is executed: The upper unit of the two-layer game model receives basic data including enterprise A's declaration of 180-230 units of computing power, enterprise B's declaration of 130-180 units of computing power, and enterprise C's declaration of 80-130 units of computing power, as well as real-time carbon price data of 11 yuan per unit. At the same time, it receives virtual environmental characteristic data including the upward trend of carbon emissions during peak business periods. The computing power usage range is divided according to the declaration requirements: enterprise A is 180-230 units, enterprise B is 130-180 units, and enterprise C is 80-130 units. The total range is calculated as 50+50+50=150 units. The initial carbon quota allocation ratio of 33.3% for each of enterprises A, B, and C is generated according to the ratio of 50 / 150≈33.3%. Next, step 142 is executed: the upper-level unit retrieves the initial ratio and extracts the trend from the virtual environment feature data that "the carbon emissions of the corresponding equipment of A exceed the standard, and the quotas of B and C are remaining" according to the initial ratio during the working days from 9:00 to 17:00 (peak period of enterprise A). After comparison, the ratio is adjusted, reducing A to 28%, increasing B to 36%, and increasing C to 36%, thus forming the final carbon quota allocation ratio. Then, step 143 is executed: the lower-level unit receives the final proportions, with associated enterprise A corresponding to server group 1 (28%), B corresponding to server group 2 (36%), and C corresponding to server group 3 (36%); the proportions are sorted as server group 2 = server group 3 > server group 1, and the baseline adjustment coefficients for groups 2 and 3 are set to 1.1, while that for group 1 is 0.9; the operating power is obtained from the basic data, the basic adjustment values ​​are calculated, and the preliminary voltage range and preliminary frequency range are obtained; combined with the carbon emission trend limit, the upper limit of voltage for groups 2 and 3 is reduced to 230 volts and the upper limit of frequency is reduced to 51 Hz, while the lower limit of voltage for group 1 is increased to 200 volts and the lower limit of frequency is increased to 49.9 Hz, forming the power supply adjustment threshold. Then, step 144 is executed: the lower-level unit retrieves the thresholds and adjusts the voltage of group 1 to 210 volts and groups 2 and 3 to 215 volts; according to the voltage adjustment frequency, group 1 is 50 Hz and groups 2 and 3 are 50.3 Hz. Finally, step 145 is executed: each server group runs according to the adjusted parameters. Group 1 meets the computing power requirements of A with 210 volts and 50 Hz and has low energy consumption and low emissions. Groups 2 and 3 match the requirements of B and C with 215 volts and 50.3 Hz and have no ineffective energy consumption. In the end, the data center computing load is flexibly adapted to the needs of enterprises, and carbon emissions are controlled within the target range to achieve low-carbon regulation.

[0060] By executing steps 141 to 145, this embodiment of the application receives basic data and virtual environment characteristic data in step 141, divides the computing power range according to the enterprise's declaration requirements, and generates an initial carbon quota ratio to ensure that the initial quota is related to the actual needs of the enterprise and avoids blind allocation; step 142 compares and adjusts the initial ratio with the carbon emission trend to form a final quota ratio that is adapted to the carbon neutrality target, solving the problem of carbon emission exceeding the standard or quota waste that may be caused by the initial quota; step 143 associates the final ratio with the power-consuming equipment, calculates and limits the power supply adjustment threshold by combining the operating power and the benchmark adjustment coefficient, so that the equipment adjustment range is in line with the carbon quota target and avoids deviation in the adjustment direction; step 144 adjusts the voltage and frequency according to the threshold to ensure that the equipment parameters are adapted and operate stably, providing a reliable equipment foundation for regulation; step 145 allows the equipment to operate according to the adjusted parameters to achieve coordinated control of computing load and carbon emissions. The synergistic effect of the above steps not only ensures the computing power needs of enterprises but also precisely controls the carbon emissions of data centers, effectively solving problems such as poor adaptability of traditional quota allocation, blind equipment adjustment, and uncontrolled carbon emissions, and ultimately achieving the accuracy, stability, and rationality of low-carbon regulation of data center computing load.

[0061] In one possible embodiment, step 143, the lower-level unit determines the power supply adjustment threshold for the electrical equipment based on the final carbon quota allocation ratio, including:

[0062] a1. The lower-level unit will link the final carbon quota allocation ratio with the corresponding electrical equipment of each enterprise, and clarify the carbon quota ratio of each electrical equipment.

[0063] The lower-level unit is the module in the two-layer game model responsible for handling the relevant operations of power equipment adjustment based on the final carbon quota allocation ratio. The final carbon quota allocation ratio is the proportion of carbon quotas that each enterprise should obtain, determined by the upper-level unit based on the enterprise's computing power needs and carbon emission trends. The power equipment corresponding to each enterprise is the equipment in the data center that provides computing power support for each enterprise, such as a server group composed of multiple servers. The carbon quota ratio corresponding to each power equipment is the quota ratio corresponding to each power equipment after associating the final carbon quota allocation ratio with the power equipment. This ratio is the key basis for determining the equipment adjustment parameters in the future. Finally, the carbon quota ratio corresponding to each power equipment is generated through association operations.

[0064] a2. Sort each electrical device according to its carbon quota percentage from high to low, and set a corresponding benchmark adjustment coefficient for each sorted electrical device.

[0065] The sorting from high to low is based on the carbon quota percentage of each device in descending order. If the percentages are the same, they are sorted together. The benchmark adjustment coefficient is a coefficient set based on the carbon quota percentage of the devices and is used to calculate the adjustment range of the devices. Generally, the higher the percentage, the larger the coefficient, so as to match the adjustment needs of different devices and finally generate the benchmark adjustment coefficient for each device.

