Logistics carbon emission data monitoring system based on block chain and edge calculation

Abstract

The invention discloses a logistics carbon emission data monitoring system based on a block chain and edge calculation, and relates to the technical field of logistics systems. According to the logistics carbon emission data monitoring system based on the block chain and edge calculation, on the aspect of data integration and estimation, multi-source heterogeneous data is fused to construct a dynamic compensation model, a lag compensation algorithm is developed, emission factors are adjusted, more accurate real-time carbon emission estimation is achieved, and full-chain data standardization and association are achieved through middleware; in the aspect of privacy protection, the blockchain and zero-knowledge proof technology are combined, enterprise data authenticity verification is ensured, original information is not leaked, a carbon emission result is subjected to chaining evidence storage, original data is encrypted and stored, and cross-enterprise federal learning ensures that data cannot be out of a domain; the logistics carbon emission data monitoring accuracy, privacy protection and auditing supervision intelligent level are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of logistics systems, and specifically to a logistics carbon emission data monitoring system based on blockchain and edge computing. Background Art

[0002] Logistics activities are a significant source of carbon emissions, encompassing fuel combustion in transportation vehicles and energy consumption in storage facilities. According to relevant research, the global logistics industry accounts for approximately 8%-10% of total global carbon emissions, and this proportion continues to rise with the booming e-commerce and international trade. Accurately monitoring logistics carbon emissions can clearly assess the impact of logistics activities on climate change, providing a basis for developing targeted emission reduction measures, thereby effectively reducing the negative impact of the logistics industry on the climate system and slowing global warming.

[0003] Publication No. CN117114542A discloses a cold chain logistics carbon emission data monitoring method and device based on blockchain and edge computing, which monitors the temperature, humidity and carbon concentration information of each transportation stage in real time, and sends the sampled data to the edge computing node for processing and calculation; the edge computing node receives and converts multiple sensor data in real time, and calculates the energy consumption and carbon emissions at each moment; the edge computing node uploads environmental data, energy consumption calculation results and carbon emission information to each node of the blockchain network; the blockchain network summarizes the data collected by each edge computing node to calculate the carbon emissions of the cold chain transportation tool in each transportation stage, and updates the data in real time; the edge computing node implements adjustments on the device side to achieve real-time, dynamic closed-loop carbon emission monitoring and control.

[0004] However, as shown above, existing technologies focus only on temperature, humidity, and carbon concentration information during the cold chain logistics transportation phase, and fail to fully integrate heterogeneous data from multiple sources, such as maritime / air transport satellite trajectories, port AIS data, storage electricity, and transportation fuel. This makes it difficult to construct a dynamic compensation model that covers the entire logistics chain and crosses transportation modes, and it is unable to accurately reflect the carbon emission correlations between different transportation modes and warehousing links. This results in limited comprehensiveness and accuracy of carbon emission estimates, and it cannot provide a reliable basis for optimizing carbon emissions throughout the entire logistics chain. Furthermore, in actual logistics scenarios, data collection is subject to lags. For example, in maritime transport, satellite trajectories are not synchronized with port loading and unloading data, and in air transport, carbon emissions fluctuate due to meteorological changes. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the present invention provides a logistics carbon emission data monitoring system based on blockchain and edge computing to solve the above problems.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a logistics carbon emission data monitoring system based on blockchain and edge computing, including:

[0007] The fusion and dynamic compensation module integrates multi-source heterogeneous data on maritime / air transport satellite trajectories, port AIS data, and warehousing power and transportation fuel to build a dynamic compensation model for cross-border transportation. A lag compensation algorithm is developed through cross-validation of satellite trajectory predictions and AIS data, combining historical data fitting to estimate carbon emissions in real time. For air transport scenarios, emission factors are dynamically adjusted by integrating aircraft model and route meteorological data. A logistics carbon data middleware is developed to standardize data formats across the entire supply chain, and an intelligent mapping engine is deployed to link warehousing power and transportation fuel data to the same carbon footprint model.

[0008] The privacy protection module, based on blockchain and zero-knowledge proof technology, generates a ZKP when enterprises submit data to verify authenticity without leaking the original information. The hash value of the carbon emission calculation result is stored on the chain, and the original data is encrypted and stored in the IPFS distributed network, requiring the enterprise's private key to authorize access. A cross-enterprise federated learning model is built, and each enterprise only uploads encrypted gradient parameters to ensure that data does not leave the domain and privacy is secure.

[0009] The intelligent audit and supervision module deploys AI audit robots to automatically compare multi-dimensional data, mark abnormal fluctuations, and issue warnings based on severity. The audit rules integrate static threshold and dynamic threshold models, establish a whitelist of third-party audit institutions for enterprises to choose from, and store the hash value of the audit report on the chain.

