Big data-based carbon trading market intelligent analysis system and method

Abstract

The application provides a big data-based carbon transaction market intelligent analysis system and method, relates to the field of carbon asset analysis, and is characterized in that: a blockchain+digital twin MRV credible closed loop is first created, data is unforgeable, traceable and automatically certified; a 'carbon-energy-economy' coupled multi-scale model is constructed, the carbon price prediction accuracy is improved; an all-industry-chain carbon footprint quota allocation is innovated, and an incentive mechanism is combined to stimulate emission reduction power; a federal learning collaborative supervision is first created, millisecond-level identification of violations, personalized early warning and CBAM compliance deduction are realized; a carbon asset digital twin and cross-chain transfer technology are first created, the linkage of 'carbon data-carbon assets-carbon finance' is promoted, and market liquidity is improved.

Description

Technical Field

[0001] This invention relates to the field of carbon asset analysis, and in particular to an intelligent analysis system and method for carbon trading markets based on big data. Background Technology

[0002] Currently, carbon trading has become a core means to promote corporate emission reduction and optimize resource allocation. Its large-scale and international development has created an urgent need for intelligent analysis technology. The current carbon trading market suffers from problems such as incomplete data collection, insufficient credibility of the MRV system, low accuracy of carbon price prediction, imprecise quota allocation, and weak cross-entity collaborative supervision. Furthermore, the high barriers to carbon asset transfer and insufficient linkage with carbon finance make it difficult to support efficient market operation and cross-border compliance requirements.

[0003] Therefore, there is a need to provide a big data-based intelligent analysis system and method for the carbon trading market to solve the above problems. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent analysis system and method for the carbon trading market based on big data. It pioneers a blockchain + digital twin MRV trusted closed loop to achieve data immutability, traceability, and automatic verification; constructs a multi-scale coupled "carbon-energy-economy" model to improve the accuracy of carbon price prediction; innovates the allocation of carbon footprint quotas across the entire industry chain, combining incentive mechanisms to stimulate emission reduction motivation; pioneers federated learning collaborative supervision to achieve millisecond-level identification of violations, personalized early warning, and CBAM compliance simulation; and pioneers carbon asset digital twin and cross-chain transfer technologies to promote the linkage of "carbon data-carbon assets-carbon finance" and improve market liquidity.

[0005] To achieve the above objectives, this invention provides a big data-based intelligent analysis system for the carbon trading market, comprising a terminal access layer, a data acquisition layer, a data preprocessing layer, a big data storage layer, an intelligent analysis layer, an application service layer, a security layer, and a carbon asset transfer layer. These layers are sequentially connected and communicate with each other to collaboratively achieve intelligent analysis, control, and efficient transfer of carbon assets throughout the entire carbon trading market process. The input end of the terminal access layer receives operation instructions from the user end and outputs collection instructions through instruction issuance or result-oriented push; The data acquisition layer receives acquisition instructions from the terminal access layer and external multi-source heterogeneous data, and outputs raw acquisition data or digital twin mirror base data; After receiving the raw data output by the data acquisition layer, the data processing layer completes feature selection through a hybrid algorithm of mRMR and PACF, removes outliers using the Z-score method, and encrypts the data using zero-knowledge proof and homomorphic encryption technology. Simultaneously, it realizes the association mapping between the processed data and the basic data of the digital twin mirror, and outputs standardized cleaned data, feature extracted data, encrypted data, and blockchain evidence hash value. The big data storage layer adopts a distributed storage architecture, which classifies and stores the standardized cleaned data, feature extraction data, encrypted data and blockchain evidence hash values ​​output by the data processing layer in both structured and unstructured categories, and builds a minute-by-minute dynamic update mechanism to synchronously complete data redundancy backup; it outputs the retrieval data required by the intelligent analysis layer, including historical retrieval data or real-time retrieval data. The intelligent analysis layer uses a multi-algorithm fusion model to complete the analysis and calculation based on the input retrieval data and the analysis instructions issued by the terminal access layer. It achieves cross-entity joint modeling through federated learning and outputs carbon asset-related data, including price forecast reports, quota allocation results, risk warning information, and violation identification results. The application service layer outputs business processing results, compliance prompts, and decision-making suggestions; The security layer adopts a four-in-one security mechanism to monitor data transmission and operation behavior in real time, and complete access permission verification, intrusion detection and data encryption protection. The carbon asset-related data output from the intelligent analysis layer and the transfer instructions from the application service layer are input into the carbon asset transfer layer. A carbon asset mirror is constructed using digital twin technology, and an equivalent conversion is completed through a value assessment model. The transfer is realized based on the Polkadot cross-chain protocol, and the carbon asset digital mirror, digital passport, and equivalent transfer results are output.

[0006] Preferably, the multi-source heterogeneous data includes various types of raw monitoring data, reported data, public data and cross-border data, and the data acquisition layer includes an IoT monitoring module, a multi-source data docking module, a data acquisition module for small and medium-sized enterprises and a cross-border data acquisition module; The IoT monitoring module collects data once per minute and transmits it in real time via the MQTT protocol; the multi-source data docking module performs preliminary conversion of heterogeneous data formats and eliminates invalid data sources; the SME data collection module performs data standardization verification; the cross-border data collection module achieves data synchronization with international platforms and completes compliant data format adaptation; and outputs the pre-verified raw heterogeneous data, which is then synchronously transmitted to the data preprocessing layer and the digital twin mirror module.

