Carbon emission monitoring method and system for data center

Abstract

The invention discloses a carbon emission monitoring method and system for a data center, and the method comprises the steps: collecting the power generation proportion data of different energy types in real time, predicting the power supply change trend of renewable energy in combination with meteorological information, calculating the power supply proportion of each energy type, and constructing a dynamic weight matrix; constructing a time sequence prediction model by using historical and real-time renewable energy data, and predicting the future renewable energy power supply capability; carbon emission factors and priorities of different data centers are calculated, a task migration feasibility matrix is constructed, calculation tasks are dynamically migrated to a low-carbon emission data center according to a scheduling revenue function, a task scheduling path is optimized, cross-regional data transmission time delay and energy consumption are reduced, and cross-regional optimization scheduling of computing power is achieved; a multi-dimensional time sequence prediction model is constructed, the carbon emission trend is predicted, and carbon emission abnormity is monitored in real time. According to the invention, the problem of matching of computing power load and renewable energy of a data center can be solved.

Description

Technical Field

[0001] The present invention relates to the technical field of carbon emissions, and particularly to a carbon emission monitoring method and system for a data center. Background Art

[0002] With the acceleration of global digital transformation, data centers have become important infrastructures to support key technologies such as cloud computing, big data, and artificial intelligence. However, the rapid development of data centers is accompanied by extremely high energy consumption.

[0003] Currently, more and more data centers are starting to introduce renewable energy sources such as wind energy and photovoltaics to reduce their dependence on traditional fossil fuels. However, due to the volatility of renewable energy generation and the uneven geographical distribution, there is a significant mismatch between the load demand of data centers and the supply of green electricity. Most data centers tend to be located in core areas with convenient communication and low network latency, while areas rich in renewable energy resources are often remote, resulting in a contradiction between the load distribution of data centers and the geographical distribution of renewable energy. In addition, wind energy and photovoltaic power generation are greatly affected by factors such as weather, seasons, and day-night changes, leading to drastic fluctuations in their power supply capabilities, while the computing tasks of data centers usually require a highly stable and reliable power supply. This time mismatch between supply and demand makes it difficult for data centers to fully utilize renewable energy, thus limiting the potential for carbon emission reduction to a certain extent.

[0004] Traditional data center energy consumption monitoring methods are mainly based on total power consumption statistics, which cannot accurately evaluate carbon emissions under different energy structures and are also difficult to dynamically adjust computing power to adapt to changes in renewable energy. Therefore, an innovative carbon emission monitoring method is needed that can not only real-time sense the energy consumption and carbon emission situation of data centers but also, in combination with the supply prediction of renewable energy, dynamically optimize the computing load scheduling strategy to achieve an adaptive match between computing power demand and renewable energy supply, thereby effectively reducing carbon emissions and improving the utilization efficiency of green energy. Summary of the Invention

[0005] To solve the above problems, the present invention provides a carbon emission monitoring method and system for a data center.

[0006] To achieve the above object, the technical solutions adopted by the present invention are as follows:

[0007] On the one hand, the present invention discloses a carbon emission monitoring method for a data center, including:

[0008] Step 1: Real-time collect the power generation proportion data of different energy types, combine meteorological information to predict the change trend of renewable energy power supply, calculate the power supply proportion of each energy type, and construct a dynamic weight matrix;

[0009] Step 2: Use historical and real-time renewable energy data to build a time series prediction model to predict the future renewable energy power supply capacity. Combine the computing load characteristics of the data center, dynamically adjust the load distribution, optimize the task execution time and resource allocation, and achieve dynamic matching with the renewable energy supply level;

[0010] Step 3: Calculate the carbon emission factors and priorities of different data centers, build a task migration feasibility matrix, and dynamically migrate computing tasks to low-carbon emission data centers according to the scheduling revenue function to optimize the task scheduling path, reduce cross-regional data transmission delay and energy consumption, and achieve cross-regional optimal scheduling of computing power;

[0011] Step 4: Build a multi-dimensional time series prediction model to predict the carbon emission trend, monitor carbon emission anomalies in real time, divide the warning levels and take corresponding measures to ensure the stable operation of the data center under the low-carbon operation goal.

[0012] Furthermore: The said Step 1 includes:

[0013] Obtain the real-time power supply data of each energy type through the power grid dispatching data interface, and combine the smart meter and SCADA system to monitor the proportion of power input sources in the area where the data center is located; at the same time, use the meteorological data interface to obtain wind speed, solar radiation intensity, reservoir water level information, and combine historical data to predict the change trend of renewable energy power supply;

[0014] Calculate the power generation proportion of a single energy type, and build a dynamic weight matrix of the power grid power generation structure;

[0015] Establish a time series change trend analysis, calculate the first derivative of each energy proportion with respect to time, build a change rate vector of the power generation structure, and quantify the change trend of different energy proportions to provide a dynamic correction factor for adaptive carbon emission calculation.

