An automated carbon accounting and intelligent correction method based on multi-source data fusion

Abstract

The application discloses an automatic carbon accounting and intelligent correction method based on multi-source data fusion, relates to the technical field of carbon emission, and solves the technical problem that it is difficult to set corresponding confidence evaluation models to dynamically adjust the confidence of different types of data according to different data types, which affects the accuracy of the final carbon emission data of an enterprise; production energy consumption data and supply chain energy consumption data are acquired through an enterprise ERP system; the data are grouped, so that different types of energy consumption data can be classified and processed, thereby improving the efficiency and accuracy of data processing; through the confidence evaluation model, the reliability and accuracy of different types of energy consumption data can be quantified, the data of each group are dynamically weighted and fused, and errors and uncertainties are reduced; the confidence evaluation model can evaluate the confidence of data when the data changes over time or conditions, so that the data processing strategy can be dynamically adjusted, and the continuous accuracy of carbon emission calculation is ensured.

Description

Technical Field

[0001] This invention belongs to the field of carbon emissions, specifically an automated carbon accounting and intelligent correction method based on multi-source data fusion. Background Technology

[0002] Carbon emissions refer to the process by which carbon dioxide (CO2) and other greenhouse gases enter the atmosphere during human activities. Traditional carbon accounting methods often rely on a single data source and manual operation, which not only leads to low accounting efficiency but may also affect the accuracy of the results due to incomplete or erroneous data. To overcome these limitations, automated carbon accounting is an automated process that uses advanced data processing technologies and algorithms to quantify and assess carbon emissions. It involves measuring and calculating greenhouse gas emissions from various human activities, such as industrial production, transportation, and agricultural activities. Automated carbon accounting can significantly improve accounting efficiency, reduce errors from manual operation, and enable real-time monitoring of carbon emissions.

[0003] Most carbon accounting and intelligent correction solutions perform unified calculation and evaluation of carbon emission data, making it difficult to set up corresponding confidence assessment models to dynamically adjust the confidence level of different types of data according to different data types, which affects the accuracy of the final carbon emission data of enterprises. Summary of the Invention

[0004] This invention aims to solve at least one of the technical problems existing in the prior art; to this end, this invention proposes an automated carbon accounting and intelligent correction method based on multi-source data fusion, which is used to solve the technical problem that it is difficult to set corresponding confidence assessment models according to different data types to dynamically adjust the confidence of different types of data, thus affecting the accuracy of the final carbon emission data of enterprises.

[0005] To address the aforementioned problems, a first aspect of the present invention provides an automated carbon accounting and intelligent correction method based on multi-source data fusion, comprising the following steps:

[0006] Obtain production energy consumption data and supply chain energy consumption data through the enterprise's ERP system;

[0007] Deploy edge data processing nodes in enterprise areas, outsourced processing areas and logistics areas to collect temperature data and greenhouse gas emission data, thermal infrared image data and regional satellite remote sensing carbon flux images of the entire supply chain, and store the collected data and the historical data of the nodes in the corresponding edge data processing nodes.

[0008] Production energy consumption data and supply chain energy consumption data are grouped, a confidence assessment model is constructed, and dynamic weight fusion is performed on the data of each group to calculate the total carbon emissions of the entire supply chain. The calculation formula is as follows:

[0009]

[0010] Among them, T carbon The total carbon emissions across the entire supply chain, m represents the total number of energy consumption data groups, and E represents the total carbon emissions across the entire supply chain. j Let Cenergy represent the energy consumption of the j-th data group within the energy consumption data grouping. 1j Let e ​​be the confidence level of the j-th group of data in the energy consumption data grouping. j Cenergy represents the carbon emission factor of the j-th data group in the energy consumption data group; n represents the total number of emission data groups. 2i E represents the confidence level of the i-th group of data in the emission data grouping. i P represents the greenhouse gas emissions of the i-th data group in the emissions data group; g represents the total number of equipment operation data groups; P represents the total number of data groups in the equipment operation data group. k Let t be the power value of the k-th device data in the device operation data group. k Cenergy represents the runtime of the k-th device data in the device runtime data group. 3k e represents the confidence level of the k-th device data in the device operation data group. k The carbon emission factor of the kth device data in the device operation data group;

[0011] Edge data processing nodes divide the collected data into structured data and unstructured data, evaluate the data similarity between the collected structured data and the historical structured data of different nodes, and establish data association links for data with similarity higher than a threshold.

[0012] By weighting and fusing structured and unstructured data, a comparison value of total carbon emissions is calculated.

[0013] An industry accounting factor library is built in the cloud. Based on the comparison data of total carbon emissions and data association links, the data on total carbon emissions of the entire supply chain is corrected and uploaded to the blockchain.

