A regional carbon emission prediction method and device based on periodic adaptation, a terminal device, and a storage medium

Abstract

The application discloses a regional carbon emission prediction method and device based on periodic self-adaption, a terminal equipment and a storage medium, and belongs to the technical field of carbon emission prediction. The method is as follows: obtaining regional carbon emission historical time series data, corresponding regional historical power consumption characteristics and regional historical meteorological characteristics of a region to be predicted, and calculating a time domain candidate period, a frequency domain candidate period and an information criterion candidate period; if the three candidate periods are different from each other, taking the information criterion candidate period as a global optimal period value; otherwise, taking the value of the same candidate period as the global optimal period value; finally, combining the regional carbon emission historical time series data, the regional historical power consumption characteristics and the regional historical meteorological characteristics, the regional predicted carbon emission of the region to be predicted in a preset future period is predicted. Through the implementation of the application, the problem that the carbon emission prediction result is inaccurate due to the fact that the prior art relies on a manually preset static period parameter or a solidified experience mode to set an optimal period can be solved.

Description

Technical Field

[0001] This invention relates to the field of carbon emission prediction technology, and in particular to a method, apparatus, terminal equipment and storage medium for predicting regional carbon emissions based on periodic adaptive methods. Background Technology

[0002] With the increasingly severe global climate change problem, achieving "carbon peaking and carbon neutrality" has become an important strategic goal. Regional carbon emissions are a core indicator for measuring the effectiveness of local emission reduction efforts. Because regional carbon emissions are influenced by multiple factors, including seasonal climate changes, holiday cycles, and macroeconomic industrial policies, their time-series data not only exhibit strong seasonal fluctuations but also show a non-linear evolution trend in the long term. Therefore, how to utilize limited historical data to construct a high-precision medium- and long-term regional carbon emission prediction model is of great practical significance for governments to formulate emission reduction quotas and for enterprises to participate in carbon market trading.

[0003] Existing carbon emission prediction schemes typically rely on seasonal characteristics to predict carbon emissions. However, when dealing with seasonal characteristics, they depend on manually preset static periodic parameters or fixed experience patterns to set the optimal period. This makes it difficult for the optimal period to dynamically adapt to changes in industrial structure and external environment in different regions, resulting in drift in the carbon emission cycle and consequently, inaccurate prediction results. Summary of the Invention

[0004] This invention provides a method, apparatus, terminal device, and storage medium for predicting regional carbon emissions based on periodic adaptive methods. It can solve the problem that existing technologies rely on manually preset static periodic parameters or fixed experience patterns to set the optimal period, which leads to inaccurate carbon emission prediction results.

[0005] An embodiment of the present invention provides a regional carbon emission prediction method based on periodic adaptive methods, comprising: Obtain historical time-series data of regional carbon emissions in the area to be predicted, as well as the corresponding historical electricity consumption characteristics and historical meteorological characteristics of the area. Based on historical time-series data of regional carbon emissions and several preset periods, the autocorrelation function, power spectral density, and Akaike information criterion value for each preset period were calculated. Based on the autocorrelation function, power spectral density, and Akaike information criterion value, the time-domain candidate period, frequency-domain candidate period, and information criterion candidate period are determined. If the time-domain candidate period, frequency-domain candidate period, and information criterion candidate period are all different, then the information criterion candidate period is taken as the globally optimal period value; otherwise, the values ​​of the same candidate periods are taken as the globally optimal period value. Based on the historical time-series data of regional carbon emissions, historical electricity consumption characteristics of the region, historical meteorological characteristics of the region, and the global optimal period value, the predicted regional carbon emissions of the region to be predicted under the preset future time period are predicted.

[0006] Furthermore, the calculation of the autocorrelation function, power spectral density, and Akaike information criterion value for each preset period, based on historical time-series data of regional carbon emissions and several preset periods, includes: The series mean was calculated based on the historical time-series data of carbon emissions in the region. The autocorrelation function is calculated based on the historical time-series data of carbon emissions in the region and the mean of the series. The historical time-series data of carbon emissions in the region are subjected to discrete Fourier transform to obtain the Fourier transform results; The power spectral density is calculated based on the Fourier transform results. Based on the historical time series data of carbon emissions in the region, a basic exponential smoothing model corresponding to each preset period is fitted. Based on the smoothing models of each basic index and the historical time series data of carbon emissions in the region, the maximum likelihood estimate corresponding to each preset period is calculated. Based on the maximum likelihood estimation, the Akaike information criterion value for each preset period is calculated.

[0007] Furthermore, determining the time-domain candidate period, frequency-domain candidate period, and information criterion candidate period based on the autocorrelation function, power spectral density, and Akaike information criterion value includes: The first lag order in the autocorrelation function that exceeds the upper limit of the preset confidence interval and is the peak value of the autocorrelation coefficient is taken as the time-domain candidate period; The frequency corresponding to the maximum value of the power spectral density is taken as the dominant frequency, and the reciprocal of the dominant frequency is rounded to obtain the frequency domain candidate period. The preset period corresponding to the smallest Akaike information criterion value is used as the alternative period for the information criterion.

[0008] Furthermore, the step of predicting the predicted regional carbon emissions of the region under a preset future period based on the region's historical carbon emission time-series data, historical electricity consumption characteristics, historical meteorological characteristics, and the global optimal periodic value includes: Obtain the regional predicted electricity consumption characteristics and regional predicted meteorological characteristics of the area to be predicted under the preset future time period; Based on the historical time-series data of carbon emissions in the region, the global optimal periodic value, the historical electricity consumption characteristics of the region, and the historical meteorological characteristics of the region, a carbon emission index smoothing model and a residual regression model are constructed. Based on the historical time-series data of carbon emissions in the region, the predicted electricity consumption characteristics of the region, the predicted meteorological characteristics of the region, the carbon emission index smoothing model, and the residual regression model, the predicted carbon emissions of the region are obtained.

[0009] Furthermore, based on the historical time-series data of regional carbon emissions, the globally optimal periodic value, the historical electricity consumption characteristics of the region, and the historical meteorological characteristics of the region, a carbon emission index smoothing model and a residual regression model are constructed, including: Based on the historical time-series data of carbon emissions in the region and the global optimal periodic value, a carbon emission index smoothing model is fitted. Based on the carbon emission index smoothing model, each historical carbon emission data in the historical time series data of the region is fitted and predicted to obtain the fitted value of each historical carbon emission data. Based on the historical carbon emission data and the corresponding fitted values, the fitting residuals of each historical carbon emission data are calculated. Based on the historical electricity consumption characteristics and historical meteorological characteristics of the region, a historical covariate feature matrix is ​​constructed; Using the historical covariate feature matrix as input and the fitting residual as output, a residual regression model is constructed to characterize the mapping relationship between the covariate feature matrix and the fitting residual.