[0007] Preferably, the data preprocessing layer includes a data cleaning module, a feature selection module, a data encryption module, a blockchain notarization module, and a digital twin mirror module; The data cleaning module identifies and removes outliers using the Z-score method and completes the data using interpolation. The feature selection module balances feature correlation and redundancy using a hybrid algorithm of mRMR and PACF, extracting key time-series and related features. The data encryption module uses the AES encryption algorithm combined with zero-knowledge proofs. The blockchain notarization module generates data hash values, synchronously writes them to the consortium blockchain node, and automatically completes data verification through smart contracts. The digital twin mirroring module completes 3D modeling of the production scenario, realizing real-time mapping between production parameters and carbon emission data. It outputs standardized feature data, encrypted data, on-chain notarization hash values, and digital twin mapping data, and synchronously pushes them to the big data storage layer and intelligent analysis layer.

[0008] Preferably, the big data storage layer adopts a Hadoop distributed architecture, storing structured data in a MySQL database and unstructured data in a MongoDB database, establishing data indexes, constructing a data synchronization and update mechanism, receiving standardized data output from the data preprocessing layer and updating it in real time, synchronously completing data redundancy backup and fault recovery; outputting the retrieval data required by the intelligent analysis layer, including historical retrieval data or real-time retrieval data.

[0009] Preferably, the intelligent analysis layer includes a price prediction module, a quota allocation module, a risk warning module, and a regulatory analysis module; The price prediction module decomposes time-series data of carbon price, energy price, and economic indicators through VMD, identifies coupling correlations through Granger causality test, divides high-frequency or low-frequency components through sequence complexity, and constructs a multi-scale adaptive prediction model of "carbon-energy-economy" coupling by using GARCH-ELM hybrid model, PSO-LSSVM model, PSO-GM model and attention mechanism. This enables accurate prediction of carbon quota prices across multiple scales and factors, and outputs carbon price prediction reports and error assessment results. The quota allocation module introduces the LCA life cycle assessment method to calculate the carbon footprint of the entire industry chain, and uses the improved entropy weight method to calculate the weight of indicators, so as to realize the refined allocation and dynamic adjustment of carbon quotas and output the quota allocation results. The risk warning module achieves cross-entity joint modeling through federated learning, quantifies risk levels using the analytic hierarchy process, customizes personalized warning thresholds by combining attention mechanisms, embeds CBAM rules to complete compliance simulations, and outputs segmented warning information. The regulatory analysis module identifies abnormal transactions and triggers blocking instructions based on a hybrid model of federated learning and CNN-LSTM-Transformer, traces the source of violations through the blockchain, and outputs the violation identification results.

[0010] Preferably, the application service layer includes enterprise clients, regulatory clients, third-party institution clients, public clients, and financial institution clients according to the user client type; it receives the analysis results output by the intelligent analysis layer according to the user client type and completes service matching in combination with user business instructions; It provides personalized emission reduction suggestions and carbon asset solutions to enterprises, regulatory data and violation warnings to regulators, carbon inclusion information to the public, and carbon asset value data to financial institutions; and outputs business processing results, compliance tips, and decision-making suggestions.

[0011] Preferably, the security layer includes a hardware security module, a software security module, a management security module, and an algorithm security module; The hardware security module employs encrypted storage and redundant backup devices; the software security module utilizes access control, intrusion detection, and data encryption technologies; the management security module establishes a hierarchical authorization and security audit system; the algorithm security module employs federated learning privacy-preserving algorithms and blockchain encryption mechanisms to ensure the security of cross-entity data collaborative analysis; and outputs security protection instructions, access permission verification results, and security audit reports.

[0012] Preferably, the carbon asset transfer layer includes a carbon asset digital twin module, a value assessment module, a cross-chain transfer module, and a carbon finance tool module. The carbon asset digital twin module constructs a digital mirror of the entire lifecycle of carbon assets and generates a unique NFT digital passport. The value assessment module calculates the value of carbon assets through a unified assessment model and completes the equivalent conversion of different types of carbon assets. The cross-chain transfer module uses the Polkadot cross-chain protocol to achieve data docking and completes the transfer of carbon assets through smart contracts. It outputs the carbon asset digital mirror, digital passport, and equivalent transfer results.

[0013] Preferably, the analysis system embeds the GHG Protocol or ISO 14064 international carbon emission accounting rules to complete the carbon emission data format adaptation and accounting method calibration, realize the comparison and mutual recognition of domestic and cross-border carbon data, output standardized cross-border carbon data, compliant accounting results, and support cross-border carbon trading and equivalent transfer of carbon assets.