[0014] Furthermore, in one example of the above aspects, the production energy consumption data includes: production equipment operation ledgers, factory energy consumption ledgers, raw material consumption ledgers, and factory greenhouse gas emission ledgers; the supply chain energy consumption data includes: transportation energy consumption ledgers, greenhouse gas emission data of outsourced factories, and energy consumption ledgers of distributors.

[0015] Furthermore, collecting temperature data and greenhouse gas emission data across the entire supply chain, thermal infrared image data, and regional satellite remote sensing carbon flux imagery includes the following steps:

[0016] By deploying IoT sensor clusters through edge data processing nodes, temperature data and greenhouse gas emissions from enterprise factories and outsourced factories are collected; by integrating satellite remote sensing data with UAV infrared imaging systems, thermal infrared image data of factories, transportation logistics and outsourced processing are collected, as well as satellite remote sensing carbon flux images of factory and outsourced processing areas.

[0017] Furthermore, in one example of the above aspects, the production energy consumption data and supply chain energy consumption data are grouped, and a confidence assessment model is constructed, including the following steps:

[0018] Production energy consumption data and supply chain energy consumption data are grouped into data groups, including energy consumption data groups, emission data groups, and equipment operation data groups. Specifically: the energy consumption data groups are further grouped according to factory energy consumption ledgers, raw material consumption ledgers, transportation energy consumption ledgers, and distributor energy consumption ledgers; the emission data groups are further grouped according to factory greenhouse gas emission ledgers and outsourced factory greenhouse gas emission data; and the equipment operation data groups are further grouped according to production equipment operation ledgers.

[0019] After the data is grouped, the data within the detection time interval of each data group is filtered to construct confidence models for energy consumption data group, emission data group, and equipment operation data group.

[0020] Furthermore, in one example of the above aspects, the energy consumption data grouping confidence model is:

[0021]

[0022] Where Cenergy1 is the confidence level of the energy consumption data group, σ1 is the variance of the corresponding data of the energy consumption data group within the detection time interval, λ is the time decay factor, t is the data detection time interval, ΔX1 is the deviation of the corresponding data of the energy consumption data group from the mean within the detection time interval, and X base1 The mean of the corresponding data for grouping energy consumption data within the detection time interval;

[0023] The confidence model for the grouped emission data is as follows:

[0024]

[0025] Where Cenergy2 is the confidence level of the emission data group, σ2 is the variance of the corresponding data for the emission data group within the detection time interval, ΔX2 is the deviation of the corresponding data for the emission data group from the mean within the detection time interval, and X base2Q1 represents the mean of the data for each emission data group within the detection time interval, and Q2 represents the proportion of missing data for each emission data group within the detection time interval.

[0026] The confidence model for the grouped equipment operation data is as follows:

[0027]

[0028] Where Cenergy3 is the confidence level of the device operation data group, σ3 is the variance of the corresponding data of the device operation data group within the detection time interval, and Q3 is the missing proportion of the device operation data group within the detection time interval.

[0029] Furthermore, in one example of the above aspects, the edge data processing node divides the collected data into structured data and unstructured data, evaluates the data similarity between the collected structured data and the historical structured data of different nodes, and establishes data association links for data with similarity higher than a threshold, including the following steps:

[0030] The edge data processing node divides real-time collected data and historical data into structured data and unstructured data. The structured data includes temperature data and greenhouse gas emissions from enterprise factories and outsourced factories. The unstructured data includes thermal infrared image data of factories, transportation logistics and outsourced processing, as well as satellite remote sensing carbon flux images of factory and outsourced processing areas.

[0031] All edge data processing nodes send the access links of structured data to the cloud data processing module. The cloud data processing module calculates the cosine similarity between the real-time collected structured data and other structured data, and evaluates the data similarity between the collected structured data and the historical structured data of different nodes.

[0032] For data with similarity exceeding a threshold, establish a link between the collected structured data and the corresponding structured data.

[0033] Furthermore, in one example of the above aspects, calculating a comparison of total carbon emissions by weighting and fusing structured and unstructured data includes the following steps:

[0034] Acquire unstructured historical data, as well as carbon emission data of factories, transportation logistics and outsourced processing at the corresponding time, and add carbon emission data as tags for the corresponding unstructured data;

[0035] Predicting carbon emissions by training a deep learning model on unstructured data with added labels includes: training a deep learning model, predicting carbon emissions data based on the collected unstructured data, and obtaining the predicted carbon emissions value of the unstructured data.

