Data center green electricity purchasing method, system and device based on multi-source power coordination and medium

Abstract

The invention discloses a data center green electricity purchasing method, system and device based on multi-source electric power coordination and a medium, and relates to the technical field of energy management, and the method comprises the steps: collecting multi-source operation data, carrying out the abnormal value detection, and correcting the abnormal value to obtain an input data set; based on the input data set, establishing a data center power consumption demand prediction model, and obtaining a data center power consumption demand prediction quantity in a target period; constructing a multi-source electric power coordination optimization model based on the electricity demand prediction quantity and the renewable energy output information, and calculating the optimal distribution proportion of each type of electric power source; generating a plurality of groups of green electricity purchasing strategies under the condition of satisfying each constraint according to the optimal distribution proportion and in combination with the external operation environment characteristics; and evaluating the green electricity purchasing strategy, and screening out an optimal target purchasing scheme. According to the method, the technical problems that a systematic optimization model is lacked and local resource endowment and multi-channel power collaborative adaptation are not fully considered in traditional research are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of energy management, in particular to a data center green electricity procurement method, system, device and medium based on multi-source power coordination. BACKGROUND

[0002] Under the dual driving of rapid iteration of artificial intelligence technology and vigorous development of digital economy, data centers, as the core carriers of computing power infrastructure, are entering a new stage of large-scale layout and intensive development. With the deepening of the "double carbon" goal, green electricity transformation policies are becoming increasingly stringent. Data centers, as high energy-consuming subjects, are required to increase the proportion of green electricity and help new energy consumption, becoming an important participant in energy transformation. Under this background, the algorithm and electricity coordination mode has become a hot research and practice direction because it can realize the dynamic matching of computing power demand and power supply and promote the local consumption of new energy.

[0003] Existing researches mainly focus on data center energy efficiency optimization, single green electricity procurement mode exploration, algorithm and electricity coordination scheduling strategy and comprehensive energy efficiency management platform construction, and there is still room for research in policy adaptive green electricity economic procurement model construction and green electricity proportion constraint under power procurement scheme optimization. On the one hand, there is a lack of linkage prediction mechanism of computing power and power demand. Existing researches mostly separately carry out computing power scheduling optimization or power load prediction, and do not dynamically optimize and adjust policy energy efficiency constraints, resulting in insufficient matching degree of power demand estimation and actual computing power running state, which is difficult to support accurate planning of green electricity procurement quantity. On the other hand, a green electricity procurement technology system that can systematically respond to multi-level policy requirements has not been established. Existing researches mostly focus on single policy index interpretation or single power purchase channel analysis, and cannot build a collaborative optimization architecture integrating policy constraints, multi-channel power supply characteristics and system operation stability In view of the above problems, the present application provides a data center green electricity procurement system and method based on multi-source power coordination, which establishes a power demand dynamic prediction model linked with data center business load, constructs a target optimization function integrating policy constraints and multi-channel green electricity procurement conditions, and realizes deep matching of new energy power generation characteristics and data center multi-element power procurement strategy based on regional resource characteristics. The system can improve the policy compliance and regional adaptability of the green electricity procurement scheme, and provide reliable technical support for power resource scheduling of data centers in complex policy environment. SUMMARY

[0004] In view of the above problems, the present application provides a data center green electricity procurement method, system, device and medium based on multi-source power coordination.

[0005] Therefore, the problem to be solved by the present application is that existing researches are mostly separately carried out for power scheduling optimization or power load prediction, without dynamic optimization adjustment of policy energy efficiency constraints, resulting in insufficient matching degree between power demand calculation and actual power running state, and difficulty in supporting accurate planning of green electricity procurement quantity.

[0006] To solve the above technical problems, the present application provides the following technical solutions: a data center green electricity procurement method based on multi-source power coordination, which comprises: collecting multi-source running data and performing outlier detection to correct outliers to obtain an input data set; based on the input data set, establishing a data center power demand prediction model to obtain a data center power demand prediction value in a target period; based on the power demand prediction value and renewable energy output information, constructing a multi-source power coordination optimization model to calculate the optimal allocation ratio of each type of power source; according to the optimal allocation ratio, combining external operating environment characteristics, generating a plurality of groups of green electricity procurement strategies under the condition of meeting each constraint; and evaluating the green electricity procurement strategies to select the optimal target procurement scheme.

[0007] As a preferred scheme of the data center green electricity procurement method based on multi-source power coordination, the input data set is obtained by: obtaining multi-source running data related to data center power consumption, and performing unified format conversion and time sequence alignment on data from different sources to construct an original data set; performing outlier identification processing on the original data set to identify abnormal data that does not match normal operation characteristics in numerical distribution through a deviation detection mechanism; and performing correction processing on the detected abnormal data to obtain the input data set.

