Establishment of high-frequency power carbon emission intensity prediction model, prediction method and system

Abstract

The present invention provides a method and system for establishing a high-frequency electricity carbon emission intensity prediction model, and relates to the technical field of electricity carbon emission intensity prediction. The method includes: obtaining low-frequency electricity carbon emission sample data and high-frequency electricity consumption sample data in a set area within a historical period; decomposing the low-frequency electricity carbon emission sample data based on the high-frequency electricity consumption sample data to generate a high-frequency electricity carbon emission sample data set; obtaining a high-frequency electricity carbon emission intensity sample data set based on the high-frequency electricity carbon emission sample data set and the high-frequency electricity consumption sample data; and training a neural network model using the high-frequency electricity carbon emission intensity sample data set to obtain a high-frequency electricity carbon emission intensity prediction model. The present invention expands the training sample data of the high-frequency electricity carbon emission intensity prediction model through the obtained high-frequency electricity carbon emission intensity sample data, meets the demand for accurate and high-frequency electricity carbon emission intensity prediction, and realizes high-frequency and accurate prediction of electricity carbon emission intensity.

Description

Technical Field

[0001] The present invention relates to the technical field of electricity carbon emission intensity prediction, and in particular to the establishment of a high-frequency electricity carbon emission intensity prediction model, a prediction method and a system. Background Art

[0002] With the continuous development of the energy industry, the power sector has gradually become a significant source of carbon emissions. According to statistics, in recent years, power carbon emissions have accounted for approximately one-third of total carbon emissions, with a trend of increasing annually. Therefore, the low-carbon transformation of the power industry has received increasing attention. In power system carbon emission control efforts, carbon emission intensity can be predicted to measure the power system's carbon emissions. Therefore, a set of accurate and high-frequency power carbon emission measurement and prediction methods is a critical foundation for implementing various carbon reduction measures.

[0003] Some scholars have proposed methods for measuring electricity carbon emissions using high-frequency electricity data. However, existing high-frequency electricity carbon emission intensity is based on statistical analysis of fuel consumption ratios on the power generation side. This involves tracing electricity consumption and extrapolating the fuel consumed to produce the corresponding amount of electricity. Consequently, this method suffers from low update frequency and lags, making it impossible to obtain high-frequency and accurate predictions of electricity carbon emissions. Furthermore, in regions where the proportion of installed renewable energy capacity is increasing, electricity carbon emission intensity struggles to reflect the temporal changes in the "electricity-carbon" relationship, easily leading to significant errors in electricity carbon emission measurement. Summary of the Invention

[0004] To overcome the above-mentioned deficiencies of the prior art, the present invention provides a method for establishing a high-frequency power carbon emission intensity prediction model, comprising:

[0005] Obtain low-frequency electricity carbon emission sample data and high-frequency electricity consumption sample data for a set region within a historical period;

[0006] Based on the high-frequency electricity consumption sample data, the low-frequency electricity carbon emission sample data is decomposed to generate a high-frequency electricity carbon emission sample data set;

[0007] According to the high-frequency electricity carbon emission sample data set and the high-frequency electricity consumption sample data set, a high-frequency electricity carbon emission intensity sample data set is obtained;

[0008] The neural network model is trained using a high-frequency electricity carbon emission intensity sample dataset to obtain a high-frequency electricity carbon emission intensity prediction model.

[0009] Preferably, the low-frequency electricity carbon emission sample data is decomposed based on the high-frequency electricity consumption sample data to generate a high-frequency electricity carbon emission sample data set, including:

[0010] Using high-frequency electricity consumption sample data as reference data, a frequency conversion generator is used to decompose the low-frequency electricity carbon emission sample data to generate a high-frequency electricity carbon emission sample data set;

[0011] The frequency conversion generator satisfies the following expression:

[0012]

[0013] Among them, minT y is the objective function of the frequency conversion generator, P i M is the high-frequency electricity consumption data at time i in the historical period, is the high-frequency electricity consumption data at time i+1 in the historical period, E Y is the low-frequency electricity carbon emission data, is the decomposed high-frequency electricity carbon emission data at time i in the historical period, is the decomposed electricity carbon emission intensity data at time i+1 in the historical period, and l is the sum of all times in the historical period.

[0014] Preferably, after decomposing the low-frequency power carbon emission sample data according to the high-frequency power consumption sample data to generate the high-frequency power carbon emission sample data set, and before obtaining the high-frequency power carbon emission intensity sample data set according to the high-frequency power carbon emission sample data set and the high-frequency power consumption sample data, the method further includes:

[0015] Based on the high-frequency data optimization model of the differential evolution algorithm, the high-frequency power carbon emission sample data in the high-frequency power carbon emission sample data set is optimized to obtain the optimized high-frequency power carbon emission sample data.

[0016] Preferably, the high-frequency data optimization model based on the differential evolution algorithm optimizes the high-frequency power carbon emission sample data in the high-frequency power carbon emission sample data set to obtain the optimized high-frequency power carbon emission sample data, including:

[0017] Set the high-frequency electricity carbon emission sample data as the initial population and set the optimization target of the high-frequency data optimization model;

[0018] Calculate the fitness of the initial population based on the optimization goal of the high-frequency data optimization model;

[0019] The initial population is used as the current population, and the differential evolution algorithm is used to optimize the current population to generate a mutant population; the fitness of the mutant population is calculated according to the optimization target of the high-frequency data optimization model;

[0020] If the fitness of the variant population is lower than that of the current population, the variant population is used as the current population, and the optimization objective of the high-frequency data optimization model is continuously used for optimization until the fitness of the variant population is higher than that of the current population. The variant population is then determined as the optimized high-frequency power carbon emission sample data.

[0021] If the fitness of the mutant population is higher than the fitness of the current population, the mutant population is determined as the optimized high-frequency power carbon emission sample data.

[0022] Preferably, a neural network model is trained using a high-frequency power carbon emission intensity sample data set to obtain a high-frequency power carbon emission intensity prediction model, including:

[0023] The high-frequency electricity carbon emission intensity sample dataset is divided into a training set and a validation set according to the rated ratio;

[0024] The neural network model is trained based on the training set to obtain an initial prediction model for high-frequency power carbon emission intensity;

[0025] Validate the initial prediction model based on the validation set, and adjust the hyperparameters of the initial prediction model to obtain a validation prediction model;

[0026] The prediction model will be verified to determine the high-frequency electricity carbon emission intensity prediction model.

[0027] Preferably, before using the high-frequency power carbon emission intensity sample data set to train a neural network model to obtain a high-frequency power carbon emission intensity prediction model, the method further includes:

[0028] Collect actual data on high-frequency electricity carbon emission intensity during a set time period;

[0029] After determining the high-frequency power carbon emission intensity prediction model according to the verification prediction model, the method further includes:

[0030] Perform predictions based on the high-frequency power carbon emission intensity prediction model to obtain high-frequency power carbon emission intensity prediction data for a set time period;

[0031] Obtaining an evaluation value of a high-frequency power carbon emission intensity prediction model based on an error between actual high-frequency power carbon emission intensity data and corresponding high-frequency power carbon emission intensity prediction data;

[0032] If the evaluation value meets the preset conditions, the high-frequency power carbon emission intensity prediction model is determined as the final high-frequency power carbon emission intensity prediction model;

[0033] If the evaluation value does not meet the preset conditions, the high-frequency power carbon emission sample data set is optimized according to the evaluation value, and then the high-frequency power carbon emission intensity prediction model is optimized according to the optimized high-frequency power carbon emission sample data set until the evaluation value corresponding to the optimized high-frequency power carbon emission intensity prediction model meets the preset conditions, and the optimized high-frequency power carbon emission intensity prediction model is determined as the final high-frequency power carbon emission intensity prediction model.