[0010] Furthermore, the step of predicting the predicted carbon emissions for the region based on the historical time-series data of regional carbon emissions, predicted regional electricity consumption characteristics, predicted regional meteorological characteristics, carbon emission index smoothing model, and residual regression model includes: Based on the predicted electricity consumption characteristics and predicted meteorological characteristics of the region, a prediction covariate feature matrix is ​​constructed; Based on the historical time-series data of carbon emissions in the region and the carbon emission index smoothing model, the baseline predicted value of carbon emissions is calculated. Based on the predicted covariate feature matrix and the residual regression model, the predicted residual values ​​are calculated. The sum of the baseline carbon emission forecast and the residual forecast is taken as the predicted carbon emission for the region.

[0011] Based on the above method embodiments, the present invention provides corresponding apparatus embodiments; This invention provides a periodically adaptive regional carbon emission prediction device, comprising: The system includes a data acquisition module, a data calculation module, a cycle calculation module, an optimal cycle determination module, and a carbon emission prediction module. The data acquisition module is used to acquire historical time-series data of regional carbon emissions in the area to be predicted, as well as the historical electricity consumption characteristics and historical meteorological characteristics of the area corresponding to the historical time-series data of regional carbon emissions. The data calculation module is used to calculate the autocorrelation function, power spectral density, and Akaike information criterion value for each preset period based on historical time-series data of regional carbon emissions and several preset periods. The period calculation module is used to determine the time-domain candidate period, frequency-domain candidate period, and information criterion candidate period based on the autocorrelation function, power spectral density, and Akaike information criterion value. The optimal period determination module is used to determine the global optimal period value when the time-domain candidate period, frequency-domain candidate period, and information criterion candidate period are different; otherwise, the value of the same candidate period is taken as the global optimal period value. The carbon emission prediction module is used to predict the regional predicted carbon emissions of the area to be predicted in a preset future period based on the historical time series data of regional carbon emissions, historical electricity consumption characteristics of the area, historical meteorological characteristics of the area, and the global optimal period value.

[0012] Furthermore, the data calculation module includes: The system includes a sequence mean calculation unit, an autocorrelation function calculation unit, a Fourier transform unit, a power spectral density calculation unit, a basic exponential smoothing model construction unit, a maximum likelihood estimation calculation unit, and an Akaike information criterion value calculation unit. The sequence mean calculation unit is used to calculate the sequence mean based on the historical time-series data of carbon emissions in the region. The autocorrelation function calculation unit is used to calculate the autocorrelation function based on the historical time series data of carbon emissions in the region and the mean of the series. The Fourier transform unit is used to perform discrete Fourier transform on the historical time-series data of carbon emissions in the region to obtain the Fourier transform result. The power spectral density calculation unit is used to calculate the power spectral density based on the Fourier transform result; The basic exponential smoothing model construction unit is used to fit the basic exponential smoothing model corresponding to each preset period based on the historical time series data of carbon emissions in the region. The maximum likelihood estimation calculation unit is used to calculate the maximum likelihood estimate corresponding to each preset period based on the smoothing models of each basic index and the historical time series data of carbon emissions in the region. The Akaike Information Criterion Value Calculation Unit is used to calculate the Akaike Information Criterion Value for each preset period based on the maximum likelihood estimation.

[0013] Based on the above method embodiments, the present invention provides a corresponding terminal device embodiment; The present invention provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the periodic adaptive regional carbon emission prediction method described in any embodiment of the present invention.

[0014] Based on the above method embodiments, the present invention provides a corresponding storage medium embodiment; The present invention provides a storage medium including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the periodic adaptive regional carbon emission prediction method described in any embodiment of the present invention.

[0015] The embodiments of the present invention have the following beneficial effects: This invention provides a method, apparatus, terminal device, and storage medium for predicting regional carbon emissions based on periodic adaptive methods. The method includes: acquiring historical time-series data of regional carbon emissions in the region to be predicted, as well as historical electricity consumption characteristics and historical meteorological characteristics of the region corresponding to the historical time-series data; calculating an autocorrelation function, a power spectral density, and an Akaike information criterion value for each preset period based on the historical time-series data of regional carbon emissions and several preset periods; determining a time-domain candidate period, a frequency-domain candidate period, and an information criterion candidate period based on the autocorrelation function, the power spectral density, and the Akaike information criterion value; if the time-domain candidate period, the frequency-domain candidate period, and the information criterion candidate period are different, then the information criterion candidate period is taken as the globally optimal period value; otherwise, the values ​​of the same candidate periods are taken as the globally optimal period value; and predicting the predicted regional carbon emissions of the region to be predicted for a preset future period based on the historical time-series data of regional carbon emissions, historical electricity consumption characteristics, historical meteorological characteristics, and the globally optimal period value. Therefore, in this invention, the time-domain candidate period, frequency-domain candidate period, and information criterion candidate period are first determined by calculating the autocorrelation function, power spectral density, and Akaike information criterion value for each preset period. Then, based on an adaptive period arbitration mechanism that combines majority voting and criterion priority, the same candidate period or the information criterion candidate period is used as the globally optimal period value and subsequently participates in the carbon emission prediction calculation. This adaptive period arbitration mechanism, compared to a preset fixed period, can better adapt to the carbon emission characteristics of the region, thereby improving the accuracy of the prediction results. Attached Figure Description

[0016] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a schematic flowchart of a regional carbon emission prediction method based on periodic adaptive method provided in an embodiment of the present invention.

[0018] Figure 2 This is a schematic diagram of the autocorrelation function provided in an embodiment of the present invention.

[0019] Figure 3 This is a schematic diagram of the power spectral density provided in an embodiment of the present invention.

[0020] Figure 4 This is a schematic diagram of the Akaike Information Criterion Curve provided in an embodiment of the present invention.

[0021] Figure 5 This is a schematic diagram of carbon emission prediction results from different models provided in an embodiment of the present invention.

[0022] Figure 6 This is a schematic diagram of a regional carbon emission prediction device based on periodic adaptive method provided in an embodiment of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0025] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0026] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0027] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0028] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0029] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.

