A carbon emission calculation method based on power big data

Abstract

A carbon emission measurement method based on electric power big data, which relates to the technical field of carbon emission measurement, selects corresponding carbon emission associated variables according to the characteristics of the object to be measured, verifies the stability, cointegration relationship and causality between the associated variables and the power consumption variable, thereby constructs a relationship model between the associated variables and the power consumption variable, calculates the relevant adjustment factors, and calculates the carbon emission data of the object to be measured according to the relationship model. The beneficial effects of the present invention are: based on high-frequency, multi-dimensional electric power big data, the present invention uses frequency conversion, model inference, seasonal adjustment and other technologies to calculate the missing high-frequency macro-statistical indicators, adopts the inventory compilation method of the IPCC, and draws on various carbon emission coefficients given in international organizations, domestic authoritative organizations and public research literature, so as to perform high-frequency measurement of carbon emissions of regions, key industries and households.

Description

Technical Field

[0001] The present invention relates to the technical field of carbon emission calculation, and in particular to a carbon emission calculation method based on power big data. Background Art

[0002] The main methods for calculating carbon emissions at home and abroad include inventory compilation method, actual measurement method, material balance method, model method, etc. (Zhang Deying et al., 2005). Among them, inventory compilation method is the most commonly used method. International authoritative institutions, such as IEA, CDIAC, EDGAR, EIA, etc., although the specific calculation methods used for carbon emissions measurement are different, basically use the inventory compilation method based on the IPCC calculation formula (Li Qingqing et al., 2018).

[0003] Specific methods for calculating household carbon emissions include input-output method, consumer lifestyle method, life cycle assessment method, carbon footprint calculation model method, and inventory compilation method (also known as emission coefficient method), among which the inventory compilation method is more widely used.

[0004] The so-called inventory compilation method is mainly based on the "2006 IPCC National Greenhouse Gas Inventory Guidelines" published by the IPCC. On the basis of clarifying the main sources of greenhouse gas emissions (the main sources of emissions are from energy activities, industrial production processes, agriculture, land use changes and forestry, and waste, etc.), the following formula is used to calculate carbon emissions:

[0005] Emissions = activity level × emission factor;

[0006] According to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories, there are three types of emission factors to choose from, namely, IPCC default emission factors, country-specific emission factors, and emission factors obtained using complex methods using model tools.

[0007] The biggest limitation of calculating carbon emissions based on the inventory compilation method is that the frequency of the initial data is low, usually annual data, so only low-frequency carbon emission calculation results can be calculated; in addition, due to the lack of low-frequency annual energy consumption data in some dimensions, such as households or industries in certain cities, it is difficult to calculate low-frequency annual carbon emissions in many dimensions. Therefore, the shortcomings of existing technologies are that the frequency of calculated carbon emission data is low, usually annual results, and relatively lagging, and the data in some dimensions is missing, making carbon emission calculation impossible. Summary of the invention

[0008] In order to overcome the shortcomings of the existing technology, the present invention proposes a carbon emission measurement method based on power big data. Based on high-frequency, multi-dimensional power big data, the high-frequency energy consumption data of provinces, cities, key industries, and households are inferred by making full use of frequency conversion, model inference, seasonal adjustment and other technologies, so as to achieve high-frequency monitoring of carbon emissions in regions, key industries, and households.

[0009] The present invention provides a carbon emission calculation method based on power big data, comprising the following steps:

[0010] The first step is to select the corresponding carbon emission related variables according to the characteristics of the object to be measured;

[0011] The second step is to test the stability of the associated variable and the electricity consumption variable. If both variables are stable series, go directly to the fourth step;

[0012] The third step is to conduct a cointegration test if both variables are non-stationary variables. If there is a cointegration relationship between the two variables, proceed to the next step of testing.

[0013] The fourth step is to test the causal relationship between the associated variable and electricity consumption. If electricity consumption is the Granger cause of the associated variable, proceed to the next step;

[0014] The fifth step is to build a relationship model between the associated variables and the electricity consumption variables;

[0015] The sixth step is to calculate the relevant adjustment factors and calculate the carbon emission data of the object to be tested based on the relationship model in the fifth step.