[0066] a3. Based on real-time computing power data, obtain the current operating power of each electrical device, calculate the operating power with the reference adjustment coefficient, and obtain the preliminary voltage adjustment range and the preliminary frequency adjustment range.

[0067] Among them, the real-time computing power data is the data previously collected, reflecting the current computing power usage of each enterprise. It includes the operating load information of electrical equipment, from which the operating power of the equipment can be extracted. The current operating power of each electrical device is the power value of the device in its current working state, which can reflect the actual load of the device. The preliminary voltage adjustment range is the initial range of adjustable voltage of the device calculated by the operating power of the device and the reference adjustment coefficient. The preliminary frequency adjustment range is the initial range of adjustable frequency of the device, which is also calculated by the above. Both serve as the basis for subsequent optimization and adjustment ranges, and finally generate the preliminary voltage and frequency adjustment ranges of each device.

[0068] a4. Based on the carbon emission change trend in the virtual environment characteristic data, limit the initial voltage adjustment range and the initial frequency adjustment range respectively to define the upper limit of voltage fluctuation, the lower limit of voltage fluctuation, and the frequency fluctuation range.

[0069] Among them, the virtual environment feature data is data containing carbon emission change information obtained by simulating different computing power scheduling scenarios through digital twin models; the carbon emission change trend is extracted from the virtual environment feature data, showing the trend of carbon emissions changing over time or computing power, such as carbon emissions increasing during peak business periods and decreasing during off-peak periods; the limiting process is to adjust the upper and lower limits of the initial adjustment range according to the carbon emission change trend, making it more in line with the carbon neutrality target; the voltage fluctuation upper limit and voltage fluctuation lower limit are the highest and lowest adjustable values ​​of the device voltage determined after the limiting process; the frequency fluctuation range is the adjustable range of the device frequency determined after the limiting process, ultimately generating the upper and lower limits of voltage fluctuation and the frequency fluctuation range for each device.

[0070] a5. Integrate the upper limit of voltage fluctuation, the lower limit of voltage fluctuation, and the upper and lower limits of frequency fluctuation range into the power supply adjustment threshold for electrical equipment.

[0071] Among them, the upper limit and lower limit of voltage fluctuation are the voltage adjustment boundaries of each device determined after the limiting process in step a4; the frequency fluctuation range is the frequency adjustment range of each device determined after the limiting process in step a4; the power supply adjustment threshold of the electrical equipment is the final adjustment standard formed by integrating the upper limit and lower limit of voltage fluctuation and the frequency fluctuation range of each device. This threshold clarifies the adjustable range of the power supply voltage and frequency of the equipment and is the direct basis for subsequent precise adjustment of equipment parameters, ultimately generating the power supply adjustment threshold of each device.

[0072] This application provides the following specific example: A data center needs to determine the power supply adjustment threshold for electrical equipment. The specific process is as follows: First, step a1 is executed. The lower-level unit receives the final carbon quota allocation ratio from the upper-level unit (Company A 28%, Company B 36%, Company C 36%), retrieves the "Company-Equipment" association record, and finds that Company A corresponds to server group 1, Company B corresponds to server group 2, and Company C corresponds to server group 3. The quota ratio is then bound to the equipment, specifying that server group 1 accounts for 28%, server group 2 accounts for 36%, and server group 3 accounts for 36%. Next, step a2 is executed. The carbon quota ratio is sorted from high to low as server group 2 = server group 3 > server group 1. A baseline adjustment coefficient of 1 is set for server groups 2 and 3. .1. Set 0.9 for server group 1; then execute step a3 to extract the current operating power of each device from the real-time computing power data, calculate the basic adjustment value, and determine the initial adjustment range with reference to the device manual; then execute step a4 to analyze the virtual environment characteristic data, extract the trend of "daily carbon emission peak from 10-16 o'clock, operation within the initial range will exceed the standard", and limit the initial range (server group 1: voltage lower limit raised to 200 volts, frequency lower limit raised to 49.9 Hz; server groups 2 and 3: voltage upper limit lowered to 230 volts, frequency upper limit lowered to 51 Hz), clarify the upper and lower limits of voltage fluctuation and frequency fluctuation range of each device; finally execute step a5 to integrate the parameters of each device to form the power supply adjustment threshold, and complete the entire threshold determination process.

[0073] By executing steps a1 to a5, this embodiment of the application achieves precise binding between carbon quotas and power-consuming equipment through step a1, preventing subsequent equipment adjustments from deviating from carbon quota targets; step a2 sets differentiated benchmark adjustment coefficients according to proportions, solving the problem of "uniform coefficients leading to mismatches between adjustment intensity and quota requirements"; step a3 calculates the initial adjustment range based on real-time operating power, ensuring that the adjustment range closely matches the actual load of the equipment and avoiding blind adjustments out of operating condition; step a4 limits the adjustment range based on carbon emission trends, enabling the adjustment range to adapt to the low-carbon control needs of different time periods, preventing carbon emission exceedances or inefficient equipment operation; step a5 integrates parameters to form clear thresholds, providing a unified and executable standard for subsequent equipment voltage and frequency adjustments. The entire process deeply integrates carbon quotas, equipment load, carbon emission trends, and adjustment ranges, ensuring that the final determined power supply adjustment thresholds not only meet the carbon neutrality target of the data center but also adapt to the actual operating needs of the equipment, laying a reliable foundation for subsequent precise equipment control and achieving carbon neutrality of computing load.