[0016] Furthermore: The said Step 2 includes:

[0017] Extract historical renewable energy power generation data from the time series database to form a time series, obtain real-time meteorological data such as current solar radiation intensity, wind speed, and reservoir water level through the meteorological API, and build a prediction input matrix;

[0018] Use a multi-modal long short-term memory network for prediction, and dynamically adjust the prediction value in combination with the historical error correction coefficient and trend correction coefficient to improve the prediction accuracy;

[0019] Build a computing load weight vector according to the task type of the data center, calculate the load adjustment coefficient, and dynamically match the renewable energy supply level;

[0020] Adjust the execution time of low-priority tasks according to renewable energy forecasts and load adjustment factors, optimize the allocation of AI computing loads, perform task migrations when necessary, and schedule computing tasks to data centers with a higher proportion of renewable energy to achieve dynamic matching between data center loads and renewable energy supplies.

[0021] Further: Step 3 includes:

[0022] Based on the dynamic weight matrix of the power grid generation structure, calculate the carbon emission factors of each data center at different times, and measure the low-carbon computing power scheduling priority of the data center through the carbon emission priority factor. Data centers with higher priorities are more suitable for receiving computing power tasks;

[0023] Construct a task migration feasibility matrix, evaluate the task migration delay between data centers, and combine with the scheduling revenue function to judge the feasibility of migrating tasks from high-carbon emission data centers to low-carbon emission data centers;

[0024] Aim to minimize the comprehensive weight of carbon emissions and network latency during task scheduling, calculate the optimal task scheduling path, and avoid overloading a single data center due to a surge in tasks;

[0025] Perform data sharding on the tasks to be migrated, reduce the occupation of transmission bandwidth and improve parallelism. Monitor the execution of tasks through edge computing nodes, and use blockchain to record the task migration path to ensure the traceability and security of scheduling, and achieve cross-regional optimal scheduling of computing power.

[0026] Further: Step 4 includes:

[0027] Combine data center loads, renewable energy supplies, power grid carbon emission factors, task migration traffic, and historical carbon emission data to construct a multi-dimensional time series prediction model, and use a two-layer time series prediction network. Among them, the short-term trend prediction model is used for detecting carbon emission fluctuations at the hourly level, and the long-term trend prediction model is used for predicting carbon emission changes at the daily / weekly level;

[0028] Set the carbon emission warning threshold, determined based on the historical carbon emission mean and standard deviation; use a two-layer adaptive anomaly detection model, combining local anomaly factor detection and Bayesian mutation detection, to monitor the carbon emission trend in real time and identify anomaly points;

[0029] Divide the carbon emission anomaly degree into mild, moderate, and severe warning levels; in the case of a mild warning, perform lightweight load adjustment; in the case of a moderate warning, combine load adjustment and cross-regional computing power scheduling to optimize the proportion of low-carbon electricity use; in the case of a severe warning, unload high-energy-consuming tasks to low-carbon data centers, enable backup clean energy, and notify the management to reduce the priority of non-critical tasks.

[0030] On the other hand, the present invention discloses a carbon emission monitoring system for a data center, comprising:

[0031] A dynamic weight matrix construction module for the power grid generation structure: By collecting in real time the power generation proportion data of different energy types, combining meteorological information to predict the change trend of renewable energy power supply, calculating the power supply proportion of each energy type and constructing a dynamic weight matrix;

[0032] A renewable energy prediction load adjustment module: Using historical and real-time renewable energy data to construct a time series prediction model, predicting the future power supply capacity of renewable energy, combining the computing load characteristics of the data center, dynamically adjusting the load distribution, optimizing the task execution time and resource allocation, and achieving dynamic matching with the renewable energy supply level;

[0033] A cross-regional low-carbon computing power scheduling module: Calculating the carbon emission factors and priorities of different data centers, constructing a task migration feasibility matrix, dynamically migrating computing tasks to low-carbon emission data centers according to the scheduling revenue function, optimizing the task scheduling path, reducing cross-regional data transmission delay and energy consumption, and achieving cross-regional optimal scheduling of computing power;

[0034] A carbon emission warning module: Constructing a multi-dimensional time series prediction model, predicting the carbon emission trend, monitoring carbon emission anomalies in real time, dividing the warning levels and taking corresponding measures to ensure the stable operation of the data center under the low-carbon operation goal.

[0035] Compared with the prior art, the technical progress achieved by the present invention lies in:

[0036] The present invention adopts an adaptive carbon emission calculation algorithm based on the real-time power structure, which can obtain in real time the energy structure of the power grid accessed by the data center and dynamically calculate the carbon emission factor. By adaptively adjusting the calculation model, the carbon emission calculation can accurately reflect the change in the proportion of renewable energy, providing a reliable data basis for subsequent optimization. It solves the problem that the traditional method is only based on total power consumption statistics and cannot accurately calculate carbon emissions, making the carbon emission assessment more refined and dynamic. The present invention adopts a load adjustment algorithm based on renewable energy prediction, uses AI to predict the power supply capacity of renewable energy such as wind energy and photovoltaic energy, and combines the adjustability of the computing tasks in the data center to increase the computing load during the energy peak period and reduce the load or perform task migration during the energy trough period. By intelligently adjusting the execution time of computing tasks, it realizes the dynamic matching of computing power demand and renewable energy supply, improves the utilization rate of renewable energy, and reduces carbon emissions. It solves the problem that the traditional load management method cannot consider the matching of renewable energy supply and demand, making the data center more low-carbon and intelligent.