[0016] The specific steps of the carbon emission calculation method based on power big data applied to regional carbon emission calculation are as follows:

[0017] S1, calculate the annual value of total energy consumption of each subordinate region and each industry E1, use E to represent the annual value of regional energy consumption and the annual value of total energy consumption of each industry, use G to represent the annual value of regional added value and the annual value of added value of each industry, and G1 to represent the annual value of gross regional product and the annual value of gross product of each industry; then:

[0018] E1=(E / G)*G1

[0019] S2, test the stability of total energy consumption and electricity consumption variables. If both variables are stable series, go directly to S4; use E to represent the annual value of regional energy consumption and total energy consumption of each industry, and perform a stability test on the E series, that is,

[0020]

[0021] Construct the following statistic for Et-1 The previous parameters are tested. If the parameters are less than 0, the sequence is stable.

[0022] or

[0023] S3, if both variables are non-stationary variables, then perform a cointegration test. If the two variables have a cointegration relationship, then proceed to S4; use Ec to represent the electricity consumption of the region and each industry, and use E to represent the annual value of the regional energy consumption and the total energy consumption of each industry. If both sequences are non-stationary and are first-order single integrals, then perform the following cointegration test:

[0024]

[0025] in, y t is a two-dimensional vector y t =(E t ,Ec t ) According to the above expression, the following statistics can be constructed to test the cointegration relationship between two variables;

[0026]

[0027] where r * is the canonical correlation dilution, M is the total number of variables, and r is the number of cointegrating vectors;

[0028] S4, test the causal relationship between total energy consumption and electricity consumption. If electricity consumption is the Granger cause of total energy consumption, proceed to the next step; use Ec to represent the electricity consumption of regions and industries, and use E to represent the annual value of regional energy consumption and total energy consumption of industries. If the null hypothesis H0 in the following model is established, it means that electricity consumption is the Granger cause of total energy consumption;

[0029]

[0030] H0:α1=α2=…α p =0

[0031] S5, estimate and screen the long-term equilibrium relationship model between the total energy consumption and electricity consumption of each region and industry; use Ec to represent the electricity consumption of each region and industry, and establish a long-term equilibrium relationship model for Ec and E series as follows:

[0032]

[0033] S6, calculate the adjustment factor F; use the above model and the monthly data of Qinghai Electric Power Company to calculate the total monthly energy consumption data of the region and each industry for 12 months each year, recorded as Then the adjustment factor F is:

[0034]

[0035] S7, using the adjustment factor F and the carbon emission coefficient h of standard coal, the monthly carbon emissions C of the region and each industry are obtained. t :

[0036]

[0037] The specific steps of the carbon emission calculation method based on power big data applied to carbon emission calculation in key industries are as follows:

[0038] S01, using the interpolation model to interpolate the missing industry output data;

[0039] S02, using the seasonal adjustment model X-12-ARIMA, seasonally adjust the original series of industry output and electricity consumption. Taking electricity consumption Ec as an example,

[0040] Ec t =T t *C t *S t *I t

[0041] Based on this, the seasonal factor S of each sequence can be separated, and the TCI sequence of each sequence can be obtained at the same time. The output TCI sequence is recorded as Y, and the electricity consumption TCI sequence is recorded as X;

[0042] S03, test the stability of the output TCI sequence Y and the electricity consumption TCI sequence X. The test method is the same as step S2 of regional carbon emission calculation. If both sequence Y and sequence X are stable sequences, directly enter the causal relationship test.

[0043] S04, if both the output TCI series Y and the electricity consumption TCI series X are non-stationary variables, a cointegration test is performed. The test method is the same as step S3 of regional carbon emission calculation; if the two variables have a cointegration relationship, proceed to the next step of testing;

[0044] S05, check the causal relationship between the output TCI sequence Y and the electricity consumption TCI sequence X, using the same method as step S4 of regional carbon emission calculation; if the electricity consumption TCI sequence X is the Granger cause of the output TCI sequence Y, proceed to the next step;

[0045] S06, estimate and screen the ADL (q, s) model of the output TCI series Y and the electricity consumption TCI series X as follows:

[0046]

[0047] S07, based on the seasonal factor S and the carbon emission coefficient w of the corresponding industry products, the carbon emissions C in the production process of each industry can be calculated. t ;

[0048]

[0049] The carbon emission calculation method based on power big data is applied to the carbon emission calculation of residential households, distinguishing between rural households and urban households. The carbon emission calculation of rural households is similar to that of urban households.