[0074] In one possible embodiment, a3, based on real-time computing power data, obtains the current operating power of each electrical device, calculates the operating power with a reference adjustment coefficient, and obtains a preliminary voltage adjustment range and a preliminary frequency adjustment range, including:

[0075] b1. Based on real-time computing power data, extract the computing power usage records corresponding to each electrical device, and obtain the current operating power of each electrical device from the computing power usage records.

[0076] Among them, real-time computing power data is data previously collected reflecting the current computing power usage of data centers by various enterprises, including information related to the operation of application equipment; computing power usage records are usage records directly associated with each piece of equipment extracted from real-time computing power data, which contain details of computing power consumption during equipment operation; the current operating power of each piece of equipment is the power value of the equipment in its current working state, which reflects the actual operating load of the equipment, and can be extracted from the computing power usage records based on the correspondence between computing power consumption and power; finally, by extracting computing power usage records and obtaining current operating power, basic data that fits the actual load is provided for subsequent adjustment of computing equipment parameters.

[0077] b2. Based on the functional type of electrical equipment, set corresponding voltage conversion ratios and frequency conversion ratios for different types of electrical equipment.

[0078] The functional type of the electrical equipment is categorized based on the core tasks it performs in the data center, such as equipment specifically designed for processing computing tasks for enterprises or equipment specifically designed to provide cooling to maintain equipment temperature in the data center. The voltage conversion ratio is a proportional value set according to the functional type of the equipment, used to convert the basic adjustment values ​​into a voltage adjustment range that the equipment can adapt to. The frequency conversion ratio is also a proportional value set according to the functional type of the equipment, used to convert the basic adjustment values ​​into a frequency adjustment range that the equipment can adapt to. These two ratios will vary depending on the function of the equipment. Ultimately, by setting differentiated ratios, a basis is provided for calculating the adjustment range for equipment with different functions.

[0079] b3. Calculate the current operating power of each electrical device and the corresponding benchmark adjustment coefficient to obtain the basic adjustment value of each electrical device.

[0080] The benchmark adjustment factor is a value previously set based on the carbon quota ratio of each device, used to reflect the difference in adjustment intensity among different devices. The basic adjustment value is the result obtained by multiplying the current operating power of each device by the corresponding benchmark adjustment factor. This value integrates the actual load of the device and the adjustment intensity requirement, and is the core basic data for subsequent calculation of the preliminary voltage and frequency adjustment range. Finally, the basic adjustment value for each device is calculated.

[0081] b4. Multiply the basic adjustment value by the corresponding voltage conversion ratio to obtain the preliminary voltage adjustment range of each electrical device. Multiply the basic adjustment value by the corresponding frequency conversion ratio to obtain the preliminary frequency adjustment range of each electrical device.

[0082] The initial voltage adjustment range is obtained by multiplying the basic adjustment value of each device by the corresponding voltage conversion ratio, and then combining it with the current voltage of the device. The initial frequency adjustment range is obtained by multiplying the basic adjustment value of each device by the corresponding frequency conversion ratio, and then combining it with the current frequency of the device. These two ranges are the initial basis for subsequent limiting processing based on carbon emission trends, and ultimately generate the initial voltage and frequency adjustment ranges for each device.

[0083] This application provides the following specific example: When a data center conducts preliminary adjustment calculations for power equipment, the overall process is as follows: First, step b1 is executed. Staff use data processing software to extract the computing power usage records of each power device from real-time computing power data. For example, the record for server group 1 for company A shows "current hourly computing power consumption 210 units," the record for server group 2 for company B shows "current hourly computing power consumption 230 units," and the record for the cooling equipment group for company C shows "current hourly computing power consumption 180 units." Then, based on the preset "computing power consumption - operating power" correspondence rule, the current operating power of server group 1 is obtained as 2100 watts, server group 22300 watts, and cooling equipment group 1800 watts, respectively. Next, step b2 is executed, classifying server groups 1 and 2 as "computing equipment" and the cooling equipment group as "cooling equipment." Referring to the equipment characteristics, the voltage conversion ratio for computing equipment is set to 0.1, and the frequency conversion ratio is set to 0.002; the voltage conversion ratio for cooling equipment is set to 0.08, and the frequency conversion ratio is set to... 0.0015; then execute step b3 to retrieve the baseline adjustment coefficients for each device and calculate the basic adjustment values: Server group 1 is 2100 × 0.9 = 1890, Server group 2 is 2300 × 1.1 = 2530, and the cooling equipment group is 1800 × 1.0 = 1800; finally, execute step b4 to confirm the current voltage and frequency of each device and calculate the preliminary voltage adjustment range: Server group 1 is 200 ± (1890 × 0.1) = 181.1 to 218.9 volts, and Server group 2 is 235.3 volts. The voltage for the cooling equipment group is 190±(1800×0.08)=175.6 to 204.4 volts; the initial frequency adjustment range is calculated as follows: server group 1 is 50±(1890×0.002)=49.622 to 50.378 Hz, server group 2 is 50±(2530×0.002)=49.494 to 50.506 Hz, and the cooling equipment group is 50±(1800×0.0015)=49.73 to 50.27 Hz, thus completing the entire initial adjustment range calculation process.