[0010] Preferably, the fusion and dynamic compensation module includes the following steps for collecting and preprocessing shipping / air transport satellite tracks and port AIS data:

[0011] A11. Satellite trajectory data acquisition: Access Sentinel-1 synthetic aperture radar satellite images to obtain the real-time position, speed and heading data of the ship's latitude and longitude, with a time resolution of Δt s minute;

[0012] A12. Data cleaning: Eliminate outliers. The speed outlier elimination condition is V>V max or V <V min , fill missing values;

[0013] A13, AIS data collection: Obtain AIS messages from the MarineTraffic platform and parse the vessel's MMSI number, draft, and load status fields;

[0014] A14. Time alignment: Match satellite data with AIS data according to UTC timestamps. Time alignment error: ∈ t <Δt a minute.

[0015] Preferably, the design correction of the lag compensation algorithm by the fusion and dynamic compensation module includes:

[0016] A2.1. Historical data fitting: Collect carbon emission data of ships on the same route over the past 12 months and build a regression model:

[0017] Carbon emissions = α·V 2 +β·W+γ·C+∈;

[0018] Among them, α, β, γ are regression coefficients, V is the ship speed, W is the deadweight tonnage, C is the sea state coefficient, and ε is the error term;

[0019] A2.2 Calculation of sea state coefficient:

[0020]

[0021] Among them, H represents wave height, U represents wind speed, k1 and k2 are weight coefficients, H ref 、U ref are the reference thresholds for wave height and wind speed, respectively;

[0022] A2.3. Calculation of 72-hour lag compensation: Input current satellite / AIS data and substitute it into the regression model to predict real-time carbon emissions. The compensation formula is:

[0023]

[0024] Among them, E comp represents the carbon emissions after compensation, represents the model prediction value, represents the average value of lagged data;

[0025] A2.4. Error control: Validate the regression model through cross-validation algorithm and adjust the parameters to control the error MAE<∈ m ;

[0026] A2.5. For post-compensation carbon emissions, Kalman filtering is used for smoothing: a state-space model is established, satellite predictions are integrated with port loading and unloading records, and carbon emission estimates are dynamically adjusted.

[0027] A2.6. Port anchor point calibration: When the ship approaches the port, the regression model is calibrated based on the port loading and unloading data to eliminate the accumulated error; Calibration trigger condition: d port <D th , where d port Indicates the distance from the port, D th Indicates the distance threshold.

[0028] Preferably, the step of dynamically adjusting the emission factor of the fusion and dynamic compensation module for the air transport scenario includes:

[0029] A3.1 Data Collection:

[0030] Aircraft model data: Establish an aircraft model database, associate engine type, and fuel efficiency η parameters;

[0031] Route weather data: access FlightRadar24API to obtain real-time wind speed V wind , wind direction, temperature and turbulence intensity;

[0032] A3.2. Feature conversion: Convert wind direction to angle θ with heading and calculate effective wind speed V eff =V wind ·cosθ.

[0033] Preferably, the dynamic adjustment of emission factors includes:

[0034] Calculation of baseline emission factor: Calculate the baseline emission factor EF according to the ICAO standard formula:

[0035]

[0036] Where k' is a constant used to adjust the baseline emission factor. Indicates fuel flow, F thrust Indicates thrust;

[0037] Meteorological correction factors include:

[0038] Headwind correction factor Where c1 is the adjustment constant of EF for meteorological conditions;

[0039] Temperature correction factor Where c2 is the adjustment constant of EF with respect to temperature;

[0040] Turbulence Correction: It means that moderate turbulence increases EF by c3, and severe turbulence increases it by c4;

[0041] The calculation formula for the dynamic emission factor based on the meteorological correction factor is:

[0042] EF dyn =EF base ×(1+Δ wind )×(1+Δ temp )×(1+Δ turb ).

[0043] Preferably, the total carbon emissions calculation method is further adjusted to take into account the different flight phases:

[0044] Flight phase division: The flight is divided into four phases: taxiing, climbing, cruising, and descending. Assume that the total carbon emissions of the flight are TCE, the dynamic emission factor is EF, the fuel consumption is expressed as FC, and the correction factor is expressed as CF. Calculate the carbon emissions separately:

[0045] Sliding phase: FC 滑行 = taxi time × taxi fuel flow;

[0046] Climbing phase: FC 爬升 = Climb time × Climb fuel flow × (1 + CF 爬升 );

[0047] Cruise phase: FC 巡航 =cruise time × taxi fuel flow;

[0048] Descending stage: FC 下降 = descent time × descent fuel flow × (1 + CF 下降 );

[0049] The formula for calculating total carbon emissions is:

[0050] TCE=∑(EF 动态 ×FC)=EF 动态,滑行 ×FC 滑行 +EF 动态,爬升 ×FC 爬升 +EF 动态,巡航 ×FC 巡航 +EF

[0051] 动态,下降 ×FC 下降 .