[0034] Based on the same inventive concept, the present invention also provides a system for establishing a high-frequency power carbon emission intensity prediction model, comprising:

[0035] Prepare a data acquisition module for acquiring low-frequency electricity carbon emission sample data and high-frequency electricity consumption sample data for a set region within a historical period;

[0036] A high-frequency electricity carbon emission data set acquisition module is used to decompose low-frequency electricity carbon emission sample data based on high-frequency electricity consumption sample data to generate a high-frequency electricity carbon emission sample data set;

[0037] A high-frequency power carbon emission intensity sample data set acquisition module is used to obtain a high-frequency power carbon emission intensity sample data set based on the high-frequency power carbon emission sample data set and the high-frequency electricity consumption sample data;

[0038] The high-frequency power carbon emission intensity prediction model determination module is used to train the neural network model using a high-frequency power carbon emission intensity sample data set to obtain a high-frequency power carbon emission intensity prediction model.

[0039] Preferably, the high-frequency electricity carbon emission data set acquisition module is specifically used to:

[0040] Using high-frequency electricity consumption sample data as reference data, a frequency conversion generator is used to decompose the low-frequency electricity carbon emission sample data to generate a high-frequency electricity carbon emission sample data set;

[0041] The frequency conversion generator satisfies the following expression:

[0042]

[0043] Among them, minT y is the objective function of the frequency conversion generator, P i M is the high-frequency electricity consumption sample data at time i in the historical period, is the high-frequency electricity consumption sample data at time i+1 in the historical period, E Y is the sample data of low-frequency power carbon emissions, is the high-frequency electricity carbon emission sample data at time i in the historical period, is the high-frequency electricity carbon emission sample data at time i+1 in the historical period, and l is the sum of all times in the historical period.

[0044] Preferably, the system further comprises:

[0045] The high-frequency power carbon emission sample data optimization module is used to optimize the high-frequency power carbon emission sample data in the high-frequency power carbon emission sample data set based on the high-frequency data optimization model of the differential evolution algorithm to obtain the optimized high-frequency power carbon emission sample data.

[0046] Preferably, the high-frequency power carbon emission sample data optimization module is specifically used to:

[0047] Set the high-frequency electricity carbon emission sample data as the initial population and set the optimization target of the high-frequency data optimization model;

[0048] Calculate the fitness of the initial population based on the optimization goal of the high-frequency data optimization model;

[0049] The initial population is used as the current population, and the differential evolution algorithm is used to optimize the current population to generate a mutant population; the fitness of the mutant population is calculated according to the optimization target of the high-frequency data optimization model;

[0050] If the fitness of the variant population is lower than that of the current population, the variant population is used as the current population, and the optimization objective of the high-frequency data optimization model is continuously used for optimization until the fitness of the variant population is higher than that of the current population. The variant population is then determined as the optimized high-frequency power carbon emission sample data.

[0051] If the fitness of the mutant population is higher than the fitness of the current population, the mutant population is determined as the optimized high-frequency power carbon emission sample data.

[0052] Preferably, the high-frequency power carbon emission intensity prediction model determination module is specifically used to:

[0053] The high-frequency electricity carbon emission intensity sample dataset is divided into a training set and a validation set according to the rated ratio;

[0054] The neural network model is trained based on the training set to obtain an initial prediction model for high-frequency power carbon emission intensity;

[0055] Validate the initial prediction model based on the validation set, and adjust the hyperparameters of the initial prediction model to obtain a validation prediction model;

[0056] The prediction model will be verified to determine the high-frequency electricity carbon emission intensity prediction model.

[0057] Preferably, the system further comprises:

[0058] A high-frequency power carbon emission intensity actual data collection unit, used to collect the high-frequency power carbon emission intensity actual data for a set time period;

[0059] The system also includes a high-frequency power carbon emission intensity prediction model optimization module;

[0060] The high-frequency power carbon emission intensity prediction model optimization module is specifically used to:

[0061] Perform predictions based on the high-frequency power carbon emission intensity prediction model to obtain high-frequency power carbon emission intensity prediction data for a set time period;

[0062] Obtaining an evaluation value of a high-frequency power carbon emission intensity prediction model based on an error between actual high-frequency power carbon emission intensity data and corresponding high-frequency power carbon emission intensity prediction data;

[0063] If the evaluation value meets the preset conditions, the high-frequency power carbon emission intensity prediction model is determined as the final high-frequency power carbon emission intensity prediction model;

[0064] If the evaluation value does not meet the preset conditions, the high-frequency power carbon emission sample data set is optimized according to the evaluation value, and then the high-frequency power carbon emission intensity prediction model is optimized according to the optimized high-frequency power carbon emission sample data set until the evaluation value corresponding to the optimized high-frequency power carbon emission intensity prediction model meets the preset conditions, and the optimized high-frequency power carbon emission intensity prediction model is determined as the final high-frequency power carbon emission intensity prediction model.

[0065] Based on the same inventive concept, the present invention also provides a method for predicting carbon emission intensity of high-frequency power, comprising:

[0066] Obtain historical data on high-frequency electricity carbon emission intensity in a specified area within a specific period;

[0067] Input the historical high-frequency power carbon emission intensity data into the high-frequency power carbon emission intensity prediction model for prediction, and obtain the high-frequency power carbon emission intensity prediction value for the future period corresponding to the specific period;

[0068] The high-frequency power carbon emission intensity prediction model is established according to the method for establishing the high-frequency power carbon emission intensity prediction model described above.

[0069] Preferably, obtaining historical data on high-frequency electricity carbon emission intensity in a set region within a specific period includes:

[0070] Obtain historical data on low-frequency electricity carbon emissions and high-frequency electricity consumption in a set area within a specific period of time;

[0071] Based on the high-frequency electricity consumption historical data, the low-frequency electricity carbon emission historical data is decomposed to generate the high-frequency electricity carbon emission historical data;

[0072] Based on the historical data of high-frequency power carbon emissions and the historical data of high-frequency electricity consumption, the historical data of high-frequency power carbon emission intensity are obtained.

[0073] Based on the same inventive concept, the present invention also provides a system for predicting high-frequency power carbon emission intensity, comprising:

[0074] A module for acquiring historical data on high-frequency power carbon emission intensity, used to acquire historical data on high-frequency power carbon emission intensity in a set area within a specific period of time;

[0075] A module for obtaining a predicted value of high-frequency power carbon emission intensity is used to input historical high-frequency power carbon emission intensity data into a high-frequency power carbon emission intensity prediction model for prediction, and obtain a predicted value of high-frequency power carbon emission intensity for a future period corresponding to a specific period.