[0008] As a preferred scheme of the data center green electricity procurement method based on multi-source power coordination, the data center power demand prediction value in the target period is obtained by: calculating a prediction value based on the input data set, including a device annual maximum load prediction value and a device maximum load annual operating hour prediction value; introducing a comprehensive energy efficiency level and an hour fluctuation value as a correction factor to dynamically adjust the prediction value; and substituting the prediction value and the correction factor into a power demand prediction model to calculate the total data center power demand in the target prediction period.

[0009] The beneficial effects of the preferred technical scheme are: the present application combines historical load, operating hours and energy efficiency data at the power demand prediction layer to construct a dynamic prediction model, and through quantitative calculation and correction, accurately predicts the data center power demand to provide dynamic data support for target optimization and scheme generation.

[0010] As a preferred scheme of the data center green electricity procurement method based on multi-source power coordination, the multi-source power coordination optimization model comprises a first optimization objective function, which is used to represent the data center power consumption performance index under different power source combinations, and the optimization direction is to minimize the energy cost; and a second optimization objective function, which is used to maximize the use proportion of renewable energy, and the part of each type of power source that can be counted into the green electricity proportion is used as the optimization direction, the contribution degree of renewable energy in the total power consumption under different combination strategies is quantified, and is used as the core evaluation index of the objective function.

[0011] The beneficial effect of the preferred technical scheme is that, in view of the limitation of the existing optimization model, the data center energy cost optimization and the maximum green electricity consumption proportion are taken as the double core objectives, multi-dimensional constraint conditions such as power balance, policy constraint threshold, energy efficiency standard requirement and the like are integrated, a multi-objective collaborative optimization model is constructed, and the collaborative balance of power procurement compliance and power supply stability is realized.

[0012] As a preferred scheme of the data center green electricity procurement method based on multi-source power coordination, the constraint conditions comprise: setting an energy efficiency constraint condition for limiting the comprehensive energy efficiency of the data center in the target period to be not more than the maximum threshold of the specified requirement; setting a green electricity proportion constraint condition for specifying that the total amount of renewable energy used by the data center accounts for not less than the specified index in the overall power consumption; setting a power balance constraint for requiring the total amount of power supply of each type of power source to be consistent with the predicted power demand; and setting a non-negativity constraint for requiring all predicted power demands to be greater than or equal to zero.

[0013] As a preferred scheme of the data center green electricity procurement method based on multi-source power coordination, the generation of the plurality of groups of green electricity procurement strategies comprises: taking the calculated optimal allocation proportion of each type of power source as a basic parameter, and discretizing the time scale of the target period to construct an initial pool of combination schemes of multi-source power supply under different procurement modes; introducing external operating environment characteristics related to the region where the data center is located for different combination schemes to construct a strategy adaptability evaluation mechanism, and evaluating the matching degree of different combination schemes under the target operating background; and generating a plurality of candidate green electricity procurement strategies under the premise that each scheme satisfies the constraint condition.

[0014] As a preferred embodiment of the data center green electricity procurement method based on multi-source power coordination described in this invention, the step of selecting the optimal target procurement scheme includes: setting a performance evaluation index system for candidate green electricity procurement strategies, including two categories of indicators: engineering feasibility and operational controllability; using a weighted scoring method to comprehensively score each candidate strategy based on preset index weights; and selecting the group with the highest score among the candidate strategies as the final green electricity procurement scheme for the target period based on the comprehensive scoring results.

[0015] To address the aforementioned technical problems, this invention provides the following technical solution: a data center green electricity procurement system based on multi-source power coordination, comprising: a data acquisition module, a demand forecasting module, an allocation ratio calculation module, and a strategy generation module; the data acquisition module is used to collect multi-source operating data, perform outlier detection, and correct outliers to obtain an input dataset; the demand forecasting module, based on the input dataset, establishes a data center electricity demand forecasting model to obtain the predicted data center electricity demand within a target period; the allocation ratio calculation module, based on the predicted electricity demand and renewable energy output information, constructs a multi-source power coordination optimization model to calculate the optimal allocation ratio for various power sources; the strategy generation module, based on the optimal allocation ratio and combined with external operating environment characteristics, generates several sets of green electricity procurement strategies under various constraints, evaluates the green electricity procurement strategies, and selects the optimal target procurement scheme.

[0016] A computer device includes a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of a data center green electricity procurement method based on multi-source power coordination as described above.

[0017] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a data center green electricity procurement method based on multi-source power coordination as described above.