[0014] A big data-based intelligent analysis method for the carbon trading market, applied to a big data-based intelligent analysis system for the carbon trading market, includes the following steps: S1: Collect multi-source heterogeneous data through the data acquisition layer, including real-time carbon emission data, carbon trading data, external influencing factor data, and cross-border data. Build a digital mirror of the enterprise's production scenario through the digital twin mirror module, synchronously mapping production process parameters and carbon emission data. Achieve trusted data storage through the blockchain notarization module. Complete digital twin mirror modeling and on-chain data notarization through real-time sensor acquisition, multi-source data integration, and preliminary verification. Output the original collected data, basic digital twin mirror data, or on-chain notarized data. S2: The data preprocessing layer processes the raw collected data output from S1 to generate standardized analysis data. The data processing methods include outlier removal, data completion, feature selection, and data encryption, and simultaneously completes the association between the data and the digital image. The output includes standardized cleaned data, feature extracted data, encrypted data, and blockchain evidence hash value. S3: Through the price prediction module of the intelligent analysis layer, a multi-scale adaptive prediction model coupled with "carbon-energy-economy" is used to predict carbon quota prices on multiple scales and with multiple factors, and generate a price prediction report; output carbon price prediction report and error assessment results; S4: By introducing the quota allocation module of the intelligent analysis layer into the LCA life cycle assessment method to calculate the carbon footprint of the entire industry chain, and using the improved entropy weight method to calculate the index weights, the carbon quota is refined and dynamically adjusted, and the quota allocation results are output. S5: The risk warning module of the intelligent analysis layer achieves cross-entity joint modeling through federated learning, uses the analytic hierarchy process to quantify risk levels, combines attention mechanisms to customize personalized warning thresholds, embeds CBAM rules to complete compliance simulations, and outputs segmented warning information. S6: The regulatory analysis module of the intelligent analysis layer monitors transaction behavior in real time, identifies violations and blocks them at the millisecond level, traces the source of violations, and outputs the violation identification results; S7: Through the carbon asset transfer layer, construct a digital image and digital passport for carbon assets, complete the valuation of carbon assets and cross-market equivalent transfer, and provide support for carbon financial instruments; output the results of carbon asset digital image, digital passport and equivalent transfer. S8: Push analysis results, decision-making suggestions, and compliance tips to various user terminals through the application service layer to support users in carrying out relevant business; output business processing results, compliance tips, and decision-making suggestions. S9: The security layer provides full-process security protection to ensure data and system security; it outputs security protection instructions, access permission verification results, and security audit reports.

[0015] Therefore, the present invention employs the above-mentioned intelligent analysis system and method for the carbon trading market based on big data, and the technical effects are as follows: (1) This invention uses blockchain + digital twin MRV full-process trusted closed-loop technology to achieve precise association between "physical scene - digital image - data storage", solves the core problems of data fraud, delayed verification and insufficient traceability in the existing technology, shortens the data verification time, reduces the data dispute rate, and builds a precise binding mechanism of "data - behavior - responsibility". At the same time, it realizes trusted sharing and privacy protection of MRV data, and meets the trusted data needs of high-quality development of the carbon trading market.

[0016] (2) This invention proposes a multi-scale adaptive prediction model that couples carbon, energy and economy, which breaks through the limitations of single carbon price prediction. It can accurately capture the market fluctuation pattern, provide reliable support for corporate trading decisions, risk prevention and control, and price regulation and policy simulation by regulatory authorities, and help investors avoid market risks and increase the value of carbon assets.

[0017] (3) This invention achieves a breakthrough improvement in carbon price prediction. It adopts a refined quota dynamic allocation technology based on the carbon footprint of the entire industry chain, breaks the traditional one-size-fits-all allocation model, incorporates carbon emissions from the supply chain into the quota allocation consideration, and combines the linkage mechanism of "quota pledging-emission reduction incentive" to ensure regional fairness and industry heterogeneity, stimulate the enthusiasm of enterprises for emission reduction across the entire industry chain, promote the formation of a virtuous cycle of "emission reduction-profit-re-emission reduction", and improve the emission reduction efficiency of the carbon market.

[0018] (4) This invention constructs a cross-entity collaborative supervision and precise risk prevention and control system, which realizes cross-entity data collaborative analysis without disclosing corporate privacy, effectively curbs illegal activities such as abnormal transactions and data fraud, avoids cross-border carbon cost risks, and ensures the stable and orderly development of the carbon trading market.

[0019] (5) This invention uses carbon asset digital twin and cross-market equivalent transfer technology to build a unified carbon asset value assessment system, realize the equivalent transfer and reliable exchange of carbon assets of different types and regions, break the barriers to carbon asset transfer, and enhance the liquidity and vitality of the carbon market.

[0020] (6) This invention can be embedded in international carbon emission accounting rules, promote cross-platform collaboration and international integration, realize cross-regional and cross-industry data collaboration through cross-chain technology, connect with the international carbon market system, and reduce the compliance costs of cross-border carbon trading for enterprises. Attached Figure Description

[0021] Figure 1 This is an organizational framework diagram of the big data-based intelligent analysis system for the carbon trading market, as described in this invention. Figure 2 This is a flowchart of the intelligent analysis method for the carbon trading market based on big data, as described in this invention. Detailed Implementation

[0022] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0023] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0024] Example like Figure 1As shown, this invention provides a big data-based intelligent analysis system for the carbon trading market, including a terminal access layer, a data acquisition layer, a data preprocessing layer, a big data storage layer, an intelligent analysis layer, an application service layer, a security layer, and a carbon asset transfer layer. The above layers are connected in sequence to achieve intelligent analysis, control, and efficient transfer of carbon assets throughout the entire carbon trading market process. The input end of the terminal access layer receives operation instructions from the user end and outputs collection instructions through instruction issuance or result-oriented push; The data acquisition layer receives acquisition instructions from the terminal access layer and external multi-source heterogeneous data, and outputs raw acquired data or digital twin mirror base data. Multi-source heterogeneous data includes various types of raw monitoring data, reported data, public data and cross-border data. The data acquisition layer includes an IoT monitoring module, a multi-source data docking module, a data acquisition module for small and medium-sized enterprises and a cross-border data acquisition module. The IoT monitoring module collects data once per minute and transmits it in real time via the MQTT protocol; the multi-source data docking module performs preliminary conversion of heterogeneous data formats and eliminates invalid data sources; the SME data collection module performs data standardization verification; the cross-border data collection module achieves data synchronization with international platforms and completes compliant data format adaptation; and outputs the pre-verified raw heterogeneous data, which is then synchronously transmitted to the data preprocessing layer and the digital twin mirror module.

