A method for accounting for carbon emissions in a megacity based on power statistical data

Abstract

The application discloses a kind of based on electric power statistical data's super city carbon emission accounting method, steps are as follows: according to time scale, the second power load data of the second block corresponding to each power consumption department of the first block is obtained;According to the second power load data, the first power consumption statistical data of each power consumption department of the first block is calculated;According to time scale, the carbon emission data corresponding to the first block is obtained;Using coefficient smooth variable model, the carbon emission data corresponding to the first block and corresponding first power consumption statistical data are fitted, and the variable coefficient corresponding to the first block on time scale is obtained;Using autoregressive moving average model, the variable coefficient on time scale is predicted;Variable coefficient is imported into coefficient smooth variable model, and the carbon emission data corresponding to the second block on time scale is calculated.The scheme can obtain more timely, accurate and representative carbon emission data for future years.

Description

Technical Field

[0001] This invention relates to the field of carbon emission accounting, and more specifically, to a method for calculating carbon emissions in megacities based on electricity statistics. Background Technology

[0002] Cities, as major sources of greenhouse gas emissions, will ultimately directly impact the development of a region's low-carbon economy. However, due to the difficulty in obtaining energy statistics for megacities and smaller administrative districts, it is impossible to distinguish the emission reduction efforts of megacities from those of other provinces.

[0003] Due to the limited sources of carbon emission data in my country, it is also difficult to compare the differences in carbon emissions across different regions. The urban carbon emission accounting method based on DMSP-OLS nighttime light data has made up for the shortcomings of incomplete statistical data and inconsistent statistical standards, but it still has problems such as insufficient ability to express the spatial and scale of carbon emissions at the urban scale.

[0004] Chinese Patent, Publication No.: CN 114169714 A, Publication Date: March 11, 2022, this application relates to the field of environmental protection management technology, and discloses a regional carbon emission accounting method, including: constructing a regional carbon emission accounting model; collecting basic data related to regional carbon emissions; cleaning the collected regional carbon emission data; and substituting the cleaned regional carbon emission data into the regional carbon emission accounting model to obtain the regional carbon emissions. This method calculates the regional carbon emissions by defining the accounting scope. However, because the data types within the accounting scope are numerous but not exhaustive, the calculation is complex and inaccurate.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this application, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to address the time lag and insufficient representation of carbon emissions from different energy consumption sectors in existing carbon emission accounting methods. It proposes a carbon emission accounting method for megacities based on electricity statistics. To obtain and predict carbon emission data for smaller blocks, it incorporates high-quality electricity statistics into the carbon emission accounting process. A smoothed variable model with matching coefficients for each block is constructed for carbon emission accounting. Furthermore, an autoregressive moving average model is used to predict the variable coefficients on the time scale, enabling the acquisition of more timely, accurate, and representative carbon emission data for future years.

[0007] In a first aspect, one technical solution provided in this embodiment of the invention is a method for calculating carbon emissions in megacities based on electricity statistics, comprising the following steps:

[0008] Obtain the second power load data of each power-consuming sector in the second-level blocks corresponding to the first-level blocks belonging to megacities based on the time scale;

[0009] The first electricity consumption statistics of each electricity-consuming sector in the first-level block were calculated based on the second electricity load data.

[0010] Obtain carbon emission data corresponding to the primary blocks based on the time scale;

[0011] A coefficient smoothing variable model was used to fit the carbon emission data and the corresponding first electricity consumption statistics of the first-level blocks to obtain the variable coefficients corresponding to the first-level blocks on the time scale, and thus obtain the variable coefficient sequence.

[0012] Based on the variable coefficient sequence, an autoregressive moving average model is used to predict the variable coefficient on the time scale;

[0013] The calculated variable coefficients are imported into the coefficient smoothing variable model to calculate the carbon emission data corresponding to the secondary blocks on the time scale.