[0018] The beneficial effects of this invention are as follows: The green electricity procurement system and method for data centers based on multi-source power coordination proposed in this invention solves the technical problems of traditional research lacking systematic optimization models and failing to fully consider local resource endowments and multi-channel power synergy adaptation. Addressing the challenges faced by data centers in green electricity compliance, multi-source power coordination, and regional resource integration, this invention constructs a technical architecture integrating data acquisition, load forecasting, target optimization, and solution generation by integrating dynamic business load forecasting, policy constraint analysis, local wind and solar resource endowments, and multi-channel power data. It enables dynamic forecasting of data center power demand, optimized generation of multi-source power coordination solutions under multiple constraints, and improves the synergistic absorption efficiency of local renewable energy and externally procured power, providing technical support for the refined scheduling of power resources in the green and low-carbon transformation of data centers. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart of a data center green electricity procurement method based on multi-source power coordination in Example 1. Detailed Implementation

[0021] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0022] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0023] Example 1, referring to Figure 1 This is the first embodiment of the present invention, which provides a method for green electricity procurement for data centers based on multi-source power coordination, including: S1: Collect multi-source runtime data, perform outlier detection, correct outliers, and obtain the input dataset.

[0024] S2: Based on the input dataset, establish a data center power demand forecasting model to obtain the data center power demand forecast within the target period.

[0025] S3: Based on electricity demand forecasts and renewable energy output information, construct a multi-source power coordination optimization model to calculate the optimal allocation ratio of various power sources.

[0026] S4: Based on the optimal allocation ratio and combined with the characteristics of the external operating environment, generate several sets of green electricity procurement strategies while satisfying various constraints.

[0027] S5: Evaluate green electricity procurement strategies and select the optimal target procurement plan.

[0028] It should be noted that existing methods are mostly based on single-objective optimization or static allocation, lacking multi-objective coordination mechanisms oriented towards system operational efficiency and renewable energy utilization. They fail to demonstrate comprehensive control over multi-source power coordinated dispatch and are ill-suited to scenarios where green electricity volatility and load flexibility coexist. Furthermore, the generation of green electricity procurement strategies is often simplistic or reliant on manual experience, frequently failing to fully integrate external operating environment characteristics such as power access capacity, dispatch path feasibility, and supply period stability. This results in poor strategy generalization and insufficient deployability, limiting their ability to be promoted in practical engineering projects.

[0029] Example 2, the second embodiment of the present invention, differs from the first embodiment in that: a data center green electricity procurement method based on multi-source power coordination further includes, in step S1, obtaining the input dataset includes the following steps A1-A3: A1: Acquire multi-source operational data related to data center power consumption, and perform unified format conversion and time-series alignment on data from different sources to construct the original dataset; A2: Perform anomaly detection processing on the original dataset, and identify abnormal data that do not conform to the normal operation characteristics in terms of numerical distribution through the deviation detection mechanism; A3: Correct the detected abnormal data to obtain the input dataset.

[0030] In step A1, the multi-source runtime data includes: Data center operation data: Annual maximum operating load (PIT) of IT equipment, annual utilization hours of maximum load. Historical comprehensive energy consumption level (PUE).

[0031] Electricity price-related data: Cost per kilowatt-hour of green electricity trading Green certificate transaction cost per kilowatt-hour Market-based transaction cost per kilowatt-hour .

[0032] Data collection under policy constraints: Requirements for the proportion of green electricity consumption in data centers Overall energy consumption reduction level (Converted to annual decline rate).

[0033] In this embodiment of the application, in step A2, the deviation detection mechanism adopts a fixed threshold identification method based on statistical distribution, including the following steps A211-A213: A211: In the collected raw dataset, calculate the mean and standard deviation of each data sequence to establish the statistical characteristic range of normal operation data.

[0034] A212: The valid data interval is set according to the three-standard-deviation (3σ) principle. When a data point is detected to exceed the range... When the range is exceeded, the data point is identified as an outlier. This represents the mean. It represents the standard deviation.

[0035] A213: Identify and record data that is deemed abnormal for subsequent correction and data quality verification to ensure the stability of the input dataset.

[0036] Specifically, for the data collected by the data acquisition module, if values ​​exceed a reasonable range (0 or far exceeding the average level of historical data), they are identified and corrected using the 3σ principle. Assuming the collected data sequence is... , and The formula is expressed as: in, Indicates the total number of data sequences. Indicates an index.

[0037] Set the data valid range as After removing outliers, linear interpolation is used to fill the gaps, and the filled values ​​are... Represented as: .

