Industrial enterprise carbon performance evaluation method based on supervision data

Abstract

The invention provides an industrial enterprise carbon performance evaluation method based on supervision data. The method comprises the following steps: 1) collecting and checking carbon data; 2) refining carbon performance management indexes; 3) constructing a carbon performance evaluation model; 4) intelligent carbon decision support; according to the invention, the precision and efficiency of enterprise carbon emission management can be effectively improved, and the enterprise is helped to comprehensively optimize the carbon management process through real-time carbon emission monitoring, intelligent decision support and cross-department collaborative management.

Description

Technical Field

[0001] The invention relates to a carbon performance evaluation method for industrial enterprises based on regulatory data, belonging to the field of carbon emission management. Background Art

[0002] As the global climate change problem intensifies, greenhouse gas (GHG) emission control has become an important issue for governments and international organizations. As the main greenhouse gas, the control of carbon dioxide (CO2) emissions is of key significance to curbing the rise in global temperatures.

[0003] Industrial enterprises are one of the main sources of carbon emissions. The effective evaluation and management of their carbon performance is of great significance to achieving national and regional carbon reduction targets. However, the existing carbon performance evaluation methods mainly rely on the self-reported data of enterprises. This dependence brings about the problem that the authenticity and integrity of the data are difficult to guarantee. Enterprises may engage in data falsification and selective disclosure in the process of self-reporting, which leads to doubts about the accuracy and credibility of the carbon performance evaluation results. In addition, the traditional carbon performance evaluation methods are often extensive in the construction of the indicator system, and fail to fully consider the carbon emission characteristics of different industries, different enterprises and their internal departments, making it difficult to fully reflect the carbon management level of enterprises. With the rapid development of information technology, the application of technologies such as big data, Internet of Things (IoT) and artificial intelligence (AI) in carbon management has gradually increased. However, the existing carbon performance evaluation methods have not yet made full use of these advanced technologies, resulting in data processing, analysis and decision support still relying on traditional means, which is inefficient and difficult to achieve intelligent management. Especially in terms of real-time monitoring and cross-departmental collaborative management, the existing methods lack systematicity and integration, and cannot achieve information sharing and collaborative work between departments within the enterprise, limiting the overall improvement of carbon performance management.

[0004] In addition, carbon performance management not only involves the monitoring and verification of carbon emission data, but also includes carbon cost management, information disclosure, carbon footprint analysis and carbon audit feedback. Existing methods are insufficient in the comprehensive management of these aspects, making it difficult to form a systematic and comprehensive carbon management system. For example, carbon cost management often lacks sophisticated analysis tools, information disclosure is opaque, carbon footprint analysis fails to fully consider carbon emissions upstream and downstream of the supply chain, and the carbon audit feedback mechanism is imperfect, making it difficult for companies to make effective adjustments and optimizations in the carbon management process. Summary of the invention

[0005] The present invention proposes a carbon performance evaluation method for industrial enterprises based on regulatory data, which aims to build a comprehensive management system of carbon data drive, carbon indicator management, carbon performance evaluation and intelligent carbon decision-making.

[0006] The technical solution of the present invention is: a method for evaluating the carbon performance of industrial enterprises based on regulatory data, the method comprising: step 1) carbon data collection and verification; step 2) refinement of carbon performance management indicators; step 3) construction of a carbon performance evaluation model; step 4) intelligent carbon decision support.

[0007] Furthermore, the step 1) carbon data collection and verification specifically includes: step 1-1) determining the enterprise's carbon emissions; step 1-2) building a data verification model; the step 2) carbon performance management indicator refinement specifically includes: step 2-1) carbon cost management; step 2-2) information disclosure index; step 2-3) carbon footprint monitoring; step 2-4) carbon audit feedback.

[0008] Furthermore, the step 1-1) of determining the carbon emissions of the enterprise specifically includes:

[0009] The enterprise carbon emissions E include direct emissions E direct , Indirect Emissions indirect and other indirect emissions other , the calculation of enterprise carbon emissions E includes the following formula (1) to formula (4):

[0010] E=E direct +E indirect +E other Formula (1);

[0011] in:

[0012]

[0013]

[0014] In the formula, A i is the activity data of direct emission source i; B j is the activity data of indirect emission source j; C k is the activity data of other indirect emission sources k; CF i is the emission factor of direct emission source i, CF j is the emission factor of indirect emission source j, CF k is the emission factor of other indirect emission sources k; EF i is the environmental impact factor of direct emission source i, EF j is the environmental impact factor of indirect emission source j, EF k is the environmental impact coefficient of other indirect emission sources k.