[0014] Preferably, the step of obtaining the second power load data of each power consumption sector in the second-level block corresponding to the first-level block of the megacity according to the time scale includes the following steps:

[0015] Obtain the second power load data of each power-consuming sector in the second-level blocks corresponding to each first-level block on an annual basis;

[0016] Obtain multi-attribute data from the second power load data and classify the multi-attribute data according to the carbon energy conversion ratio.

[0017] Preferably, the second power load data includes industrial power load data, regional power load data, and industry power load data.

[0018] Preferably, the carbon-to-energy conversion ratio of the industrial power load data is the ratio of the annual carbon emissions of the industry to the annual gross domestic product of the industry;

[0019] The carbon-to-energy conversion ratio of the regional electricity load data is the ratio of the region's annual carbon emissions to its annual GDP.

[0020] The carbon-to-energy conversion ratio of the industry's electricity load data is the ratio of the industry's annual carbon emissions to its annual GDP.

[0021] Preferably, the step of calculating the first electricity consumption statistics of each electricity-consuming sector in the primary block based on the second electricity load data includes the following steps:

[0022] The first-level block corresponds to several second-level blocks; the second power load data of each power-consuming department corresponding to each second-level block is obtained in sequence, and the obtained second power load data is classified according to attributes and then accumulated to obtain the first power consumption statistics data of each power-consuming department belonging to the first-level block.

[0023] Preferably, the step of obtaining carbon emission data corresponding to the primary block based on the time scale includes the following steps: obtaining carbon emission data corresponding to the primary block and the secondary block based on the annual statistical yearbooks corresponding to each level of block; and correcting the carbon emission data of the primary block through the secondary block to obtain corrected carbon emission data.

[0024] Preferably, the step of correcting the carbon emission data of the primary block through the secondary block to obtain the carbon emission data includes the following steps:

[0025] The carbon emission prediction model is obtained by training the Prophet prediction model based on the carbon emission data of the primary block and the corresponding secondary block.

[0026] The annual carbon emission forecast is obtained by using a trained carbon emission forecasting model, and the average of the forecast and the actual carbon emission is taken as the carbon emission correction data.

[0027] As a preferred embodiment, the coefficient smoothing variable model has the following formula:

[0028] C=β0(z)+x′β(z)+ε=X′δ(z)+ε

[0029] Where, matrix x′=(E 1t ,…,E nt E nt This represents the electricity load data consumed by the nth sector of the first-level block in year t; matrix X′=(1,x′), matrix δ(z)=(β0(z),β1(z),…,β n (z)), C is the carbon emission vector; β0(z) represents the intercept parameter function; β i (z) represents the slope parameter function; i represents the number of electricity consuming departments and i = 1, 2, 3, ..., n; t is the time range selected in the case; and ε is the error coefficient.

[0030] The beneficial effects of this invention are as follows: This invention provides a method for calculating carbon emissions in megacities based on electricity statistics. To obtain and predict carbon emission data for smaller blocks, it incorporates electricity statistics with excellent statistical quality into the carbon emission calculation. A smoothed variable model with matching coefficients for each block is constructed for carbon emission calculation. Furthermore, an autoregressive moving average model is used to predict the variable coefficients on the time scale, enabling the acquisition of more timely, accurate, and representative carbon emission data for future years. The carbon emission calculation method for megacities based on electricity statistics designed in this invention has more accurate estimation capabilities and possesses real-time performance and accuracy for carbon emission calculation in megacities. It solves the problems of difficulty in collecting statistical data and complex, redundant models in previous carbon emission estimation methods.

[0031] The above description of the invention is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0032] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings. The drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings.

[0033] Figure 1 This is a flowchart of a method for calculating carbon emissions in megacities based on electricity statistics, according to the present invention. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only one preferred embodiment of this invention and are only used to explain this invention. They do not limit the scope of protection of this invention. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0035] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the operations (or steps) as sequential processes, many of the operations (or steps) can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but it may also have additional steps not included in the figures; the process may correspond to a method, function, procedure, subroutine, subroutine, etc.

[0036] Example: Figure 1 As shown, a method for calculating carbon emissions in megacities based on electricity statistics includes the following steps:

[0037] S1. Obtain the second power load data of each power consumption sector in the second-level block corresponding to the first-level block of the megacity based on the time scale.