[0036] The coverage of structured data from statistical edge data processing nodes on the industry chain;

[0037] Greenhouse gas emissions were revised based on temperature data and greenhouse gas emissions from both the factory and outsourced facilities:

[0038]

[0039] Among them, E corr For the corrected greenhouse gas emissions, E raw For greenhouse gas emissions, T act For the actual measured temperature, T std Standard operating temperature;

[0040] Calculate the carbon emissions from the structured data based on the corrected greenhouse gas emissions and the corresponding GWP values ​​for the greenhouse gases:

[0041] T co =∑(E corr ×E GWP )

[0042] Among them, T co For carbon emissions from structured data, E GWP The GWP value corresponding to greenhouse gases;

[0043] Based on the carbon emissions data from structured data, and combined with the coverage of the industrial chain by the structured data, a comparative value of total carbon emissions is calculated:

[0044] T0=p0*T co +(1-p0)T eo

[0045] Where T0 represents the total carbon emissions comparison value, p0 represents the coverage rate of structured data to the industrial chain, and T eo This is a predicted carbon emission value based on unstructured data.

[0046] Furthermore, in one example of the above aspects, building an industry accounting factor library in the cloud includes the following steps:

[0047] Regularly acquire industry classification data, energy type data, and energy type emission factor data;

[0048] Based on industry classification data, energy type data, and emission factor data of energy types, an industry accounting factor library is built in the cloud, which includes a main table, sub-tables, and related tables. The main table contains industry classification data, the sub-tables contain energy type data, and the related tables contain the correspondence between emission factor data and energy types.

[0049] Regularly update the data in the main table, child tables, and related tables.

[0050] Furthermore, in one example of the above, the data on total carbon emissions across the entire supply chain is corrected based on comparative data of total carbon emissions and data linkages, including the following steps:

[0051] By using the industry accounting factor library, we can check whether the emission factors used to calculate the total carbon emissions of the entire supply chain have been updated.

[0052] If the emission factors used in the calculation are not updated, update the corresponding emission factors and recalculate the total carbon emissions of the entire supply chain; otherwise, do nothing.

[0053] Based on the data links, obtain the total carbon emission comparison data of the related data as an auxiliary total carbon emission comparison value;

[0054] The total carbon emissions data for the entire supply chain are revised based on the total carbon emissions comparison data and the auxiliary total carbon emissions comparison data.

[0055] Furthermore, in one example of the above, the total carbon emissions data for the entire supply chain is corrected based on the total carbon emissions comparison data and the auxiliary total carbon emissions comparison data, using the following formula:

[0056]

[0057] Among them, T ne For the revised total carbon emissions across the entire supply chain, T carbon T0 represents the total carbon emissions across the entire supply chain, while T0 represents the total carbon emissions comparison value. 02 To assist in comparing total carbon emissions.

[0058] Compared with the prior art, the beneficial effects of the present invention are:

[0059] I. This invention improves the efficiency and accuracy of data processing by grouping data, facilitating the classification and processing of different types of energy consumption data. Through a confidence assessment model, the reliability and accuracy of different types of energy consumption data can be quantified, and dynamic weight fusion of data in each group reduces errors and uncertainties.

[0060] Second, the confidence assessment model constructed in this invention can evaluate the confidence level of data as it changes over time or under different conditions, thereby dynamically adjusting the data processing strategy and ensuring the continuous accuracy of carbon emission calculations. Dynamic weight fusion facilitates the automatic adjustment of the weights of each data group based on its importance and reliability, thereby optimizing the data fusion effect and improving the accuracy of carbon emission calculations.

[0061] Third, this invention, through data association and linking, facilitates the evaluation and verification of the accuracy of newly collected data using historical data, providing a reliable guarantee for subsequent data analysis and application. The establishment of data association and linking allows data from different nodes to be easily integrated for analysis, forming a more complete and comprehensive dataset, providing data support for subsequent data analysis. Attached Figure Description

[0062] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0063] Figure 1 A schematic diagram of the method flow for an automated carbon accounting and intelligent correction method based on multi-source data fusion provided in an embodiment of the present invention;

[0064] Figure 2 A schematic diagram of data grouping for an automated carbon accounting and intelligent correction method based on multi-source data fusion provided in an embodiment of the present invention;

[0065] Figure 3 This diagram illustrates a comparison of carbon emission data accuracy using an automated carbon accounting and intelligent correction method based on multi-source data fusion, as provided in an embodiment of the present invention. Detailed Implementation

[0066] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0067] Please see Figures 1 to 3 The first aspect of this invention provides an automated carbon accounting and intelligent correction method based on multi-source data fusion, comprising the following steps:

[0068] Obtain production energy consumption data and supply chain energy consumption data through the enterprise's ERP system;

[0069] Deploy edge data processing nodes in enterprise areas, outsourced processing areas and logistics areas to collect temperature data and greenhouse gas emission data, thermal infrared image data and regional satellite remote sensing carbon flux images of the entire supply chain, and store the collected data and the historical data of the nodes in the corresponding edge data processing nodes.