[0084] By executing steps b1 to b4, this embodiment of the application extracts the device computing power usage record from the real-time computing power data and obtains the current operating power in step b1, establishing a direct correlation between computing power data and the actual load of the device, avoiding parameter failure due to subsequent calculations being detached from the device status; step b2 sets differentiated voltage and frequency conversion ratios according to the device function type, solving the problem of mismatch between a uniform ratio and device function, ensuring that the adjustment range can adapt to the operating characteristics of different devices; step b3 multiplies the current power of the device with the benchmark adjustment coefficient to obtain the basic adjustment value, integrating the adjustment intensity requirements related to the device load and carbon quota, providing core data that takes into account both needs for subsequent calculations; step b4 calculates the preliminary adjustment range based on the basic adjustment value and the function adaptation ratio, combined with the current voltage and frequency of the device. The generated range not only fits the actual operating status of the device, but also matches the carbon neutrality target and the device functional requirements, avoiding the problem of blindly setting the adjustment range. The entire process deeply integrates computing power, device load, carbon quota adjustment intensity, and device function, providing a reasonable and accurate initial basis for subsequent optimization of the adjustment range based on carbon emission trends, ensuring that subsequent regulation work can be carried out in an orderly manner around the actual device situation and the carbon neutrality target.

[0085] In one possible embodiment, S13, based on a digital twin model, simulates changes in carbon emissions under a preset computing power scheduling strategy to generate virtual environment characteristic data, including:

[0086] Step 131: Based on the digital twin model, according to the allocated time periods in the preset computing power scheduling strategy, simulate the changes in power consumption and cooling consumption corresponding to the enterprise's computing power usage in each allocated time period.

[0087] The digital twin model is a virtual model that simulates the actual operation of a data center, reflecting changes in equipment operating parameters, resource allocation, and enterprise computing power consumption. The preset computing power scheduling strategy is a pre-defined plan for allocating computing power, including allocation periods that divide a day or scheduling cycle into multiple specific time intervals. Allocation periods are specific time intervals defined in the computing power scheduling strategy for phased allocation of enterprise computing power. Enterprise computing power usage refers to the scale and changes in actual computing power used by each enterprise within each allocation period. Power consumption change data is real-time information on the power consumption of data center equipment as enterprise computing power usage changes. Cooling consumption change data is real-time information on the cooling system's cooling capacity consumption as it changes with computing power usage to maintain normal equipment operating temperatures. Finally, through time-segmented simulation using the digital twin model, power consumption change data and cooling consumption change data corresponding to enterprise computing power usage are generated for each allocation period.

[0088] Step 132: Calculate the carbon emission change value for each allocated time period based on the electricity consumption change data and the cooling consumption change data for each allocated time period.

[0089] The carbon emission change values ​​are calculated based on the changes in electricity and cooling consumption in each time period, combined with a pre-determined conversion coefficient of "carbon emissions per unit of electricity / cooling". The real-time change values ​​of carbon emissions in each allocated time period are calculated and generated, providing core content for subsequent integration into complete environmental data.

[0090] Step 133: Integrate the carbon emission change values ​​for each allocation period to form virtual environmental characteristic data.

[0091] The carbon emission change values ​​for each allocation period are calculated in step 132, representing the specific changes in carbon emissions over time within each allocation period, including carbon emission data at key time points such as the beginning, peak, and end of the period. The virtual environment characteristic data is formed by integrating the carbon emission change values ​​for all allocation periods in chronological order and supplementing the data of key nodes within each period, resulting in a dataset that fully reflects the carbon emission change trend throughout the entire scheduling cycle. Finally, by integrating the data from each period, a virtual environment characteristic data set covering the entire scheduling cycle and with coherent data is formed, providing a trend basis for subsequent optimization of carbon quota allocation and equipment adjustment.

[0092] This application provides the following specific examples: First, step 131 is executed to call the digital twin model and retrieve the preset computing power scheduling strategy; simulating the period from 8:00 to 12:00, computing power is allocated to enterprises A (200 units), B (150 units), and C (100 units) according to the strategy. Based on the rule that "every 10 units of computing power corresponds to 10 watts of electricity and 6 units of cooling capacity", the electricity is calculated to increase from 500 watts to 500 + (200 + 150 + 100) / 10 × 10 = 1000 watts, and the cooling capacity is calculated to increase from 300 units to 300 + (200 + 150 + 100) / 10 × 6 = 600 units; similarly, simulating the period from 12:00 to 18:00, the electricity is obtained as 1000-1600 watts and the cooling capacity as 600-900 units; simulating the period from 18:00 to 8:00 the next day, the electricity is obtained as 1600-600 watts and the cooling capacity as 900-360 units. Next, step 132 is executed, setting the carbon emission factor for electricity to 0.5 units of carbon per watt and for cooling to 0.3 units of carbon per unit of cooling. Carbon emissions are calculated for the period 8:00-12:00: initial 500×0.5 + 300×0.3 = 340 units, ending 1000×0.5 + 600×0.3 = 680 units. For the period 12:00-18:00, initial 1000×0.5 + 600×0.3 = 680 units, ending 1600×0.5 + 900×0.3 = 1070 units. For the period 18:00-8:00 the next day, initial 1600×0.5 + 900×0.3 = 1070 units, ending 600×0.5 + 360×0.3 = 408 units. Finally, step 133 is executed, connecting the carbon emission data from each period in chronological order to form virtual environmental characteristic data.

[0093] By executing steps 131 to 133, this embodiment of the application uses a digital twin model in step 131 to simulate the changes in electricity and cooling consumption corresponding to enterprise computing power usage in different time periods. This transforms the abstract computing power scheduling strategy into specific resource consumption data, providing basic information that fits the actual operating scenario for subsequent carbon emission calculations and avoiding carbon emission calculation errors caused by a lack of consumption data. Step 132 converts the consumption data into carbon emission change values ​​by setting conversion coefficients, establishing a direct correlation between resource consumption and carbon emissions. This makes the originally scattered consumption data a core indicator that can intuitively reflect environmental impact, providing key content for subsequent trend analysis. Step 133 integrates carbon emission data from different time periods and supplements key node information to form virtual environmental characteristic data that covers the entire scheduling cycle and is data-coherent. This clearly presents the overall trend of carbon emission changes and avoids the one-sidedness of trend analysis caused by data dispersion. The entire process, from consumption simulation to carbon emission conversion and then to data integration, forms a complete environmental characteristic data chain. This provides a comprehensive and accurate basis for optimizing carbon quota allocation ratios and adjusting equipment operating parameters based on carbon emission trends, ensuring that subsequent control work can align with the carbon emission change patterns of the data center and contribute to the achievement of carbon neutrality goals.