[0037] The present invention adopts a cross - regional low - carbon computing power scheduling algorithm to perform intelligent task migration among multiple data centers, dynamically allocate computing tasks to regions with a higher proportion of low - carbon electricity, and achieve cross - regional green computing resource scheduling. Combining edge computing and distributed computing architectures, it maximally reduces the use of high - carbon electricity and avoids the increase in carbon emissions caused by insufficient green electricity supply in a single data center. It solves the problem of the mismatch between the load distribution of data centers and the geographical distribution of renewable energy, making the operation of data centers more flexible, efficient, and low - carbon. The present invention adopts a carbon emission early - warning algorithm based on intelligent detection. Through time - series analysis, machine learning, and Bayesian anomaly detection techniques, it predicts future carbon emission trends and gives an early warning before exceeding the preset threshold. Combining a multi - level response mechanism, it executes different levels of regulation measures according to the degree of carbon emission anomalies, such as load adjustment, task migration, release of clean energy storage, etc., to ensure the low - carbon and stable operation of data centers. It solves the problems of strong lag in traditional monitoring methods and lack of an early - warning mechanism, enables data centers to have self - adaptive regulation capabilities, and effectively reduces carbon emission risks.

[0038] In summary, the present invention can solve the problem of the mismatch between computing power load and renewable energy in data centers. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention, but do not constitute a limitation to the present invention.

[0040] In the drawings:

[0041] Figure 1 is the flow chart of the present invention;

[0042] Figure 2 is the system structure diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0043] The following specific embodiments can be combined with each other. For the same or similar concepts or processes, they may not be repeated in some embodiments. The embodiments of the present invention will be described below in conjunction with the drawings.

[0044] Embodiment 1

[0045] As Figure 1 shown, the present invention discloses a carbon emission monitoring method for data centers, including:

[0046] Step 1: Real - time collect the power generation proportion data of different energy types, combine meteorological information to predict the change trend of renewable energy power supply, calculate the power supply proportion of each energy type, and construct a dynamic weight matrix;

[0047] Step 2: Use historical and real-time renewable energy data to build a time series prediction model to predict the future renewable energy power supply capacity. Combine the computing load characteristics of the data center, dynamically adjust the load distribution, optimize the task execution time and resource allocation, and achieve dynamic matching with the renewable energy supply level;

[0048] Step 3: Calculate the carbon emission factors and priorities of different data centers, build a task migration feasibility matrix, and dynamically migrate computing tasks to low-carbon emission data centers according to the scheduling revenue function. Optimize the task scheduling path, reduce cross-regional data transmission delay and energy consumption, and achieve cross-regional optimal scheduling of computing power;

[0049] Step 4: Build a multi-dimensional time series prediction model to predict the carbon emission trend, monitor carbon emission anomalies in real time, divide the warning levels and take corresponding measures to ensure the stable operation of the data center under the goal of low-carbon operation.

[0050] Specifically, Step 1 includes:

[0051] To achieve adaptive carbon emission calculation based on the real-time power structure, first, it is necessary to build a dynamic weight matrix W of the power grid generation structure t , which is used to describe the power supply ratio of different energy types in the power grid at a certain moment t, so as to provide accurate input for subsequent carbon emission calculations. The specific implementation steps are as follows:

[0052] 1.1 Data collection: Obtain the real-time generation proportion of different energy types

[0053] Through the power grid dispatching data interface (such as the API or open data platform provided by the State Grid or regional power companies), obtain the power supply data of each energy type (coal-fired, gas-fired, hydropower, wind power, photovoltaic, nuclear power, etc.) at the current moment t, denoted as P i (t), where i represents different energy types.

[0054] Combine the smart meter and the SCADA system to monitor the power input source in the area where the data center is located in real time, and obtain the actual proportion data of various power supplies to improve the accuracy and timeliness of the data.

[0055] Through the meteorological data interface (such as satellite remote sensing, weather stations, etc.), obtain information such as wind speed, solar radiation intensity, reservoir water level, etc., and combine historical data to predict the power supply change trend of renewable energy in the short term in the future, providing auxiliary reference for subsequent optimization.