[0070] The production energy consumption data and supply chain energy consumption data are grouped, a confidence assessment model is constructed, and the data of each group are dynamically weighted and fused to calculate the total carbon emissions of the entire supply chain.

[0071] Edge data processing nodes divide the collected data into structured data and unstructured data, evaluate the data similarity between the collected structured data and the historical structured data of different nodes, and establish data association links for data with similarity higher than a threshold.

[0072] By weighting and fusing structured and unstructured data, a comparison value of total carbon emissions is calculated.

[0073] An industry accounting factor library is built in the cloud. Based on the comparison data of total carbon emissions and data association links, the data on total carbon emissions of the entire supply chain is corrected and uploaded to the blockchain.

[0074] Specifically, in this embodiment, production energy consumption data and supply chain energy consumption data are obtained through the enterprise's ERP system. The manufacturing enterprise calculates carbon emissions from its own factories, core outsourced component processing plants, and multiple logistics trunk lines covering the whole country and records the data in the ledger.

[0075] ERP system integration includes:

[0076] Production energy consumption: The MES system acquires real-time energy consumption data for processes such as compressor welding and injection molding;

[0077] Supply chain energy consumption: Synchronize CNC machining data from outsourced factories and refueling records of logistics fleets via EDI interface;

[0078] Data storage: Outsourced processing data is encrypted and stored using the Hyperledger Fabric framework to ensure that the data cannot be tampered with.

[0079] Edge data processing nodes are deployed in enterprise areas, outsourced processing areas, and logistics areas, as shown in Table 1 below:

[0080]

[0081] Table 1: Edge Node Deployment Table

[0082] Edge storage strategy:

[0083] High-frequency data (>1Hz): Data cached on the local SSD for the last 24 hours;

[0084] Low-frequency data: Compressed storage (Zstandard algorithm, compression ratio ≥5:1) and then synchronized to the regional central node.

[0085] Production energy consumption data and supply chain energy consumption data are grouped, and a confidence assessment model is constructed. Dynamic weight fusion is applied to the data in each group to calculate the total carbon emissions of the entire supply chain. Grouping the data facilitates the classification and processing of different types of energy consumption data, thereby improving the efficiency and accuracy of data processing. The confidence assessment model facilitates the quantification of the reliability and accuracy of different types of energy consumption data, and the dynamic weight fusion of data in each group reduces errors and uncertainties. Furthermore, the confidence assessment model can evaluate the confidence level of data as it changes over time or under different conditions, thereby dynamically adjusting the data processing strategy to ensure the continuous accuracy of carbon emission calculations. Dynamic weight fusion automatically adjusts the weights of each data group based on its importance and reliability, thus optimizing the data fusion effect and improving the accuracy of carbon emission calculations.

[0086] Edge data processing nodes categorize collected data into structured and unstructured data. They assess the similarity between the collected structured data and historical structured data from different nodes, establishing data association links for data with similarity exceeding a threshold. By comparing the similarity between newly collected structured data and historical data, highly similar data can be quickly identified. Establishing association links for highly similar data facilitates subsequent analysis and evaluation of data accuracy. Furthermore, data association links allow for the verification of the accuracy of newly collected data through historical data evaluation, providing reliable assurance for subsequent data analysis and applications. The establishment of data association links enables the easy fusion of data from different nodes for analysis, forming a more complete and comprehensive dataset, providing data support for subsequent data analysis.

[0087] By weighting and fusing structured and unstructured data, a comparative value for total carbon emissions is calculated. An industry accounting factor library is built in the cloud. Based on the carbon emission comparison data and data links, the total carbon emissions data for the entire supply chain is corrected and uploaded to the blockchain. The weighted fusion of structured data, such as energy consumption records and production data, and unstructured data facilitates the comprehensive consideration of information from multiple data sources, thereby improving the accuracy of carbon emission calculations. Furthermore, the industry accounting factor library contains carbon emission-related factors and parameters for various industries. Regularly updating the factor library data for comparison and correction of total carbon emissions further ensures the accuracy of the calculation results. Uploading the corrected carbon emission data to the blockchain ensures the immutability and transparency of the data, providing strong support for its subsequent use and supervision.

[0088] In one embodiment of the present invention, the production energy consumption data includes: production equipment operation ledger, factory energy consumption ledger, raw material consumption ledger and factory greenhouse gas emission ledger; the supply chain energy consumption data includes: transportation energy consumption ledger, outsourced factory greenhouse gas emission data and distributor energy consumption ledger.