[0030] See Figure 1 To address the problem that existing technologies rely on manually preset static period parameters or fixed empirical models to set the optimal period, leading to inaccurate carbon emission prediction results, an embodiment of the present invention provides a period-adaptive regional carbon emission prediction method, comprising: Step S101: Obtain the historical time-series data of regional carbon emissions in the area to be predicted, as well as the historical electricity consumption characteristics and historical meteorological characteristics of the area corresponding to the historical time-series data of regional carbon emissions. Specifically, raw historical carbon emission data can be obtained through a regional carbon emission monitoring platform and timestamped according to a fixed time frequency (such as monthly) to form initial regional carbon emission historical time series data.

[0031] As an illustration, the initial regional carbon emission historical time series data can be defined as follows: In the formula, This represents the initial regional carbon emission historical time series data. This represents historical carbon emission data corresponding to different time points in the sequence. This represents the total number of observed samples in the sequence.

[0032] Specifically, the historical electricity consumption characteristics and historical meteorological characteristics of the aforementioned regions are all data that are strictly aligned with the historical time series data of regional carbon emissions.

[0033] Specifically, the historical electricity consumption characteristics of the aforementioned regions include: regional industrial electricity consumption; the historical meteorological characteristics of the aforementioned regions include: monthly maximum temperature, average humidity, etc.

[0034] Preferably, after obtaining the initial regional historical time-series carbon emission data, regional historical electricity consumption characteristics, and regional historical meteorological characteristics, to prevent interference from acquisition failures or extreme noise on subsequent periodic extraction and trend modeling, the Raida criterion is adopted. The criteria are used to detect abnormal fluctuations in these data, remove abnormal jump values, fill in missing values, and achieve standardized data processing.

[0035] Specifically, taking the initial regional carbon emission historical time series data as an example, the sample mean and sample standard deviation of the initial regional carbon emission historical time series data are calculated, and an outlier threshold is constructed based on the sample standard deviation.

[0036] Then, each data point in the sequence is traversed to determine whether the deviation between the data point and the sample mean exceeds the above outlier threshold. If it does, the data point is determined to be an outlier value and removed from the sequence. Then, this time point is marked as a vacancy.

[0037] The sample mean and sample standard deviation are calculated using the following formulas: In the formula, Represents the sample mean. This represents the i-th data point in the sequence. This represents the sample standard deviation.

[0038] Preferably, the outlier threshold is 3 times the sample standard deviation, and the judgment condition for outlier jump values ​​can be expressed by the following formula: Specifically, after removing anomalous jump values, linear interpolation is used to reconstruct the numerical data to address the gaps caused by the removal of data points and the missing points inherent in the original data collection, in order to ensure the mathematical continuity of the time series and ultimately form the historical time series data of carbon emissions in the aforementioned region.

[0039] Indicative, assuming a time point There is a missing value. By optimizing forward and backward, we can identify the two nearest valid observations located at the time nodes. and (satisfy The corresponding effective historical carbon emissions are respectively and .

[0040] Then time node The formula for calculating the interpolation repair value at the location is: In the formula, Indicates time node Interpolation repair value at the location.

[0041] It should be noted that the data corresponding to the region's historical electricity consumption characteristics and regional historical meteorological characteristics are processed using the same standardized process as the initial regional carbon emission historical time series data. After standardization, the data are then used in the subsequent carbon emission prediction process.

[0042] Step S102: Based on the historical time series data of regional carbon emissions and several preset periods, calculate the autocorrelation function, power spectral density and the Akaike information criterion value in each preset period. Specifically, for historical time-series data of regional carbon emissions, periodic features are extracted from three independent dimensions: time domain, frequency domain, and information theory, to obtain the autocorrelation function, power spectral density, and Akaike information criterion value under each preset period.

[0043] In a preferred embodiment, the step of calculating the autocorrelation function, power spectral density, and Akaike information criterion value for each preset period based on regional carbon emission historical time-series data and several preset periods includes: The series mean was calculated based on the historical time-series data of carbon emissions in the region. The autocorrelation function is calculated based on the historical time-series data of carbon emissions in the region and the mean of the series. Specifically, the aforementioned autocorrelation function is a function with lag order as the independent variable and autocorrelation coefficient as the dependent variable. The autocorrelation coefficient is calculated as follows: In the formula, This represents the autocorrelation coefficient corresponding to the lag order k. This represents the historical carbon emission data at time t in the historical time series data of regional carbon emissions. Represents the mean of the sequence. This represents the historical carbon emission data at time t+k in the historical time series data of regional carbon emissions.

[0044] An illustrative diagram of the autocorrelation function, based on historical carbon emission data of a certain province, is shown below. Figure 2 As shown, Figure 2 The horizontal axis represents the lag period, i.e., the lag order, and the vertical axis, "correlation coefficient," represents the autocorrelation coefficient corresponding to each lag order.

[0045] The historical time-series data of carbon emissions in the region are subjected to discrete Fourier transform to obtain the Fourier transform results; The power spectral density is calculated based on the Fourier transform results. Specifically, in the frequency domain, the discrete Fourier transform is used to convert the historical time-series data of regional carbon emissions in the time domain to the frequency domain in order to capture the high-energy oscillation frequencies hidden in complex time-series fluctuations and obtain the power spectral density at different frequencies.

[0046] The power spectral density is calculated using the following formula: In the formula, This represents the power spectral density corresponding to frequency f. Indicates frequency.

[0047] An illustrative diagram of the power spectral density, drawn based on historical carbon emission data of a certain province, is shown below. Figure 3 As shown, Figure 3 The horizontal axis represents frequency, and the vertical axis, "density," represents power spectral density.

[0048] Based on the historical time series data of carbon emissions in the region, a basic exponential smoothing model corresponding to each preset period is fitted. Specifically, in terms of information theory and model structure goodness, a basic exponential smoothing model is instantiated for each preset period, and its parameters are fitted using historical time-series data of regional carbon emissions to obtain the basic exponential smoothing model corresponding to each preset period. The expression of the above basic exponential smoothing model is as follows: In the formula, This represents the actual observed carbon emissions at time t; The smoothing level term at time t is represented. The trend term at time t This represents the seasonal term at time t. , and All are smoothing coefficients, where m represents the preset period. This is a single-step prediction value generated at time t based on the state at the previous time step.