[0094] In one possible embodiment, step 132, calculating the carbon emission change value for the corresponding allocation period based on the electricity consumption change data and cooling consumption change data for each allocation period, includes:

[0095] c1. Based on the electricity consumption change data for each allocation period, calculate the change in electricity carbon emissions for the corresponding period according to the proportion of carbon emissions generated by electricity during the allocation period.

[0096] The data on power consumption changes in each allocation period were previously simulated using a digital twin model. This data includes information on the changes in power consumption by data center equipment within each allocation period, including the power consumption at the beginning and end of the period, as well as changes in between. The carbon emission ratio generated by electricity during each allocation period represents the carbon emissions generated per unit of electricity consumed during that period. This ratio varies depending on the type of energy used (e.g., green energy, conventional energy) within the period. The change in electricity carbon emissions is calculated by combining the power consumption change data for each allocation period with the corresponding electricity carbon emission ratio, representing the change in carbon emissions due to changes in power consumption within that period. The final generated change in electricity carbon emissions for each allocation period will provide carbon emission data related to power consumption for the subsequent aggregation of total carbon emission changes for each period.

[0097] c2. Based on the data on the change in cooling consumption during each allocation period, the change in carbon emissions of cooling during the corresponding period is calculated according to the proportion of carbon emissions generated by cooling during the allocation period.

[0098] The data on power consumption changes in each allocation period were previously simulated using a digital twin model. This data includes information on the changes in power consumption by data center equipment within each allocation period, including the power consumption at the beginning and end of the period, as well as changes in between. The carbon emission ratio generated by electricity during each allocation period represents the carbon emissions generated per unit of electricity consumed during that period. This ratio varies depending on the type of energy used (e.g., green energy, conventional energy) within the period. The change in electricity carbon emissions is calculated by combining the power consumption change data for each allocation period with the corresponding electricity carbon emission ratio, representing the change in carbon emissions due to changes in power consumption within that period. The final generated change in electricity carbon emissions for each allocation period will provide carbon emission data related to power consumption for the subsequent aggregation of total carbon emission changes for each period.

[0099] c3. Add the changes in carbon emissions from electricity and cooling during the same allocation period to obtain the carbon emission change value for the corresponding allocation period.

[0100] The carbon emission change value is obtained by adding the change in electricity carbon emissions and the change in cooling carbon emissions in the same allocated time period, representing the change in the total carbon emissions of the data center during that time period. The carbon emission change values ​​generated for each allocated time period will provide complete information on carbon emission changes during the time period for subsequent integration into virtual environment feature data in chronological order, ensuring that the overall carbon emission changes of the data center during that time period are reflected.

[0101] This application provides the following specific example: A data center conducts numerical calculations of carbon emission changes during various time periods. The overall process is as follows: First, step c1 is executed to retrieve power consumption change data for three time periods: 500 watts from 8:00 to 12:00, 1000 watts from 12:00 to 18:00, and 1600 watts from 18:00 to 8:00 the next day; the carbon emission ratio of electricity for each time period is determined based on energy usage records. The carbon emission rate for electricity is set at 0.4 for the period from 00:00 to 12:00 due to the high proportion of green energy, 0.6 for the period from 12:00 to 18:00 due to the high proportion of conventional energy, and 0.3 for the period from 18:00 to 8:00 the next day due to the rebound in the proportion of green energy. The changes in carbon emissions for electricity in each period are calculated as (1000-500)×0.4=200 units of carbon, (1600-1000)×0.6=360 units of carbon, and (600-1600)×0.3=-300 units of carbon. Next, execute step c2 to retrieve the data on the changes in cooling capacity consumption for each time period: 300 units from 8:00 to 12:00, 600 units from 12:00 to 18:00, 900 units from 18:00 to 360 units from 18:00 to 8:00 the next day. Determine the carbon emission ratio of cooling capacity for each time period based on the operating efficiency of the refrigeration system: 0.3 for the period with high efficiency from 8:00 to 12:00, 0.4 for the period with decreased efficiency from 12:00 to 18:00, and 0.2 for the period with increased efficiency from 18:00 to 8:00 the next day. Calculate the changes in carbon emissions for each time period: (600-300)×0.3=90 units of carbon, (900-600)×0.4=120 units of carbon, and (360-900)×0.2=-108 units of carbon. Finally, step c3 is executed, which adds up the changes in carbon emissions for electricity and cooling during the same time period. The change is 200 + 90 = 290 units of carbon for the period from 8:00 to 12:00, 360 + 120 = 480 units of carbon for the period from 12:00 to 18:00, and -300 + (-108) = -408 units of carbon for the period from 18:00 to 8:00 the next day. This gives the carbon emission change values ​​for each allocated time period.