[0052] Preferably, the step of converting electricity and fuel consumption into carbon footprint in the fusion and dynamic compensation module includes:

[0053] A4.1. Data preprocessing: Clean outliers, unify time granularity, and correlate warehousing and transportation data.

[0054] A4.2. Parameter design:

[0055] E wh : Annual electricity consumption of storage facilities;

[0056] F trans : Annual fuel consumption of transportation vehicles;

[0057] EF elec : Electricity carbon emission factor, i.e., the average value of the regional power grid;

[0058] EF fuel : fuel carbon emission factor;

[0059] D: Total annual transport mileage, used to allocate fuel consumption;

[0060] η: Transport load utilization rate = actual load / maximum load;

[0061] A4.3. Dynamic configuration of parameters:

[0062] Select electricity carbon emission factor EF according to regional power grid type elec ;

[0063] Select the fuel carbon emission factor EF according to the fuel type fuel ;

[0064] Monitor load data through IoT devices and update the load utilization rate η of transport vehicles in real time;

[0065] A4.4. Convert electricity and fuel consumption into carbon footprints. The calculation process includes:

[0066] A4.4.1. Calculate the carbon footprint of storage electricity: CF wh =E wh ×EF elec ;

[0067] A4.4.2. Calculate the carbon footprint of transportation fuels:

[0068] A4.4.3. Calculate the total carbon footprint: CF total =CF wh +CF trans ;

[0069] A4.5. Model validation and optimization: Verify model accuracy by comparing historical data; introduce machine learning to predict future carbon footprint.

[0070] Preferably, the privacy protection module specifically includes:

[0071] B1. Blockchain + Zero-Knowledge Proof Storage: Enterprise data is first processed through homomorphic encryption and then generated into a ZKP, which supports selective disclosure without exposing the entire data. The hash value of the carbon emission calculation result is uploaded to the blockchain, and the original data is stored in the IPFS distributed network, requiring the enterprise's private key to authorize access.

[0072] B2. Federated Learning Privacy Computing: Build a cross-enterprise federated learning model where each enterprise only uploads encrypted gradient parameters to collaboratively optimize the carbon emissions prediction model without data leaving the domain.

[0073] B3. NFTization of carbon credits: Convert emission reductions into NFTs, record metadata of emission reduction behavior, time, and location, and automatically trade through smart contracts to improve the liquidity of carbon assets; the NFT hash value is linked to blockchain evidence to ensure that carbon credits cannot be tampered with and are traceable.

[0074] Preferably, the enterprise uploads encrypted gradient parameter operations specifically including:

[0075] B2.1. Local training: Each enterprise uses its own data to train the model locally and obtain gradient parameters;

[0076] Encryption: Gradient parameters are encrypted to ensure data security during transmission. Noise is added to the gradient parameters to prevent the original data from being inferred through the model while maintaining model accuracy.

[0077] Parameter upload: Upload the encrypted gradient parameters to the federated learning server.

[0078] Preferably, the work content of the intelligent audit and supervision module specifically includes:

[0079] C1. AI Anomaly Detection: Deploy AI audit robots to automatically compare multi-dimensional data on fuel consumption, load, and GPS trajectory, flag abnormal fluctuations, and trigger manual review. Manual review then provides graded warnings based on the severity of the anomaly. Audit rules include static and dynamic rules. Static rules set specific thresholds for audit items, while dynamic rules create thresholds using threshold models trained with historical data.

[0080] C2. Third-party audit pool: Establish a whitelist of third-party audit institutions. Enterprises can independently select institutions for verification. The hash value of the audit report is uploaded to the chain, and the entire audit process is recorded.