[0015] Further, the data verification model includes: step 1-2-1) consistency check; step 1-2-2) outlier detection; step 1-2-3) data integrity verification;

[0016] The step 1-2-1) consistency check specifically includes the following formula (5):

[0017]

[0018] E in the formula reported For enterprises to report their carbon emissions, E calculated is the carbon emissions calculated using activity data and emission factors, ε is the allowable error range, and C is the relative error between the company’s self-reported carbon emissions and the actual calculated carbon emissions;

[0019] The step 1-2-2) outlier detection specifically includes using a statistical method to detect outliers and using Z-Score for judgment, as shown in the following formula (6):

[0020]

[0021] In the formula, μ E is the average carbon emission, E calculated is the carbon emissions calculated using activity data and emission factors, σ E is the standard deviation of carbon emissions, Z is the multiple of the standard deviation between a data point and the mean of the data set, if |Z|>Z threshold , it is determined as an outlier and the corresponding E calculated is an outlier, Z threshold The threshold for outlier determination.

[0022] Furthermore, the step 1-2-3) of verifying data integrity includes using a missing data interpolation algorithm to ensure the integrity of the data set; the method of using a missing data interpolation algorithm to ensure the integrity of the data set specifically includes the following formula (7):

[0023] E complete =Impute(E incomplete ) formula (7);

[0024] In the formula, E complete represents the complete data set after interpolation, E incomplete is the original data set, which refers to the various activity data collected by the enterprise in carbon performance management; if there are missing values ​​in some records, they must be processed through the interpolation algorithm; Impute means the operation of interpolating the original data set through K-nearest neighbor interpolation;

[0025] In order to ensure the integrity of carbon emission data, we first identify the dataset E incomplete There are data records with missing values ​​in the ; for data points x containing missing values i , based on the remaining complete data points x jThe known information of , calculates the Euclidean distance between the two, and then selects the K data points with the smallest distance to the data point as the reference points for interpolation:

[0026]

[0027] In the formula, x i and x j Represent two data points, x im and x jm is the value on the mth dimension, and n represents the number of dimensions of the data point;

[0028] According to the distance d(x i ,x j ) size, select the same as x i The most recent K data points x k For reference, the K neighbor data points are interpolated by weighting by the inverse of the distance, and the formula is as follows (7-2):

[0029]

[0030] in the formula is the interpolated value, xk is the value of the kth neighbor, is the weight, d k is the distance to the kth neighbor; the interpolated value calculated Replace the missing values ​​in the original data to generate a complete data set E complete .

[0031] Furthermore, the step 2-1) carbon cost management includes carbon cost calculation, and the carbon cost calculation formula takes into account the carbon tax and carbon trading price; the carbon cost calculation includes the following formula (8):

[0032]

[0033] In the formula, C carbon is the carbon cost, P tax is the carbon tax price, P trade,t is the carbon trading price at time t, E direct,t is the carbon trading volume of the enterprise at time t, E trade,t is the carbon trading volume of the enterprise at time t; the information disclosure index of step 2-2) specifically includes determining the weight by using the hierarchical analysis method and calculating it by using the fuzzy comprehensive evaluation method as shown in formula (9):

[0034]

[0035] In the formula, I is the information disclosure index, q is the number of information disclosure items, and S l is the fuzzy evaluation value of information disclosure item l, ωl,updated is the weight of information disclosure item l determined by the analytic hierarchy process (AHP), μ(S l ) is a fuzzy comprehensive evaluation function, which is used to convert the fuzzy evaluation results into deterministic scores;

[0036] The step 2-3) of carbon footprint monitoring includes using a weighted carbon intensity model to calculate a carbon footprint monitoring index, specifically the following formula (10):

[0037]

[0038] In the formula, F is the carbon footprint monitoring index, D is the number of monitoring dimensions, ω d is the weight of dimension d, E d is the carbon emission of dimension d, R d is the relevant economic indicator of dimension d;

[0039] The step 2-4) carbon audit feedback includes calculating the carbon audit feedback index using a dynamic adjustment model, specifically as follows:

[0040]

[0041] In the formula, A is the carbon audit feedback index, N is the number of audit cycles, and ΔE n is the change in carbon emissions in the nth period, E base,n is the baseline carbon emissions in the nth period, λ n is the adjustment coefficient of the nth cycle, λ n Reflect the importance of audit feedback.

[0042] Furthermore, the step 3) constructs a carbon performance evaluation model, specifically including: step 3-1) constructs a performance scoring model; step 3-2) standardization processing; step 3-3) weight determination; step 3-4) uses a comprehensive scoring formula to obtain a comprehensive score; the step 4) intelligent carbon decision support specifically includes: step 4-1) optimization model; step 4-2) decision variables; step 4-3) constraints.