[0050] The specific steps of the carbon emission calculation method based on power big data applied to carbon emission calculation of urban households are as follows:

[0051] S-1, test the stationarity of the per capita consumption PC series and the per capita GDP series. The test method is the same as step S2 of regional carbon emission calculation. If both the per capita consumption PC series and the per capita GDP series are stationary series, directly proceed to the causal relationship test.

[0052] S-2, if both the per capita consumption PC series and the per capita GDP series are non-stationary variables, a cointegration test is performed. The test method is the same as step S3 of regional carbon emission measurement; if the per capita consumption PC series and the per capita GDP series have a cointegration relationship, proceed to the next step of the test;

[0053] S-3, test the causal relationship between the per capita consumption PC sequence and the per capita GDP sequence, using the same method as step S4 of regional carbon emission estimation; if the per capita consumption PC sequence and the per capita GDP sequence Granger cause, proceed to the next step;

[0054] S-4, using the four categories of urban consumption data and the corresponding urban GDP data, establish a long-term equilibrium relationship model between per capita consumption (PC) and per capita GDP (PG); and use population data to calculate the consumption data TC of each district and county, where i represents the consumption category and j represents the district and county;

[0055] TC ij =(α0+α1*PG ij )*P ij

[0056] S-5, using the interpolation model, interpolate the analysis indicators to obtain the monthly series indicator TC ijt ; Estimate and screen the consumption TC of urban households in the area ijt The ADL(1,0) model with its monthly electricity consumption X is as follows:

[0057]

[0058] S-6, according to the Sturges grouping method, the electricity consumption of urban households in each district and county is used to group them, and the emission comparison coefficient GF of different groups is determined; at the same time, the fixed-base price adjustment factor PF is calculated; referring to the emission factor CF of each type of consumption, the indirect carbon emissions IC of each type of consumption is calculated ijt , the formula is as follows:

[0059]

[0060] S-7, using the proportion of the four major types of consumption carbon emissions to the total consumption carbon emissions of households p1, calculate the total indirect carbon emissions of households in each district and county IC jt :

[0061]

[0062] S-8, using the emission comparison coefficient GF, electricity emission coefficient EF, the proportion of electricity carbon emissions to direct carbon emissions p2, and household electricity consumption data X1 of different groups of urban household electricity consumption in each district and county jt Calculate the direct carbon emissions DC of households in each district and county jt :

[0063]

[0064] S-9, add up the direct carbon emissions and indirect carbon emissions of the household to get the total carbon emissions of the household C jt

[0065] C jt =DC jt +IC jt .

[0066] Explanation of relevant terms:

[0067] Electric power big data: refers to the high-frequency electricity consumption data of each electricity user. Specifically, the high-frequency data here refers to daily data.

[0068] Long-term equilibrium model: refers to the general regression model. When the error term meets the assumption of the equilibrium error term, the corresponding regression model is the long-term equilibrium model.

[0069] ADL model: refers to the autoregressive distributed lag model, in which the independent variables include the lag terms of the dependent variable and the lag terms of the independent variables.

[0070] X-12-ARIMA: It is a seasonal adjustment method that uses repeated moving averages and combines the ARIMA model to extend the sequence at both ends, thereby achieving decomposition of the sequence.

[0071] Denton model: refers to a time decomposition method that maintains the changing trend of the original low-frequency data as the criterion, based on numerical smoothing and repeated optimization.

[0072] Carbon emission coefficient: refers to the carbon emissions generated per unit of energy during the combustion or use of a certain energy, proposed by international authoritative organizations (such as IPCC) or researchers based on a large amount of empirical facts.