[0038] In an optional implementation, the deviation detection mechanism may also employ an interval restriction judgment method based on empirical rules, including the following steps A221-A223: A221: Based on equipment operation specifications or historical operating experience, set reasonable value ranges for different types of data parameters (e.g., equipment load should fluctuate between 10% and 90%).

[0039] A222: During the data acquisition process, the system compares each item of real-time data to determine whether it falls within the corresponding reasonable range.

[0040] A223: If the value exceeds the upper limit of the range or falls below the lower limit, the data will be marked as abnormal and included in the subsequent correction process.

[0041] In another alternative implementation, the deviation detection mechanism may also employ a trend abrupt change detection method based on a sliding window, including the following steps A231-A233: A231: Establish a sliding detection window for continuously acquired time series data and calculate the average rate of change or slope of adjacent windows.

[0042] A232: Determine whether the change in the current window compared to the previous window exceeds a preset threshold. If the rate of change deviates from the normal fluctuation range, it is determined to be an abnormal trend.

[0043] A233: Mark the detected trend abrupt change points so that the anomaly correction module can perform dynamic smoothing and data consistency restoration.

[0044] It should be noted that this step, by performing systematic outlier detection and correction on multi-source operational data, ensures that the data foundation upon which the downstream prediction and optimization models rely has stronger consistency, accuracy, and completeness.

[0045] Furthermore, in step S2, obtaining the predicted data center power demand for the target period includes the following steps B1-B3: B1: Predicted values ​​are derived from the input dataset, including the predicted annual maximum operating load of the equipment and the predicted annual operating hours of the equipment at maximum load.

[0046] B2: Introduce the comprehensive energy efficiency level and the fluctuation value of the number of hours as correction factors to dynamically adjust the predicted value.

[0047] B3: Substitute the predicted value and correction factor into the power demand forecasting model to calculate the total power demand of the data center within the target forecast period.

[0048] In this embodiment of the application, step B1, the predicted value is calculated using a linear trend extrapolation correction strategy, including the following steps B111-B113: B111: Several years prior to the statistical target period (e.g.) to Year, The annual change in the maximum load and operating hours of IT equipment (in years) is used to calculate the average annual growth (or decline) rate.

[0049] B112: with Using a year as a baseline, a linear extrapolation method is used to predict the maximum operating load and operating hours of IT equipment in advance.

[0050] B113: Based on the predicted values, the annual decline rate of comprehensive energy efficiency is introduced as an adjustment factor to perform a one-time linear correction on the results, thus obtaining the final predicted parameters for the target cycle.

[0051] In an optional implementation, the predicted value can also be calculated using an exponential decay correction strategy, including the following steps B121-B123: B121: Select historical operating data from the past N years, and set the corresponding exponential decay weight for each year, so that the weight of more recent years is higher.

[0052] B122: A baseline forecast value is obtained by using a weighted average method to predict the maximum load and operating hours of IT equipment.

[0053] B123: Based on the energy efficiency improvement trend of the system operation, set the energy efficiency correction coefficient and make an exponential decrease adjustment to the prediction parameters to obtain the prediction results for the target period.

[0054] In another alternative implementation, the predicted value can also be calculated using a segmented interval correction strategy, including the following steps B131-B133: B131: Based on the data center type or load level, the operating load forecast interval is divided into multiple level segments (such as high load area, medium load area, low load area), and each segment corresponds to a set of correction rules.

[0055] B132: Determine the load segment to which the current forecast value belongs, and call the correction factor or empirical adjustment ratio that matches the segment.

[0056] B133: Apply the selected correction factor to the maximum load forecast and the operating hours forecast to complete the parameter adjustment for the target period.

[0057] Specifically, the principle of the power demand forecasting model is based on the maximum load of IT equipment in the data center in historical years and the number of hours of operation of the maximum load in the year, while taking into account the dynamic correction factor of comprehensive energy efficiency, so as to achieve accurate prediction of the power consumption of the data center throughout the entire life cycle.

[0058] The dynamic forecasting model for data center power demand can be expressed as: in, This represents the total electricity demand of the data center in year T. This represents the predicted maximum annual operating load of IT equipment in year T. This represents the predicted annual operating hours of IT equipment at maximum load in year T. This represents the overall energy consumption level of the data center in year T. This represents the fluctuation value of the annual maximum load operating hours of IT equipment in year T.

[0059] It should be noted that by introducing key feature variables from multi-source operational data to construct an electricity demand forecasting model, it is possible to comprehensively consider influencing factors such as business load fluctuations and changes in the status of key equipment, and establish a dynamic forecasting mechanism that reflects the actual operational behavior of the data center. This not only improves the responsiveness of the forecast results to emergencies, resource bottlenecks, or operational strategy adjustments, but also enhances the advance notice and fine granularity of electricity demand trend capture.