[0049] Based on the smoothing models of each basic index and the historical time series data of carbon emissions in the region, the maximum likelihood estimate corresponding to each preset period is calculated. Specifically, for each basic index smoothing model, the prediction error is calculated after fitting and predicting each historical carbon emission data point in the regional carbon emission historical time series data, resulting in the likelihood function corresponding to each preset period. Taking one preset period as an example, its corresponding likelihood function is expressed as follows: In the formula, Represents the likelihood function. This represents the model parameters of the basic exponential smoothing model. The variance represents the prediction error. This represents the prediction error at time t.

[0050] Subsequently, taking the logarithm of the above likelihood function yields the log-likelihood function: In the formula, This represents the log-likelihood function.

[0051] For any one Based on the sum of squared errors, the objective variance that maximizes the log-likelihood function is calculated. This process can be expressed as: In the formula, Indicates the target variance. express The sum of squared errors.

[0052] Then, substituting the target variance into the log-likelihood function, we obtain the log-likelihood of the set as follows: In the formula, express The corresponding lumped log-likelihood.

[0053] Based on the above formula, by minimizing The optimal model parameters and the optimal sum of squared errors corresponding to these optimal model parameters can then be obtained.

[0054] Finally, by substituting the optimal sum of squared errors into the formula for the log-likelihood function, the maximum likelihood estimate for the preset period can be obtained.

[0055] Based on the maximum likelihood estimation, the Akaike information criterion value for each preset period is calculated.

[0056] Specifically, the Akaike Information Criterion value is calculated using the following formula: In the formula, This represents the Akaike information criterion value corresponding to the preset period m. This represents the maximum likelihood estimate for a preset period m. This represents the total number of independent parameters to be estimated in the basic exponential smoothing model corresponding to the preset period m.

[0057] An illustrative diagram illustrating the Akaike Information Criterion Curve under different preset periods, based on historical carbon emission data of a certain province. Figure 4 As shown, Figure 4 The horizontal axis, "Test Period Length," represents each preset period, while the vertical axis, "Akaike Information Criterion," represents the corresponding Akaike Information Criterion value.

[0058] In this preferred embodiment, the autocorrelation function, power spectral density, and Akaike information criterion value for each preset period are calculated based on the historical time-series data of regional carbon emissions and several preset periods.

[0059] Step S103: Determine the time-domain candidate period, frequency-domain candidate period, and information criterion candidate period based on the autocorrelation function, power spectral density, and Akaike information criterion value; In a preferred embodiment, determining the time-domain candidate period, frequency-domain candidate period, and information criterion candidate period based on the autocorrelation function, power spectral density, and Akaike information criterion value includes: The first lag order in the autocorrelation function that exceeds the upper limit of the preset confidence interval and is the peak value of the autocorrelation coefficient is taken as the time-domain candidate period; Specifically, based on the preset maximum lag order (like (months), traversal calculation All within the range value.

[0060] Subsequently, the envoy was identified. The first major lag order that breaks through the upper limit of the confidence interval and exhibits the most significant local peak is used as the periodic feature indicator in the time domain dimension, and is output and recorded as the time domain candidate period.

[0061] The frequency corresponding to the maximum value of the power spectral density is taken as the dominant frequency, and the reciprocal of the dominant frequency is rounded to obtain the frequency domain candidate period. Specifically, the dominant frequency that maximizes the power spectral density within the effective frequency range is searched. Since frequency and period are reciprocals, the reciprocal of this dominant frequency is taken and rounded to the nearest integer to obtain the periodic characteristic indicator in the frequency domain. This indicator is then output and recorded as the candidate period in the frequency domain. In the formula, Indicates the frequency domain candidate period, Indicates the dominant frequency. This indicates the rounding operation.

[0062] The preset period corresponding to the smallest Akaike information criterion value is used as the alternative period for the information criterion.

[0063] Specifically, based on the Akaike information criterion values ​​for all preset periods, a [data / method] is generated. Curve, extraction The preset period during which the curve reaches its global trough is used as a feature indicator in the information theory dimension, and is output and recorded as the alternative period for the information criterion: In the formula, Indicates the information criterion selection period. This represents the Akaike information criterion value corresponding to the preset period m.

[0064] In this preferred embodiment, the time-domain candidate period, frequency-domain candidate period, and information criterion candidate period are determined based on the autocorrelation function, power spectral density, and Akaike information criterion value, respectively.

[0065] Step S104: If the time-domain candidate period, frequency-domain candidate period, and information criterion candidate period are different from each other, then the information criterion candidate period is taken as the globally optimal period value; otherwise, the values ​​of the same candidate periods are taken as the globally optimal period value. Specifically, this invention employs an arbitration logic mechanism based on majority voting and criterion priority to arbitrate and match the three alternative cycles obtained in the aforementioned steps, and confirm and output the globally optimal cycle value.

[0066] Specifically, firstly, a candidate period set and feature majority orientation determination are constructed. The three candidate periods obtained in the previous steps are merged to construct a candidate period feature set: In the formula, Represents the set of candidate periodic features. Indicates the alternative time-domain period.

[0067] Then, the majority voting principle is used to scan the set and determine whether there are elements with the same value in the set. That is, if the frequency of a certain candidate period in the set is ≥2, then the value of this candidate period is directly confirmed as the global optimal seasonal period (i.e., the global optimal period mentioned above).

[0068] Indicative, Figure 2 , Figure 3 and Figure 4 For example, the alternative periods corresponding to the three attached figures all point to 12, so 12 is taken as the globally optimal period in this invention.

[0069] Preferably, this rule ensures that, in regular and well-defined carbon emission data, the model can adaptively lock onto the most confident physical cycle.

[0070] If there are no elements with equal values ​​in the set, that is, the values ​​of the three candidate periods are not equal, the arbitration downgrade mechanism with absolute priority of information criteria is triggered. At this time, it is determined that the carbon emission data of the current region to be predicted is seriously affected by abnormal policy intervention or extreme weather noise interference, resulting in logical conflicts in the local fluctuation characteristics of the pure time domain or pure frequency domain.

[0071] At this point, the majority vote is automatically abandoned, triggering the arbitration downgrade mechanism. The absolute priority principle of information criteria is implemented, and the alternative period of information criteria is unconditionally adopted as the judgment result, that is, the globally optimal period value is forcibly assigned to the alternative period of information criteria.

[0072] Specifically, the globally optimal periodic value determined in this step will be used as a dynamic structure hyperparameter, directly injected into and determining the step size and recursive span of the seasonal state update equation of the carbon emission index smoothing model in subsequent steps.