[0076] The high-frequency power carbon emission intensity prediction model is established according to the method for establishing the high-frequency power carbon emission intensity prediction model described above.

[0077] Based on the same inventive concept, the present invention further provides an electronic device, comprising: at least one processor and a memory; the memory and the processor are connected via a bus;

[0078] The memory is used to store one or more programs;

[0079] When the one or more programs are executed by the at least one processor, the method for establishing a high-frequency power carbon emission intensity prediction model or the method for predicting high-frequency power carbon emission intensity as described above is implemented.

[0080] Based on the same inventive concept, the present invention also provides a readable storage medium having an execution program stored thereon. When the execution program is executed, it implements the method for establishing a high-frequency power carbon emission intensity prediction model or the method for predicting high-frequency power carbon emission intensity as described above.

[0081] Compared with the closest prior art, the present invention has the following beneficial effects:

[0082] The present invention provides a method for establishing a high-frequency power carbon emission intensity prediction model, comprising: obtaining low-frequency power carbon emission sample data and high-frequency electricity consumption sample data in a set area within a historical period; decomposing the low-frequency power carbon emission sample data based on the high-frequency electricity consumption sample data to generate a high-frequency power carbon emission sample data set; obtaining a high-frequency power carbon emission intensity sample data set based on the high-frequency power carbon emission sample data set and the high-frequency electricity consumption sample data; and training a neural network model using the high-frequency power carbon emission intensity sample data set to obtain a high-frequency power carbon emission intensity prediction model. The present invention does not rely on power generation side fuel consumption statistics to obtain power carbon emission intensity data, but rather expands the sample data set for training the high-frequency power carbon emission intensity prediction model by obtaining a large amount of high-frequency power carbon emission intensity sample data by decomposing the low-frequency power carbon emission sample data, and then trains the high-frequency power carbon emission intensity prediction model using the large amount of high-frequency power carbon emission intensity sample data in the sample data set, thereby improving the prediction accuracy of the high-frequency power carbon emission intensity prediction model by increasing the number of training samples. It solves the problems of low updating frequency and lag in the existing technology of electricity carbon emission intensity data obtained through statistical analysis of fuel consumption ratios, as well as the problem of difficulty in quickly reflecting changes in electricity carbon emission intensity data caused by changes in installed capacity in a set area. It meets the demand for accurate and high-frequency prediction of electricity carbon emission intensity and realizes high-frequency and accurate prediction of electricity carbon emission intensity.

[0083] The present invention also provides a method for predicting high-frequency power carbon emission intensity, comprising: obtaining historical data on high-frequency power carbon emission intensity within a specific time period in a set region; inputting the historical high-frequency power carbon emission intensity data into a high-frequency power carbon emission intensity prediction model for prediction, and obtaining a predicted value of high-frequency power carbon emission intensity for a future time period corresponding to the specific time period; the high-frequency power carbon emission intensity prediction model is established according to a method for establishing a high-frequency power carbon emission intensity prediction model. The present invention obtains an accurate and high-frequency power carbon emission intensity prediction value by inputting the historical high-frequency power carbon emission intensity data into the high-frequency power carbon emission intensity prediction model for prediction, thereby meeting the demand for accurate and high-frequency power carbon emission intensity prediction and achieving high-frequency and accurate prediction of power carbon emission intensity. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] Figure 1 A schematic flow chart of a method for establishing a high-frequency power carbon emission intensity prediction model provided by the present invention;

[0085] Figure 2 A schematic flow chart of a method for predicting carbon emission intensity of high-frequency power provided by the present invention;

[0086] Figure 3 A specific example flow chart of the method for predicting high-frequency power carbon emission intensity provided by the present invention;

[0087] Figure 4 A structural diagram of a system for establishing a high-frequency power carbon emission intensity prediction model provided by the present invention;

[0088] Figure 5 This is a structural diagram of a high-frequency power carbon emission intensity prediction system provided by the present invention;

[0089] Figure 6 A schematic diagram of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0090] The following describes embodiments of the present invention in detail, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and are not to be construed as limiting the present invention.

[0091] Example 1:

[0092] The present invention provides a method for establishing a high-frequency power carbon emission intensity prediction model, specifically, Figure 1 A flow chart of a method for establishing a high-frequency power carbon emission intensity prediction model provided in an embodiment of the present invention includes the following steps as shown in the figure:

[0093] S101: Obtain low-frequency electricity carbon emission sample data and high-frequency electricity consumption sample data for a set region within a historical period;

[0094] S102: Decomposing the low-frequency electricity carbon emission sample data based on the high-frequency electricity consumption sample data to generate a high-frequency electricity carbon emission sample data set;

[0095] S103: Obtain a high-frequency power carbon emission intensity sample data set based on the high-frequency power carbon emission sample data set and the high-frequency power consumption sample data set;

[0096] S104: Using a high-frequency electricity carbon emission intensity sample data set to train a neural network model to obtain a high-frequency electricity carbon emission intensity prediction model.

[0097] The present invention expands the sample data of the high-frequency electricity carbon emission intensity prediction model by decomposing the low-frequency electricity carbon emission sample data to obtain high-frequency electricity carbon emission sample data, meets the demand for accurate and high-frequency electricity carbon emission intensity prediction, and realizes high-frequency and accurate prediction of electricity carbon emission intensity.

[0098] The feasibility of using electricity carbon emission intensity data to measure high-frequency electricity carbon emissions rests on the following premises: 1. Due to the fuel consumption characteristics of power generation units, electricity data has a significant correlation with carbon emissions in the power industry; 2. Electricity data offers advantages such as strong real-time performance, high data quality, and a solid measurement foundation, making it possible to use electricity data for low-cost, high-frequency carbon emissions accounting. Electricity carbon emission data achieves high-frequency carbon emission measurement or forecasting by multiplying electricity carbon emission intensity data by current or projected electricity consumption.

[0099] However, due to the current state of carbon measurement, technical researchers only obtain annual electricity carbon emissions data for a specific region. However, the annual measurement spans a long period of time, resulting in a limited number of electricity carbon emissions samples, making it difficult to develop an accurate electricity carbon emissions intensity prediction model. To enable the electricity carbon emissions intensity prediction model to accurately and frequently predict electricity carbon emissions intensity, it is necessary to expand the electricity carbon emissions sample dataset.

[0100] The specific method of expanding the electricity carbon emission sample dataset is described in detail below.

[0101] In this disclosure, the terms "low-frequency" and "high-frequency" refer to frequencies on a time scale, used to distinguish time precision. Specifically, low-frequency electricity carbon emission sample data can be electricity carbon emission sample data on an annual time scale; high-frequency electricity carbon emission sample data can be electricity carbon emission sample data on a monthly time scale; and high-frequency electricity consumption sample data can be electricity consumption sample data on a monthly time scale. It is understood that with the continuous development of the carbon measurement field, high frequency can be used to achieve finer time precision on a time scale, such as weekly or daily.