[0025] After receiving the raw data output by the data acquisition layer, the data processing layer completes feature selection through a hybrid algorithm of mRMR and PACF, removes outliers using the Z-score method, and encrypts the data using zero-knowledge proof and homomorphic encryption technology. Simultaneously, it realizes the association mapping between the processed data and the basic data of the digital twin mirror, and outputs standardized cleaned data, feature extracted data, encrypted data, and blockchain evidence hash value. The data preprocessing layer includes a data cleaning module, a feature selection module, a data encryption module, a blockchain notarization module, and a digital twin mirroring module; The data cleaning module identifies and removes outliers using the Z-score method and completes the data using interpolation. The feature selection module balances feature correlation and redundancy using a hybrid algorithm of mRMR and PACF to extract key time-series features and related features. The core formula for mRMR feature selection is used to balance feature relevance and redundancy. The calculation formula is as follows: ; In the formula Features With target variable mutual information, Features With selected features mutual information, This represents the number of selected features. The PACF partial correlation function formula is used to identify the lagged correlations in time series data and to help screen key time series features; the specific formula is as follows: ; In the formula, This represents the partial correlation coefficient for a time series with lag order k. For time series in t The value of the moment. For time series in tk The value of the moment. This indicates the calculation of conditional covariance. This indicates the calculation of conditional variance; The outlier identification formula in the Z-score method is expressed as follows: ; In the formula, For a single data value, The mean of the data. For standard deviation, when A value greater than 3 is considered an outlier.

[0026] The data encryption module uses the AES encryption algorithm combined with zero-knowledge proof; the blockchain notarization module generates data hash values, synchronously writes them to the consortium chain node, and automatically completes data verification through smart contracts; the digital twin mirror module completes 3D modeling of the production scene, realizing real-time mapping between production parameters and carbon emission data; it outputs standardized feature data, encrypted data, on-chain notarization hash values, and digital twin mapping data, and synchronously pushes them to the big data storage layer and intelligent analysis layer.

[0027] The big data storage layer adopts a Hadoop distributed architecture, storing structured data in a MySQL database and unstructured data in a MongoDB database. It establishes data indexes, builds a data synchronization and update mechanism, receives standardized data output from the data preprocessing layer and updates it in real time, and synchronously completes data redundancy backup and fault recovery. It outputs the retrieval data required by the intelligent analysis layer, including historical retrieval data or real-time retrieval data.

[0028] The intelligent analysis layer includes a price prediction module, a quota allocation module, a risk warning module, and a regulatory analysis module; The price forecasting module decomposes time-series data of carbon prices, energy prices, and economic indicators using VMD (Virtual Machine Decomposition), and identifies the coupling correlation characteristics among carbon prices, energy prices, and economic indicators through Granger causality tests. The formula is as follows: ; In the formula, To constrain the sum of squared regression residuals, For the sum of squared residuals of unconstrained regression, The lag order is... The number of samples; The VMD decomposition constraint formula is: ; in K The preset total number of modal components, Let be the partial derivative with respect to time t. Let j be the Dirac delta function, and j be the imaginary unit. Pi For the k-th modal component, For the first k The center frequency of each modal component For the complex exponential term of the Fourier transform, The square of the L2 norm, This is the allowable error threshold.

[0029] Then, the high-frequency or low-frequency components are divided using the Lempel-Ziv sequence complexity algorithm. The Lempel-Ziv sequence complexity formula is as follows: ; In the formula, For sequence length, The number of distinct subsequences in the sequence is used to classify components with a complexity ≥ 0.6 as high-frequency components and < 0.6 as low-frequency components. The GARCH-ELM mixture model is used to predict short-term price fluctuations. The model formula is as follows: ; In the formula Let be the conditional variance at time t. This is a constant term (greater than 0). The coefficient of the ARCH term (greater than or equal to 0). The coefficients of the GARCH term are greater than or equal to 0, and satisfy the following conditions: (To ensure model stability) The square of the residual at time t-1; The PSO-LSSVM model is used to predict long-term price trends. The radial basis kernel formula is as follows: ; in The first i The, the j 1 sample data; among which The kernel function parameter (greater than 0, controlling the width of the kernel function) is used to solve for the optimal value by the particle swarm optimization algorithm (PSO).

[0030] For cross-correlated data, a PSO-GM model is used for prediction. Finally, an attention mechanism is employed to optimize the multi-dimensional prediction results through nonlinear integral fusion, thereby improving prediction accuracy. The nonlinear integral fusion formula is as follows: ; in, This is the final predicted value. These are short-term, medium-term, and long-term forecasts, respectively. To integrate weights, These are attention weights (reflecting the importance of the prediction result for the i-th class); satisfying... The weights are calculated inversely based on the prediction error. ; For the first i The mean absolute error of the prediction results; For the first j The mean absolute error of the prediction results.