[0043] Furthermore, the step 3-1) constructs a performance scoring model, specifically including adopting a multi-index comprehensive evaluation model and combining principal component analysis to perform dimensionality reduction processing to obtain a comprehensive carbon performance score, specifically as follows:

[0044]

[0045] In the formula, P is the comprehensive carbon performance score, K is the number of main components, φ k is the contribution rate of the principal component k, Z k is the score of principal component k;

[0046] Step 3-2) Standardization processing includes standardizing each index using the range standardization method to eliminate the dimension effect, as shown in the following formula (13):

[0047]

[0048] Among them, X represents any indicator, and after standardization, we get X', X min Represents the minimum value of any indicator, X max Represents the maximum value of any indicator;

[0049] The step 3-3) weight determination specifically includes using the entropy weight method to determine the weight of each indicator, and the calculation process includes the following:

[0050] Step 3-3-1) Use formula (14) to calculate the weight of the indicator:

[0051]

[0052] In the formula, X′ ij is the standardized value of the i-th sample on the j-th index, N is the total number of samples, p ij It indicates the relative weight of a sample on a certain indicator, that is, the proportion of the standardized value of the sample to the sum of the standardized values ​​of all samples;

[0053] Step 3-3-2) Calculate the entropy value using formula (15)-formula (16):

[0054]

[0055] In the formula, H j is the entropy value;

[0056] Step 3-3-3) Calculate the weight using formula (17):

[0057]

[0058] In the formula, M is the total number of indicators, ω j is the weight;

[0059] The step 3-4) uses a comprehensive scoring formula to obtain a comprehensive score, which specifically includes combining the standardized indicators with the weights to obtain a comprehensive score, specifically as follows:

[0060]

[0061] In the formula, P i is the comprehensive carbon performance score of the i-th sample, ω j is the weight, M is the total number of indicators, X′ ij is the value of the i-th sample on the j-th standardized index.

[0062] Furthermore, the step 4-1) optimizes the model, specifically including using a multi-objective optimization model to balance the carbon performance score P and the enterprise economic benefits E econmic , specifically as follows: Formula (19) - Formula (22):

[0063]

[0064] Subject to E≤E target Formula (21);

[0065]

[0066] In the formula, L is the number of carbon emission reduction measures, R l is the contribution rate of carbon emission reduction measures l to economic benefits, Q l To strengthen the implementation of carbon reduction measures, C l is the cost of carbon reduction measure l, B is the carbon management budget, E target is the carbon emission target value, and E is the carbon emission of the enterprise.

[0067] Furthermore, the decision variables in step 4-2) include the implementation strength of each carbon emission reduction measure Q = {Q1, Q2, ..., Q L}, and carbon trading volume E trade :

[0068] Q={Q1,Q2,...,Q L},E trade ≥0;

[0069] In addition to the constraints in the above optimization model, the constraints in step 4-3) also need to consider the following constraints:

[0070]

[0071] where ΔE l is the emission reduction of carbon reduction measure l, E base is the carbon emissions in the base period, E trade is the carbon trading volume, E is the carbon emissions of the enterprise, Q l To strengthen the implementation of carbon emission reduction measures.

[0072] Beneficial effects of the present invention:

[0073] 1) The present invention can effectively improve the accuracy and efficiency of corporate carbon emission management, and help companies comprehensively optimize their carbon management processes through real-time monitoring of carbon emissions, intelligent decision support, and cross-departmental collaborative management;

[0074] 2) The present invention can not only acquire and analyze carbon emission data in real time, but also provide carbon compliance strategy optimization solutions based on historical and real-time data, ensuring that enterprises reduce carbon emission intensity and improve resource utilization efficiency while fulfilling their carbon emission reduction obligations. It can also avoid the risks of exceeding carbon emissions by dynamically adjusting carbon quotas, thereby effectively promoting the low-carbon and sustainable development of enterprises. DETAILED DESCRIPTION

[0075] A carbon performance evaluation method for industrial enterprises based on regulatory data, the method comprising:

[0076] Step 1) Carbon data collection and verification;

[0077] Step 2) Refinement of carbon performance management indicators;

[0078] Step 3) Constructing a carbon performance evaluation model;

[0079] Step 4) Smart carbon decision support.

[0080] The step 1) carbon data collection and verification specifically includes:

[0081] Step 1-1) Determine the carbon emissions of the enterprise;

[0082] Step 1-2) Build a data verification model.

[0083] The step 1-1) determines the carbon emissions of the enterprise, specifically including:

[0084] The enterprise carbon emissions E include direct emissions E direct , Indirect Emissions indirect and other indirect emissions other , the calculation of enterprise carbon emissions E includes the following formula (1) to formula (4):

[0085] E=E direct +E indirect +E other Formula (1);

[0086] in:

[0087]

[0088] In the formula, A i is the activity data of direct emission source i (such as fuel consumption); B j is the activity data of indirect emission source j (such as electricity consumption); C k is the activity data of other indirect emission sources k (such as supply chain emissions); CF i is the emission factor of direct emission source i, CF j is the emission factor of indirect emission source j, CF kis the emission factor of other indirect emission sources k; EF i is the environmental impact factor of direct emission source i, EF j is the environmental impact factor of indirect emission source j, EF k is the environmental impact factor of other indirect emission sources k, EF i EF j EF k Used to adjust the environmental impact weights of different emission sources.