[0102] By executing steps c1 through c3, this embodiment calculates changes in electricity carbon emissions by combining energy type differences across different time periods in step c1, avoiding calculation errors caused by ignoring the influence of energy type and ensuring that the impact of electricity consumption on carbon emissions is accurately reflected. Step c2 calculates changes in cooling capacity carbon emissions by combining differences in cooling system operating efficiency, supplementing the impact of cooling capacity consumption—a core energy consumption of data centers—on carbon emissions, and avoiding incomplete carbon emission calculations due to the lack of cooling capacity factors. Step c3 integrates the changes in electricity and cooling capacity carbon emissions within the same time period by summing them, forming complete data reflecting the total carbon emission changes over the period, avoiding the loss of overall carbon emission trends caused by splitting data. The entire process, from accurate calculation of individual factors to overall data integration, provides accurate and complete time-period carbon emission data for subsequent integration into virtual environment characteristic data in chronological order, ensuring that subsequent analysis of data center carbon emission trends can truly reflect the overall operating situation and providing reliable data support for formulating practical low-carbon control strategies.

[0103] In one possible embodiment, S12, based on real-time computing power data and real-time carbon price data, a digital twin model is constructed, including:

[0104] Step 121: Associate the real-time computing power data with the operation records of each power-consuming device in the data center to determine the relationship between the computing power usage of different enterprises and the power-consuming devices.

[0105] Real-time computing power data is a collection of information reflecting the current computing power usage of various enterprises within the data center, including enterprise name, computing power scale used, and usage time. The operation records of each power device within the data center are a collection of information recording the working status of each power device (such as server groups providing computing services or cooling equipment maintaining equipment temperature), including device number, running time, load status, and the enterprise it serves. The correlation between the computing power usage of different enterprises and the power devices they use refers to clarifying the specific power devices supporting a particular enterprise's computing power usage. Finally, by associating real-time computing power data with the power device operation records, this correlation is generated, laying the foundation for establishing a link between computing power and energy consumption.

[0106] Step 122: Based on the correlation, set the corresponding ratio between computing power usage and power consumption of the applied electrical equipment so that changes in computing power reflect changes in power consumption, and set the conversion rules between power consumption and cooling consumption.

[0107] Among them, the correlation relationship is the correspondence between enterprise computing power usage and power consumption equipment determined in step 121, which is the basis for setting subsequent ratios and rules; the correspondence ratio between computing power usage and power consumption of the applied power equipment refers to clarifying the numerical relationship of how much power the applied power equipment will consume for each certain scale of computing power used. Through this ratio, changes in computing power can be directly reflected in changes in power consumption; the conversion rule between power consumption and cooling consumption refers to clarifying the rule of how much cooling energy needs to be added to maintain the normal operating temperature of the equipment for each certain amount of power consumed. Through this rule, changes in power consumption can be reflected in changes in cooling consumption; finally, these two ratios and rules are generated to establish the correspondence between computing power, power consumption, and cooling consumption, so that the changes of the three are interconnected and can be derived from each other.

[0108] Step 123: Introduce real-time carbon price data and set the conversion method between electricity consumption, cooling consumption and carbon emissions according to the energy carbon emission level corresponding to the carbon price.

[0109] Among them, real-time carbon price data is the trading price information of each unit of carbon emission rights in the carbon trading market. This data affects the carbon emission level of the energy used by the data center. The energy carbon emission level corresponding to the carbon price refers to the amount of carbon emissions generated in the process of producing a unit of electricity or generating a unit of cooling capacity under the current real-time carbon price. The conversion method between electricity consumption, cooling capacity consumption and carbon emissions refers to the calculation method that clarifies how much carbon emissions are generated for each certain amount of electricity or cooling capacity consumed. Finally, by introducing real-time carbon price data and combining it with the corresponding energy carbon emission level, this conversion method is set up to establish a direct link between electricity consumption, cooling capacity consumption and carbon emissions, so that energy consumption data can be converted into carbon emission data.

[0110] Step 124: Integrate the corresponding ratios, conversion rules, and conversion methods to generate a digital twin model that can reflect electricity consumption, cooling consumption, and carbon emissions in real time.

[0111] Among them, the digital twin model is a virtual model that can simulate the actual operating state of a data center. By integrating the corresponding proportions, conversion rules and conversion methods mentioned above, the model can automatically calculate and display the corresponding power consumption, cooling consumption and carbon emissions in real time after inputting the enterprise's real-time computing power data. Finally, this digital twin model that can reflect the changes of the three in real time provides a visual and computable virtual tool for the subsequent operation monitoring and control of the data center.

[0112] This application provides the following specific example: The overall process of building a digital twin model for a data center is as follows: First, step 121 is executed, which retrieves real-time computing power data and power equipment operation records through the data management system. By matching enterprise names, the association relationships between enterprise A and server group 1, enterprise B and server group 2, and enterprise C and server group 3 are determined. Next, step 122 is executed, which statistically analyzes server group 1 for three consecutive days based on this association relationship, recording the computing power and power consumption at different times. It is found that every 10 units of computing power corresponds to 100 watts of power, and the ratio of "10 units of computing power = 100 watts of power" is set. Then, the power and cooling consumption are statistically analyzed, and it is found that every 100 watts of power requires 50 units of cooling to maintain the equipment temperature, and the rule of "100 watts of power = 50 units of cooling" is set. The ratios and rules for server groups 2 and 3 are set in the same way. Then execute steps 123 to obtain the real-time carbon price of 10 yuan / unit carbon emission through the official interface of the carbon trading market. Query the energy carbon emission level under this carbon price: 100 watts of electricity generates 0.05 units of carbon emission, and 50 units of cooling generates 0.03 units of carbon emission. Based on this, set the conversion method: electricity carbon emission = electricity consumption ÷ 100 × 0.05, cooling carbon emission = cooling consumption ÷ 50 × 0.03. Finally, execute step 124, input the ratio, rules, and conversion method into the virtual modeling tool, and write the linkage logic of "computing power → electricity → cooling capacity → carbon emissions"; during testing, input the current computing power of enterprise A as 200 units. The model first calculates the electricity: 200 ÷ 10 × 100 = 2000 watts, then calculates the cooling capacity: 2000 ÷ 100 × 50 = 1000 units, and finally calculates the carbon emissions: (2000 ÷ 100 × 0.05) + (1000 ÷ 50 × 0.03) = 1 + 0.6 = 1.6 units. The results are displayed in real time, completing the construction of the digital twin model.