[0073] Preferably, the algorithm's principle lies in the fact that, compared to autocorrelation function and power spectral density, which only focus on local numerical features of the data, the AIC information criterion starts from the global model structure goodness and performs a mathematically optimal penalty and trade-off between fitting residuals and overfitting risk. Therefore, when feature conflicts occur, using the information criterion as an alternative period can serve as an absolute safety boundary to ensure the generalization ability of the subsequent prediction process.

[0074] Preferably, existing technologies, when dealing with seasonal fluctuations, often set static periods (such as a fixed one-year period) or rely heavily on manual parameter fine-tuning. When the carbon emission cycle patterns drift due to different industrial structures, their model applicability is poor. This invention, through the quantitative extraction and intelligent arbitration of multi-dimensional mathematical indicators, achieves data-driven and adaptive generation of cycle hyperparameters. When the carbon emission patterns of the target region change, the solution of this invention can accurately match the optimal model structure, significantly reducing the cost of manual parameter tuning and structural bias in engineering deployment.

[0075] Step S105: Based on the historical time series data of regional carbon emissions, historical electricity consumption characteristics of the region, historical meteorological characteristics of the region, and the global optimal periodic value, predict the regional predicted carbon emissions of the region to be predicted under the preset future time period.

[0076] Specifically, based on historical time-series data of regional carbon emissions and the globally optimal periodic value, a carbon emission index smoothing model with a damped term is constructed and the residuals are separated. Subsequently, combined with historical electricity consumption characteristics and historical meteorological characteristics of the region, a residual regression model is constructed. Finally, the carbon emission index smoothing model and the residual regression model are used to predict the carbon emissions of the region to be predicted.

[0077] In a preferred embodiment, predicting the predicted regional carbon emissions of the region to be predicted for a preset future period based on the region's historical carbon emission time-series data, historical electricity consumption characteristics, historical meteorological characteristics, and the global optimal periodic value includes: Obtain the regional predicted electricity consumption characteristics and regional predicted meteorological characteristics of the area to be predicted under the preset future time period; Specifically, based on meteorological forecast data provided by external meteorological departments and electricity consumption plan data for the forecast period provided by the power grid dispatching system, the predicted electricity consumption characteristics and predicted meteorological characteristics of the above-mentioned regions are obtained.

[0078] Based on the historical time-series data of carbon emissions in the region, the global optimal periodic value, the historical electricity consumption characteristics of the region, and the historical meteorological characteristics of the region, a carbon emission index smoothing model and a residual regression model are constructed. Based on the historical time-series data of carbon emissions in the region, the predicted electricity consumption characteristics of the region, the predicted meteorological characteristics of the region, the carbon emission index smoothing model, and the residual regression model, the predicted carbon emissions of the region are obtained.

[0079] In this preferred embodiment, the predicted regional carbon emissions of the region under a preset future time period are predicted based on the region's historical carbon emission time series data, the region's historical electricity consumption characteristics, the region's historical meteorological characteristics, and the global optimal period value.

[0080] In another preferred embodiment, the step of constructing a carbon emission index smoothing model and a residual regression model based on the historical time-series data of regional carbon emissions, the global optimal periodic value, the historical electricity consumption characteristics of the region, and the historical meteorological characteristics of the region includes: Based on the historical time-series data of carbon emissions in the region and the global optimal periodic value, a carbon emission index smoothing model is fitted. The specific carbon emission index smoothing model is an exponential smoothing model with a nonlinear damping mechanism. This model is designed to address the nonlinear characteristics of regional carbon emission evolution. It breaks through the limitations of traditional unconstrained models where the trend term is linear or exponentially divergent. It is innovatively constructed by explicitly introducing a nonlinear damping coefficient in the iteration of the trend component.

[0081] Specifically, the carbon emission index smoothing model is based on the real carbon emission data at time t, and combines... The state at time t is recursively updated by considering the previous state and the previous seasonal state with the global optimal periodicity. The core state update equations are constructed as follows: First is the horizontal update equation, whose expression is: In the formula, express The baseline level after smoothing is the smoothing level term in the carbon emission index smoothing model. This represents the smoothing coefficient corresponding to the horizontal update equation, with a value between 0 and 1. This represents the seasonal fluctuation amplitude at time tm after smoothing, and is the seasonal term in the carbon emission index smoothing model. express The baseline level after smoothing. Indicates the damping coefficient. express The growth or decline slope after smoothing at time -1 is the trend term in the carbon emission index smoothing model.

[0082] Secondly, there is the trend update equation, whose expression is: In the formula, express The slope of growth or decline after smoothing at any given moment. This represents the smoothing coefficient corresponding to the trend update equation, and its value is between 0 and 1.

[0083] Finally, the seasonal factor update equation is expressed as follows: In the formula, This represents the seasonal fluctuation amplitude at time t after smoothing. This represents the smoothing coefficient corresponding to the seasonal factor update equation, with a value between 0 and 1.

[0084] Specifically, in all the above-mentioned items that include the trend from the previous moment... In the expressions, the linear momentum is proportionally attenuated by multiplying by the damping coefficient.

[0085] Specifically, the aforementioned carbon emission index smoothing model uses historical time-series data of regional carbon emissions and employs nonlinear numerical optimization algorithms such as L-BFGS (memory-constrained quasi-Newton method) to globally optimize the three smoothing coefficients and damping coefficients in the model.

[0086] Specifically, the objective loss function is set as the sum of squared residuals from single-step predictions over the historical fitting interval. Through iterative approximation, the optimal set of smoothing coefficients and the optimal damping coefficient that minimize the objective loss function are calculated, thus obtaining the carbon emission index smoothing model. In this process, due to... The trend slope of the carbon emission sequence decreases marginally in each time iteration. This invention also sets strict mathematical feasible region constraints on the damping coefficient, namely: This constraint mechanism, at a mathematical level, fits the nonlinear asymptotic convergence characteristics of regional carbon emissions.

[0087] Based on the carbon emission index smoothing model, each historical carbon emission data in the historical time series data of the region is fitted and predicted to obtain the fitted value of each historical carbon emission data. Specifically, after completing the optimal parameter fitting, the basic fitted value sequence is output based on the carbon emission index smoothing model within the time interval corresponding to the historical time series data of carbon emissions in the entire region. Each data point in this basic fitted value sequence is the above-mentioned fitted value obtained based on the carbon emission index smoothing model.