[0073] Indirect carbon emissions: refers to the indirect carbon emissions caused by residents' activities such as food, tobacco, alcohol, clothing, education and cultural consumption.

[0074] Direct carbon emissions: refers to the direct carbon emissions caused by the use of electricity, natural gas, gasoline and other energy sources by households.

[0075] Sturges grouping method: refers to the empirical formula for calculating the number of groups k based on the number of data n. The final number of groups is obtained after rounding the calculated result.

[0076] The beneficial effects of the present invention are as follows: the present invention establishes a relationship model between electricity consumption and external macro-statistical variables, and then can use high-frequency electricity consumption data to infer the total energy consumption of regions and industries and the high-frequency data of industry output, and at the same time can infer the high-frequency data of various types of consumption of households, and finally infer the high-frequency data of carbon emissions.

[0077] The present invention is based on high-frequency and multi-dimensional power big data, and utilizes a variety of frequency conversion technologies, seasonal adjustment technologies, and model inference technologies to convert low-frequency energy consumption data and other macro-statistical indicator data into high-frequency data. Combined with the carbon emission coefficients of IPCC and related literature, the present invention solves the problem of high-frequency carbon emission measurement for regions, key industries, and households. DETAILED DESCRIPTION

[0078] Embodiment 1, the present invention provides a carbon emission calculation method based on power big data, comprising the following steps:

[0079] The first step is to select the corresponding carbon emission related variables according to the characteristics of the object to be measured;

[0080] The second step is to test the stability of the associated variable and the electricity consumption variable. If both variables are stable series, go directly to the fourth step;

[0081] The third step is to conduct a cointegration test if both variables are non-stationary variables. If there is a cointegration relationship between the two variables, proceed to the next step of testing.

[0082] The fourth step is to test the causal relationship between the associated variable and electricity consumption. If electricity consumption is the Granger cause of the associated variable, proceed to the next step;

[0083] The fifth step is to build a relationship model between the associated variables and the electricity consumption variables;

[0084] The sixth step is to calculate the relevant adjustment factors and calculate the carbon emission data of the object to be tested based on the relationship model in the fifth step.

[0085] The specific steps of the carbon emission calculation method based on power big data applied to regional carbon emission calculation are as follows:

[0086] S1, calculate the annual value of total energy consumption of each subordinate region and each industry E1, use E to represent the annual value of regional energy consumption and the annual value of total energy consumption of each industry, use G to represent the annual value of regional added value and the annual value of added value of each industry, and G1 to represent the annual value of gross regional product and the annual value of gross product of each industry; then:

[0087] E1=(E / G)*G1

[0088] S2, test the stability of total energy consumption and electricity consumption variables. If both variables are stable series, go directly to S4; use E to represent the annual value of regional energy consumption and total energy consumption of each industry, and perform a stability test on the E series, that is,

[0089]

[0090] Construct the following statistic for E t-1 The previous parameters are tested. If the parameters are less than 0, the sequence is stable.

[0091] or

[0092] S3, if both variables are non-stationary variables, then perform a cointegration test. If the two variables have a cointegration relationship, then proceed to S4; use Ec to represent the electricity consumption of the region and each industry, and use E to represent the annual value of the regional energy consumption and the total energy consumption of each industry. If both sequences are non-stationary and are first-order single integrals, then perform the following cointegration test:

[0093]

[0094] in, y t is a two-dimensional vector y t =(E t ,Ec t ) According to the above expression, the following statistics can be constructed to test the cointegration relationship between two variables;

[0095]

[0096] where r * is the canonical correlation dilution, M is the total number of variables, and r is the number of cointegrating vectors;

[0097] S4, test the causal relationship between total energy consumption and electricity consumption. If electricity consumption is the Granger cause of total energy consumption, proceed to the next step; use Ec to represent the electricity consumption of regions and industries, and use E to represent the annual value of regional energy consumption and total energy consumption of industries. If the null hypothesis H0 in the following model is established, it means that electricity consumption is the Granger cause of total energy consumption;

[0098]