[0089] The step 1-2) constructs a data verification model, specifically including:

[0090] In order to ensure the authenticity and accuracy of the data, a multi-level data verification model is adopted, which includes: step 1-2-1) consistency check; step 1-2-2) outlier detection; step 1-2-3) data integrity verification.

[0091] The step 1-2-1) consistency check specifically includes the following formula (5):

[0092]

[0093] E in the formula reported For enterprises to report their carbon emissions, E calculated is the carbon emissions calculated using activity data and emission factors, ε is the allowable error range, and C is the relative error between the company’s self-reported carbon emissions and the actual calculated carbon emissions.

[0094] The step 1-2-2) outlier detection specifically includes using a statistical method to detect outliers and using Z-Score for judgment, as shown in the following formula (6):

[0095]

[0096] In the formula, μ E is the average carbon emission, E calculated is the carbon emissions calculated using activity data and emission factors, σ E is the standard deviation of carbon emissions, Z is the multiple of the standard deviation between a data point and the mean of the data set, if |Z|>Z threshold , it is determined as an outlier and the corresponding E calculated is an outlier, Z threshold It is the threshold for outlier determination, usually 2 or 3, indicating the degree of deviation of the data point from the mean.

[0097] The step 1-2-3) of verifying data integrity includes using a missing data interpolation algorithm to ensure the integrity of the data set; the missing data interpolation algorithm is preferably K-nearest neighbor interpolation (K-NN); the method of using a missing data interpolation algorithm to ensure the integrity of the data set specifically includes the following formula (7):

[0098] E complete =Impute(E incomplete ) formula (7);

[0099] In the formula, E complete represents the complete data set after interpolation, E incomplete The original data set refers to the various activity data collected by enterprises in carbon performance management, which usually include carbon emission data, energy consumption data and related economic indicators. If there are missing values ​​in some records, they must be processed through interpolation algorithms. Impute means the operation of interpolating the original data set through K-nearest neighbor interpolation.

[0100] In order to ensure the integrity of carbon emission data, we first identify the dataset E incomplete (Dataset E incomplete Including direct emissions E direct , Indirect Emissions indirect and other indirect emissions other , emission factors, etc.) contains missing values; for data points x containing missing values i , based on the remaining complete data points x j The known information of , calculates the Euclidean distance between the two, and then selects the K data points with the smallest distance to the data point as the reference points for interpolation:

[0101]

[0102] In the formula, x i and x j Represent two data points, x im and x jm is the value of the mth dimension, and n represents the number of dimensions of the data point.

[0103] According to the distance d(x i ,x j ) size, select the same as x i The most recent K data points x k For reference, the K neighbor data points are interpolated by weighting by the inverse of the distance, and the formula is as follows (7-2):

[0104]

[0105] in the formula is the interpolated value, xk is the value of the kth neighbor, is the weight, d k is the distance to the kth neighbor; the interpolated value calculated Replace the missing values ​​in the original data to generate a complete data set E complete .

[0106] The step 2) refines the carbon performance management indicators, specifically including:

[0107] Step 2-1) Carbon cost management;

[0108] Step 2-2) Information disclosure index;

[0109] Step 2-3) Carbon footprint monitoring;

[0110] Step 2-4) Carbon audit feedback.

[0111] The step 2-1) carbon cost management includes carbon cost calculation, and the carbon cost calculation formula takes into account carbon tax and carbon trading price; the carbon cost calculation is as follows (8):

[0112]

[0113] In the formula, C carbon is the carbon cost, P tax is the carbon tax price (unit: yuan / ton CO2), P trade,t is the carbon trading price (unit: yuan / ton CO2) at time t, E direct,t is the carbon trading volume of the enterprise at time t, E trade,t is the carbon trading volume of the enterprise at time t.

[0114] The step 2-2) of information disclosure index specifically includes determining weights using the analytic hierarchy process (AHP) and calculating using the fuzzy comprehensive evaluation method as shown in formula (9):

[0115]

[0116] In the formula, I is the information disclosure index, q is the number of information disclosure items, and S l is the fuzzy evaluation value of information disclosure item l, ω l,updated is the weight of information disclosure item l determined by the analytic hierarchy process (AHP), μ(S l ) is a fuzzy comprehensive evaluation function, which is used to convert the fuzzy evaluation results into deterministic scores.