[0056] 1.2 Data processing: Build a dynamic weight matrix of the power grid generation structure

[0057] Calculate the generation proportion of a single energy type:

[0058] Set the total power generation of the power grid where the data center is located as Ptotal For each energy source \(i\) at time \(t\), the power supply ratio \(R_{i}(t)\) i is calculated as follows:

[0059]

[0060] where, represents the sum of all energy types, ensuring that the ratios of all energies satisfy the normalization constraint:

[0061]

[0062] Construct the dynamic weight matrix of the power generation structure of the power grid:

[0063] After calculating the power supply ratios of each energy type, construct the weight matrix \(W_t\) of the power generation structure:

[0064] W t = [R coal (t) R gas (t) R hydro (t) R wind (t) R solar (t) R nuclear (t)]

[0065] where:

[0066] R coal (t) represents the proportion of coal-fired power generation;

[0067] R gas (t) represents the proportion of gas-fired power generation;

[0068] R hydro (t) represents the proportion of hydropower generation;

[0069] R wind (t) represents the proportion of wind power generation;

[0070] R solar (t) represents the proportion of photovoltaic power generation;

[0071] R nuclear (t) represents the proportion of nuclear power generation.

[0072] 1.3 Calculate the change rate of the power generation structure: Establish a time series change trend analysis

[0073] Calculate the first-order time derivative of the power generation ratio:

[0074] To analyze the dynamic change trend of the power generation structure of the power grid, calculate the change rate of each energy proportion with respect to time:

[0075]

[0076] Among them, Δt is the sampling time interval, usually set at the minute level to ensure the real-time nature of the calculation.

[0077] Construct the power generation structure change rate vector:

[0078]

[0079] This vector is used to quantify the change trend of the proportion of different energy sources and provide a dynamic correction factor for subsequent adaptive carbon emission calculations.

[0080] In this step, through real-time data acquisition, grid power supply ratio calculation, construction of a dynamic weight matrix for the power generation structure, time series trend analysis, and data storage and invocation, a complete dynamic weight matrix W of the grid power generation structure is formed t , providing accurate input data for subsequent adaptive carbon emission calculations.

[0081] Specifically, step 2 includes:

[0082] After completing step 1 (construction of the dynamic weight matrix W of the grid power generation structure t ), the data center can obtain the power supply proportion of various energy sources at the current moment in real time. However, due to the volatility of renewable energy, there are still limitations in optimizing carbon emissions relying solely on the current power supply structure. Therefore, it is necessary to predict the future power supply capacity of renewable energy and, combined with the computing load characteristics of the data center, dynamically adjust the load distribution to maximize the matching of the supply level of renewable energy, thereby optimizing carbon emissions.

[0083] 2.1 Data acquisition: Obtain historical and real-time data of renewable energy

[0084] To predict the power supply capacity of renewable energy at the future time t + τ, the following key data needs to be collected:

[0085] Obtain historical time series data

[0086] Extract the renewable energy power generation data of the past n time windows from the time series database (such as InfluxDB, TimescaleDB) to form a time series ε = {E solar (t), E wind (t), E hydro (t)}.

[0087] Among them, E solar (t) represents the photovoltaic power generation, E wind (t) represents the wind power generation, and E hydro (t) represents the hydropower generation.

[0088] Obtain real-time meteorological data

[0089] Obtain the solar radiation intensity S(t), wind speed V w (t), reservoir water level H w (t) and other data at the current moment through the meteorological API as the prediction input.

[0090] Construct the prediction input matrix

[0091] Set the prediction window length T f , and define the input matrix X t :

[0092]

[0093] This matrix will be used as the input for the renewable energy prediction model.

[0094] 2.2 Renewable energy power generation prediction: Construct a time series prediction model

[0095] Use a multi-modal long short-term memory network for prediction

[0096] Since the power supply of renewable energy is greatly affected by meteorological factors and the time series data characteristics of different energy types are different, a multi-modal long short-term memory network is used for modeling.

[0097] Set the prediction model:

[0098]

[0099] Among them, is the power supply of the i-th type of energy at the future τ moment in the prediction, and θ represents the model parameters.

[0100] Construct a dynamic correction mechanism for prediction errors

[0101] Since prediction errors may lead to load adjustment deviations, a dynamic correction factor is introduced:

[0102]

[0103] Among them:

[0104] α is the historical error correction coefficient,

[0105] β is the trend correction coefficient,

[0106] represents the change rate of renewable energy power generation.

[0107] Calculate the finally corrected predicted value:

[0108]

[0109] 2.3 Calculate the load adjustment coefficient: Match the renewable energy supply level

[0110] Construct a computational load weight vector and set the computational load proportion vector for different types of tasks:

[0111] L t =[W batch (t) W realtime (t) W AI (t)]

[0112] Where:

[0113] W batch (t) represents low-priority tasks that can be deferred for execution (such as AI model training, data analysis).

[0114] W realtime (t) represents high-priority real-time computational tasks (such as video rendering, database transactions).

[0115] W AI (t) represents AI computational tasks that can be optimized (such as inference computing).