[0113] By executing steps 121 to 124, this embodiment of the application clarifies the relationship between enterprise computing power and electrical equipment in step 121, solving the problem of lacking clear equipment attribution in subsequent rule settings and laying a precise data correspondence foundation for the entire process; step 122 establishes the linkage relationship between computing power and electricity, and electricity and cooling capacity, transforming abstract computing power requirements into specific energy consumption data, filling the gap in the connection between "computing power requirements - actual energy consumption", and allowing changes in energy consumption to be derived from changes in computing power; step 123 introduces real-time carbon prices and sets the conversion method between energy consumption and carbon emissions, enabling energy consumption data to directly reflect environmental impact, aligning with the core objective of low-carbon regulation, and avoiding the disconnect between energy consumption control and carbon emission targets; step 124 integrates all rules to construct a digital twin model, realizing the visualized linkage of "computing power input - multi-indicator output", avoiding the regulation difficulties caused by data fragmentation. By closely integrating enterprise needs, equipment operation, energy consumption control, and carbon emission targets, the generated model can intuitively reflect the data center's operating status, providing a reliable virtual tool for subsequent precise control of computing power allocation and carbon emission control. This ensures that control decisions are aligned with actual operating conditions and helps data centers gradually achieve carbon neutrality goals.

[0114] Figure 3 A schematic diagram of a low-carbon control system for data center computing load provided in this application embodiment is shown below. Figure 3 As shown, the system includes:

[0115] The data acquisition module 31 is used to collect real-time computing power data from various enterprises and obtain real-time carbon price data from the carbon trading market. The real-time computing power data and real-time carbon price data together serve as the basic data for low-carbon regulation of the data center's computing load.

[0116] Module 32 is used to build a digital twin model based on real-time computing power data and real-time carbon price data. The digital twin model maps the power consumption, cooling consumption and carbon emissions of the data center in real time.

[0117] The generation module 33 is used to simulate the changes in carbon emissions under a preset computing power scheduling strategy based on the digital twin model, so as to generate virtual environment feature data.

[0118] The adjustment module 34 is used to design a two-layer game model. The upper unit of the two-layer game model receives basic data and virtual environment characteristic data, and determines the final carbon quota allocation ratio according to the computing power needs of each enterprise. The lower unit dynamically adjusts the power supply voltage and power frequency of the electrical equipment in the data center according to the final carbon quota allocation ratio, so as to achieve low-carbon control of the computing load of the data center.

[0119] The low-carbon control system for data center computing load in this application embodiment is used to implement the aforementioned low-carbon control method for data center computing load. Therefore, the specific implementation of the low-carbon control system for data center computing load can be found in the embodiment section of the low-carbon control method for data center computing load above. The specific implementation can be referred to the description of the corresponding embodiments, which will not be repeated here.

[0120] This application also provides an electronic device, comprising: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of any of the above-described methods for low-carbon control of data center computing load.

[0121] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of any of the above-described methods for low-carbon control of data center computing load.

[0122] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, random access memory, portable hard drives, magnetic disks, or optical disks.

[0123] Embodiments of the present invention also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the above embodiments of the low-carbon control method for data center computing load.

[0124] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0125] The above provides a detailed description of a low-carbon control method, system, device, and storage medium for data center computing load provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A low-carbon control method for data center computing load, characterized in that, include: Real-time computing power data from various enterprises is collected, and real-time carbon price data from the carbon trading market is obtained. The real-time computing power data and the real-time carbon price data together serve as the basic data for low-carbon regulation of the data center's computing load. Based on the real-time computing power data and the real-time carbon price data, a digital twin model is constructed, which maps the power consumption, cooling consumption and carbon emissions of the data center in real time. Based on the digital twin model, carbon emission changes under a preset computing power scheduling strategy are simulated to generate virtual environment characteristic data. A two-layer game model is designed. The upper-layer unit of the two-layer game model receives the basic data and the virtual environment feature data, and determines the final carbon quota allocation ratio according to the computing power needs of each enterprise. The lower-layer unit dynamically adjusts the power supply voltage and power frequency of the electrical equipment in the data center according to the final carbon quota allocation ratio, so as to achieve low-carbon control of the computing load of the data center.

2. The low-carbon control method for data center computing load according to claim 1, characterized in that, The design of the two-layer game model involves the upper-layer unit receiving the basic data and the virtual environment feature data, determining the final carbon quota allocation ratio based on the computing power needs of each enterprise, and the lower-layer unit dynamically adjusting the power supply voltage and frequency of the electrical equipment in the data center according to the final carbon quota allocation ratio, in order to achieve low-carbon control of the computing load of the data center, including: The upper unit of the two-layer game model receives the basic data and the virtual environment feature data, divides the computing power usage range of different enterprises according to the computing power demand declared by each enterprise, and generates the initial carbon quota allocation ratio according to the proportion of the computing power usage range. The upper-level unit compares the initial carbon quota allocation ratio with the carbon emission change trend in the virtual environment feature data, and adjusts the initial carbon quota allocation ratio of each enterprise according to the comparison results to form the final carbon quota allocation ratio. The lower-level unit determines the power supply adjustment threshold of the electrical equipment based on the final carbon quota allocation ratio; Based on the power supply adjustment threshold, the power supply voltage of the electrical equipment is adjusted to a preset threshold range, and the power consumption frequency is adjusted according to the adjusted power supply voltage. Based on the adjusted power supply voltage and power consumption frequency, the computing load of the data center is controlled in a low-carbon manner.