[0117] Based on the original AHP, combined with big data analysis, we can use actual data to analyze the weights ω lThe analytic hierarchy process (AHP) was used to determine the initial weights of each item of information disclosure (such as transparency, timeliness, and accuracy), which were the basis for the subsequent fuzzy comprehensive evaluation.

[0118] In order to determine the weight of each information disclosure item, it is necessary to establish a judgment matrix A, in which each element represents the comparison of the importance of each disclosure item.

[0119]

[0120] In the matrix, each element a ij Indicates information disclosure item S i Relative to S j For example, if transparency disclosure is more important than timeliness disclosure, then a 透明度，及时性 >1.

[0121] Standardize each column of the judgment matrix A and calculate the standardized matrix A normalizedij

[0122]

[0123] Take the average of the elements in each row of the standardized matrix to obtain the standardized weight of each disclosure item:

[0125]

[0126] where w i is the weight of each disclosure item after the judgment matrix is ​​standardized, and q represents the total number of information disclosure items.

[0127] After obtaining the initial weight ω l,first Afterwards, the historical data correction process based on big data analysis and machine learning models can further improve the weights to dynamically adapt to changes in information disclosure behavior; the weight calculation formula at the revised disclosure item level is:

[0128] ω l,updated =ω l,first ×g(historical_data)

[0129] In the formula, g(historical_data) is a weight correction function based on historical data.

[0130] For μ(S l ), the current fuzzy comprehensive evaluation value only considers the quality of information disclosure in a single dimension, and expands the fuzzy evaluation to a multi-dimensional evaluation, such as the transparency, timeliness and accuracy of disclosure. The improved fuzzy evaluation formula is:

[0131] μ(Sl )=ω1×μ 透明度 (S l )+ω2×μ 及时性 (S l )+ω3×μ 准确性 (S l )

[0132] ω1+ω2+ω3=1

[0133] In the formula, ω1 is the weight of the transparency dimension, ω2 is the weight of the timeliness dimension, ω3 is the weight of the accuracy dimension, and μ 透明度 (S l ) is the fuzzy evaluation function of the transparency dimension, μ 及时性 (S l ) is the fuzzy evaluation function of timeliness dimension, μ 准确性 (S l ) is the fuzzy evaluation function of accuracy dimension.

[0134] The step 2-3) of carbon footprint monitoring includes using a weighted carbon intensity model to calculate a carbon footprint monitoring index, specifically the following formula (10):

[0135]

[0136] In the formula, F is the carbon footprint monitoring index, D is the number of monitoring dimensions (such as production, transportation, supply chain, etc.), ω d is the weight of dimension d, E d is the carbon emission of dimension d, R d is the relevant economic indicator of dimension d (such as output value, output).

[0137] The step 2-4) carbon audit feedback includes calculating the carbon audit feedback index using a dynamic adjustment model, specifically as follows:

[0138]

[0139] In the formula, A is the carbon audit feedback index, N is the number of audit cycles, and ΔE n is the change in carbon emissions in the nth period, E base,n is the baseline carbon emissions in the nth period, λ n is the adjustment coefficient of the nth cycle, λ n Reflect the importance of audit feedback.

[0140] The step 3) constructing a carbon performance evaluation model specifically includes:

[0141] Step 3-1) Constructing a performance scoring model;

[0142] Step 3-2) Standardization processing;

[0143] Step 3-3) weight determination;

[0144] Step 3-4) Use the comprehensive scoring formula to get the comprehensive score.

[0145] The step 3-1) constructs a performance scoring model, specifically including adopting a multi-index comprehensive evaluation model and combining principal component analysis (PCA) to perform dimensionality reduction processing to obtain a comprehensive carbon performance score, specifically as follows:

[0146]

[0147] In the formula, P is the comprehensive carbon performance score, K is the number of main components, φ k is the contribution rate of the principal component k, Z k is the score of principal component k, Z k Calculated by principal component analysis (PCA).

[0148] Step 3-2) Standardization processing includes standardizing each index using the range standardization method to eliminate the dimension effect, as shown in the following formula (13):

[0149]

[0150] Where X represents any indicator (such as C carbon Carbon cost, I information disclosure index, F carbon footprint monitoring index, A carbon audit feedback index), after standardization, we get X', X min Represents the minimum value of any indicator, X max Represents the maximum value of any indicator.

[0151] The step 3-3) weight determination specifically includes using the entropy weight method to determine the weight of each indicator, and the calculation process includes the following:

[0152] Step 3-3-1) Use formula (14) to calculate the weight of the indicator:

[0153]

[0154] In the formula, X′ ij is the standardized value of the i-th sample on the j-th index, N is the total number of samples, p ij It indicates the relative weight of a sample on a certain indicator, that is, the ratio of the standardized value of the sample to the total standardized values ​​of all samples.