[0102] In S101 , low-frequency electricity carbon emission sample data and high-frequency electricity consumption sample data of a set region within a historical period are obtained.

[0103] It is understandable that the low-frequency electricity carbon emission sample data E of the region in the historical period is set Y , high-frequency electricity consumption sample data P M This includes electricity carbon emissions data and high-frequency electricity consumption data for various sectors and entities, which can be obtained through collection or measurement. The specific acquisition methods are not detailed here.

[0104] For example, the low-frequency electricity carbon emission sample data of a certain region in each of the eight years from 2016 to 2023, as well as the high-frequency electricity consumption sample data of each month in the eight years from 2016 to 2023 can be obtained.

[0105] In S102, the low-frequency electricity carbon emission sample data is decomposed based on the high-frequency electricity consumption sample data to generate a high-frequency electricity carbon emission sample data set, including:

[0106] Taking high-frequency electricity consumption sample data as reference data, a frequency conversion generator is used to decompose the low-frequency electricity carbon emission sample data to generate a high-frequency electricity carbon emission sample data set.

[0107] Among them, the frequency conversion generator satisfies the following expression:

[0108]

[0109] Among them, minT y is the objective function of the frequency conversion generator, P i M is the high-frequency electricity consumption sample data at time i in the historical period, is the high-frequency electricity consumption sample data at time i+1 in the historical period, E Y is the sample data of low-frequency power carbon emissions, is the high-frequency electricity carbon emission sample data at time i in the historical period, is the high-frequency electricity carbon emission sample data at time i+1 in the historical period, and l is the sum of all times in the historical period.

[0110] In the present invention, data decomposition is a process of decomposing low-frequency electricity carbon emission sample data to obtain high-frequency electricity carbon emission sample data, that is, obtaining electricity carbon emission sample data with higher time accuracy.

[0111] Specifically, under the condition that the objective function is satisfied, the frequency conversion generator decomposes the low-frequency power carbon emission sample data using the high-frequency power consumption sample data as reference data, obtaining the high-frequency power carbon emission sample data at each moment in the decomposed historical period, and generating a high-frequency power carbon emission sample dataset. It will be understood that the decomposed high-frequency power carbon emission sample data has the same frequency as the corresponding high-frequency power consumption sample data. In this embodiment, high frequency refers to monthly.

[0112] in, It is a constraint condition. The sum of the high-frequency electricity carbon emission sample data at all moments in the historical period constitutes the low-frequency electricity carbon emission sample data of the historical period, and the high-frequency electricity carbon emission sample data at each moment is greater than 0.

[0113] For example, by decomposing the low-frequency power carbon emission sample data, we can obtain the high-frequency power carbon emission sample data for each month in the eight years from 2016 to 2023, that is, we obtain 96 high-frequency power carbon emission sample data to form a high-frequency power carbon emission sample data set.

[0114] In this embodiment, a Denton (frequency conversion) generator is used as the data generator f(·). It is understandable that other frequency conversion methods can also be used according to actual conditions, which will not be described in detail here.

[0115] It's understandable that the high-frequency power carbon emission sample data used in this invention isn't actual high-frequency power carbon emission data, but rather is derived by decomposing low-frequency power carbon emission sample data using a frequency conversion generator. Therefore, the accuracy of this high-frequency power carbon emission sample data is questionable and may contain significant errors.

[0116] In order to improve the accuracy of the high-frequency power carbon emission sample data, further, after S102, the high-frequency power carbon emission sample data in the high-frequency power carbon emission sample data set is optimized based on the high-frequency data optimization model of the differential evolution algorithm to obtain optimized high-frequency power carbon emission sample data. In the present invention, the differential evolution algorithm is used as the core algorithm of the high-frequency data optimization model h(·). Other genetic optimization algorithms with similar effects can also be used to optimize the high-frequency power carbon emission sample data, which will not be described in detail here.

[0117] Specifically, the high-frequency electricity carbon emission sample data is set as the initial population And set the optimization goal of the high-frequency data optimization model;

[0118] Calculate the fitness of the initial population based on the optimization goal of the high-frequency data optimization model;

[0119] The initial population is used as the current population Use the differential evolution algorithm to optimize the current population and generate a mutant population Calculate the fitness of the mutant population based on the optimization goal of the high-frequency data optimization model;

[0120] If the fitness of the variant population is lower than that of the current population, that is, when the fitness decreases, the variant population is used as the current population, and the optimization objective of the high-frequency data optimization model is continued to be optimized until the fitness of the variant population is higher than that of the current population. The variant population is determined as the optimized high-frequency power carbon emission sample data;

[0121] If the fitness of the mutant population is higher than that of the current population, that is, when the fitness increases, the mutant population is determined as the optimized high-frequency power carbon emission sample data.

[0122] The high-frequency power carbon emission sample data is optimized by the above-mentioned specific method, so that the error between the optimized high-frequency power carbon emission sample data and the actual high-frequency power carbon emission data is as small as possible.

[0123] According to the optimization target of the high-frequency data optimization model, the fitness of the initial population is calculated, that is, the fitness of the high-frequency data optimization model is initialized. Specific initialization methods include but are not limited to random initialization and Gaussian initialization, which are not limited here.

[0124] In S103 , a high-frequency electricity carbon emission intensity sample data set is obtained according to the high-frequency electricity carbon emission sample data set and the high-frequency electricity consumption sample data.

[0125] The high-frequency electricity carbon emission sample data at each moment in the historical period is divided by the high-frequency electricity consumption sample data at the corresponding moment to obtain the high-frequency electricity carbon emission intensity sample data, and then generate the high-frequency electricity carbon emission intensity sample data set.

[0126] High-frequency power carbon emission intensity data represents the ratio of high-frequency power carbon emission data to high-frequency electricity consumption data. In some specific implementations, high-frequency power carbon emission intensity data may also be represented using a high-frequency power carbon emission factor or high-frequency power carbon intensity. In this disclosure, the unit of high-frequency power carbon emission data is million tons (Mt), and the unit of high-frequency electricity consumption data is gigawatt-hour (GWh).

[0127] In the specific implementation process, high-frequency electricity carbon emission sample data Or optimized high-frequency electricity carbon emission sample data The high-frequency power consumption sample data P at the corresponding time i M Divide by to obtain the high-frequency power carbon emission intensity sample data F at that moment i M The specific formula is as follows:

[0128] or

[0129] Among them, F i M is the sample data of high-frequency electricity carbon emission intensity at time i, is the high-frequency electricity carbon emission sample data at time i, is the optimized high-frequency power carbon emission sample data at time i, P i M is the high-frequency electricity consumption sample data at time i.

[0130] After completing the expansion of the power carbon emission sample data set through steps S101 to S103, in S104, the high-frequency power carbon emission sample data set is used to train the neural network model to obtain a high-frequency power carbon emission intensity prediction model.

[0131] The high-frequency power carbon emission intensity prediction model in the present invention is an end-to-end optimized high-frequency power carbon emission intensity prediction model from the input end to the prediction end. The end-to-end optimization is a process that optimizes both the high-frequency power carbon emission sample data and the high-frequency power carbon emission prediction data.