[0116] The computational load adjustment coefficient λ t , using the normalization method, calculates the ratio of renewable energy power supply relative to the historical maximum value:

[0117]

[0118] Where:

[0119] represents the total predicted amount of renewable energy after correction,

[0120] γ controls the weight of renewable energy,

[0121] δ controls the computational load weight,

[0122] λ t is used to adjust the task load to dynamically match the renewable energy supply level.

[0123] 2.4 Dynamic load adjustment: Execute task migration and power optimization

[0124] Adjust the computational time of low-priority tasks

[0125] When λ t > 1 (abundant renewable energy), execute batch tasks in advance to utilize high green power periods.

[0126] When λ t < 1 (insufficient renewable energy), defer or reduce the allocation of computational resources for batch tasks.

[0127] 2. Optimize the allocation of AI computational load

[0128] For AI computing tasks, according to λ t Adjust the operating modes of GPUs and TPUs to reduce the computing load during high carbon emission periods.

[0129] 3. Distributed task migration

[0130] In combination with the subsequent cross-regional low-carbon computing power scheduling algorithm, migrate some computing tasks to data centers with a higher proportion of renewable energy.

[0131] This step is based on the grid power generation structure data in Step 1, and uses time series prediction technology, dynamic correction mechanism and load adjustment coefficient calculation method to achieve dynamic matching of the data center load and renewable energy supply, thereby reducing carbon emissions and improving energy utilization efficiency.

[0132] Specifically, Step 3 includes:

[0133] After completing Step 1 (construction of the dynamic weight matrix W of the grid power generation structure t ) and Step 2 (load adjustment algorithm based on renewable energy prediction), the data center can already dynamically adjust the computing load based on the local renewable energy power supply capacity. However, due to the spatial mismatch of the renewable energy supply levels of data centers in different regions, the optimization of a single data center still cannot fully reduce the overall carbon emissions. Therefore, the core goal of Step 3 is:

[0134] 1. Calculate the carbon emission factors of different data centers based on the proportion of renewable energy in the cross-regional power grid;

[0135] 2. Dynamically migrate computing tasks between multiple data centers to transfer the computing power demand from high carbon emission regions to low carbon emission regions;

[0136] 3. Optimize the task scheduling path to reduce cross-regional data transmission delay and energy consumption loss to ensure efficient execution.

[0137] 3.1 Calculation of carbon emission factors of cross-regional data centers

[0138] Establish a carbon emission factor calculation model for data centers, and set the carbon emission factor C j (t) of the jth data center at time t:

[0139]

[0140] Where:

[0141] R i,j (t) represents the proportion of the ith type of energy used by data center j at time t, and this data is from the dynamic weight matrix W of the grid power generation structure in Step 1 t .

[0142] EF i is the unit carbon emission factor (kg CO2 / kWh) of the i-th type of energy, which can be provided by the International Energy Agency (IEA) or local grid operators.

[0143] Calculate the carbon emission priority factor of the data center and set the carbon emission priority factor P j (t) of the data center to measure the low-carbon computing power scheduling priority:

[0144]

[0145] where:

[0146] C max (t) is the highest carbon emission factor among all data centers at the current moment;

[0147] C min (t) is the lowest carbon emission factor among all data centers at the current moment.

[0148] When P j (t) is larger, it indicates that data center j has a cleaner power supply and is more suitable for receiving computing power tasks.

[0149] 3.2 Calculate task migration decision: Cross-regional computing power optimization

[0150] Construct a task migration feasibility matrix. Set the data center set as D = {D1, D2,..., D m}, and the task migration feasibility matrix M between different data centers is defined as follows:

[0151]

[0152] where T jk represents the transmission delay of migrating tasks from data center D j to data center D k , which is determined by factors such as network bandwidth, link status between data centers, and task data size. Only migrations with T jk within the threshold range will be considered to ensure task execution efficiency.

[0153] Task migration decision function, calculate the scheduling benefit function of migrating tasks from a high-carbon emission data center D j to a low-carbon emission data center D k :

[0154] G jk (t) = α·(C j (t) - C k (t)) - β·T jk

[0155] Among them:

[0156] α is the weight factor for carbon emission reduction;

[0157] β is the weight factor for task transmission delay;

[0158] When G jk (t)>0, task migration is beneficial to the overall carbon emission optimization, and task scheduling is allowed to be executed.

[0159] Calculation of the optimal task scheduling path. In order to perform optimal low-carbon computing power migration between multiple data centers, through the following algorithm, the comprehensive weight of carbon emission and network delay during task scheduling is minimized.

[0160] Set the objective function:

[0161]

[0162] Where is the set of task migration paths.

[0163] This step can ensure that the task can be executed on time at the target data center, and the task is preferentially scheduled to the data center with a lower carbon emission factor to prevent the problem of load overload caused by a surge in tasks in a single data center.