[0155] Step 3-3-2) Calculate the entropy value using formula (15)-formula (16):

[0156]

[0157] In the formula, Hj is the entropy value;

[0158] Step 3-3-3) Calculate the weight using formula (17):

[0159]

[0160] In the formula, M is the total number of indicators, ω j is the weight.

[0161] The step 3-4) uses a comprehensive scoring formula to obtain a comprehensive score, which specifically includes combining the standardized indicators with the weights to obtain a comprehensive score, specifically as follows:

[0162]

[0163] In the formula, P i is the comprehensive carbon performance score of the i-th sample, ω j is the weight, M is the total number of indicators, X′ ij is the value of the i-th sample on the j-th standardized index.

[0164] The step 4) smart carbon decision support specifically includes:

[0165] Step 4-1) Optimizing the model;

[0166] Step 4-2) Decision variables;

[0167] Step 4-3) Constraints.

[0168] The step 4-1) optimizes the model, specifically including using a multi-objective optimization model to balance the carbon performance score P and the enterprise economic benefits E econmic , specifically as follows: Formula (19) - Formula (22):

[0169]

[0170] Subject to E≤E target Formula (21);

[0171]

[0172] In the formula, L is the number of carbon emission reduction measures, R l is the contribution rate of carbon emission reduction measures l to economic benefits, Q l To strengthen the implementation of carbon reduction measures, C l is the cost of carbon reduction measure l, B is the carbon management budget, E target is the carbon emission target value, and E is the carbon emission of the enterprise.

[0173] The decision variables in step 4-2) include the implementation strength of each carbon emission reduction measure Q = {Q1, Q2, ..., Q L}, and carbon trading volume E trade :

[0174] Q={Q1,Q2,...,Q L},E trade ≥0.

[0175] In addition to the constraints in the above optimization model, the constraints in step 4-3) also need to consider the following constraints:

[0176]

[0177] where ΔE l is the emission reduction of carbon reduction measure l, E base is the carbon emissions in the base period, E trade is the carbon trading volume, E is the carbon emissions of the enterprise, Q l To strengthen the implementation of carbon emission reduction measures.

[0178] The present invention covers carbon verification, carbon compliance and carbon monitoring, systematically obtains direct emissions, indirect emissions and other indirect emissions data of industrial enterprises in the supply chain, introduces environmental impact coefficients, and adjusts the weights of the impacts of different emission sources to ensure the accuracy of carbon emissions; at the same time, a multi-level data verification model is adopted, including consistency checks, outlier detection and data integrity verification, and statistical methods and interpolation algorithms are used to ensure the authenticity and reliability of the data; the present invention refines the carbon performance management indicator system and constructs a comprehensive evaluation model; the weights of each indicator are determined by principal component analysis and entropy weight method, and standardized processing is performed to comprehensively reflect the carbon management level of the enterprise; the intelligent carbon decision support system balances carbon performance and economic benefits through a multi-objective optimization model, and adopts advanced optimization algorithms to optimize the implementation of carbon emission reduction measures, thereby realizing dynamic optimization and continuous improvement of the enterprise's carbon management strategy.

[0179] The present invention covers the links of carbon data collection and verification, refinement of carbon performance management indicators, construction of carbon performance evaluation model and intelligent decision support; the present invention comprehensively covers key links such as carbon verification, carbon compliance and carbon monitoring. By refining carbon performance management indicators, coordinating various departments within the enterprise, and focusing on carbon cost management, information disclosure, carbon footprint monitoring and carbon audit feedback, a comprehensive management system of carbon data drive, carbon indicator management, carbon performance evaluation and intelligent carbon decision-making is constructed, which can effectively improve the accuracy and efficiency of enterprise carbon emission management, and help enterprises comprehensively optimize carbon management processes through real-time monitoring of carbon emissions, intelligent decision support and cross-departmental collaborative management.

[0180] Example 1

[0181] Carbon emission data verification:

[0182] An industrial enterprise reports that its annual carbon emissions are 5,000 tons of CO2. By collecting its fuel consumption, electricity consumption and supply chain emission data and applying the carbon emissions calculation formula of the present invention, it is calculated that:

[0183]

[0184] To perform a consistency check:

[0185]

[0186] Meet the consistency requirements; further use Z-Score detection, if the Z value is within the allowable range, the data passes the verification.

[0187] Example 2

[0188] Carbon Performance Score:

[0189] The standardized scores of enterprises on the four indicators are:

[0190] C c ' arbon =0.8, I'=0.6, F'=0.9, A'=0.7;

[0191] The weights are determined by the entropy weight method:

[0192] ω carbon =0.3,ω disclosure =0.2,ω footprint =0.4,ω audit =0.1;

[0193] Overall performance rating:

[0194] P=0.3×0.8+0.2×0.6+0.4×0.9+0.1×0.7=0.79.