[0132] In S104, a neural network model g(·) framework is first constructed; the high-frequency power carbon emission sample data of the set period in the high-frequency power carbon emission sample data set is used. As input data, the high-frequency electricity carbon emission sample data of the period after the set period in the high-frequency electricity carbon emission sample data set is used. As output data, the high-frequency power carbon emission intensity prediction model can be expressed as Where n1 represents the total number of moments in the set time period.

[0133] For example, there are 96 optimized high-frequency power carbon emission sample data for each month in the eight years from 2016 to 2023. A total of 60 input data of optimized high-frequency power carbon emission sample data for the five years from 2016 to 2020 can be selected as the set period. The high-frequency power carbon emission sample data in January 2021 can be used as the output data to iteratively train the neural network model to obtain a high-frequency power carbon emission intensity prediction model.

[0134] The specific implementation process of iteratively training the neural network model to obtain the high-frequency power carbon emission intensity prediction model includes:

[0135] The high-frequency electricity carbon emission intensity sample dataset is divided into a training set and a validation set according to the rated ratio;

[0136] The neural network model is trained based on the training set to obtain an initial prediction model for high-frequency power carbon emission intensity;

[0137] Validate the initial prediction model based on the validation set, and adjust the hyperparameters of the initial prediction model to obtain a validation prediction model;

[0138] The prediction model will be verified to determine the high-frequency electricity carbon emission intensity prediction model.

[0139] In the present invention, the high-frequency electricity carbon emission intensity sample data set is divided into a training set and a validation set according to a rated ratio of 7:3. It can be understood that different rated ratios can also be selected to divide the high-frequency electricity carbon emission intensity sample data set according to actual model training requirements.

[0140] After training the neural network model based on the training set and obtaining the initial prediction model of high-frequency power carbon emission intensity, the initial prediction model is verified on the validation set and the hyperparameters of the initial prediction model are adjusted to obtain the validation prediction model gvalid (·), the verification prediction model can be determined as a high-frequency electricity carbon emission intensity prediction model.

[0141] It is understandable that using a neural network model to train a high-frequency electricity carbon emission intensity prediction model is only an example. In the specific implementation process, linear regression, ridge regression or other available machine learning models can also be used for training, without limitation here.

[0142] Thus, the present invention provides expanded high-frequency power carbon emission sample data, and performs iterative training on a neural network model based on the expanded high-frequency power carbon emission sample data to obtain a high-frequency power carbon emission intensity prediction model.

[0143] With the rapid development of carbon measurement in recent years, the accuracy of collected electricity carbon emissions data has gradually increased, although it still lags behind. For example, monthly electricity carbon emissions data is now available. It is understandable that with the continued development of carbon measurement, daily electricity carbon emissions data may be collected in the future.

[0144] In order to further improve the accuracy of the obtained high-frequency power carbon emission intensity prediction model, the high-frequency power carbon emission intensity prediction model can be further optimized using the actual data of high-frequency power carbon emission intensity in a set time period collected in recent years.

[0145] Before using the high-frequency power carbon emission intensity sample data set to train the neural network model to obtain the high-frequency power carbon emission intensity prediction model, in some specific implementations, actual data on high-frequency power carbon emission intensity in a set time period can also be collected.

[0146] Specifically, after determining the high-frequency power carbon emission intensity prediction model based on the verification prediction model, it also includes:

[0147] Performing a prediction based on a high-frequency power carbon emission intensity prediction model to obtain high-frequency power carbon emission intensity prediction data for the set time period;

[0148] Obtaining an evaluation value of a high-frequency power carbon emission intensity prediction model based on an error between actual high-frequency power carbon emission intensity data and corresponding high-frequency power carbon emission intensity prediction data;

[0149] If the evaluation value meets the preset conditions, the high-frequency power carbon emission intensity prediction model is determined as the final high-frequency power carbon emission intensity prediction model;

[0150] If the evaluation value does not meet the preset conditions, the high-frequency power carbon emission sample data set is optimized according to the evaluation value, and then the high-frequency power carbon emission intensity prediction model is optimized according to the optimized high-frequency power carbon emission sample data set until the evaluation value corresponding to the optimized high-frequency power carbon emission intensity prediction model meets the preset conditions, and the optimized high-frequency power carbon emission intensity prediction model is determined as the final high-frequency power carbon emission intensity prediction model.

[0151] The calculation formula of the evaluation value can be expressed as:

[0152]

[0153] Among them, ε is the evaluation value, F i M is the actual data of high-frequency power carbon emission intensity at time i in the set time period, The high-frequency power carbon emission intensity forecast data at time i in the set time period, and l1 is all the times in the set time period.

[0154] In some optional implementations, the preset condition is that the evaluation value does not decrease for ten consecutive iterations.

[0155] For example, if the actual data of high-frequency power carbon emission intensity for each month in 2023 is collected, the high-frequency power carbon emission intensity prediction data for each month in 2023 can be predicted based on the high-frequency power carbon emission intensity prediction model in S104, and the evaluation value of the high-frequency power carbon emission intensity prediction model for 2023 can be determined based on the error between the actual data of high-frequency power carbon emission intensity for each month in 2023 and the corresponding high-frequency power carbon emission intensity prediction data. After determining the evaluation value, determine whether the evaluation value is within the preset error range. If so, determine that the high-frequency power carbon emission intensity prediction model is the final high-frequency power carbon emission intensity prediction model; if not, optimize the high-frequency power carbon emission sample data set according to the evaluation value to obtain the optimized high-frequency power carbon emission sample data for each month in 2023, and then obtain the optimized high-frequency power carbon emission sample data set based on the optimized high-frequency power carbon emission sample data set and the high-frequency electricity consumption sample data. Re-train the high-frequency power carbon emission intensity sample prediction model based on the optimized high-frequency power carbon emission intensity sample data set to obtain the evaluation value corresponding to the high-frequency power carbon emission intensity prediction model, until it is determined that the evaluation value is within the preset error range, and determine that the optimized high-frequency power carbon emission intensity prediction model is the final high-frequency power carbon emission intensity prediction model.

[0156] The method for establishing a high-frequency power carbon emission intensity prediction model provided by the present invention obtains low-frequency power carbon emission sample data and high-frequency electricity consumption sample data in a set area within a historical period; decomposes the low-frequency power carbon emission sample data based on the high-frequency electricity consumption sample data to generate a high-frequency power carbon emission sample data set; obtains a high-frequency power carbon emission intensity sample data set based on the high-frequency power carbon emission sample data set and the high-frequency electricity consumption sample data; and trains a neural network model using the high-frequency power carbon emission intensity sample data set to obtain a high-frequency power carbon emission intensity prediction model. The present invention expands the sample data of the high-frequency power carbon emission intensity prediction model through the obtained high-frequency power carbon emission intensity sample data, meets the demand for accurate and high-frequency power carbon emission intensity prediction, and realizes high-frequency and accurate prediction of power carbon emission intensity.