[0031] ; in, The predicted value at time t The actual value at time t. For the sample size, the smaller the MAE, the higher the prediction accuracy; Construct a multi-scale adaptive prediction model that couples "carbon-energy-economy" to achieve accurate prediction of carbon quota prices across multiple scales and factors, and output carbon price prediction reports and error assessment results; ; ; In the formula, The average of the actual values ​​is the mean. The higher the accuracy, the better the prediction effect.

[0032] The quota allocation module introduces the LCA life cycle assessment method to calculate the carbon footprint of the entire industry chain, and uses the improved entropy weight method to calculate the weight of indicators, so as to realize the refined allocation and dynamic adjustment of carbon quotas and output the quota allocation results. The formula for calculating the carbon footprint of the entire industry chain is as follows: ; in For the carbon footprint of the entire industry chain, Carbon emissions from upstream raw material procurement Carbon emissions from midstream production and processing Carbon emissions from downstream product transportation; The basic formula for calculating corporate carbon allowances is as follows: ; in The basic quota for enterprise i Let j be the benchmark value for industry j to which company i belongs. This represents the production capacity (per unit of product) of company i. For enterprise emission reduction performance coefficient, The regional adjustment coefficient is R = 0.9-1.0 for the central and western regions and R = 1.0 for the eastern region. The green supply chain performance score (value range 0-1).

[0033] The formula for calculating the emission reduction performance coefficient is as follows: ; in For the carbon emission intensity of enterprise i, This represents the industry's average carbon emission intensity. The value ranges from 0.8 to 1.2; the lower the carbon emission intensity, the lower the carbon emission intensity. The larger the value, the greater the quota reward.

[0034] The formula for the comprehensive score of multi-dimensional quota allocation is as follows: ; in To improve the calculation of the k-th index weight using the entropy weight method (satisfying...) ), The score of enterprise i on the k-th indicator (range 0-100) is used for the final quota adjustment.

[0035] The risk warning module achieves cross-entity joint modeling through federated learning, quantifies risk levels using the analytic hierarchy process, customizes personalized warning thresholds by combining attention mechanisms, embeds CBAM rules to complete compliance simulations, and outputs segmented warning information. The regulatory analysis module identifies abnormal transactions and triggers blocking instructions based on a hybrid model of federated learning and CNN-LSTM-Transformer, traces the source of violations through the blockchain, and outputs the violation identification results.

[0036] Based on the prediction results of the intelligent analysis layer and the real-time carbon emission data of enterprises, the annual quotas of enterprises are dynamically adjusted. Enterprises that exceed their emission reduction targets are rewarded with quotas, while those that fail to meet their targets are penalized and have their quotas reduced. Combined with the carbon inclusive mechanism, the emission reductions of individuals and SMEs are included in the quota allocation system, expanding the scope of participants in the carbon market. An innovative quota pledging mechanism is introduced, allowing enterprises to pledge excess quotas to financial institutions to obtain carbon financing. At the same time, enterprises with excellent green supply chain performance are given additional quota rewards, promoting emission reduction across the entire industrial chain.

[0037] Simplified carbon emission accounting formula for SMEs: ,in For the total carbon emissions of small and medium-sized enterprises, For the first i Fuel consumption For the first i Net calorific value of fuels For the first i Various fuel carbon emission factors The number of types of fuel used by the company; Carbon Inclusive Emissions Reduction Conversion Formula: ,in For convertible quotas, Emission reductions for carbon inclusion projects This is the conversion factor (values ​​range from 0.8 to 1.0).

[0038] The big data storage layer adopts a distributed storage architecture, which classifies and stores the standardized cleaned data, feature extraction data, encrypted data and blockchain evidence hash values ​​output by the data processing layer in both structured and unstructured categories, and builds a minute-by-minute dynamic update mechanism to synchronously complete data redundancy backup; it outputs the retrieval data required by the intelligent analysis layer, including historical retrieval data or real-time retrieval data. The intelligent analysis layer uses a multi-algorithm fusion model to complete the analysis and calculation based on the input retrieval data and the analysis instructions issued by the terminal access layer. It achieves cross-entity joint modeling through federated learning and outputs carbon asset-related data, including price forecast reports, quota allocation results, risk warning information, and violation identification results. The application service layer outputs business processing results, compliance prompts, and decision-making suggestions; the application service layer includes enterprise clients, regulatory clients, third-party institution clients, public clients, and financial institution clients according to user client type; it receives analysis results output by the intelligent analysis layer according to user client type and completes service matching in combination with user business instructions; It provides personalized emission reduction suggestions and carbon asset solutions to enterprises, regulatory data and violation warnings to regulators, carbon inclusion information to the public, and carbon asset value data to financial institutions; and outputs business processing results, compliance tips, and decision-making suggestions.

[0039] The security layer includes a hardware security module, a software security module, a management security module, and an algorithm security module; The hardware security module employs encrypted storage and redundant backup devices; the software security module utilizes access control, intrusion detection, and data encryption technologies; the management security module establishes a hierarchical authorization and security audit system; the algorithm security module employs federated learning privacy-preserving algorithms and blockchain encryption mechanisms to ensure the security of cross-entity data collaborative analysis; and outputs security protection instructions, access permission verification results, and security audit reports.