[0195] Example 3

[0196] Intelligent optimization decision:

[0197] The company set a carbon emission target of 4,500 tons of CO2 and a carbon management budget of 1 million yuan; by building a multi-objective optimization model and applying the particle swarm optimization algorithm, the optimal implementation intensity of each carbon emission reduction measure was determined. This maximizes the carbon performance score while ensuring that the economic benefits do not fall below the preset threshold.

Claims

1. A carbon performance evaluation method for industrial enterprises based on regulatory data, characterized by include: Step 1) Carbon data collection and verification; Step 2) Refinement of carbon performance management indicators; Step 3) Construct carbon performance evaluation model; Step 4) Intelligent carbon decision support.

2. The method for evaluating the carbon performance of industrial enterprises based on regulatory data according to claim 1 is characterized by: The step 1) carbon data collection and verification specifically includes: step 1-1) determining the carbon emissions of the enterprise; step 1-2) building a data verification model; the step 2) carbon performance management indicator refinement specifically includes: step 2-1) carbon cost management; step 2-2) information disclosure index; step 2-3) carbon footprint monitoring; step 2-4) carbon audit feedback.

3. The method for evaluating the carbon performance of industrial enterprises based on regulatory data according to claim 2 is characterized by: The step 1-1) determines the carbon emissions of the enterprise, specifically including: The enterprise carbon emissions E include direct emissions E direct , Indirect Emissions indirect and other indirect emissions other , the calculation of enterprise carbon emissions E includes the following formula (1) to formula (4): E = E direct +E indirect +E other Formula (1): in: In the formula, A i is the activity data of direct emission source i; B j is the activity data of indirect emission source j; C k is the activity data of other indirect emission sources k; CF i is the emission factor of direct emission source i, CF j is the emission factor of indirect emission source j, CF k is the emission factor of other indirect emission sources k; EF i is the environmental impact factor of direct emission source i, EF j is the environmental impact factor of indirect emission source j, EF k is the environmental impact coefficient of other indirect emission sources k.

4. The method for evaluating the carbon performance of industrial enterprises based on regulatory data according to claim 2 is characterized by: The data verification model includes: step 1-2-1) consistency check; step 1-2-2) outlier detection; step 1-2-3) data integrity verification; The step 1-2-1) consistency check specifically includes the following formula (5): E in the formula reported For enterprises to report their carbon emissions, E calculated is the carbon emissions calculated using activity data and emission factors, ε is the allowable error range, and C is the relative error between the company’s self-reported carbon emissions and the actual calculated carbon emissions; The step 1-2-2) outlier detection specifically includes using a statistical method to detect outliers and using Z-Score for judgment, as shown in the following formula (6): In the formula, μ E is the average carbon emission, E calculated is the carbon emissions calculated using activity data and emission factors, σ E is the standard deviation of carbon emissions, Z is the multiple of the standard deviation between a data point and the mean of the data set, if |Z|>Z threshold , it is determined as an outlier and the corresponding E calculated is an outlier, Z threshold The threshold for outlier determination.

5. The method for evaluating the carbon performance of industrial enterprises based on regulatory data according to claim 4 is characterized by: The step 1-2-3) data integrity verification includes using a missing data interpolation algorithm to ensure the integrity of the data set; the method of using a missing data interpolation algorithm to ensure the integrity of the data set specifically includes the following formula (7): E complete =Impute(E incomplete ) Formula (7); In the formula, E complete represents the complete data set after interpolation, E incomplete is the original data set, which refers to the various activity data collected by the enterprise in carbon performance management; if there are missing values ​​in some records, they must be processed through the interpolation algorithm; Impute means the operation of interpolating the original data set through K-nearest neighbor interpolation; In order to ensure the integrity of carbon emission data, we first identify the dataset E incomplete There are data records with missing values ​​in the ; for data points x containing missing values i , based on the remaining complete data points x j The known information of , calculates the Euclidean distance between the two, and then selects the K data points with the smallest distance to the data point as the reference points for interpolation: In the formula, x i and x j Represent two data points, x im and x jm is the value on the mth dimension, and n represents the number of dimensions of the data point; According to the distance d(x i ,x j ) size, select the same as x i The most recent K data points x k For reference, the K neighbor data points are interpolated by weighting by the inverse of the distance, and the formula is as follows (7-2): in the formula is the interpolated value, x k is the value of the kth neighbor, is the weight, d k is the distance to the kth neighbor; the interpolated value calculated Replace the missing values ​​in the original data to generate a complete data set E complete .