[0164] 3.3 Task migration execution: Intelligent issuance of scheduling instructions

[0165] Task packaging and splitting

[0166] Perform data sharding on the computing tasks to be migrated, reduce the bandwidth occupation during data transmission, and improve the task parallelism at the same time. The task packaging scheme:

[0167] T package ={T1, T2,..., T n}

[0168] Among them,

[0169] The scheduling instruction issuance mechanism monitors the task execution through the edge computing node to ensure that the execution of critical tasks is not affected during the scheduling process. Adopt a scheduling confirmation mechanism based on smart contracts, and use the blockchain to record the task migration path to ensure the traceability and security of task scheduling.

[0170] This step realizes the dynamic cross-regional scheduling of computing power while ensuring the execution efficiency of computing tasks, maximizes the utilization of low-carbon power resources, and thus effectively reduces the overall carbon emission of the data center.

[0171] Specifically, step 4 includes:

[0172] After completing Step 1 (Adaptive Carbon Emission Calculation Algorithm Based on Real-time Power Structure), Step 2 (Load Adjustment Algorithm Based on Renewable Energy Forecast), and Step 3 (Cross-regional Low-carbon Computing Power Scheduling Algorithm), the data center has been able to optimize in terms of carbon emission measurement, dynamic load adjustment, and cross-regional low-carbon computing power scheduling. However, due to the instability of the power grid structure, the uncertainty of renewable energy supply, and the volatility of the computing task load in the data center, there is still a risk of unexpected increase in carbon emissions. Therefore, a carbon emission warning algorithm based on intelligent detection is needed to identify high carbon emission risks in advance and provide automated intervention measures to ensure the stable operation of the data center under the goal of low-carbon operation.

[0173] 4.1 Prediction Model Construction: Carbon Emission Trend Prediction Based on Multidimensional Data

[0174] To achieve accurate carbon emission warning, first, a multidimensional time series prediction model needs to be constructed, which comprehensively considers factors such as data center load, renewable energy supply, power grid carbon emission factor, and task migration situation.

[0175] Input Variable Selection

[0176] Set the set of key factors affecting the carbon emissions of the data center at time t:

[0177] X t ={L t ,R t ,W t ,M t ,H t ,T t}

[0178] Where:

[0179] L t : The current computing load of the data center (CPU, GPU utilization rate).

[0180] R t : The proportion of renewable energy (wind energy, photovoltaic, etc.).

[0181] W t : The carbon emission factor matrix in the current power grid structure (calculated by Step 1).

[0182] M t : The cross-regional task migration flow (calculated by the cross-regional low-carbon computing power scheduling algorithm in Step 3).

[0183] H t : Historical carbon emission data (used for time series modeling).

[0184] T t : External environmental factors such as ambient temperature and humidity.

[0185] A double-layer time series prediction network is adopted, which is respectively used for short-term trend prediction (hour level) and long-term trend prediction (day / week level):

[0186]

[0187] Short-term trend prediction model (LSTM): It is used to predict the carbon emission changes within the next 16 hours and is applicable to short-term fluctuation detection.

[0188] Long-term trend prediction model (Transformer): It is used to predict the carbon emission changes at the daily level to discover potential structural carbon emission anomalies.

[0189] Final carbon emission prediction value:

[0190]

[0191] Among them, λ is the weighted coefficient of short-term prediction and long-term prediction, which is dynamically adjusted according to actual needs.

[0192] 4.2 Carbon emission anomaly detection mechanism

[0193] Set the carbon emission warning threshold. According to historical data and policy goals, set the warning threshold for the carbon emission of the data center:

[0194] C thresh = C avg + γ·σ C

[0195] Among them:

[0196] C avg : Historical average carbon emission.

[0197] σ C : Standard deviation of carbon emissions.

[0198] γ: Adjustment factor (can be dynamically adjusted according to the data center strategy).

[0199] Intelligent detection of carbon emission anomalies. A double-layer adaptive anomaly detection model is adopted, combining local outlier factor and Bayesian-based mutation detection to monitor the carbon emission trend in real time:

[0200] Local outlier factor detection formula:

[0201]

[0202] When LOF(C t) When δ (threshold) is exceeded, it is determined that the current carbon emissions are an abnormal point.

[0203] Bayesian mutation detection:

[0204]

[0205] When P(change|C t ) > τ, it is considered that the data center may enter an abnormal carbon emissions state.

[0206] 4.3 Early warning level division and response mechanism

[0207] According to the degree of carbon emissions abnormality, the early warning levels are divided, and different levels of response measures are taken:

[0208] Minor warning (Level1):

[0209] Trigger condition: C t exceeds C thresh but does not exceed 1.2C thresh .

[0210] Treatment measures: Lightweight load adjustment (reduce the CPU frequency using the load adjustment algorithm based on renewable energy prediction in step 2); Send warning information to the data center operation and maintenance personnel.

[0211] Medium warning (Level2):

[0212] Trigger condition: 1.2C thresh < C t < 1.5C thresh .