[0088] Based on the historical carbon emission data and the corresponding fitted values, the fitting residuals of each historical carbon emission data are calculated. Specifically, the historical time-series data of regional carbon emissions is subtracted from the baseline fitted value sequence at each corresponding time point to obtain the fitted residual sequence. Each data point in the fitted residual sequence is the fitted residual of the corresponding historical carbon emission data, which can be calculated using the following formula: In the formula, This represents the fitting residual at time t. This represents the baseline fitted value at time t.

[0089] Preferably, the above-mentioned fitted residual sequence filters out the inherent time autocorrelation, periodicity and convergence trend of the regional carbon emission historical time series data, and effectively separates the unresolved biases caused by external meteorological and electricity consumption characteristic fluctuations.

[0090] Based on the historical electricity consumption characteristics and historical meteorological characteristics of the region, a historical covariate feature matrix is ​​constructed; Specifically, the aforementioned historical covariate feature matrix is ​​represented as follows: In the formula, Represents the historical covariate feature matrix. N represents the total sequence length of the historical time sample, and D represents the total number of types of multidimensional covariates (e.g., including multidimensional features such as total electricity consumption, temperature, wind speed, and precipitation). Therefore, the element in the i-th row and j-th column of the matrix... This represents the specific quantitative value of the j-th covariate at the i-th historical observation time.

[0091] Using the historical covariate feature matrix as input and the fitting residual as output, a residual regression model is constructed to characterize the mapping relationship between the covariate feature matrix and the fitting residual.

[0092] Specifically, a machine learning-based residual regression model is constructed using the historical covariate feature matrix as the input feature and the fitted residual as the target variable. By inputting the historical covariate feature matrix and the target variable into the model for supervised learning training, a nonlinear mapping relationship is established. The model construction process is existing technology and will not be elaborated further here. This mapping relationship can be expressed as: In the formula, Indicates the mapping relationship. Represents the historical covariate feature matrix. This represents the fitting residual.

[0093] Preferably, this residual regression model is used to quantify the actual residual impact of fluctuations in external physical characteristics such as regional industrial electricity consumption and meteorological features on carbon emissions.

[0094] Preferably, existing technologies, when fitting evolutionary trends, mainly rely on unconstrained time-dependent linear or pure autoregressive extrapolation, leading to long-term predictions that easily deviate from real-world patterns and exhibit unidirectional divergence. This invention innovatively introduces a damped convergence mechanism when constructing a carbon emission index smoothing model, forcing the trend slope to marginally decrease over time. This mathematical treatment objectively reflects the nonlinear asymptotic convergence characteristic of emission reduction potential approaching its limit under macro-level "dual carbon" targets, ensuring that the output long-term benchmark curve is strictly controlled within physical soft boundaries, thus improving the reliability of medium- and long-term macro-level decision-making.

[0095] In this preferred embodiment, a carbon emission index smoothing model and a residual regression model are constructed based on regional historical carbon emission time series data, global optimal periodic values, regional historical electricity consumption characteristics, and regional historical meteorological characteristics.

[0096] In another preferred embodiment, the step of predicting the predicted carbon emissions of the region based on the historical time-series data of regional carbon emissions, predicted regional electricity consumption characteristics, predicted regional meteorological characteristics, a carbon emission index smoothing model, and a residual regression model includes: Based on the predicted electricity consumption characteristics and predicted meteorological characteristics of the region, a prediction covariate feature matrix is ​​constructed; Specifically, the predicted covariate feature matrix is ​​constructed using the same method as the historical covariate feature matrix.

[0097] Based on the historical time-series data of carbon emissions in the region and the carbon emission index smoothing model, the baseline predicted value of carbon emissions is calculated. Specifically, for the next h steps (i.e., the prediction period is...) The regional carbon emissions at a given time (i.e., at a specific point in time) are used to generate a baseline prediction value using a fitted carbon emission index smoothing model. The mathematical analytical expression for this forward multi-step prediction is as follows: In the formula, The number of prediction steps for outward extrapolation ( ), express The baseline carbon emission forecast for that time period. This represents the globally optimal period value. For the periodic rotation control variable, the value satisfies The smallest non-negative integer is used to ensure that the model can iteratively extract seasonal factors corresponding to the same historical period. , This represents the trend term at time N. This represents the smoothing level term corresponding to time N.

[0098] Preferably, since the damping coefficient is constrained to be When predicting the number of steps As the value approaches infinity, the cumulative trend terms, including the damping coefficient, form a geometric series that, when summed, strictly converges to the finite constant limit. This mechanism establishes the nonlinear asymptotic convergence characteristics of the long-term evolution of carbon emissions in a physical sense, effectively avoiding unbounded overshoot errors caused by unconstrained prediction.

[0099] Based on the predicted covariate feature matrix and the residual regression model, the predicted residual values ​​are calculated. Specifically, the feature matrix of the predictive covariates for the corresponding prediction period is input into the trained residual regression model, and the corresponding residual prediction value is calculated and output.

[0100] The sum of the baseline carbon emission forecast and the residual forecast is taken as the predicted carbon emission for the region.

[0101] Specifically, the baseline carbon emission forecast based on time-series evolution patterns is linearly superimposed with the residual forecast based on objective physical factors to obtain the final regional predicted carbon emissions constrained by natural laws: In the formula, express Predicted carbon emissions for the region at the given time. express The predicted residual value at time point.

[0102] Schematic diagrams illustrating carbon emission prediction results from different models are shown below. Figure 5 As shown, Figure 5 The horizontal axis represents the date, and the vertical axis represents the actual total carbon emissions of a province. "MAPE" indicates the average absolute percentage error. Figure 5 This paper illustrates the comparison between the actual carbon emission predictions and the predicted values ​​for the same raw data—the total carbon emissions of the province—based on three methods: exponential smoothing, damped exponential smoothing, and the "damped exponential smoothing + residual regression" mechanism used in this invention. Figure 5 As can be seen, the scheme of this invention is significantly superior to the other two carbon emission prediction schemes, and the long-term carbon emission prediction results of this scheme have a smaller error compared with the actual values ​​over time compared with the other two schemes.

[0103] Preferably, the present invention visualizes the predicted regional carbon emissions, which integrate short-term cyclical oscillations, long-term convergence boundary constraints, and external objective law constraints, in the form of a time series. After receiving the predicted data, the regional integrated energy management system or the local government's "dual carbon" platform can use it to generate medium- and long-term macro-emission reduction strategies.