[0040] The analysis system embeds the GHG Protocol or ISO 14064 international carbon emission accounting rules to complete carbon emission data format adaptation and accounting method calibration, enabling comparison and mutual recognition of domestic and cross-border carbon data, and outputting standardized cross-border carbon data, compliant accounting results, and support for cross-border carbon trading and equivalent carbon asset transfer. The carbon asset transfer layer includes a carbon asset digital twin module, a value assessment module, a cross-chain transfer module, and a carbon finance tool module. The carbon asset digital twin module constructs a digital mirror of the entire lifecycle of carbon assets, generating a unique NFT digital passport. The value assessment module calculates the value of carbon assets through a unified assessment model, completing equivalent conversions of different types of carbon assets. The cross-chain transfer module uses the Polkadot cross-chain protocol to achieve data integration and completes carbon asset transfer through smart contracts, outputting carbon asset digital mirrors, digital passports, and equivalent transfer results.

[0041] The security layer adopts a four-in-one security mechanism to monitor data transmission and operation behavior in real time, and complete access permission verification, intrusion detection and data encryption protection. The carbon asset-related data output from the intelligent analysis layer and the transfer instructions from the application service layer are input into the carbon asset transfer layer. A carbon asset mirror is constructed using digital twin technology, and an equivalent conversion is completed through a value assessment model. The transfer is realized based on the Polkadot cross-chain protocol, and the carbon asset digital mirror, digital passport, and equivalent transfer results are output.

[0042] The formula for valuing carbon assets is as follows: ; in For carbon asset value, This is the predicted carbon price. For the quantity of carbon assets, This refers to the carbon asset appreciation rate (annual growth rate, ranging from 0 to 0.2). This is the liquidity coefficient (reflecting the activity level of carbon asset trading, with a value ranging from 0.7 to 1.0). The formula for converting carbon assets to equivalent value is as follows: ; in , The quantity of carbon assets of different types (or different regions) , This formula ensures that different carbon assets are equivalent in value when exchanged, corresponding to the unit value of the carbon assets.

[0043] like Figure 2 As shown, a big data-based intelligent analysis method for the carbon trading market, applied to a big data-based intelligent analysis system for the carbon trading market, includes the following steps: S1: Collect multi-source heterogeneous data through the data acquisition layer, including real-time carbon emission data, carbon trading data, external influencing factor data, and cross-border data. Build a digital mirror of the enterprise's production scenario through the digital twin mirror module, synchronously mapping production process parameters and carbon emission data. Achieve trusted data storage through the blockchain notarization module. Complete digital twin mirror modeling and on-chain data notarization through real-time sensor acquisition, multi-source data integration, and preliminary verification. Output the original collected data, basic digital twin mirror data, or on-chain notarized data. S2: The data preprocessing layer processes the raw collected data output from S1 to generate standardized analysis data. The data processing methods include outlier removal, data completion, feature selection, and data encryption, and simultaneously completes the association between the data and the digital image. The output includes standardized cleaned data, feature extracted data, encrypted data, and blockchain evidence hash value. S3: Through the price prediction module of the intelligent analysis layer, a multi-scale adaptive prediction model coupled with "carbon-energy-economy" is used to predict carbon quota prices on multiple scales and with multiple factors, and generate a price prediction report; output carbon price prediction report and error assessment results; S4: By introducing the quota allocation module of the intelligent analysis layer into the LCA life cycle assessment method to calculate the carbon footprint of the entire industry chain, and using the improved entropy weight method to calculate the index weights, the carbon quota is refined and dynamically adjusted, and the quota allocation results are output. S5: The risk warning module of the intelligent analysis layer achieves cross-entity joint modeling through federated learning, uses the analytic hierarchy process to quantify risk levels, combines attention mechanisms to customize personalized warning thresholds, embeds CBAM rules to complete compliance simulations, and outputs segmented warning information. S6: The regulatory analysis module of the intelligent analysis layer monitors transaction behavior in real time, identifies violations and blocks them at the millisecond level, traces the source of violations, and outputs the violation identification results; S7: Through the carbon asset transfer layer, construct a digital image and digital passport for carbon assets, complete the valuation of carbon assets and cross-market equivalent transfer, and provide support for carbon financial instruments; output the results of carbon asset digital image, digital passport and equivalent transfer. S8: Push analysis results, decision-making suggestions, and compliance tips to various user terminals through the application service layer to support users in carrying out relevant business; output business processing results, compliance tips, and decision-making suggestions. S9: The security layer provides full-process security protection to ensure data and system security; it outputs security protection instructions, access permission verification results, and security audit reports.

[0044] Therefore, the present invention adopts the above-mentioned intelligent analysis system and method for carbon trading market based on big data, including a terminal access layer, a data acquisition layer, a data preprocessing layer, a big data storage layer, an intelligent analysis layer, an application service layer, a security layer, and a carbon asset transfer layer. Each layer works in concert to realize intelligent analysis, control and efficient transfer of carbon assets throughout the entire carbon trading market process.