[0060] Furthermore, in step S3, constructing the multi-source power coordination optimization model includes the following steps C1-C2: C2: Construct the first optimization objective function to characterize the power consumption performance index of the data center under different power source combinations, with the goal of minimizing energy costs.

[0061] C2: Construct a second optimization objective function to maximize the proportion of renewable energy use. The portion of various electricity sources that can be counted as green electricity is taken as the optimization direction. The contribution of renewable energy to the total electricity consumption under different combination strategies is quantified and used as the core evaluation index of the objective function.

[0062] In this embodiment of the application, in step C1, the method for constructing the first optimization objective function adopts a linear performance function based on multi-source power composition, including the following steps C111-C113: C111: Divide the data center's power demand within the target period into several power sources, including at least three categories: external green electricity, direct supply of local green electricity, and non-green electricity sources.

[0063] C112: Calculate the supply volume and corresponding cost of each type of electricity separately.

[0064] C113: Construct a linear combination expression with the objective function of the sum of "multiple types of electricity usage × corresponding costs" to measure the overall power consumption performance level of the entire system under this combination.

[0065] Specifically, the first optimization objective function is constructed to optimize the energy cost of the data center, expressed by the formula as follows: in, This indicates that the goal is to minimize. This indicates the annual electricity cost of a data center; , , , These represent the data center green electricity transaction cost, market-based transaction cost, green certificate purchase cost, and total cost of direct green electricity connection in year T, respectively. , , , These represent the electricity price per kWh for data center green electricity transactions, green certificate electricity price per kWh, market-based transaction electricity price per kWh, and direct green electricity connection price per kWh in year T, respectively. , , These represent the data center green electricity trading volume, market-based trading volume, and green electricity direct connection volume in year T, respectively. The proportion of electricity purchased using green certificates in the total electricity traded through market mechanisms.

[0066] in: in, , These represent the project investment cost and the power transmission and distribution fee, respectively. This indicates the percentage of annual maintenance costs to initial investment. Indicates the number of years the new energy power station has been in operation; , These represent the installed capacity of wind power and photovoltaic power in green electricity direct connection, respectively; , , , These represent the cost per watt of wind power, the cost per watt of photovoltaic power, the cost of maximum monthly demand, and the electricity price, respectively.

[0067] In an optional implementation, the method for constructing the first optimization objective function may also employ a weighted performance optimization function, including the following steps C121-C123: C121: Based on the operating characteristics and access methods of different power sources, assign weight coefficients representing the performance of each source. The weights can be set according to technical parameters such as system stability and energy conversion efficiency.

[0068] C122: The usage ratio of various power sources is used as a model variable, and the ratio is weighted and summed with the corresponding weight coefficients to form an efficiency evaluation function under the target period.

[0069] C123: Using the weighted result as the optimization objective, seek the power allocation combination strategy that minimizes (or optimizes) the weighted efficiency evaluation value, thereby achieving quantitative optimization of the multi-source power utilization effect.

[0070] In another alternative implementation, the method for constructing the first optimization objective function can also employ a multi-source scheduling objective function based on a nonlinear performance index, including the following steps C131-C133: C131: Set the corresponding nonlinear performance function for each type of power source. The performance function can be set in combination with its technical characteristics such as volatility, load response time, and operational safety margin.

[0071] C132: Substitute the usage of various types of electricity into the corresponding nonlinear functions to construct the overall performance objective function as a nonlinear superposition of the sub-functions.

[0072] C133: Using the constructed nonlinear performance function as the optimization objective, it determines the dispatch ratio of various types of power to ensure that the overall system performance is in the optimal state. It is suitable for scenarios that pursue system responsiveness and security.

[0073] To further explain, the second optimization objective function aims to maximize the proportion of green electricity consumption in data centers, and the formula is as follows: .

[0074] It should be noted that this step introduces a dual objective function of maximizing system operating efficiency and maximizing green electricity utilization. This ensures that the strategy not only pursues the optimal state of power utilization efficiency but also achieves efficient integration of renewable energy resources, thereby improving the overall system operational coordination and resource utilization. This avoids the local optima problem faced by traditional single-objective power dispatch models in multi-source environments, and possesses stronger global system adaptability and resource coordination and control capabilities.

[0075] Furthermore, in step S4, the conditions for each constraint include: As required, the average power utilization efficiency of data centers nationwide should be reduced to below 1.5, and that of newly built and expanded large and super-large data centers should be reduced to within 1.25, while that of national hub nodes should be ≤1.2.