[0104] Preferably, existing forecasting schemes are mostly limited to univariate time-series extrapolation, failing to incorporate core variables directly driving carbon emissions, such as electricity consumption and meteorology, into the modeling. This invention overcomes this limitation by constructing a residual regression model based on regional industrial electricity consumption characteristics and meteorological characteristics. When faced with extreme weather (such as a surge in cooling load due to high temperatures) or production schedule adjustments, this model can respond to fluctuations in external covariates and dynamically correct baseline forecast biases.

[0105] In this preferred embodiment, the predicted regional carbon emissions are obtained based on historical time-series data of regional carbon emissions, predicted regional electricity consumption characteristics, predicted regional meteorological characteristics, carbon emission index smoothing model, and residual regression model.

[0106] Based on the above method embodiments, the present invention provides corresponding apparatus embodiments.

[0107] like Figure 6 As shown, an embodiment of the present invention provides a regional carbon emission prediction device based on periodic adaptive emission, comprising: The system includes a data acquisition module, a data calculation module, a cycle calculation module, an optimal cycle determination module, and a carbon emission prediction module. The data acquisition module is used to acquire historical time-series data of regional carbon emissions in the area to be predicted, as well as the historical electricity consumption characteristics and historical meteorological characteristics of the area corresponding to the historical time-series data of regional carbon emissions. The data calculation module is used to calculate the autocorrelation function, power spectral density, and Akaike information criterion value for each preset period based on historical time-series data of regional carbon emissions and several preset periods. The period calculation module is used to determine the time-domain candidate period, frequency-domain candidate period, and information criterion candidate period based on the autocorrelation function, power spectral density, and Akaike information criterion value. The optimal period determination module is used to determine the global optimal period value when the time-domain candidate period, frequency-domain candidate period, and information criterion candidate period are different; otherwise, the value of the same candidate period is taken as the global optimal period value. The carbon emission prediction module is used to predict the regional predicted carbon emissions of the area to be predicted in a preset future period based on the historical time series data of regional carbon emissions, historical electricity consumption characteristics of the area, historical meteorological characteristics of the area, and the global optimal period value.

[0108] In a preferred embodiment, the data calculation module includes: The system includes a sequence mean calculation unit, an autocorrelation function calculation unit, a Fourier transform unit, a power spectral density calculation unit, a basic exponential smoothing model construction unit, a maximum likelihood estimation calculation unit, and an Akaike information criterion value calculation unit. The sequence mean calculation unit is used to calculate the sequence mean based on the historical time-series data of carbon emissions in the region. The autocorrelation function calculation unit is used to calculate the autocorrelation function based on the historical time series data of carbon emissions in the region and the mean of the series. The Fourier transform unit is used to perform discrete Fourier transform on the historical time-series data of carbon emissions in the region to obtain the Fourier transform result. The power spectral density calculation unit is used to calculate the power spectral density based on the Fourier transform result; The basic exponential smoothing model construction unit is used to fit the basic exponential smoothing model corresponding to each preset period based on the historical time series data of carbon emissions in the region. The maximum likelihood estimation calculation unit is used to calculate the maximum likelihood estimate corresponding to each preset period based on the smoothing models of each basic index and the historical time series data of carbon emissions in the region. The Akaike Information Criterion Value Calculation Unit is used to calculate the Akaike Information Criterion Value for each preset period based on the maximum likelihood estimation.

[0109] It should be noted that the device embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without creative effort. The above schematic diagrams are merely examples of a periodically adaptive regional carbon emission prediction device and do not constitute a limitation on a periodically adaptive regional carbon emission prediction device. It may include more or fewer components than illustrated, or combine certain components, or use different components.

[0110] Based on the above method embodiments, the present invention provides corresponding terminal device embodiments.

[0111] Another embodiment of the present invention provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the periodic adaptive regional carbon emission prediction method described in any embodiment of the present invention.

[0112] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the device. The aforementioned terminal devices may be computing devices such as desktop computers, laptops, handheld computers, and cloud servers. These devices may include, but are not limited to, processors and memory. The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. This processor is the control center of the device, connecting various parts of the device via various interfaces and lines. The aforementioned memory can be used to store the aforementioned computer programs and / or modules. The aforementioned processor implements various functions of the aforementioned device by running or executing the computer programs and / or modules stored in the aforementioned memory, and by calling data stored in the memory. The aforementioned memory may mainly include a program storage area and a data storage area, wherein the program storage area may store the operating system, at least one application program required for a function, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0113] Based on the above method embodiments, the present invention provides corresponding storage medium embodiments.

[0114] Another embodiment of the present invention provides a storage medium including a stored computer program, wherein, when the computer program is running, it controls the device where the storage medium is located to execute the regional carbon emission prediction method based on periodic adaptive method described in any embodiment of the present invention.

[0115] In this embodiment, the storage medium is a computer-readable storage medium, and the computer program includes computer program code, which may be in the form of source code, object code, executable file, or some intermediate form. The computer-readable medium may include any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0116] The above are preferred embodiments of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A regional carbon emission prediction method based on periodic adaptive methods, characterized in that, include: Obtain historical time-series data of regional carbon emissions in the area to be predicted, as well as the corresponding historical electricity consumption characteristics and historical meteorological characteristics of the area. Based on historical time-series data of regional carbon emissions and several preset periods, the autocorrelation function, power spectral density, and Akaike information criterion value for each preset period were calculated. Based on the autocorrelation function, power spectral density, and Akaike information criterion value, the time-domain candidate period, frequency-domain candidate period, and information criterion candidate period are determined. If the time-domain candidate period, frequency-domain candidate period, and information criterion candidate period are all different, then the information criterion candidate period is taken as the globally optimal period value; otherwise, the values ​​of the same candidate periods are taken as the globally optimal period value. Based on the historical time-series data of regional carbon emissions, historical electricity consumption characteristics of the region, historical meteorological characteristics of the region, and the global optimal period value, the predicted regional carbon emissions of the region to be predicted under the preset future time period are predicted.

2. The regional carbon emission prediction method based on periodic adaptive method according to claim 1, characterized in that, The calculation of the autocorrelation function, power spectral density, and Akaike information criterion value for each preset period, based on historical time-series data of regional carbon emissions and several preset periods, includes: The series mean was calculated based on the historical time-series data of carbon emissions in the region. The autocorrelation function is calculated based on the historical time-series data of carbon emissions in the region and the mean of the series. The historical time-series data of carbon emissions in the region are subjected to discrete Fourier transform to obtain the Fourier transform results; The power spectral density is calculated based on the Fourier transform results. Based on the historical time series data of carbon emissions in the region, a basic exponential smoothing model corresponding to each preset period is fitted. Based on the smoothing models of each basic index and the historical time series data of carbon emissions in the region, the maximum likelihood estimate corresponding to each preset period is calculated. Based on the maximum likelihood estimation, the Akaike information criterion value for each preset period is calculated.