[0213] Treatment measures: Enhanced load adjustment (the load adjustment algorithm based on renewable energy prediction in step 2 is combined with the continuous cross-regional low-carbon computing power scheduling algorithm in step 3 to migrate some computing tasks to low-carbon data centers); Dynamically adjust the task scheduling strategy to increase the proportion of low-carbon power usage.

[0214] Severe warning (Level3):

[0215] Trigger condition C t > 1.5C thresh .

[0216] Treatment measures: Trigger emergency task offloading, migrate high-energy-consuming tasks to low-carbon data centers; Enable backup clean energy (such as energy storage systems); Notify the management and, if necessary, reduce the execution priority of some non-critical computing tasks.

[0217] The carbon emissions early warning algorithm based on intelligent detection mainly realizes the real-time detection and intelligent early warning of carbon emissions through the following steps:

[0218] This step can issue an early warning when carbon emissions are about to exceed the standard, and perform intelligent regulation in combination with the previous algorithms (adaptive carbon emission calculation algorithm for real-time power structure, load adjustment algorithm based on renewable energy prediction, and cross-regional low-carbon computing power scheduling algorithm), so as to achieve low-carbon operation of the data center while ensuring computing power requirements.

[0219] Embodiment 2

[0220] As Figure 2 shown, the present invention discloses a carbon emission monitoring system for a data center, including:

[0221] Grid power generation structure dynamic weight matrix construction module: By collecting real-time power generation proportion data of different energy types, combining meteorological information to predict the change trend of renewable energy power supply, calculating the power supply proportion of each energy type and constructing a dynamic weight matrix;

[0222] Renewable energy prediction load adjustment module: Using historical and real-time renewable energy data to construct a time series prediction model, predicting the future renewable energy power supply capacity, combining the computing load characteristics of the data center, dynamically adjusting the load distribution, optimizing the task execution time and resource allocation, and achieving dynamic matching with the renewable energy supply level;

[0223] Cross-regional low-carbon computing power scheduling module: Calculating the carbon emission factors and priorities of different data centers, constructing a task migration feasibility matrix, dynamically migrating computing tasks to low-carbon emission data centers according to the scheduling benefit function, optimizing the task scheduling path, reducing cross-regional data transmission delay and energy consumption, and achieving cross-regional optimal scheduling of computing power;

[0224] Carbon emission warning module: Constructing a multi-dimensional time series prediction model, predicting the carbon emission trend, real-time monitoring of carbon emission anomalies, dividing the warning levels and taking corresponding measures to ensure the stable operation of the data center under the low-carbon operation goal.

[0225] The modules in Embodiment 2 are used to implement the functions in Embodiment 1. This embodiment can be implemented by a system, which includes a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the carbon emission monitoring method for a data center described in Embodiment 1 of the present application is implemented. The system also includes other components well-known to those skilled in the art such as a communication bus and a communication interface, and their settings and functions are known in the art, so they will not be described in detail here.

[0226] In this application, the aforementioned memory can be any tangible medium that contains or stores a program, which can be used by or in conjunction with an instruction execution system, system, or device. For example, a computer-readable storage medium can be any suitable magnetic storage medium or magneto-optical storage medium, such as resistive random access memory (RRAM), dynamic random access memory (DRAM), static random access memory (SRAM), enhanced dynamic random access memory (EDRAM), high-bandwidth memory (HBM), hybrid memory cube (HMC), etc., or any other medium that can be used to store the required information and can be accessed by an application, module, or both. Any such computer storage medium can be part of the device or accessible or connectable to the device. Any application or module described in this application can be implemented using computer-readable / executable instructions that can be stored or otherwise maintained by such a computer-readable medium.

[0227] Finally, it should be noted that the above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the scope of protection of the claims of the present invention.

Claims

1. A method and system for monitoring carbon emissions in a data center, characterized in that: include: Step 1: Collect data on the proportion of power generation of different energy types in real time, combine meteorological information to predict the trend of renewable energy power supply, calculate the proportion of power supply of each energy type and construct a dynamic weight matrix; Step 2: Use historical and real-time renewable energy data to build a time series prediction model to predict future renewable energy power supply capacity. Combined with the computing load characteristics of the data center, dynamically adjust the load distribution, optimize task execution time and resource allocation, and achieve dynamic matching with the renewable energy supply level. Step 3: Calculate the carbon emission factors and priorities of different data centers, build a task migration feasibility matrix, dynamically migrate computing tasks to low-carbon emission data centers based on the scheduling benefit function, optimize task scheduling paths, reduce cross-regional data transmission latency and energy consumption, and achieve cross-regional optimized scheduling of computing power; Step 4: Build a multi-dimensional time series prediction model to predict carbon emission trends, monitor carbon emission anomalies in real time, divide warning levels and take corresponding measures to ensure that the data center operates stably under the low-carbon operation target.