[0045] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. 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 still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A big data-based intelligent analysis system for the carbon trading market, characterized in that: It includes a terminal access layer, a data acquisition layer, a data preprocessing layer, a big data storage layer, an intelligent analysis layer, an application service layer, a security layer, and a carbon asset transfer layer. These layers are connected in sequence to work together to achieve intelligent analysis, control, and efficient transfer of carbon assets throughout the entire carbon trading market process. The input end of the terminal access layer receives operation instructions from the user end and outputs collection instructions through instruction issuance or result-oriented push; The data acquisition layer receives acquisition instructions from the terminal access layer and external multi-source heterogeneous data, and outputs raw acquisition data or digital twin mirror base data; After receiving the raw data output by the data acquisition layer, the data processing layer completes feature selection through a hybrid algorithm of mRMR and PACF, removes outliers using the Z-score method, and encrypts the data using zero-knowledge proof and homomorphic encryption technology. Simultaneously, it realizes the association mapping between the processed data and the basic data of the digital twin mirror, and outputs standardized cleaned data, feature extracted data, encrypted data, and blockchain evidence hash value. The big data storage layer adopts a distributed storage architecture, which classifies and stores the standardized cleaned data, feature extraction data, encrypted data and blockchain evidence hash values ​​output by the data processing layer in both structured and unstructured categories, and builds a minute-by-minute dynamic update mechanism to synchronously complete data redundancy backup; it outputs the retrieval data required by the intelligent analysis layer, including historical retrieval data or real-time retrieval data. The intelligent analysis layer uses a multi-algorithm fusion model to complete the analysis and calculation based on the input retrieval data and the analysis instructions issued by the terminal access layer. It achieves cross-entity joint modeling through federated learning and outputs carbon asset-related data, including price forecast reports, quota allocation results, risk warning information, and violation identification results. The application service layer outputs business processing results, compliance tips, and decision-making suggestions; The security layer adopts a four-in-one security mechanism to monitor data transmission and operation behavior in real time, and complete access permission verification, intrusion detection and data encryption protection. The carbon asset-related data output from the intelligent analysis layer and the transfer instructions from the application service layer are input into the carbon asset transfer layer. A carbon asset mirror is constructed using digital twin technology, and an equivalent conversion is completed through a value assessment model. The transfer is realized based on the Polkadot cross-chain protocol, and the carbon asset digital mirror, digital passport, and equivalent transfer results are output.

2. The intelligent analysis system for the carbon trading market based on big data as described in claim 1, characterized in that, Multi-source heterogeneous data includes various types of raw monitoring data, reported data, public data and cross-border data. The data acquisition layer includes an IoT monitoring module, a multi-source data docking module, a data acquisition module for small and medium-sized enterprises and a cross-border data acquisition module. The IoT monitoring module collects data once per minute and transmits it in real time via the MQTT protocol; the multi-source data docking module performs preliminary conversion of heterogeneous data formats and eliminates invalid data sources; the SME data collection module performs data standardization verification; the cross-border data collection module achieves data synchronization with international platforms and completes compliant data format adaptation; and outputs the pre-verified raw heterogeneous data, which is then synchronously transmitted to the data preprocessing layer and the digital twin mirror module.

3. The intelligent analysis system for the carbon trading market based on big data according to claim 2, characterized in that, The data preprocessing layer includes a data cleaning module, a feature selection module, a data encryption module, a blockchain notarization module, and a digital twin mirroring module; The data cleaning module identifies and removes outliers using the Z-score method and completes the data using interpolation. The feature selection module balances feature correlation and redundancy using a hybrid algorithm of mRMR and PACF to extract key time-series features and related features. The data encryption module uses the AES encryption algorithm combined with zero-knowledge proof; the blockchain notarization module generates data hash values, synchronously writes them to the consortium chain node, and automatically completes data verification through smart contracts; the digital twin mirror module completes 3D modeling of the production scene, realizing real-time mapping between production parameters and carbon emission data; it outputs standardized feature data, encrypted data, on-chain notarization hash values, and digital twin mapping data, and synchronously pushes them to the big data storage layer and intelligent analysis layer.

4. The intelligent analysis system for the carbon trading market based on big data according to claim 1, characterized in that, The big data storage layer adopts a Hadoop distributed architecture, storing structured data in a MySQL database and unstructured data in a MongoDB database. It establishes data indexes, builds a data synchronization and update mechanism, receives standardized data output from the data preprocessing layer and updates it in real time, and synchronously completes data redundancy backup and fault recovery. It outputs the retrieval data required by the intelligent analysis layer, including historical retrieval data or real-time retrieval data.

5. The intelligent analysis system for the carbon trading market based on big data according to claim 1, characterized in that, The intelligent analysis layer includes a price prediction module, a quota allocation module, a risk warning module, and a regulatory analysis module; The price prediction module decomposes time-series data of carbon price, energy price, and economic indicators through VMD, identifies coupling correlations through Granger causality test, divides high-frequency or low-frequency components through sequence complexity, and constructs a multi-scale adaptive prediction model of "carbon-energy-economy" coupling by using GARCH-ELM hybrid model, PSO-LSSVM model, PSO-GM model and attention mechanism. This enables accurate prediction of carbon quota prices across multiple scales and factors, and outputs carbon price prediction reports and error assessment results. The quota allocation module introduces the LCA life cycle assessment method to calculate the carbon footprint of the entire industry chain, and uses the improved entropy weight method to calculate the weight of indicators, so as to realize the refined allocation and dynamic adjustment of carbon quotas and output the quota allocation results. The risk warning module achieves cross-entity joint modeling through federated learning, quantifies risk levels using the analytic hierarchy process, customizes personalized warning thresholds by combining attention mechanisms, embeds CBAM rules to complete compliance simulations, and outputs segmented warning information. The regulatory analysis module identifies abnormal transactions and triggers blocking instructions based on a hybrid model of federated learning and CNN-LSTM-Transformer, traces the source of violations through the blockchain, and outputs the violation identification results.