[0157] Example 2:

[0158] Based on the same inventive concept, the present invention also provides a method for predicting high-frequency power carbon emission intensity. Specifically, Figure 2 The flowchart of the method for predicting high-frequency power carbon emission intensity provided by an embodiment of the present invention includes the following steps as shown in the figure:

[0159] S201, obtaining historical data on high-frequency electricity carbon emission intensity in a set region within a specific period;

[0160] S202: Inputting historical high-frequency power carbon emission intensity data into a high-frequency power carbon emission intensity prediction model for prediction to obtain a predicted value of high-frequency power carbon emission intensity for a future period corresponding to a specific period;

[0161] The high-frequency power carbon emission intensity prediction model is established according to the method for establishing the high-frequency power carbon emission intensity prediction model described above.

[0162] In the method S201 for obtaining the historical data of high-frequency electricity carbon emission intensity in a set area within a specific time period, if the high-frequency electricity carbon emission sample data contains high-frequency electricity carbon emission sample data for a specific time period, the obtained high-frequency electricity carbon emission sample data can be directly determined as the historical data of high-frequency electricity carbon emission intensity; if the high-frequency electricity carbon emission sample data does not contain high-frequency electricity carbon emission sample data for a specific time period, the following method can be adopted to obtain the historical data of high-frequency electricity carbon emission intensity in a set area within a specific time period.

[0163] Obtaining historical data on high-frequency electricity carbon emission intensity for a specific period of time in a specified region includes:

[0164] Obtain historical data on low-frequency electricity carbon emissions and high-frequency electricity consumption in a set area within a specific period of time;

[0165] Based on the high-frequency electricity consumption historical data, the low-frequency electricity carbon emission historical data is decomposed to generate the high-frequency electricity carbon emission historical data;

[0166] Based on the historical data of high-frequency power carbon emissions and the historical data of high-frequency electricity consumption, the historical data of high-frequency power carbon emission intensity are obtained.

[0167] The method for obtaining the historical data of high-frequency power carbon emission intensity in a specific period of time in a given region is the same as that for obtaining the sample data of high-frequency power carbon emission intensity in Example 1, and will not be described in detail here. Specifically, the sample data of high-frequency power carbon emission intensity in Example 1 may be referred to.

[0168] In this embodiment, historical data on high-frequency power carbon emission intensity in a set area within a specific time period is obtained, and the historical data on high-frequency power carbon emission intensity is input into a previously established high-frequency power carbon emission intensity prediction model to predict the high-frequency power carbon emissions in the future time period corresponding to the specific time period, and a predicted value of high-frequency power carbon emission intensity in the future time period corresponding to the specific time period is obtained.

[0169] The high-frequency electricity carbon emission intensity prediction method provided in this embodiment inputs historical high-frequency electricity carbon emission intensity data into a high-frequency electricity carbon emission intensity prediction model for prediction, and obtains accurate and high-frequency electricity carbon emission intensity prediction values, thereby meeting the demand for accurate and high-frequency electricity carbon emission intensity prediction and realizing high-frequency and accurate prediction of electricity carbon emission intensity.

[0170] like Figure 3 The figure shows a specific example schematic diagram of the high-frequency power carbon emission intensity prediction method provided by the present invention, which is explained below.

[0171] S301: Establish a data generator based on the Denton method, and use high-frequency electricity consumption data as a reference to guide the decomposition of low-frequency electricity carbon emission data to obtain the decomposed high-frequency electricity carbon emission data;

[0172] S302: Build a high-frequency data optimization model based on the differential evolution algorithm to adjust the high-frequency electricity carbon emission data;

[0173] S303: Divide the adjusted high-frequency power carbon emission data by the corresponding high-frequency power consumption data to obtain high-frequency power carbon emission intensity data;

[0174] S304: Construct a high-frequency power carbon emission intensity prediction model framework;

[0175] S305: Using the high-frequency power carbon emission intensity data to train a high-frequency power carbon emission intensity prediction model framework, and using the actual high-frequency power carbon emission intensity data to evaluate the trained high-frequency power carbon emission intensity prediction model, to obtain an evaluation value of the high-frequency power carbon emission intensity prediction model;

[0176] S306: Determine whether the evaluation value meets the preset conditions; if so, execute S307 to determine the final high-frequency power carbon emission intensity prediction model; if not, return to S302, adjust the high-frequency power carbon emission data, and continue to execute S302-S306 until the evaluation value meets the preset conditions, and execute S307;

[0177] S308: Inputting the historical data of high-frequency power carbon emission intensity into the final high-frequency power carbon emission intensity prediction model to obtain a predicted value of high-frequency power carbon emission intensity for the period to be predicted.

[0178] Existing methods for measuring carbon emissions from electricity are divided into the fuel emission factor method, the material balance method, and the actual measurement method. These methods are based on statistical analysis of the fuel consumption ratio on the power generation side, and it is difficult to achieve high-frequency, high-accuracy, and low-cost carbon emission measurement.

[0179] In the present invention, instead of relying on statistics based on the fuel consumption ratio on the power generation side, a large amount of high-frequency power carbon emission data is obtained by decomposing the low-frequency power carbon emission data, ensuring that the machine learning model can be fully trained, thereby improving the prediction accuracy of high-frequency power carbon emission intensity.

[0180] In addition, the performance evaluation of the downstream prediction task is regarded as the optimization target of the upstream decomposition task, and is fed back into the step of decomposing the high-frequency power carbon emission data to guide the optimization of the decomposed high-frequency power carbon emission data, thereby ensuring the authenticity and effectiveness of the decomposition of the historical low-frequency power carbon emission data in the absence of real historical high-frequency power carbon emission data, and improving the robustness of the high-frequency power carbon emission intensity prediction model.

[0181] Example 3:

[0182] Based on the same inventive concept, the present invention also provides a system 400 for establishing a high-frequency power carbon emission intensity prediction model. The system structure is as follows: Figure 4 As shown, the system 400 includes:

[0183] Prepare a data acquisition module 401 for acquiring low-frequency electricity carbon emission sample data and high-frequency electricity consumption sample data for a set region within a historical period;

[0184] The high-frequency power carbon emission data set acquisition module 402 is used to decompose the low-frequency power carbon emission sample data based on the high-frequency power consumption sample data to generate a high-frequency power carbon emission sample data set;

[0185] The high-frequency power carbon emission intensity sample data set acquisition module 403 is used to obtain the high-frequency power carbon emission intensity sample data set according to the high-frequency power carbon emission sample data set and the high-frequency electricity consumption sample data;

[0186] The high-frequency power carbon emission intensity prediction model determination module 404 is used to train a neural network model using a high-frequency power carbon emission intensity sample data set to obtain a high-frequency power carbon emission intensity prediction model.

[0187] Preferably, the high-frequency power carbon emission data set acquisition module 402 is specifically configured to:

[0188] Using high-frequency electricity consumption sample data as reference data, a frequency conversion generator is used to decompose the low-frequency electricity carbon emission sample data to generate a high-frequency electricity carbon emission sample data set;

[0189] The frequency conversion generator satisfies the following expression:

[0190]

[0191] Among them, minT y is the objective function of the frequency conversion generator, P i M is the high-frequency electricity consumption sample data at time i in the historical period, is the high-frequency electricity consumption sample data at time i+1 in the historical period, E Y is the sample data of low-frequency power carbon emissions, is the high-frequency electricity carbon emission sample data at time i in the historical period, is the high-frequency electricity carbon emission sample data at time i+1 in the historical period, and l is the sum of all times in the historical period.