[0089] Furthermore, in one example of the above aspects, collecting temperature data and greenhouse gas emission data across the entire supply chain, thermal infrared image data, and regional satellite remote sensing carbon flux imagery includes the following steps:

[0090] By deploying IoT sensor clusters through edge data processing nodes, temperature data and greenhouse gas emissions from enterprise factories and outsourced factories are collected; by integrating satellite remote sensing data with UAV infrared imaging systems, thermal infrared image data of factories, transportation logistics and outsourced processing are collected, as well as satellite remote sensing carbon flux images of factory and outsourced processing areas.

[0091] In one embodiment of the present invention, production energy consumption data and supply chain energy consumption data are grouped to construct a confidence assessment model, including the following steps:

[0092] The production energy consumption data and supply chain energy consumption data are grouped as follows:

[0093] Energy consumption data is grouped into: factory energy consumption ledger, raw material consumption ledger, transportation energy consumption ledger, and retailer energy consumption ledger.

[0094] Emissions data grouping: Greenhouse gas emissions ledger of factories, greenhouse gas emissions data of outsourced factories;

[0095] Equipment operation data grouping: Production equipment operation ledger;

[0096] Filter the data within the most recent detection time interval for each data group;

[0097] For grouping energy consumption data, establish confidence model parameters:

[0098]

[0099] Where Cenergy1 is the confidence level of the energy consumption data group, σ1 is the variance of the corresponding data of the energy consumption data group within the detection time interval, λ is the time decay factor, t is the data detection time interval, ΔX1 is the deviation of the corresponding data of the energy consumption data group from the mean within the detection time interval, and X base1 The mean of the corresponding data for grouping energy consumption data within the detection time interval;

[0100] For the emission data grouping, establish the confidence model parameters:

[0101]

[0102] Where Cenergy2 is the confidence level of the emission data group, σ2 is the variance of the corresponding data for the emission data group within the detection time interval, ΔX2 is the deviation of the corresponding data for the emission data group from the mean within the detection time interval, and X base2 Q1 represents the mean of the data for each emission data group within the detection time interval, and Q2 represents the proportion of missing data for each emission data group within the detection time interval.

[0103] For grouping equipment operation data, establish confidence model parameters:

[0104]

[0105] Where Cenergy3 is the confidence level of the device operation data group, σ3 is the variance of the corresponding data of the device operation data group within the detection time interval, and Q3 is the missing proportion of the device operation data group within the detection time interval.

[0106] In one embodiment of the present invention, dynamic weighted fusion of data from each group is performed to calculate the total carbon emissions of the entire supply chain using the following formula:

[0107]

[0108] Among them, T carbon The total carbon emissions across the entire supply chain, m represents the total number of energy consumption data groups, and E represents the total carbon emissions across the entire supply chain. j Let Cenergy represent the energy consumption of the j-th data group within the energy consumption data grouping. 1j Let e ​​be the confidence level of the j-th group of data in the energy consumption data grouping. j Cenergy represents the carbon emission factor of the j-th data group in the energy consumption data group; n represents the total number of emission data groups. 2i E represents the confidence level of the i-th group of data in the emission data grouping. i P represents the greenhouse gas emissions of the i-th data group in the emissions data group; g represents the total number of equipment operation data groups; P represents the total number of data groups in the equipment operation data group. k Let t be the power value of the k-th device data in the device operation data group. k Cenergy represents the runtime of the k-th device data in the device runtime data group. 3k e represents the confidence level of the k-th device data in the device operation data group. k The carbon emission factor of the kth device data in the device operation data group.

[0109] Specifically, in this embodiment, the carbon emission data certified by power battery companies on the blockchain is analyzed and used as standard data. Data obtained from the ERP systems of companies at different production stages is used to directly calculate carbon emissions by statistically analyzing the carbon emission data of each production stage. Simultaneously, the carbon emissions are calculated using the fused data according to the above-described scheme of this embodiment. The accuracy of carbon emission calculations using different methods is compared by calculating the absolute value of the difference between the carbon emissions obtained from the two methods and the standard data, and the ratio of this difference to the standard data. See Table 2 and... Figure 3 It can be seen that by implementing the above scheme, the carbon emission data obtained by calculating carbon emissions through the fused data is more accurate.