[0099] H0:α1=α2=…α p =0

[0100] S5, estimate and screen the long-term equilibrium relationship model between the total energy consumption and electricity consumption of each region and industry; use Ec to represent the electricity consumption of each region and industry, and establish a long-term equilibrium relationship model for Ec and E series as follows:

[0101]

[0102] S6, calculate the adjustment factor F; use the above model and the monthly data of Qinghai Electric Power Company to calculate the total monthly energy consumption data of the region and each industry for 12 months each year, recorded as Then the adjustment factor F is:

[0103]

[0104] S7, using the adjustment factor F and the carbon emission coefficient h of standard coal, the monthly carbon emissions C of the region and each industry are obtained. t :

[0105]

[0106] The specific steps of the carbon emission calculation method based on power big data applied to carbon emission calculation in key industries are as follows:

[0107] S01, using the interpolation model to interpolate the missing industry output data;

[0108] S02, using the seasonal adjustment model X-12-ARIMA, seasonally adjust the original series of industry output and electricity consumption. Taking electricity consumption Ec as an example,

[0109] Ec t =T t *C t *S t *It

[0110] Based on this, the seasonal factor S of each sequence can be separated, and the TCI sequence of each sequence can be obtained at the same time. The output TCI sequence is recorded as Y, and the electricity consumption TCI sequence is recorded as X;

[0111] S03, test the stability of the output TCI sequence Y and the electricity consumption TCI sequence X. The test method is the same as step S2 of regional carbon emission calculation. If both sequence Y and sequence X are stable sequences, directly enter the causal relationship test.

[0112] S04, if both the output TCI series Y and the electricity consumption TCI series X are non-stationary variables, a cointegration test is performed. The test method is the same as step S3 of regional carbon emission calculation; if the two variables have a cointegration relationship, proceed to the next step of testing;

[0113] S05, check the causal relationship between the output TCI sequence Y and the electricity consumption TCI sequence X, using the same method as step S4 of regional carbon emission calculation; if the electricity consumption TCI sequence X is the Granger cause of the output TCI sequence Y, proceed to the next step;

[0114] S06, estimate and screen the ADL (q, s) model of the output TCI series Y and the electricity consumption TCI series X as follows:

[0115]

[0116] S07, based on the seasonal factor S and the carbon emission coefficient w of the corresponding industry products, the carbon emissions C in the production process of each industry can be calculated. t ;

[0117]

[0118] The carbon emission calculation method based on power big data is applied to the carbon emission calculation of residential households, distinguishing between rural households and urban households. The carbon emission calculation of rural households is similar to that of urban households.

[0119] The specific steps of the carbon emission calculation method based on power big data applied to carbon emission calculation of urban households are as follows:

[0120] S-1, test the stationarity of the per capita consumption PC series and the per capita GDP series. The test method is the same as step S2 of regional carbon emission calculation. If both the per capita consumption PC series and the per capita GDP series are stationary series, directly proceed to the causal relationship test.

[0121] S-2, if both the per capita consumption PC series and the per capita GDP series are non-stationary variables, a cointegration test is performed. The test method is the same as step S3 of regional carbon emission calculation; if the per capita consumption PC series and the per capita GDP series have a cointegration relationship, proceed to the next step of the test;

[0122] S-3, test the causal relationship between the per capita consumption PC sequence and the per capita GDP sequence, using the same method as step S4 of regional carbon emission estimation; if the per capita consumption PC sequence and the per capita GDP sequence Granger cause, proceed to the next step;

[0123] S-4, using the four categories of urban consumption data and the corresponding urban GDP data, establish a long-term equilibrium relationship model between per capita consumption (PC) and per capita GDP (PG); and use population data to calculate the consumption data TC of each district and county, where i represents the consumption category and j represents the district and county;

[0124] TC ij =(α0+α1*PG ij )*P ij

[0125] S-5, using the interpolation model, interpolate the analysis indicators to obtain the monthly series indicator TC ijt ; Estimate and screen the consumption TC of urban households in the area ijt The ADL(1,0) model with its monthly electricity consumption X is as follows:

[0126]