[0192] Preferably, the system further comprises:

[0193] The high-frequency power carbon emission sample data optimization module is used to optimize the high-frequency power carbon emission sample data in the high-frequency power carbon emission sample data set based on the high-frequency data optimization model of the differential evolution algorithm to obtain the optimized high-frequency power carbon emission sample data.

[0194] Preferably, the high-frequency power carbon emission sample data optimization module is specifically used to:

[0195] Set the high-frequency electricity carbon emission sample data as the initial population and set the optimization target of the high-frequency data optimization model;

[0196] Calculate the fitness of the initial population based on the optimization goal of the high-frequency data optimization model;

[0197] The initial population is used as the current population, and the differential evolution algorithm is used to optimize the current population to generate a mutant population; the fitness of the mutant population is calculated according to the optimization target of the high-frequency data optimization model;

[0198] If the fitness of the variant population is lower than that of the current population, the variant population is used as the current population, and the optimization objective of the high-frequency data optimization model is continuously used for optimization until the fitness of the variant population is higher than that of the current population. The variant population is then determined as the optimized high-frequency power carbon emission sample data.

[0199] If the fitness of the mutant population is higher than the fitness of the current population, the mutant population is determined as the optimized high-frequency power carbon emission sample data.

[0200] Preferably, the high-frequency power carbon emission intensity prediction model determination module 404 is specifically configured to:

[0201] The high-frequency electricity carbon emission intensity sample dataset is divided into a training set and a validation set according to the rated ratio;

[0202] The neural network model is trained based on the training set to obtain an initial prediction model for high-frequency power carbon emission intensity;

[0203] Validate the initial prediction model based on the validation set, and adjust the hyperparameters of the initial prediction model to obtain a validation prediction model;

[0204] The prediction model will be verified to determine the high-frequency electricity carbon emission intensity prediction model.

[0205] Preferably, the system further comprises:

[0206] A high-frequency power carbon emission intensity actual data collection unit, used to collect the high-frequency power carbon emission intensity actual data for a set time period;

[0207] The system also includes a high-frequency power carbon emission intensity prediction model optimization module;

[0208] The high-frequency power carbon emission intensity prediction model optimization module is specifically used to:

[0209] Perform predictions based on the high-frequency power carbon emission intensity prediction model to obtain high-frequency power carbon emission intensity prediction data for a set time period;

[0210] Obtaining an evaluation value of a high-frequency power carbon emission intensity prediction model based on an error between actual high-frequency power carbon emission intensity data and corresponding high-frequency power carbon emission intensity prediction data;

[0211] If the evaluation value meets the preset conditions, the high-frequency power carbon emission intensity prediction model is determined as the final high-frequency power carbon emission intensity prediction model;

[0212] If the evaluation value does not meet the preset conditions, the high-frequency power carbon emission sample data set is optimized according to the evaluation value, and then the high-frequency power carbon emission intensity prediction model is optimized according to the optimized high-frequency power carbon emission sample data set until the evaluation value corresponding to the optimized high-frequency power carbon emission intensity prediction model meets the preset conditions, and the optimized high-frequency power carbon emission intensity prediction model is determined as the final high-frequency power carbon emission intensity prediction model.

[0213] Example 4:

[0214] Based on the same inventive concept, the present invention also provides a high-frequency power carbon emission intensity prediction system 500, the system structure is as follows: Figure 5 As shown, the system 500 includes:

[0215] The high-frequency power carbon emission intensity historical data acquisition module 501 is used to acquire the high-frequency power carbon emission intensity historical data of a set area within a specific period of time;

[0216] The high-frequency power carbon emission intensity prediction value obtaining module 502 is used to input historical high-frequency power carbon emission intensity data into the high-frequency power carbon emission intensity prediction model for prediction, and obtain the high-frequency power carbon emission intensity prediction value for the future period corresponding to the specific period;

[0217] The high-frequency power carbon emission intensity prediction model is established according to the method for establishing the high-frequency power carbon emission intensity prediction model described above.

[0218] Example 5:

[0219] Based on the same inventive concept, Figure 6 As shown, the present invention also provides an electronic device, which may be a computer, a single-chip microcomputer, a smart mobile device, or the like. The electronic device in this embodiment may include a processor, a memory, a transceiver component, and the like. The memory, processor, and transceiver component are connected via a bus; the memory may be used to store an execution program, which may include instructions; and the processor may be used to execute the instructions stored in the memory. The memory may also be used to store data, which may be accessed and / or modified during the execution of the instructions.

[0220] The processor may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in a readable storage medium to implement the corresponding method flow or corresponding function, so as to realize the method for establishing a high-frequency power carbon emission intensity prediction model or the steps of the high-frequency power carbon emission intensity prediction method in the above-mentioned embodiment.

[0221] Example 6:

[0222] Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory), which is a memory device in the electronic device for storing programs and data. It can be understood that the readable storage medium here can include both the built-in storage medium in the electronic device and, of course, the extended storage medium supported by the electronic device. The storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space. These instructions can be one or more execution programs (including program codes). It should be noted that the storage medium here can be a high-speed RAM memory or a non-volatile memory (non-volatile memory), such as at least one disk storage. The processor can load and execute one or more instructions stored in the storage medium to implement the method for establishing a high-frequency power carbon emission intensity prediction model or the steps of the high-frequency power carbon emission intensity prediction method in the above embodiment.

[0223] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0224] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0225] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0226] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0227] Finally, 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 its scope of protection. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that after reading the present invention, those skilled in the art may still make various changes, modifications or equivalent substitutions to the specific implementation methods of the application, but these changes, modifications or equivalent substitutions are all within the scope of protection of the pending claims.

Claims

1. A method for establishing a high-frequency power carbon emission intensity prediction model, characterized in that: include: Obtain low-frequency electricity carbon emission sample data and high-frequency electricity consumption sample data for a set region within a historical period; Decomposing the low-frequency electricity carbon emission sample data according to the high-frequency electricity consumption sample data to generate a high-frequency electricity carbon emission sample data set; Obtaining a high-frequency power carbon emission intensity sample data set based on the high-frequency power carbon emission sample data set and the high-frequency power consumption sample data set; The high-frequency power carbon emission intensity sample data set is used to train a neural network model to obtain a high-frequency power carbon emission intensity prediction model; The step of decomposing the low-frequency electricity carbon emission sample data based on the high-frequency electricity consumption sample data to generate a high-frequency electricity carbon emission sample data set includes: Using the high-frequency electricity consumption sample data as reference data, a frequency conversion generator is used to decompose the low-frequency electricity carbon emission sample data to generate a high-frequency electricity carbon emission sample data set; The frequency conversion generator satisfies the following expression: Among them, minT y is the objective function of the frequency conversion generator, is the high-frequency electricity consumption sample data at time i in the historical period, is the high-frequency electricity consumption sample data at time i+1 in the historical period, E Y is the sample data of low-frequency power carbon emissions, is the high-frequency electricity carbon emission sample data at time i in the historical period, is the high-frequency electricity carbon emission sample data at time i+1 in the historical period, and l is the sum of all times in the historical period.