[0110]

[0111] Table 2: Comparison of the Accuracy of Carbon Emission Data

[0112] In one embodiment of the present invention, the edge data processing node divides the collected data into structured data and unstructured data, evaluates the data similarity between the collected structured data and the historical structured data of different nodes, and establishes data association links for data with similarity higher than a threshold, including the following steps:

[0113] Edge data processing nodes categorize real-time collected data and historical data into structured data and unstructured data;

[0114] All edge data processing nodes send the access links of structured data to the cloud data processing module. The cloud data processing module calculates the cosine similarity between the real-time collected structured data and other structured data, and evaluates the data similarity between the collected structured data and the historical structured data of different nodes.

[0115] For data with similarity exceeding a threshold, establish a link between the collected structured data and the corresponding structured data.

[0116] In one embodiment of the present invention, the comparison value of total carbon emissions is calculated by weighting and fusing structured and unstructured data, including the following steps:

[0117] The structured data collected by the edge data processing nodes includes temperature data and greenhouse gas emissions from enterprise factories and outsourced factories; the unstructured data collected includes thermal infrared image data of factories, transportation logistics and outsourced processing, as well as satellite remote sensing carbon flux images of factory and outsourced processing areas.

[0118] Acquire unstructured historical data, as well as carbon emission data of factories, transportation logistics and outsourced processing at the corresponding time, and use the carbon emission data as the label for the corresponding unstructured data;

[0119] Train a deep learning model using labeled data to predict carbon emissions;

[0120] Train a deep learning model to predict carbon emissions based on the collected unstructured data, and obtain the predicted carbon emissions value of unstructured data.

[0121] The coverage of structured data from statistical edge data processing nodes with data from factories, transportation logistics, and outsourced processing areas;

[0122] Greenhouse gas emissions were revised based on temperature data and greenhouse gas emissions from both the factory and outsourced facilities: , of which E corr For the corrected greenhouse gas emissions, E raw For greenhouse gas emissions, T act For the actual measured temperature, T std Standard operating temperature;

[0123] Calculate the carbon emissions of structured data: T co =∑(E corr ×E GWP ), where T co For carbon emissions from structured data, E GWP The GWP value corresponding to greenhouse gases;

[0124] Calculate the total carbon emissions comparison: T0 = p0 * T co +(1-p0)T eo Where T0 is the total carbon emission comparison value, p0 is the coverage rate of structured data to the industrial chain, and T eo This is a predicted value for carbon emissions based on unstructured data.

[0125] Specifically, in this embodiment, the GWP for the corresponding time range is obtained through an authoritative organization such as the IPCC.

[0126] ;

[0127] ;

[0128] ;

[0129] total: ;

[0130] Meanwhile, in this embodiment, the predicted carbon emissions from unstructured data are 1200 tons, while the structured data covers 70% of the industrial chain.

[0131] Calculate the total carbon emissions comparison: In this embodiment, the total carbon emission comparison value is calculated as follows: .

[0132] In one embodiment of the present invention, building an industry accounting factor library in the cloud includes the following steps:

[0133] Regularly acquire industry classification data, energy type data, and energy type emission factor data;

[0134] Build an industry accounting factor library in the cloud, including:

[0135] Main table: Industry classification data;

[0136] Sub-table: Energy type data;

[0137] Correlation table: Correspondence between emission factors and energy types;

[0138] Update the data in the data tables regularly.

[0139] In one embodiment of the present invention, the data on total carbon emissions across the entire supply chain is corrected based on carbon emission comparison data and data association links, including the following steps:

[0140] By using the industry accounting factor library, we can check whether the emission factors used to calculate the total carbon emissions of the entire supply chain are correct.

[0141] If the emission factors used in the calculation are not updated in a timely manner, update the corresponding emission factors and recalculate the total carbon emissions of the entire supply chain; otherwise, there is no need to recalculate.

[0142] Based on the data links, obtain the total carbon emission comparison data of the related data as an auxiliary total carbon emission comparison value;

[0143] The total carbon emissions data for the entire supply chain are revised based on the total carbon emissions comparison data and the auxiliary total carbon emissions comparison data.

[0144] In one embodiment of the present invention, the total carbon emissions data of the entire supply chain is corrected based on the total carbon emissions comparison data and the auxiliary total carbon emissions comparison data, using the following formula:

[0145]

[0146] Among them, T ne For the revised total carbon emissions across the entire supply chain, T carbon T0 represents the total carbon emissions across the entire supply chain, while T0 represents the total carbon emissions comparison value. 02 To assist in comparing total carbon emissions.

[0147] Specifically, in this embodiment, the corrected total carbon emissions of the entire supply chain are stored in the blockchain through a data on-chain mechanism. The data on-chain mechanism is as follows: edge nodes generate Merkle root hashes; cloud verifies data consistency; and the corrected emission values ​​are written through the Hyperledger Fabric channel.