6. The intelligent analysis system for the carbon trading market based on big data according to claim 1, characterized in that, The application service layer includes enterprise clients, regulatory clients, third-party institution clients, public clients, and financial institution clients based on user client type; it receives analysis results output by the intelligent analysis layer based on user client type and completes service matching in combination with user business instructions; Personalized emission reduction suggestions and carbon asset solutions are pushed to enterprises; regulatory data and violation warnings are pushed to regulators; carbon inclusion information is pushed to the public; and carbon asset value data is pushed to financial institutions. Output business processing results, compliance tips, and decision-making suggestions.

7. The intelligent analysis system for the carbon trading market based on big data according to claim 1, characterized in that, The security layer includes a hardware security module, a software security module, a management security module, and an algorithm security module; The hardware security module employs encrypted storage and redundant backup devices; the software security module utilizes access control, intrusion detection, and data encryption technologies; the management security module establishes a hierarchical authorization and security audit system; the algorithm security module employs federated learning privacy-preserving algorithms and blockchain encryption mechanisms to ensure the security of cross-entity data collaborative analysis; and outputs security protection instructions, access permission verification results, and security audit reports.

8. The intelligent analysis system for the carbon trading market based on big data according to claim 1, characterized in that, The carbon asset circulation layer includes a carbon asset digital twin module, a value assessment module, a cross-chain circulation module, and a carbon finance instrument module. The carbon asset digital twin module constructs a digital mirror of the entire life cycle of carbon assets and generates a unique NFT digital passport. The value assessment module calculates the value of carbon assets through a unified assessment model and completes the equivalent conversion of different types of carbon assets. The cross-chain transfer module uses the Polkadot cross-chain protocol to achieve data docking and completes carbon asset transfer through smart contracts; it outputs carbon asset digital images, digital passports, and equivalent transfer results.

9. The intelligent analysis system for the carbon trading market based on big data according to claim 1, characterized in that, The analysis system embeds the GHG Protocol or ISO 14064 international carbon emission accounting rules to complete the carbon emission data format adaptation and accounting method calibration, realize the comparison and mutual recognition of domestic and cross-border carbon data, output standardized cross-border carbon data, compliant accounting results, and support cross-border carbon trading and equivalent transfer of carbon assets.

10. A method for intelligent analysis of the carbon trading market based on big data, characterized in that, The intelligent analysis system for the carbon trading market based on big data, as described in any one of claims 1-9, includes the following steps: S1: Collect multi-source heterogeneous data through the data acquisition layer, including real-time carbon emission data, carbon trading data, external influencing factor data, and cross-border data. Build a digital mirror of the enterprise's production scenario through the digital twin mirror module, synchronously mapping production process parameters and carbon emission data. Achieve trusted data storage through the blockchain notarization module. Complete digital twin mirror modeling and on-chain data notarization through real-time sensor acquisition, multi-source data integration, and preliminary verification. Output the original collected data, basic digital twin mirror data, or on-chain notarized data. S2: The data preprocessing layer processes the raw collected data output from S1 to generate standardized analysis data. The data processing methods include outlier removal, data completion, feature selection, and data encryption, and simultaneously completes the association between the data and the digital image. The output includes standardized cleaned data, feature extracted data, encrypted data, and blockchain evidence hash value. S3: Through the price prediction module of the intelligent analysis layer, a multi-scale adaptive prediction model coupled with "carbon-energy-economy" is used to predict carbon quota prices on multiple scales and with multiple factors, and generate a price prediction report; output carbon price prediction report and error assessment results; S4: By introducing the quota allocation module of the intelligent analysis layer into the LCA life cycle assessment method to calculate the carbon footprint of the entire industry chain, and using the improved entropy weight method to calculate the index weights, the carbon quota is refined and dynamically adjusted, and the quota allocation results are output. S5: The risk warning module of the intelligent analysis layer achieves cross-entity joint modeling through federated learning, uses the analytic hierarchy process to quantify risk levels, combines attention mechanisms to customize personalized warning thresholds, embeds CBAM rules to complete compliance simulations, and outputs segmented warning information. S6: The regulatory analysis module of the intelligent analysis layer monitors transaction behavior in real time, identifies violations and blocks them at the millisecond level, traces the source of violations, and outputs the violation identification results; S7: Through the carbon asset transfer layer, construct a digital image and digital passport for carbon assets, complete the valuation of carbon assets and cross-market equivalent transfer, and provide support for carbon financial instruments; output the results of carbon asset digital image, digital passport and equivalent transfer. S8: Push analysis results, decision-making suggestions, and compliance tips to various user terminals through the application service layer to support users in carrying out relevant business; output business processing results, compliance tips, and decision-making suggestions. S9: The security layer provides full-process security protection to ensure data and system security; it outputs security protection instructions, access permission verification results, and security audit reports.

Images