[0038] Specifically, the second power load data of each power-consuming sector in the second-level blocks corresponding to each first-level block is obtained on an annual basis;

[0039] Obtain multi-attribute data from the second power load data and classify the multi-attribute data according to the carbon energy conversion ratio.

[0040] Specifically, the second power load data includes industrial power load data, regional power load data, and industry power load data.

[0041] Specifically, the carbon-to-energy conversion ratio of industrial electricity load data is the ratio of the industry's annual carbon emissions to its annual gross domestic product.

[0042] Specifically, the carbon-to-energy conversion ratio of regional electricity load data is the ratio of the region's annual carbon emissions to its annual GDP.

[0043] Specifically, the carbon-to-energy conversion ratio of industry electricity load data is the ratio of the industry's annual carbon emissions to its annual GDP.

[0044] Furthermore, classifying multi-attribute data based on carbon-energy conversion ratio can be understood as classifying the current attribute data according to the category of the largest carbon-energy conversion ratio among the attribute data. For example, the power load data of a department belongs to industrial power load data, regional power load data, and industry power load data. When the carbon-energy conversion ratio of industrial power load data is greater than that of regional power load data, and the carbon-energy conversion ratio of regional power load data is greater than that of industry power load data, then the power load data of that department is industrial power load data.

[0045] Understandably, based on administrative divisions, data is collected on electricity load from various sectors, including super-regions (such as Shanghai as a first-level region), industrial electricity load data (e.g., primary and tertiary industry data; further, based on industry classification: agriculture, forestry, fishery, and animal husbandry are classified as primary industry; industry and construction as secondary industry; and everything else as tertiary industry), regional electricity load data (e.g., electricity load data for urban and non-urban residents), and industry-specific electricity load data (construction, manufacturing, light industry, etc.). λt , where i represents each province and i = 1, 2, 3, ..., 23; λ represents the number of electricity-consuming sectors selected in the case and λ = 1, 2, 3, ..., n; t represents the year in which the data was collected.

[0046] S2. Calculate the first electricity consumption statistics of each electricity-consuming sector in the first-level block based on the second power load data.

[0047] Specifically, the primary block corresponds to several secondary blocks; the second power load data of each power-consuming department corresponding to each secondary block is obtained in sequence, and the obtained second power load data is classified according to attributes and then accumulated to obtain the first power consumption statistics data of each power-consuming department belonging to the primary block.

[0048] It is understandable that, apart from the difference in administrative level (such as Jing'an District in Shanghai), the carbon emission accounting and the sectoral composition of electricity load are the same for primary and secondary blocks. Therefore, based on the above records, the electricity consumption statistics of primary blocks can be deduced from the electricity load data and carbon emission data of each sector in secondary blocks.

[0049] S3. Obtain carbon emission data corresponding to the primary blocks based on the time scale.

[0050] Specifically, carbon emission data for primary and secondary blocks are obtained from the annual statistical yearbooks corresponding to each level of block; carbon emission data for primary blocks is then corrected using secondary blocks to obtain corrected carbon emission data.

[0051] Specifically, the carbon emission correction data is obtained by correcting the carbon emission data of the primary block through the secondary block, including the following steps:

[0052] The carbon emission prediction model is obtained by training the Prophet prediction model based on the carbon emission data of the primary block and the corresponding secondary block.

[0053] The annual carbon emission forecast is obtained by using a trained carbon emission forecasting model, and the average of the forecast and the actual carbon emission is taken as the carbon emission correction data.

[0054] Understandably, since statistical values ​​are subject to delays and predicted values ​​are affected by various factors, using the average of predicted and actual carbon emissions as the carbon emission correction value can compensate for the reliability issues caused by time delays in order to ensure the reliability of the obtained data.

[0055] S4. A coefficient smoothing variable model is used to fit the carbon emission data and the corresponding first electricity consumption statistics of the first-level block to obtain the variable coefficients corresponding to the first-level block on the time scale, and thus obtain the variable coefficient sequence.