[0076] Based on the local policy requirements where the data center is located, determine First, energy efficiency constraints are set to limit the overall energy efficiency level of the data center within the target period from exceeding the specified maximum threshold. The formula is as follows: in, This represents the predicted electricity consumption of IT equipment in the data center in year T.

[0077] A green electricity ratio constraint is set to stipulate that the proportion of renewable energy used by data centers in the total electricity consumption must not be less than a specified target. The formula is as follows: A power balance constraint is set, requiring the total power supply from all power sources to remain consistent with the predicted power demand. This is expressed by the formula: A non-negativity constraint is set, requiring that all predicted electricity consumption must be greater than or equal to zero, expressed by the formula: .

[0078] To further explain, , , The allocation is random, as long as the requirements are met. In this invention, the electricity volume consists of three parts: green electricity trading (green electricity), market-based electricity (non-green electricity), and direct green electricity connection (green electricity). If the green electricity ratio requirement is not met, additional green certificates need to be purchased. Therefore, the cost includes the costs of green electricity trading, market-based electricity, and direct green electricity connection, as well as the cost of green certificates. The amount of green certificates purchased is calculated based on the proportion of market-based electricity, that is, market-based electricity × This leads to different procurement strategies in the subsequent plans, resulting in different total costs for different quantities.

[0079] Furthermore, in step S4, generating several sets of green electricity procurement strategies includes the following steps D1-D3: D1: Using the calculated optimal allocation ratio of various power sources as the basic parameter, and combining it with the time scale of the target period for discretization, an initial pool of combined schemes for multi-source power supply under different procurement methods is constructed; the procurement methods include local green electricity direct supply, centralized green electricity trading, market-based power purchase and green certificate offsetting, etc.

[0080] D2: For different combination schemes, introduce the characteristics of the external operating environment related to the region where the data center is located, build a strategy adaptability assessment mechanism, and evaluate the degree of matching of different combination schemes under the target operating background.

[0081] D3: Generate multiple candidate green electricity procurement strategies while ensuring that each scheme meets the constraints.

[0082] In this embodiment of the application, in step D2, the strategy adaptability evaluation mechanism adopts an environment matching evaluation method based on static indicators, including the following steps D211-D213: D211: Extract static operating environment parameters of the area where the data center is located, including but not limited to: the access capacity level of the power grid in the area (e.g., whether it is a backbone network, regional network or microgrid), the resource endowment intensity of renewable energy (e.g., annual average wind speed, solar irradiance level), the coverage rate of green electricity infrastructure (e.g., whether centralized photovoltaic or wind farms have been built), and the type and intensity of support for green electricity trading or green certificates in local policies.

[0083] D212: Transform the above parameters into standardized scoring indicators. For example, the power grid structure can be divided into three levels (high, medium, and low adaptability), resource endowment can be graded according to the average annual available hours, and policy support can be quantified according to the mandatory nature of documents and the magnitude of fiscal subsidies. For each power combination scheme, multiple indicators are scored based on the type and configuration structure of the power sources it depends on, in accordance with regional characteristics, and the weight of each indicator is set. For example, the weight of resource endowment can be higher than that of policy factors.

[0084] D213: The scores of each combination of solutions are summarized into a matching index, which serves as an important reference for evaluating whether each strategy is suitable for the current regional environment. Solutions with higher matching indices will be prioritized for subsequent screening, while those with low matching indices are considered difficult to implement in this environment and can be marked as low-priority solutions or not recommended at this time.

[0085] In an optional implementation, the strategy adaptability evaluation mechanism may also employ a rule-matching-based conditional judgment evaluation method, including the following steps D221-D223: D221: Based on the power system access standards, renewable energy dispatch policies, and operational experience of different regions, a set of technical compatibility rules should be pre-established. For example, the green electricity direct supply scheme may be set to apply only to areas with a local voltage level of 110kV or above and the conditions for dedicated feeder access; or it may be required that when the proportion of wind power access exceeds a certain threshold, corresponding energy storage buffer configurations must be available.

[0086] D222: Extract the structure and power supply path information of each candidate green energy combination strategy and compare them one by one with the above rules. For example, if a scheme relies on local wind power direct supply but the regional grid capacity only supports low-voltage access, the scheme is considered to not meet the physical conditions; if a scheme intends to use a large proportion of green certificates for offsetting, but the local policy does not include this type of green certificate in the scope of recognition, it is also considered to be non-compliant.

[0087] D223: Based on the rule matching results, each strategy is marked with its compliance level: those that fully meet all rules are "highly compatible", those that partially meet the conditions are "mediumly compatible", and those that obviously conflict with or lack key conditions are "lowly compatible" or "infeasible". The evaluation results will be used as mandatory constraints or priority references when selecting or recommending strategies in the next step.