[0127] S-6, according to the Sturges grouping method, the electricity consumption of urban households in each district and county is used to group them, and the emission comparison coefficient GF of different groups is determined; at the same time, the fixed-base price adjustment factor PF is calculated; referring to the emission factor CF of each type of consumption, the indirect carbon emissions IC of each type of consumption is calculated ijt , the formula is as follows:

[0128]

[0129] S-7, using the proportion of the four major types of consumption carbon emissions to the total consumption carbon emissions of households p1, calculate the total indirect carbon emissions of households in each district and county IC jt :

[0130]

[0131] S-8, using the emission comparison coefficient GF, electricity emission coefficient EF, the proportion of electricity carbon emissions to direct carbon emissions p2, and household electricity consumption data X1 of different groups of urban household electricity consumption in each district and county jtCalculate the direct carbon emissions DC of households in each district and county jt :

[0132]

[0133] S-9, add up the direct carbon emissions and indirect carbon emissions of the household to get the total carbon emissions of the household C jt C jt =DC jt +IC jt .

Claims

1. A carbon emission calculation method based on power big data, characterized in that: The following steps are involved: The first step is to select the corresponding carbon emission related variables according to the characteristics of the object to be measured; The second step is to test the stability of the associated variable and the electricity consumption variable. If both variables are stable series, go directly to the fourth step; The third step is to conduct a cointegration test if both variables are non-stationary variables. If there is a cointegration relationship between the two variables, proceed to the next step of testing. The fourth step is to test the causal relationship between the associated variable and electricity consumption. If electricity consumption is the Granger cause of the associated variable, proceed to the next step; The fifth step is to construct a relationship model between the associated variables and the electricity consumption variables; Step 6: Calculate the relevant adjustment factors and calculate the carbon emission data of the object to be measured according to the relationship model in step 5; The specific steps of applying the carbon emission calculation method based on power big data to regional carbon emission calculation are as follows: S1, calculate the annual value of total energy consumption of each subordinate region and each industry E1, use E to represent the annual value of regional energy consumption and the annual value of total energy consumption of each industry, use G to represent the annual value of regional added value and the annual value of added value of each industry, and G1 to represent the annual value of gross regional product and the annual value of gross product of each industry; then: E1=(E / G)*G1 S2, test the stability of total energy consumption and electricity consumption variables. If both variables are stable sequences, go directly to S4; use E to represent the annual value of regional energy consumption and total energy consumption of each industry, and perform a stability test on the E sequence, that is, E t =α+βt+γE t-1 +β1ΔE t-1 +…+b m DE t-m +e t Construct the following statistic for E t-1 The previous parameters are tested. If the parameters are less than 0, the sequence is stable. or S3, if both variables are non-stationary variables, then perform a cointegration test. If the two variables have a cointegration relationship, then proceed to S4; use Ec to represent the electricity consumption of the region and each industry, and use E to represent the annual value of the regional energy consumption and the total energy consumption of each industry. If both sequences are non-stationary and are first-order single integrals, then perform the following cointegration test: in, y t is a two-dimensional vector y t =(E t ,Ec t ) According to the above expression, the following statistics can be constructed to test the cointegration relationship between two variables; where r * is the canonical correlation dilution, M is the total number of variables, and r is the number of cointegrating vectors; S4, test the causal relationship between total energy consumption and electricity consumption. If electricity consumption is the Granger cause of total energy consumption, proceed to the next step; use Ec to represent the electricity consumption of regions and industries, and use E to represent the annual value of regional energy consumption and total energy consumption of industries. If the null hypothesis H0 in the following model is established, it means that electricity consumption is the Granger cause of total energy consumption; H0:α1＝α2＝…α p ＝0 S5, estimate and screen the long-term equilibrium relationship model between the total energy consumption and electricity consumption of each region and industry; use Ec to represent the electricity consumption of each region and industry, and establish a long-term equilibrium relationship model for Ec and E series as follows: S6, calculate the adjustment factor F; use the above model and the monthly data of Qinghai Electric Power Company to calculate the total monthly energy consumption data of the region and each industry for 12 months each year, recorded as Then the adjustment factor F is: S7, using the adjustment factor F and the carbon emission coefficient h of standard coal, the monthly carbon emissions C of the region and each industry are obtained. t :