[0056] Understandably, the coefficient smoothed variable model (SVCM), as a semi-parametric estimation model, includes both parametric and non-parametric parts. It combines the advantages of the completeness of parametric models with the flexibility of non-parametric models, making it suitable for solving heterogeneity problems in empirical studies. Taking the study of carbon emissions in Shanghai as an example, heterogeneity refers to the different impacts of electricity consumption in different sectors on carbon emissions in Shanghai in different years (economic development periods). Therefore, this model can better reflect the heterogeneity between carbon emissions and electricity consumption data in different years (economic development periods).

[0057] Specifically, the following benchmark model is constructed:

[0058]

[0059] Wherein, E(ε) t |E it ,z)=0

[0060] Transforming equation (1) yields:

[0061] C=β0(z)+x′β(z)+ε=X′δ(z)+ε (2)

[0062] Where, matrix x′=(E 1t ,…,E nt E nt Electricity consumption statistics for the nth sector of the first-level block in year t; matrix X′=(1,x′), matrix δ(z)=(β0(z),β1(z),…,β n (z)), C is the carbon emission vector; β0(z) represents the intercept parameter function; β i (z) represents the slope parameter function; i represents the number of electricity consuming departments and i = 1, 2, 3, ..., n; t is the time range selected in the case; and ε is the error coefficient.

[0063] In one specific embodiment of this application, time is selected as the smoothing variable z = (tt) as the non-parametric part. 初 ) / (t 末 -t 初), that is, z∈(0,1).

[0064] Specifically, the linear relationship between carbon emissions and electricity statistics is a nonlinear function based on time z, meaning that carbon emissions are a linear function of electricity statistics under time conditions, and the coefficient βi(z) is... z The nonlinear function; compared with nonparametric estimation, SVCM reduces the "curse of dimensionality" by setting the linear impact of electricity load data on carbon emissions as a time-based condition, thus better estimating the time-varying trend of the impact of electricity statistics on carbon emissions in various sectors.

[0065] In one specific embodiment of this application, the SVCM contains only one smoothing variable. For any index variable z∈(0,1), the bandwidth vector h is a constant, and β(z) is a semi-parametric smoothing function vector. According to the kernel function method, the local constant estimate (LC) is obtained as follows:

[0066]

[0067] Where m and j represent the sample size and observation index, respectively, and k(·) and h represent the Gaussian kernel function and the corresponding kernel function window width vector, respectively.

[0068] Furthermore, considering that the local constant estimator (LC) may have a large bias near the Z boundary, the local linear estimator (LL) can be used in this case, and the model changes from (2) to the following equation:

[0069]

[0070] Equation (4) indicates that the approximate c can be obtained through a linear extension of β(z), and z j The closer the value is to zi, the more accurate the approximation.

[0071] Furthermore, define Let x be the j-th row of matrix x. That is E j Each variable in the equation is multiplied by the auxiliary variable (z). j -z); calculate β(z) and The local linear estimator (LL) is:

[0072] Where C is the carbon emission vector; k is the kernel function vector; and h represents the smoothing parameter related to z, often referred to as the "bandwidth". The most important parameter in SVCM estimation is the bandwidth. A larger bandwidth indicates that more observations are used, resulting in a smaller variance in the estimator. However, this also reduces the model's degrees of freedom and increases the bias of the estimator. Conversely, a smaller bandwidth results in a smaller bias but a larger variance in the estimator. In this embodiment, the optimal bandwidth h is obtained by using the one-out-of-one cross-validation (LOO-CV) method. * :

[0073] in It is the LOO-CV coefficient estimate obtained by dividing the given bandwidth h by the i-th observation. This is the corresponding fitted value, w(z) i ) is the weight used to adjust for problems with uneven Z-distribution.

[0074] Furthermore, the optimal bandwidth h is obtained. * Then, model (4) is fitted, and h is... * The smoothed variable value z is substituted into the kernel function to obtain the weights, and then the coefficient LL estimates of each variable are obtained according to equation (5). This part is the variable coefficient value of the known year.