[0088] In another alternative implementation, the strategy adaptability evaluation mechanism can also employ a dynamic simulation-based operational characteristic evaluation method, including the following steps D231-D233: D231: Load the power grid topology, typical load curves, and historical operating data of the target area into a system model with simulation capabilities to construct a realistic or semi-realistic operating environment. Simultaneously import typical output curves of renewable energy resources, such as hourly wind power fluctuation curves, seasonal photovoltaic output patterns, and the data center's own load characteristic models (such as peak-valley load ratio, diurnal load variations, etc.).

[0089] D232: For each power combination strategy, conduct typical daily or weekly operational simulations in the simulation environment. The focus is on observing whether different schemes can provide stable power supply during peak load periods, whether fluctuations in wind and solar power output lead to short-term power shortages, and whether they cause operational anomalies such as voltage fluctuations or frequency disturbances in the region. Simultaneously, simulate dispatch response capabilities, such as whether various backup power sources can promptly intervene and maintain normal data center operation when local green power is suddenly interrupted.

[0090] D233: The simulation results are transformed into multiple operational adaptability indicators, such as load fulfillment rate (i.e., the proportion of time that the supply and demand balance is maintained during the simulation period), power fluctuation mitigation capability, backup dispatch intervention time, and the number of power outages during operation. These indicators are used to comprehensively evaluate the operational adaptability level of each scheme. Schemes with excellent indicators are classified as high adaptability, those with medium indicators are classified as average adaptability, and those with large-scale power shortages or system disturbances are considered low adaptability and are eliminated or restricted from recommendation.

[0091] It should be noted that this step integrates the calculated multi-source power allocation ratio with the characteristics of the external operating environment to generate multiple sets of alternative green electricity procurement strategies that meet technical constraints, thereby providing multiple alternatives and switching space for energy management under different operating scenarios.

[0092] Furthermore, in step S5, selecting the optimal target procurement plan includes the following steps E1-E3: E1: Establish a performance evaluation index system for candidate green electricity procurement strategies, including two categories of indicators: engineering feasibility and operational controllability.

[0093] E2: Based on the preset indicator weights, a weighted scoring method is used to comprehensively score each candidate strategy.

[0094] E3: Based on the comprehensive scoring results, select the group with the highest score from the candidate strategies as the final green electricity procurement plan for the target period.

[0095] Specifically, engineering feasibility indicators are mainly used to reflect the degree of technical coordination and deployment convenience of the strategy during implementation. For example, it can be considered whether the target power source has an existing or adjustable grid access channel, and whether it is necessary to build additional transmission lines or adjust the existing power supply structure.

[0096] Operational controllability indicators focus on the stability and operation and maintenance management capabilities of the strategy during the actual execution phase, and can be evaluated from multiple perspectives: for example, whether energy output fluctuations are controllable, especially for intermittent resources such as wind power and photovoltaics, it is necessary to determine whether there is a buffer mechanism or energy storage means to support power changes; to know the response capability under abnormal conditions, such as whether there is the ability to quickly switch to the backup plan after power failure or interruption, and whether the overall strategy structure has a certain degree of redundancy.

[0097] All candidate procurement strategies are scored on their performance across various evaluation dimensions using a unified scoring standard, combined with indicator weights, to calculate the weighted total score for each strategy.

[0098] The weighted scores of all strategies are sorted, and the strategy with the highest score is selected as the target procurement plan. If necessary, a minimum score threshold can be set to eliminate marginal strategies.

[0099] Example 3, the third embodiment of the present invention, differs from the previous two embodiments in that it provides a data center green electricity procurement system based on multi-source power coordination, comprising a data acquisition module, a demand forecasting module, an allocation ratio calculation module, and a strategy generation module. The data acquisition module collects multi-source operational data, performs outlier detection, corrects outliers to obtain an input dataset. The demand forecasting module, based on the input dataset, establishes a data center electricity demand forecasting model to obtain the predicted data center electricity demand within a target period. The allocation ratio calculation module, based on the predicted electricity demand and renewable energy output information, constructs a multi-source power coordination optimization model to calculate the optimal allocation ratio for various power sources. The strategy generation module, based on the optimal allocation ratio and considering external operating environment characteristics, generates several sets of green electricity procurement strategies under various constraints, evaluates the green electricity procurement strategies, and selects the optimal target procurement scheme.