[0148] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. An automated carbon accounting and intelligent correction method based on multi-source data fusion, characterized in that, Includes the following steps: Obtain production energy consumption data and supply chain energy consumption data through the enterprise's ERP system; Deploy edge data processing nodes in enterprise areas, outsourced processing areas and logistics areas to collect temperature data and greenhouse gas emission data, thermal infrared image data and regional satellite remote sensing carbon flux images of the entire supply chain, and store the collected data and the historical data of the nodes in the corresponding edge data processing nodes. Production energy consumption data and supply chain energy consumption data are grouped, a confidence assessment model is constructed, and dynamic weight fusion is performed on the data of each group to calculate the total carbon emissions of the entire supply chain. The calculation formula is as follows: , among which, T carbon The total carbon emissions across the entire supply chain, m represents the total number of energy consumption data groups, and E represents the total carbon emissions across the entire supply chain. j Let Cenergy represent the energy consumption of the j-th data group within the energy consumption data grouping. 1j Let e ​​be the confidence level of the j-th group of data in the energy consumption data grouping. j Cenergy represents the carbon emission factor of the j-th data group in the energy consumption data group; n represents the total number of emission data groups. 2i E represents the confidence level of the i-th group of data in the emission data grouping. i P represents the greenhouse gas emissions of the i-th data group in the emissions data group; g represents the total number of equipment operation data groups; P represents the total number of data groups in the equipment operation data group. k Let t be the power value of the k-th device data in the device operation data group. k Cenergy represents the runtime of the k-th device data in the device runtime data group. 3k e represents the confidence level of the k-th device data in the device operation data group. k The carbon emission factor of the kth device data in the device operation data group; The process of grouping production energy consumption data and supply chain energy consumption data and constructing a confidence assessment model includes the following steps: Production energy consumption data and supply chain energy consumption data are grouped into data groups, including energy consumption data groups, emission data groups, and equipment operation data groups. Specifically: the energy consumption data groups are further grouped according to factory energy consumption ledgers, raw material consumption ledgers, transportation energy consumption ledgers, and distributor energy consumption ledgers; the emission data groups are further grouped according to factory greenhouse gas emission ledgers and outsourced factory greenhouse gas emission data; and the equipment operation data groups are further grouped according to production equipment operation ledgers. After the data is grouped, the data within the detection time interval of each data group is filtered, and confidence models for energy consumption data group, emission data group, and equipment operation data group are constructed. The confidence model for the energy consumption data grouping is as follows: ； Where Cenergy1 is the confidence level of the energy consumption data group, σ1 is the variance of the corresponding data of the energy consumption data group within the detection time interval, λ is the time decay factor, t is the data detection time interval, ΔX1 is the deviation of the corresponding data of the energy consumption data group from the mean within the detection time interval, and X base1 The mean of the corresponding data for grouping energy consumption data within the detection time interval; The confidence model for the grouped emission data is as follows: ； Where Cenergy2 is the confidence level of the emission data group, σ2 is the variance of the corresponding data for the emission data group within the detection time interval, ΔX2 is the deviation of the corresponding data for the emission data group from the mean within the detection time interval, and X base2 Q1 represents the mean of the data for each emission data group within the detection time interval, and Q2 represents the proportion of missing data for each emission data group within the detection time interval. The confidence model for the grouped equipment operation data is as follows: ； Wherein, Cenergy3 is the confidence level of the device operation data group, σ3 is the variance of the corresponding data of the device operation data group within the detection time interval, and Q3 is the missing proportion of the device operation data group within the detection time interval; Edge data processing nodes divide the collected data into structured data and unstructured data, evaluate the data similarity between the collected structured data and the historical structured data of different nodes, and establish data association links for data with similarity higher than a threshold. By weighting and fusing structured and unstructured data, a comparison value of total carbon emissions is calculated. An industry accounting factor library is built in the cloud. Based on the comparison data of total carbon emissions and data association links, the data on total carbon emissions of the entire supply chain is corrected and uploaded to the blockchain.

2. The automated carbon accounting and intelligent correction method based on multi-source data fusion according to claim 1, characterized in that, The production energy consumption data includes: production equipment operation ledger, factory energy consumption ledger, raw material consumption ledger, and factory greenhouse gas emission ledger; the supply chain energy consumption data includes: transportation energy consumption ledger, greenhouse gas emission data of outsourced factories, and energy consumption ledger of distributors.

3. The automated carbon accounting and intelligent correction method based on multi-source data fusion according to claim 1, characterized in that, Collecting temperature and greenhouse gas emission data, thermal infrared image data, and regional satellite remote sensing carbon flux images across the entire supply chain includes the following steps: By deploying IoT sensor clusters through edge data processing nodes, temperature data and greenhouse gas emissions from enterprise factories and outsourced factories are collected; by integrating satellite remote sensing data with UAV infrared imaging systems, thermal infrared image data of factories, transportation logistics and outsourced processing are collected, as well as satellite remote sensing carbon flux images of factory and outsourced processing areas.