[0075] Furthermore, this embodiment also requires a goodness-of-fit test on the model, wherein... For the standard goodness of fit, since It may be a negative value, so it is still used. As a measure of goodness of fit for nonnegative fitting:

[0076]

[0077]

[0078] S5. Based on the variable coefficient sequence, an autoregressive moving average model is used to predict the variable coefficient on the time scale.

[0079] Specifically, the Autoregressive Moving Average (ARMA) model is a short-term forecasting model with relatively high prediction accuracy in time series analysis. The basic principle of this model is to treat the historical data sequence of the forecast indicator as a set of random sequences, which is an effective combination of the autoregressive model and the moving average model.

[0080] The basic form of the ARMA model is as follows:

[0081]

[0082] in For constant coefficients, {ε tLet} be a sequence of random disturbance terms. The parameter p represents the lag order of the autoregressive component. Here are the parameters of the autoregressive model; parameter q represents the lag order of the moving average component, and φ is the parameter of the autoregressive model. i (i = 1, 2, 3, ... q) are the coefficients of the moving average model.

[0083] Furthermore, the variable coefficient sequence β of the electricity load data for each year is obtained. it The time series data β were analyzed using the ADF unit root method. it Stationarity tests are performed separately. If the original sequence is not stationary, it needs to be differencing and then tested for stationarity again, using Dβ. it This represents the first difference of the sequence.

[0084] Furthermore, the orders p and q of the ARMA(p,q) model are determined using the Bayesian Information Criterion (BIC). The variance σ of the fitting residuals is calculated based on the obtained time series data. 2 Valuation Let L be the highest order of the fitted model and t be the number of samples. Based on experience, L is generally taken as the mean. Operations can be performed within the range that satisfies 0≤p≤L and 0≤q≤L.

[0085] The BIC criterion function is defined as follows:

[0086]

[0087] If a certain order (p) * q * ) satisfies p = p * , q = q * hour Therefore, it can be shown that the best-fitting model is ARMA(p) * q * The order of the best-fit model for each group of stationary time series was calculated and determined.

[0088] Furthermore, the parameters will be obtained through the BIC method. Substitute the data into the ARMA model, and then use the time series data β that has passed the stationarity test. it or Dβ it Input the model respectively

[0089] Furthermore, after determining the estimation model, to verify its accuracy, a unit root test needs to be performed on the residuals. If the residuals pass the stationarity test, then the model can be used for prediction. Using Lβ... itThe values ​​represent the residuals of each model group. If the residuals of each group pass the stationarity test and the p-values ​​are all less than 0.05, it indicates that the model has good predictive ability.

[0090] S6. Import the calculated variable coefficients into the coefficient smoothing variable model to calculate the carbon emission data corresponding to the secondary blocks on the time scale.

[0091] Understandably, since statistical data on the types of energy consumed by each administrative district in Shanghai is unavailable, the actual accounting process can only calculate the relevant energy consumption and carbon emissions of Shanghai as a whole, based on the available basic statistical data. Therefore, a statistical relationship model between Shanghai's electricity statistics and its carbon emissions is constructed, and then the carbon emissions of each administrative district are derived using the electricity statistics of each district.

[0092] Specifically, since the variable coefficient values ​​of each variable in the coefficient smoothing variable model are the same in the same period, the electricity load data of each administrative district in Shanghai and the variable coefficient values ​​of each year obtained from steps S4 and S5 are brought into the baseline model (1) to deduce the carbon emissions of each administrative district in Shanghai in previous years.

[0093] The specific embodiments described above are preferred embodiments of the carbon emission accounting method for megacities based on electricity statistics data of the present invention, and are not intended to limit the specific scope of the present invention. The scope of the present invention includes but is not limited to the specific embodiments described above. All equivalent changes made in accordance with the shape and structure of the present invention are within the protection scope of the present invention.