2. A carbon emission monitoring method for a data center according to claim 1, characterized in that: The step 1 comprises: The real-time power supply data of each energy type is obtained through the power grid dispatching data interface, and the proportion of power input sources in the area where the data center is located is monitored in combination with the smart meter and SCADA system. At the same time, the meteorological data interface is used to obtain wind speed, solar radiation intensity, and reservoir water level information, and historical data are combined to predict the trend of renewable energy power supply changes; Calculate the power generation proportion of a single energy type and construct a dynamic weight matrix for the power grid power generation structure; Establish a time series change trend analysis, calculate the first-order derivative of each energy proportion with respect to time, construct a power generation structure change rate vector, quantify the change trend of different energy proportions, and provide a dynamic correction factor for adaptive carbon emission calculations.

3. A carbon emission monitoring method for a data center according to claim 2, characterized in that: The step 2 comprises: Extract historical renewable energy generation data from the time series database to form a time series, obtain current real-time meteorological data on solar radiation intensity, wind speed, and reservoir water level through the meteorological API, and construct a prediction input matrix; Use multimodal long and short-term memory networks for prediction, and dynamically adjust the prediction value by combining historical error correction coefficients and trend correction coefficients to improve prediction accuracy; Build a computing load weight vector based on the data center task type, calculate the load adjustment coefficient, and dynamically match the renewable energy supply level; According to the renewable energy forecast and load adjustment coefficient, the execution time of low-priority tasks is adjusted, the AI ​​computing load distribution is optimized, and task migration is performed when necessary. The computing tasks are scheduled to data centers with a higher proportion of renewable energy, so as to achieve dynamic matching of data center load and renewable energy supply.

4. A method for monitoring carbon emissions in a data center according to claim 3, characterized in that: The step 3 comprises: Based on the dynamic weight matrix of the power generation structure of the power grid, the carbon emission factors of each data center at different times are calculated, and the low-carbon computing power scheduling priority of the data center is measured by the carbon emission priority factor. The data center with a higher priority is more suitable for receiving computing power tasks. Construct a task migration feasibility matrix to evaluate the task migration latency between data centers, and combine the scheduling benefit function to determine the feasibility of migrating tasks from high-carbon emission data centers to low-carbon emission data centers; With the goal of minimizing the combined weight of carbon emissions and network latency during task scheduling, the optimal task scheduling path is calculated to avoid overload of a single data center due to a surge in tasks. Data sharding is performed on migration tasks to reduce transmission bandwidth usage and increase parallelism. Task execution is monitored through edge computing nodes, and blockchain is used to record task migration paths to ensure the traceability and security of scheduling and achieve cross-regional optimized scheduling of computing power.

5. A method for monitoring carbon emissions in a data center according to claim 4, characterized in that: The step 4 comprises: Combining data center load, renewable energy supply, grid carbon emission factors, task migration traffic and historical carbon emission data, a multi-dimensional time series prediction model is constructed, using a two-layer time series prediction network. The short-term trend prediction model is used to detect carbon emission fluctuations at the hourly level, and the long-term trend prediction model is used to predict carbon emission changes at the daily / weekly level. Set carbon emission warning thresholds based on the historical carbon emission mean and standard deviation; adopt a two-layer adaptive anomaly detection model, combined with local anomaly factor detection and Bayesian mutation detection, to monitor carbon emission trends in real time and identify anomalies; The warning levels are divided into mild, moderate and severe levels according to the degree of carbon emission abnormalities; when a mild warning is issued, a lightweight load adjustment is performed; when a moderate warning is issued, the proportion of low-carbon electricity use is optimized by combining load adjustment and cross-regional computing power scheduling; when a severe warning is issued, high-energy-consuming tasks are offloaded to low-carbon data centers, backup clean energy is enabled, and management is notified to lower the priority of non-critical tasks.

6. A carbon emission monitoring system for a data center, characterized in that: include: The module for constructing the dynamic weight matrix of the power generation structure of the power grid: by collecting the power generation proportion data of different energy types in real time, combining meteorological information to predict the trend of renewable energy power supply, calculating the power supply proportion of each energy type and constructing a dynamic weight matrix; Renewable energy prediction load adjustment module: Use historical and real-time renewable energy data to build a time series prediction model to predict future renewable energy power supply capacity. Combined with the data center computing load characteristics, it dynamically adjusts load distribution, optimizes task execution time and resource allocation, and achieves dynamic matching with renewable energy supply levels. Cross-regional low-carbon computing power scheduling module: calculates the carbon emission factors and priorities of different data centers, builds a task migration feasibility matrix, dynamically migrates computing tasks to low-carbon emission data centers according to the scheduling benefit function, optimizes task scheduling paths, reduces cross-regional data transmission latency and energy consumption, and realizes cross-regional optimized scheduling of computing power; Carbon emission early warning module: Build a multi-dimensional time series prediction model to predict carbon emission trends, monitor carbon emission anomalies in real time, divide the early warning levels and take corresponding measures to ensure the stable operation of the data center under the low-carbon operation target.

Images