3. The regional carbon emission prediction method based on periodic adaptive method according to claim 2, characterized in that, The step of determining the time-domain candidate period, frequency-domain candidate period, and information criterion candidate period based on the autocorrelation function, power spectral density, and Akaike information criterion value includes: The first lag order in the autocorrelation function that exceeds the upper limit of the preset confidence interval and is the peak value of the autocorrelation coefficient is taken as the time-domain candidate period; The frequency corresponding to the maximum value of the power spectral density is taken as the dominant frequency, and the reciprocal of the dominant frequency is rounded to obtain the frequency domain candidate period. The preset period corresponding to the smallest Akaike information criterion value is used as the alternative period for the information criterion.

4. The regional carbon emission prediction method based on periodic adaptive method according to claim 3, characterized in that, The step of predicting the predicted regional carbon emissions of the region under a preset future period based on the region's historical carbon emission time-series data, historical electricity consumption characteristics, historical meteorological characteristics, and the global optimal period value includes: Obtain the regional predicted electricity consumption characteristics and regional predicted meteorological characteristics of the area to be predicted under the preset future time period; Based on the historical time-series data of carbon emissions in the region, the global optimal periodic value, the historical electricity consumption characteristics of the region, and the historical meteorological characteristics of the region, a carbon emission index smoothing model and a residual regression model are constructed. Based on the historical time-series data of carbon emissions in the region, the predicted electricity consumption characteristics of the region, the predicted meteorological characteristics of the region, the carbon emission index smoothing model, and the residual regression model, the predicted carbon emissions of the region are obtained.

5. The regional carbon emission prediction method based on periodic adaptive method according to claim 4, characterized in that, The carbon emission index smoothing model and residual regression model are constructed based on the historical time-series data of regional carbon emissions, the global optimal periodic value, the historical electricity consumption characteristics of the region, and the historical meteorological characteristics of the region, including: Based on the historical time-series data of carbon emissions in the region and the global optimal periodic value, a carbon emission index smoothing model is fitted. Based on the carbon emission index smoothing model, each historical carbon emission data in the historical time series data of the region is fitted and predicted to obtain the fitted value of each historical carbon emission data. Based on the historical carbon emission data and the corresponding fitted values, the fitting residuals of each historical carbon emission data are calculated. Based on the historical electricity consumption characteristics and historical meteorological characteristics of the region, a historical covariate feature matrix is ​​constructed; Using the historical covariate feature matrix as input and the fitting residual as output, a residual regression model is constructed to characterize the mapping relationship between the covariate feature matrix and the fitting residual.

6. The regional carbon emission prediction method based on periodic adaptive method according to claim 5, characterized in that, The process of predicting the predicted carbon emissions for the region based on historical time-series data of regional carbon emissions, predicted regional electricity consumption characteristics, predicted regional meteorological characteristics, a carbon emission index smoothing model, and a residual regression model includes: Based on the predicted electricity consumption characteristics and predicted meteorological characteristics of the region, a prediction covariate feature matrix is ​​constructed; Based on the historical time-series data of carbon emissions in the region and the carbon emission index smoothing model, the baseline predicted value of carbon emissions is calculated. Based on the predicted covariate feature matrix and the residual regression model, the predicted residual values ​​are calculated. The sum of the baseline carbon emission forecast and the residual forecast is taken as the predicted carbon emission for the region.

7. A regional carbon emission prediction device based on periodic adaptive emission forecasting, characterized in that, include: The system includes a data acquisition module, a data calculation module, a cycle calculation module, an optimal cycle determination module, and a carbon emission prediction module. The data acquisition module is used to acquire historical time-series data of regional carbon emissions in the area to be predicted, as well as the historical electricity consumption characteristics and historical meteorological characteristics of the area corresponding to the historical time-series data of regional carbon emissions. The data calculation module is used to calculate the autocorrelation function, power spectral density, and Akaike information criterion value for each preset period based on historical time-series data of regional carbon emissions and several preset periods. The period calculation module is used to determine the time-domain candidate period, frequency-domain candidate period, and information criterion candidate period based on the autocorrelation function, power spectral density, and Akaike information criterion value. The optimal period determination module is used to determine the global optimal period value when the time-domain candidate period, frequency-domain candidate period, and information criterion candidate period are different; otherwise, the value of the same candidate period is taken as the global optimal period value. The carbon emission prediction module is used to predict the regional predicted carbon emissions of the area to be predicted in a preset future period based on the historical time series data of regional carbon emissions, historical electricity consumption characteristics of the area, historical meteorological characteristics of the area, and the global optimal period value.

8. A regional carbon emission prediction device based on periodic adaptive emission forecasting according to claim 7, characterized in that, The data calculation module includes: The system includes a sequence mean calculation unit, an autocorrelation function calculation unit, a Fourier transform unit, a power spectral density calculation unit, a basic exponential smoothing model construction unit, a maximum likelihood estimation calculation unit, and an Akaike information criterion value calculation unit. The sequence mean calculation unit is used to calculate the sequence mean based on the historical time-series data of carbon emissions in the region. The autocorrelation function calculation unit is used to calculate the autocorrelation function based on the historical time series data of carbon emissions in the region and the mean of the series. The Fourier transform unit is used to perform discrete Fourier transform on the historical time-series data of carbon emissions in the region to obtain the Fourier transform result. The power spectral density calculation unit is used to calculate the power spectral density based on the Fourier transform result; The basic exponential smoothing model construction unit is used to fit the basic exponential smoothing model corresponding to each preset period based on the historical time series data of carbon emissions in the region. The maximum likelihood estimation calculation unit is used to calculate the maximum likelihood estimate corresponding to each preset period based on the smoothing models of each basic index and the historical time series data of carbon emissions in the region. The Akaike Information Criterion Value Calculation Unit is used to calculate the Akaike Information Criterion Value for each preset period based on the maximum likelihood estimation.

9. A terminal device, characterized in that, The method includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements a periodically adaptive regional carbon emission prediction method as described in any one of claims 1 to 6.

10. A storage medium, characterized in that, The storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device where the storage medium is located to perform a periodically adaptive regional carbon emission prediction method as described in any one of claims 1 to 6.

Images