Claims

1. A method for calculating carbon emissions in megacities based on electricity statistics, characterized in that: Includes the following steps: Obtain the second power load data of each power-consuming sector in the second-level blocks corresponding to the first-level blocks belonging to megacities based on the time scale; The first electricity consumption statistics of each electricity-consuming sector in the first-level block were calculated based on the second electricity load data. Obtain carbon emission data corresponding to the primary blocks based on the time scale; A coefficient smoothing variable model was used to fit the carbon emission data and the corresponding first electricity consumption statistics of the first-level blocks to obtain the variable coefficients corresponding to the first-level blocks on the time scale, and thus obtain the variable coefficient sequence. Based on the variable coefficient sequence, an autoregressive moving average model is used to predict the variable coefficient on the time scale; The calculated variable coefficients are imported into the coefficient smoothing variable model to calculate the carbon emission data corresponding to the secondary blocks on the time scale.

2. The method for calculating carbon emissions of megacities based on electricity statistics as described in claim 1, characterized in that: The process of obtaining the second power load data of each power-consuming sector in the second-level blocks corresponding to the first-level blocks of megacities based on a time scale includes the following steps: Obtain the second power load data of each power-consuming sector in the second-level blocks corresponding to each first-level block on an annual basis; Obtain multi-attribute data from the second power load data and classify the multi-attribute data according to the carbon energy conversion ratio.

3. The method for calculating carbon emissions of megacities based on electricity statistics as described in claim 2, characterized in that: The second power load data includes industrial power load data, regional power load data, and industry power load data.

4. The method for calculating carbon emissions of megacities based on electricity statistics as described in claim 3, characterized in that: The carbon-to-energy conversion ratio of the industrial electricity load data is the ratio of the industry's annual carbon emissions to its annual gross domestic product. The carbon-to-energy conversion ratio of the regional electricity load data is the ratio of the region's annual carbon emissions to its annual GDP. The carbon-to-energy conversion ratio of the industry's electricity load data is the ratio of the industry's annual carbon emissions to its annual GDP.

5. The method for calculating carbon emissions of megacities based on electricity statistics as described in claim 1, characterized in that: The calculation of the first electricity consumption statistics of each electricity-consuming sector in the primary block based on the second electricity load data includes the following steps: The first-level block corresponds to several second-level blocks; the second power load data of each power-consuming department corresponding to each second-level block is obtained in sequence, and the obtained second power load data is classified according to attributes and then accumulated to obtain the first power consumption statistics data of each power-consuming department belonging to the first-level block.

6. The method for calculating carbon emissions of megacities based on electricity statistics as described in claim 1, characterized in that: The process of obtaining carbon emission data corresponding to the primary blocks based on a time scale includes the following steps: Carbon emission data for primary and secondary blocks are obtained from the annual statistical yearbooks corresponding to each level of block; carbon emission data for primary blocks is corrected by adjusting the carbon emission data for secondary blocks to obtain corrected carbon emission data.

7. The method for calculating carbon emissions of megacities based on electricity statistics as described in claim 6, characterized in that: The process of correcting the carbon emission data of the primary block through the secondary block to obtain the carbon emission data includes the following steps: The carbon emission prediction model is obtained by training the Prophet prediction model based on the carbon emission data of the primary block and the corresponding secondary block. The annual carbon emission forecast is obtained by using a trained carbon emission forecasting model, and the average of the forecast and the actual carbon emission is taken as the carbon emission correction data.

8. The method for calculating carbon emissions of megacities based on electricity statistics as described in claim 1, characterized in that: The coefficient smoothing variable model is defined by the following formula: C=β0(z)+x′β(z)+ε=X′δ(z)+ε Where, matrix x′=(E 1t ,…,E nt E nt This represents the electricity consumption statistics of the nth sector in the first-level block in year t; matrix X′=(1,x′), matrix δ(z)=(β0(z),β1(z),…,β n (z)), C is the carbon emission vector; β0(z) represents the intercept parameter function; β i (z) represents the slope parameter function; i represents the number of electricity consuming departments and i = 1, 2, 3, ..., n; t is the time range selected in the case; and ε is the error coefficient.

Images