[0100] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0101] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0102] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0103] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented in combination with any of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0104] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for green electricity procurement for data centers based on multi-source power coordination, characterized in that: include, Collect multi-source operational data, perform outlier detection, correct outliers, and obtain the input dataset; Based on the input dataset, a data center power demand forecasting model is established to obtain the predicted data center power demand within the target period. Based on the predicted electricity demand and renewable energy output information, a multi-source power coordination optimization model is constructed to calculate the optimal allocation ratio of various power sources. Based on the optimal allocation ratio and considering the characteristics of the external operating environment, several sets of green electricity procurement strategies are generated under the condition of satisfying various constraints. Evaluate green electricity procurement strategies and select the optimal target procurement plan.

2. The data center green electricity procurement method based on multi-source power coordination as described in claim 1, characterized in that: The obtained input dataset includes, Acquire multi-source operational data related to data center power consumption, and perform unified format conversion and time-series alignment on data from different sources to construct the original dataset; Anomaly detection processing is performed on the original dataset. An aberration detection mechanism is used to identify anomalous data whose numerical distribution does not conform to the normal operating characteristics. The detected abnormal data is corrected to obtain the input dataset.

3. The data center green electricity procurement method based on multi-source power coordination as described in claim 2, characterized in that: The predicted data center power demand for the target period includes... Predicted values ​​are derived from the input dataset, including the predicted annual maximum operating load of the equipment and the predicted annual operating hours of the equipment at maximum load. The comprehensive energy efficiency level and the fluctuation value of the number of hours are introduced as correction factors to dynamically adjust the predicted value; By substituting the predicted values ​​and correction factors into the power demand forecasting model, the total power demand of the data center within the target forecast period is calculated.

4. The data center green electricity procurement method based on multi-source power coordination as described in claim 3, characterized in that: The construction of the multi-source power coordination optimization model includes, A first optimization objective function is constructed to characterize the power consumption performance index of the data center under different power source combinations, with the optimization direction being the lowest energy cost; A second optimization objective function is constructed to maximize the proportion of renewable energy use. The portion of various electricity sources that can be counted as green electricity is taken as the optimization direction. The contribution of renewable energy to total electricity consumption under different combination strategies is quantified and used as the core evaluation index of the objective function.

5. A data center green electricity procurement method based on multi-source power coordination as described in claim 4, characterized in that: The conditions of each constraint include, Set energy efficiency constraints to limit the overall energy efficiency level of the data center within the target period from exceeding the specified maximum threshold; Set green electricity ratio constraints to stipulate that the proportion of renewable energy used by data centers in the total electricity consumption shall not be less than the specified target; Set power balance constraints to require that the total power supply from all types of power sources be consistent with the predicted power demand; Set a nonnegativity constraint that requires all predicted electricity consumption to be greater than or equal to zero.

6. The data center green electricity procurement method based on multi-source power coordination as described in claim 5, characterized in that: The generation of several sets of green electricity procurement strategies includes... The calculated optimal allocation ratio of various power sources is used as the basic parameter, and discretized in combination with the time scale of the target period to construct an initial pool of combined schemes for multi-source power supply under different procurement methods. For different combination schemes, external operating environment characteristics related to the region where the data center is located are introduced to construct a strategy adaptability assessment mechanism to evaluate the degree of matching of different combination schemes under the target operating background; Multiple candidate green electricity procurement strategies are generated, provided that each scheme meets the constraints.

7. A data center green electricity procurement method based on multi-source power coordination as described in claim 6, characterized in that: The selection of the optimal target procurement plan includes... A performance evaluation index system was established for candidate green energy procurement strategies, including two categories of indicators: engineering feasibility and operational controllability. Based on the preset indicator weights, a weighted scoring method is used to comprehensively score each candidate strategy. Based on the comprehensive scoring results, the group with the highest score among the candidate strategies will be selected as the final green electricity procurement plan for the target period.

8. A data center green electricity procurement system based on multi-source power coordination, employing the data center green electricity procurement method based on multi-source power coordination as described in any one of claims 1 to 7, characterized in that: It includes a data acquisition module, a demand forecasting module, an allocation ratio calculation module, and a strategy generation module; The data acquisition module is used to collect multi-source operational data, perform outlier detection, correct outliers, and obtain the input dataset. The demand forecasting module establishes a data center electricity demand forecasting model based on the input dataset to obtain the data center electricity demand forecast within the target period. The allocation ratio calculation module, based on the electricity demand forecast and renewable energy output information, constructs a multi-source power coordination optimization model to calculate the optimal allocation ratio for various power sources. The strategy generation module is used to generate several sets of green electricity procurement strategies based on the optimal allocation ratio and the characteristics of the external operating environment, while satisfying various constraints. The green electricity procurement strategies are then evaluated to select the optimal target procurement scheme.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the data center green electricity procurement method based on multi-source power coordination as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the data center green electricity procurement method based on multi-source power coordination as described in any one of claims 1 to 7.

Images