2. According to claim 1, a carbon emission calculation method based on power big data is characterized in that: The specific steps of the carbon emission calculation method based on power big data applied to carbon emission calculation in key industries are as follows: S01, using the interpolation model to interpolate the missing industry output data; S02, using the seasonal adjustment model X-12-ARIMA, seasonally adjust the original series of industry output and electricity consumption. Taking electricity consumption Ec as an example, Ec t =T t *C t *S t *I t Based on this, the seasonal factor S of each sequence can be separated, and the TCI sequence of each sequence can be obtained at the same time. The output TCI sequence is recorded as Y, and the electricity consumption TCI sequence is recorded as X; S03, test the stability of the output TCI sequence Y and the electricity consumption TCI sequence X. The test method is the same as step S2 of regional carbon emission calculation. If both sequence Y and sequence X are stable sequences, directly enter the causal relationship test. S04, if both the output TCI series Y and the electricity consumption TCI series X are non-stationary variables, a cointegration test is performed. The test method is the same as step S3 of regional carbon emission calculation; if the two variables have a cointegration relationship, proceed to the next step of testing; S05, check the causal relationship between the output TCI sequence Y and the electricity consumption TCI sequence X, using the same method as step S4 of regional carbon emission calculation; if the electricity consumption TCI sequence X is the Granger cause of the output TCI sequence Y, proceed to the next step; S06, estimate and screen the ADL (q, s) model of the output TCI series Y and the electricity consumption TCI series X as follows: S07, based on the seasonal factor S and the carbon emission coefficient w of the corresponding industry products, the carbon emissions C in the production process of each industry can be calculated. t ; 3. According to claim 1, a carbon emission calculation method based on power big data is characterized in that: The specific steps of the carbon emission calculation method based on power big data applied to carbon emission calculation of urban households are as follows: S-1, test the stationarity of the per capita consumption PC series and the per capita GDP series. The test method is the same as step S2 of regional carbon emission measurement. If both the per capita consumption PC series and the per capita GDP series are stationary series, directly proceed to the causal relationship test. S-2, if both the per capita consumption PC series and the per capita GDP series are non-stationary variables, a cointegration test is performed. The test method is the same as step S3 of regional carbon emission calculation; if the per capita consumption PC series and the per capita GDP series have a cointegration relationship, proceed to the next step of the test; S-3, test the causal relationship between the per capita consumption PC series and the per capita GDP series, using the same method as step S4 of regional carbon emissions measurement; if the per capita consumption PC series and the per capita GDP series Granger cause, proceed to the next step; S-4, using the four categories of urban consumption data and the corresponding urban GDP data, establish a long-term equilibrium relationship model between per capita consumption (PC) and per capita GDP (PG); And using population data, calculate the consumption data TC of each district and county, where i represents the consumption category and j represents the district and county; TC ij =(α0+α1*PG ij )*P ij S-5, using the interpolation model, interpolate the analysis indicators to obtain the monthly series indicator TC ijt ; Estimate and screen the consumption TC of urban households in the area ijt The ADL(1,0) model with its monthly electricity consumption X is as follows: S-6, according to the Sturges grouping method, the electricity consumption of urban households in each district and county is used to group them, and the emission comparison coefficient GF of different groups is determined; at the same time, the fixed-base price adjustment factor PF is calculated; referring to the emission factor CF of each type of consumption, the indirect carbon emissions IC of each type of consumption is calculated ijt , the formula is as follows: S-7, using the proportion of the four major types of consumption carbon emissions to the total consumption carbon emissions of households p1, calculate the total indirect carbon emissions of households in each district and county IC jt : S-8, using the emission comparison coefficient GF, electricity emission coefficient EF, the proportion of electricity carbon emissions to direct carbon emissions p2, and household electricity consumption data X1 of different groups of urban household electricity consumption in each district and county jt Calculate the direct carbon emissions DC of households in each district and county jt : S-9, add up the direct carbon emissions and indirect carbon emissions of the household to get the total carbon emissions of the household C jt C jt =DC jt +IC jt 。

Images