6. The method for evaluating the carbon performance of industrial enterprises based on regulatory data according to claim 2 is characterized by: The step 2-1) carbon cost management includes carbon cost calculation, and the carbon cost calculation includes the following formula (8): In the formula, C carbon is the carbon cost, P tax is the carbon tax price, P trade,t is the carbon trading price at time t, E direct,t is the carbon trading volume of the enterprise at time t, E trade,t is the carbon trading volume of the enterprise at time t; The step 2-2) of information disclosure index specifically includes determining weights by using the hierarchical analysis method and calculating by using the fuzzy comprehensive evaluation method as shown in formula (9): In the formula, I is the information disclosure index, q is the number of information disclosure items, and S l is the fuzzy evaluation value of information disclosure item l, ω l,updated is the weight of information disclosure item l determined by AHP, μ(S l ) is the fuzzy comprehensive evaluation function; The step 2-3) of carbon footprint monitoring includes using a weighted carbon intensity model to calculate a carbon footprint monitoring index, specifically the following formula (10): In the formula, F is the carbon footprint monitoring index, D is the number of monitoring dimensions, ω d is the weight of dimension d, E d is the carbon emission of dimension d, R d is the relevant economic indicator of dimension d; The step 2-4) carbon audit feedback includes calculating the carbon audit feedback index using a dynamic adjustment model, specifically as follows: In the formula, A is the carbon audit feedback index, N is the number of audit cycles, and ΔE n is the change in carbon emissions in the nth period, E base,n is the baseline carbon emissions in the nth period, λ n is the adjustment coefficient of the nth cycle, λ n Reflect the importance of audit feedback.

7. The method for evaluating the carbon performance of industrial enterprises based on regulatory data according to claim 1 is characterized by: The step 3) constructs a carbon performance evaluation model, specifically including: step 3-1) constructs a performance scoring model; step 3-2) standardization processing; step 3-3) weight determination; step 3-4) uses a comprehensive scoring formula to obtain a comprehensive score; the step 4) intelligent carbon decision support specifically includes: step 4-1) optimization model; step 4-2) decision variables; step 4-3) constraints.

8. The method for evaluating the carbon performance of industrial enterprises based on regulatory data according to claim 7 is characterized by: The step 3-1) constructs a performance scoring model, specifically including adopting a multi-index comprehensive evaluation model and combining principal component analysis to perform dimensionality reduction processing to obtain a comprehensive carbon performance score, specifically as follows: In the formula, P is the comprehensive carbon performance score, K is the number of main components, φ k is the contribution rate of the principal component k, Z k is the score of principal component k; Step 3-2) Standardization processing includes standardizing each index using the range standardization method to eliminate the dimension effect, as shown in the following formula (13): Among them, X represents any indicator, and after standardization, we get X', X min Represents the minimum value of any indicator, X max Represents the maximum value of any indicator; The step 3-3) weight determination specifically includes using the entropy weight method to determine the weight of each indicator, and the calculation process includes the following: Step 3-3-1) Use formula (14) to calculate the weight of the indicator: In the formula, X' ij is the standardized value of the i-th sample on the j-th index, N is the total number of samples, p ij Indicates the relative weight of a sample on a certain indicator; Step 3-3-2) Calculate the entropy value using formula (15)-formula (16): In the formula, H j is the entropy value; Step 3-3-3) Calculate the weight using formula (17): In the formula, M is the total number of indicators, ω j is the weight; The step 3-4) uses a comprehensive scoring formula to obtain a comprehensive score, which specifically includes combining the standardized indicators with the weights to obtain a comprehensive score, specifically as follows: In the formula, P i is the comprehensive carbon performance score of the i-th sample, ω j is the weight, M is the total number of indicators, X' ij is the value of the i-th sample on the j-th standardized index.

9. The method for evaluating the carbon performance of industrial enterprises based on regulatory data according to claim 7 is characterized by: The step 4-1) optimizes the model, specifically including using a multi-objective optimization model to balance the carbon performance score P and the enterprise economic benefit E econmic , specifically as follows: Formula (19) - Formula (22): Subject to E≤E target Formula (21): In the formula, L is the number of carbon emission reduction measures, R l is the contribution rate of carbon emission reduction measures l to economic benefits, Q l To strengthen the implementation of carbon reduction measures, C l is the cost of carbon reduction measure l, B is the carbon management budget, E target is the carbon emission target value, and E is the carbon emission of the enterprise.

10. The method for evaluating the carbon performance of industrial enterprises based on regulatory data according to claim 7 is characterized by: The decision variables in step 4-2) include the implementation strength of each carbon reduction measure Q = {Q1, Q2, ..., Q L }, and carbon trading volume E trade : Q={Q1,Q2,...,Q L },E trade ≥0; In addition to the constraints in the above optimization model, the constraints in step 4-3) also need to consider the following constraints: where ΔE l is the emission reduction of carbon reduction measure l, E base is the carbon emissions in the base period, E trade is the carbon trading volume, E is the carbon emissions of the enterprise, Q l To strengthen the implementation of carbon emission reduction measures.

Images