2. The method according to claim 1, characterized in that After decomposing the low-frequency power carbon emission sample data according to the high-frequency power consumption sample data to generate a high-frequency power carbon emission sample data set, and before obtaining a high-frequency power carbon emission intensity sample data set according to the high-frequency power carbon emission sample data set and the high-frequency power consumption sample data, the method further includes: Based on a high-frequency data optimization model of a differential evolution algorithm, the high-frequency power carbon emission sample data in the high-frequency power carbon emission sample data set is optimized to obtain optimized high-frequency power carbon emission sample data.

3. The method according to claim 2, characterized in that The high-frequency data optimization model based on the differential evolution algorithm optimizes the high-frequency power carbon emission sample data in the high-frequency power carbon emission sample data set to obtain optimized high-frequency power carbon emission sample data, including: Setting the high-frequency electricity carbon emission sample data as an initial population and setting an optimization target for the high-frequency data optimization model; Calculating the fitness of the initial population according to the optimization target of the high-frequency data optimization model; Taking the initial population as the current population, optimizing the current population using a differential evolution algorithm to generate a mutant population; calculating the fitness of the mutant population according to the optimization target of the high-frequency data optimization model; If the fitness of the variant population is lower than the fitness of the current population, the variant population is used as the current population, and the optimization objective of the high-frequency data optimization model is continuously used for optimization until the fitness of the variant population is higher than the fitness of the current population, and the variant population is determined as the optimized high-frequency power carbon emission sample data; If the fitness of the variant population is higher than the fitness of the current population, the variant population is determined as the optimized high-frequency power carbon emission sample data.

4. The method according to claim 1, wherein The neural network model is trained using the high-frequency power carbon emission intensity sample data set to obtain a high-frequency power carbon emission intensity prediction model, including: Dividing the high-frequency electricity carbon emission intensity sample dataset into a training set and a validation set according to a rated ratio; Training the neural network model according to the training set to obtain an initial prediction model for high-frequency power carbon emission intensity; Validating the initial prediction model based on the validation set, and adjusting the hyperparameters of the initial prediction model to obtain a validation prediction model; The verification prediction model is used to determine the high-frequency electricity carbon emission intensity prediction model.

5. The method according to claim 4, characterized in that Before the neural network model is trained using the high-frequency power carbon emission intensity sample data set to obtain a high-frequency power carbon emission intensity prediction model, the method further includes: Collect actual data on high-frequency electricity carbon emission intensity during a set time period; After determining the high-frequency power carbon emission intensity prediction model according to the verification prediction model, the method further includes: Performing a prediction based on the high-frequency power carbon emission intensity prediction model to obtain high-frequency power carbon emission intensity prediction data for the set time period; Obtaining an evaluation value of the high-frequency power carbon emission intensity prediction model based on an error between the actual data of the high-frequency power carbon emission intensity and the corresponding predicted data of the high-frequency power carbon emission intensity; If the evaluation value meets the preset conditions, determining the high-frequency power carbon emission intensity prediction model as the final high-frequency power carbon emission intensity prediction model; If the evaluation value does not meet the preset conditions, the high-frequency power carbon emission sample data set is optimized according to the evaluation value, and then the high-frequency power carbon emission intensity prediction model is optimized according to the optimized high-frequency power carbon emission sample data set, until the evaluation value corresponding to the optimized high-frequency power carbon emission intensity prediction model meets the preset conditions, and the optimized high-frequency power carbon emission intensity prediction model is determined to be the final high-frequency power carbon emission intensity prediction model.

6. A system for establishing a high-frequency power carbon emission intensity prediction model, characterized in that: include: Prepare a data acquisition module for acquiring low-frequency electricity carbon emission sample data and high-frequency electricity consumption sample data for a set region within a historical period; a high-frequency electricity carbon emission data set acquisition module, configured to decompose the low-frequency electricity carbon emission sample data according to the high-frequency electricity consumption sample data to generate a high-frequency electricity carbon emission sample data set; A high-frequency power carbon emission intensity sample data set acquisition module, configured to obtain a high-frequency power carbon emission intensity sample data set based on the high-frequency power carbon emission sample data set and the high-frequency electricity consumption sample data; a high-frequency power carbon emission intensity prediction model determination module, configured to train a neural network model using the high-frequency power carbon emission intensity sample data set to obtain a high-frequency power carbon emission intensity prediction model; The high-frequency electricity carbon emission data set acquisition module is specifically used to: Using the high-frequency electricity consumption sample data as reference data, a frequency conversion generator is used to decompose the low-frequency electricity carbon emission sample data to generate a high-frequency electricity carbon emission sample data set; The frequency conversion generator satisfies the following expression: Among them, minT y is the objective function of the frequency conversion generator, is the high-frequency electricity consumption sample data at time i in the historical period, is the high-frequency electricity consumption sample data at time i+1 in the historical period, E Y is the sample data of low-frequency power carbon emissions, is the high-frequency electricity carbon emission sample data at time i in the historical period, is the high-frequency electricity carbon emission sample data at time i+1 in the historical period, and l is the sum of all times in the historical period.

7. A method for predicting carbon emission intensity of high-frequency electricity, characterized in that: include: Obtain historical data on high-frequency electricity carbon emission intensity in a specified area within a specific period; Inputting the historical high-frequency power carbon emission intensity data into a high-frequency power carbon emission intensity prediction model for prediction, and obtaining a high-frequency power carbon emission intensity prediction value for a future period corresponding to a specific period; The high-frequency power carbon emission intensity prediction model is established according to the method for establishing a high-frequency power carbon emission intensity prediction model according to any one of claims 1-5.

8. The method according to claim 7, characterized in that The acquisition of historical data on high-frequency electricity carbon emission intensity in a specific period of time in a set region includes: Obtain historical data on low-frequency electricity carbon emissions and high-frequency electricity consumption in a set area within a specific period of time; Decomposing the low-frequency electricity carbon emission historical data according to the high-frequency electricity consumption historical data to generate high-frequency electricity carbon emission historical data; The high-frequency power carbon emission intensity historical data is obtained based on the high-frequency power carbon emission historical data and the high-frequency power consumption historical data.

9. A prediction system for high-frequency power carbon emission intensity, characterized in that: include: A module for acquiring historical data on high-frequency power carbon emission intensity, used to acquire historical data on high-frequency power carbon emission intensity in a set area within a specific period of time; A high-frequency power carbon emission intensity prediction value acquisition module is used to input the historical high-frequency power carbon emission intensity data into a high-frequency power carbon emission intensity prediction model for prediction, and obtain a high-frequency power carbon emission intensity prediction value for a future period corresponding to a specific period; The high-frequency power carbon emission intensity prediction model is established according to the method for establishing a high-frequency power carbon emission intensity prediction model according to any one of claims 1-5.

Images