4. The automated carbon accounting and intelligent correction method based on multi-source data fusion according to claim 1, characterized in that, Edge data processing nodes categorize collected data into structured and unstructured data, evaluate the similarity between the collected structured data and historical structured data from different nodes, and establish data association links for data with similarity exceeding a threshold. This includes the following steps: The edge data processing node divides real-time collected data and historical data into structured data and unstructured data. The structured data includes temperature data and greenhouse gas emissions from enterprise factories and outsourced factories. The unstructured data includes thermal infrared image data of factories, transportation logistics and outsourced processing, as well as satellite remote sensing carbon flux images of factory and outsourced processing areas. All edge data processing nodes send the access links of structured data to the cloud data processing module. The cloud data processing module calculates the cosine similarity between the real-time collected structured data and other structured data, and evaluates the data similarity between the collected structured data and the historical structured data of different nodes. For data with similarity exceeding a threshold, establish a link between the collected structured data and the corresponding structured data.

5. The automated carbon accounting and intelligent correction method based on multi-source data fusion according to claim 4, characterized in that, The total carbon emissions are compared by weighting and fusing structured and unstructured data, including the following steps: Acquire unstructured historical data, as well as carbon emission data of factories, transportation logistics and outsourced processing at the corresponding time, and add carbon emission data as tags for the corresponding unstructured data; Predicting carbon emissions by training a deep learning model on unstructured data with added labels includes: training a deep learning model, predicting carbon emissions data based on the collected unstructured data, and obtaining the predicted carbon emissions value of the unstructured data. The coverage of structured data from statistical edge data processing nodes on the industry chain; Greenhouse gas emissions were revised based on temperature data and greenhouse gas emissions from both the factory and outsourced facilities: E corr =E raw ×[1+0.02×(T act -T std )] Among them, E corr For the corrected greenhouse gas emissions, E raw For greenhouse gas emissions, T act For the actual measured temperature, T std Standard operating temperature; Calculate the carbon emissions from the structured data based on the corrected greenhouse gas emissions and the corresponding GWP values ​​for the greenhouse gases: T co =∑(E corr ×E GWP ) Among them, T co For carbon emissions from structured data, E GWP The GWP value corresponding to greenhouse gases; Based on the carbon emissions data from structured data, and combined with the coverage of the industrial chain by the structured data, a comparative value of total carbon emissions is calculated: T0=p0*T co +(1-p0)T eo Where T0 represents the total carbon emissions comparison value, p0 represents the coverage rate of structured data to the industrial chain, and T eo This is a predicted value for carbon emissions based on unstructured data.

6. The automated carbon accounting and intelligent correction method based on multi-source data fusion according to claim 1, characterized in that, Building an industry accounting factor library in the cloud includes the following steps: Regularly acquire industry classification data, energy type data, and energy type emission factor data; Based on industry classification data, energy type data, and emission factor data of energy types, an industry accounting factor library is built in the cloud, which includes a main table, sub-tables, and related tables. The main table contains industry classification data, the sub-tables contain energy type data, and the related tables contain the correspondence between emission factor data and energy types. Regularly update the data in the main table, child tables, and related tables.

7. The automated carbon accounting and intelligent correction method based on multi-source data fusion according to claim 1, characterized in that, Based on the comparative data of total carbon emissions and data links, the data on total carbon emissions across the entire supply chain is corrected, including the following steps: By using the industry accounting factor library, we can check whether the emission factors used to calculate the total carbon emissions of the entire supply chain have been updated. If the emission factors used in the calculation are not updated, update the corresponding emission factors and recalculate the total carbon emissions of the entire supply chain; otherwise, do nothing. Based on the data links, obtain the total carbon emission comparison data of the related data as an auxiliary total carbon emission comparison value; The total carbon emissions data for the entire supply chain are revised based on the total carbon emissions comparison data and the auxiliary total carbon emissions comparison data.

8. The automated carbon accounting and intelligent correction method based on multi-source data fusion according to claim 7, characterized in that, The total carbon emissions data for the entire supply chain is corrected based on the total carbon emissions comparison data and auxiliary total carbon emissions comparison data using the following formula: ； Among them, T ne For the revised total carbon emissions across the entire supply chain, T carbon T0 represents the total carbon emissions across the entire supply chain, while T0 represents the total carbon emissions comparison value. 02 To assist in comparing total carbon emissions.

Images