3. The low-carbon control method for data center computing load according to claim 2, characterized in that, The lower-level unit determines the power supply adjustment threshold for the electrical equipment based on the final carbon quota allocation ratio, including: The lower-level unit associates the final carbon quota allocation ratio with the corresponding electrical equipment of each enterprise, and clarifies the carbon quota percentage corresponding to each electrical equipment. The electrical equipment is sorted from high to low according to the carbon quota ratio, and a corresponding benchmark adjustment coefficient is set for each sorted electrical equipment. Based on the real-time computing power data, the current operating power of each of the electrical devices is obtained, and the operating power is calculated with the reference adjustment coefficient to obtain the preliminary voltage adjustment range and the preliminary frequency adjustment range. Based on the carbon emission change trend in the virtual environment feature data, the initial voltage adjustment range and the initial frequency adjustment range are respectively limited to define the upper limit of voltage fluctuation, the lower limit of voltage fluctuation, and the frequency fluctuation range. The upper limit of voltage fluctuation, the lower limit of voltage fluctuation, and the upper and lower limits of frequency fluctuation range are integrated into the power supply adjustment threshold of the electrical equipment.

4. The low-carbon control method for data center computing load according to claim 3, characterized in that, Based on the real-time computing power data, the current operating power of each of the electrical devices is obtained. The operating power is then calculated with the reference adjustment coefficient to obtain the preliminary voltage adjustment range and the preliminary frequency adjustment range, including: Based on the real-time computing power data, the computing power usage records corresponding to each of the electrical devices are extracted, and the current operating power of each of the electrical devices is obtained from the computing power usage records; Based on the functional type of electrical equipment, corresponding voltage conversion ratios and frequency conversion ratios are set for different types of electrical equipment; The current operating power of each of the electrical devices is calculated with the corresponding benchmark adjustment coefficient to obtain the basic adjustment value of each of the electrical devices; Multiplying the basic adjustment value by the corresponding voltage conversion ratio yields the preliminary voltage adjustment range for each of the electrical devices. Multiplying the basic adjustment value by the corresponding frequency conversion ratio yields the preliminary frequency adjustment range for each of the electrical devices.

5. The low-carbon control method for data center computing load according to claim 1, characterized in that, The process of simulating carbon emission changes under a preset computing power scheduling strategy based on the digital twin model to generate virtual environment characteristic data includes: Based on the digital twin model, according to the allocated time periods in the preset computing power scheduling strategy, the changes in power consumption and cooling consumption corresponding to the enterprise's computing power usage in each allocated time period are simulated sequentially. Based on the changes in electricity consumption and cooling consumption for each allocation period, calculate the changes in carbon emissions for the corresponding allocation period. The carbon emission change values ​​for each allocation period are integrated to form virtual environmental characteristic data.

6. The low-carbon control method for data center computing load according to claim 5, characterized in that, The step of calculating the carbon emission change value for each allocated time period based on the electricity consumption change data and the cooling consumption change data for each allocated time period includes: Based on the data on changes in electricity consumption during each allocation period, the change in carbon emissions from electricity during that allocation period is calculated. Based on the data on changes in cooling consumption during each allocation period, and according to the proportion of carbon emissions generated by cooling during each allocation period, the change in carbon emissions of cooling during the corresponding period is calculated. The change in carbon emissions from electricity and the change in carbon emissions from cooling are added together during the same allocation period to obtain the carbon emission change value for the corresponding allocation period.

7. The low-carbon control method for data center computing load according to claim 1, characterized in that, The construction of a digital twin model based on the real-time computing power data and the real-time carbon price data includes: The real-time computing power data is correlated with the operation records of various electrical equipment in the data center to determine the correlation between the computing power usage of different enterprises and the application of electrical equipment. Based on the aforementioned relationship, a corresponding ratio is set between computing power usage and the power consumption of the applied electrical equipment, so that changes in computing power reflect changes in power consumption, and a conversion rule is set between power consumption and cooling consumption. By introducing the real-time carbon price data, and setting the conversion method between electricity consumption, cooling consumption and carbon emissions according to the energy carbon emission level corresponding to the carbon price; By integrating the corresponding ratios, the conversion rules, and the conversion methods, a digital twin model can be generated that reflects the electricity consumption, the cooling consumption, and the carbon emissions in real time.

8. A low-carbon control system for data center computing load, characterized in that, include: The data acquisition module is used to collect real-time computing power data from various enterprises and obtain real-time carbon price data from the carbon trading market. The real-time computing power data and the real-time carbon price data together serve as the basic data for low-carbon regulation of the data center's computing load. A construction module is used to construct a digital twin model based on the real-time computing power data and the real-time carbon price data. The digital twin model maps the power consumption, cooling consumption and carbon emissions of the data center in real time. The generation module is used to simulate changes in carbon emissions under a preset computing power scheduling strategy based on the digital twin model, so as to generate virtual environment feature data. An adjustment module is used to design a two-layer game model. The upper-layer unit of the two-layer game model receives the basic data and the virtual environment feature data, and determines the final carbon quota allocation ratio according to the computing power needs of each enterprise. The lower-layer unit dynamically adjusts the power supply voltage and power frequency of the electrical equipment in the data center according to the final carbon quota allocation ratio, so as to achieve low-carbon control of the computing load of the data center.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, used to execute the computer program to implement the steps of the low-carbon control method for data center computing load as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, enables the implementation of a low-carbon control method for data center computing load as described in any one of claims 1 to 7.

Images