[0081] This invention provides a logistics carbon emissions data monitoring system based on blockchain and edge computing. Compared with existing technologies, it has the following advantages:

[0082] 1. This logistics carbon emission data monitoring system based on blockchain and edge computing, in terms of data integration and estimation, integrates multi-source heterogeneous data to build a dynamic compensation model, develops a lag compensation algorithm and adjusts the emission factor to achieve more accurate real-time carbon emission estimation, and realizes full-chain data standardization and correlation through middleware; in terms of privacy protection, it combines blockchain and zero-knowledge proof technology to ensure the authenticity of enterprise data without leaking original information, carbon emission results are stored on the chain, original data is encrypted and stored, and cross-enterprise federated learning ensures that data does not leave the domain; in the audit and supervision link, AI audit robots are deployed to automatically compare data, issue graded warnings, integrate multi-threshold models, provide a third-party audit whitelist, and store audit reports on the chain, which improves the accuracy of logistics carbon emission data monitoring, privacy protection and the intelligence level of audit and supervision.

[0083] 2. This logistics carbon emissions data monitoring system, based on blockchain and edge computing, collects satellite trajectory, AIS, and meteorological data from multiple channels in sea and air transport scenarios. Through pre-processing such as cleaning and alignment, data quality is ensured, laying the foundation for accurate analysis. For lagged sea transport data, a regression model is constructed, taking into account sea condition coefficients. Combined with Kalman filtering and port calibration, this effectively compensates for lag errors and improves real-time estimation accuracy. In air transport scenarios, emission factors are dynamically adjusted, taking into account flight phase and meteorological factors, making carbon emissions calculations more realistic. Warehousing electricity and transportation fuel consumption are uniformly converted into carbon footprints. Through dynamic parameter configuration, taking into account regional power grids, fuel types, and load utilization, scientific calculations are achieved, and the optimization model is verified to ensure accurate and reliable results.

[0084] 3. This logistics carbon emissions data monitoring system, based on blockchain and edge computing, utilizes blockchain and zero-knowledge proof technology. Enterprise data is homomorphically encrypted and ZKP generated, enabling selective disclosure to prevent raw data leakage. Carbon emissions results are stored on-chain, and raw data is encrypted and stored on IPFS to ensure data security. In a cross-enterprise federated learning model, enterprises only upload encrypted gradient parameters, and data does not leave the domain. Encryption and noise processing ensure transmission security, prevent data back-propagation, maintain model accuracy, and achieve a balance between privacy and efficiency. Carbon credits are NFT-ified to enhance carbon asset liquidity. NFTs are linked to blockchain evidence to ensure that carbon credits are tamper-proof and traceable.

[0085] 4. The logistics carbon emission data monitoring system based on blockchain and edge computing uses an AI audit robot to automatically compare multi-dimensional data, accurately mark abnormal fluctuations and issue graded warnings. It combines static and dynamic audit rules to improve the comprehensiveness and accuracy of anomaly detection; establishes a whitelist of third-party audit institutions for enterprises to choose from, and the audit reports are uploaded to the chain and the process is recorded, which not only ensures the fairness and professionalism of the audit, but also enhances the credibility and traceability of the results, and optimizes the logistics carbon emission review and supervision process. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] Figure 1 This is a block diagram of the overall system module of the present invention;

[0087] Figure 2 Schematic diagram of the framework of the fusion and dynamic compensation module of the present invention;

[0088] Figure 3 Schematic diagram of the framework of the privacy protection module of the present invention;

[0089] Figure 4 This is a schematic diagram of the framework of the intelligent audit and supervision module of the present invention. DETAILED DESCRIPTION

[0090] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0091] See Figures 1-4 , the present invention provides the following four technical solutions:

[0092] The first implementation method: a logistics carbon emissions data monitoring system based on blockchain and edge computing, including:

[0093] The fusion and dynamic compensation module integrates multi-source heterogeneous data on maritime / air transport satellite trajectories, port AIS data, and warehousing power and transportation fuel to build a dynamic compensation model for cross-border transportation. A lag compensation algorithm is developed through cross-validation of satellite trajectory predictions and AIS data, combining historical data fitting to estimate carbon emissions in real time. For air transport scenarios, emission factors are dynamically adjusted by integrating aircraft model and route meteorological data. A logistics carbon data middleware is developed to standardize data formats across the entire supply chain, and an intelligent mapping engine is deployed to link warehousing power and transportation fuel data to the same carbon footprint model.

[0094] The privacy protection module, based on blockchain and zero-knowledge proof technology, generates a ZKP when enterprises submit data to verify authenticity without leaking the original information. The hash value of the carbon emission calculation result is stored on the chain, and the original data is encrypted and stored in the IPFS distributed network, requiring the enterprise's private key to authorize access. A cross-enterprise federated learning model is built, and each enterprise only uploads encrypted gradient parameters to ensure that data does not leave the domain and privacy is secure.

[0095] The intelligent audit and supervision module deploys AI audit robots to automatically compare multi-dimensional data, mark abnormal fluctuations, and issue warnings based on severity. The audit rules integrate static threshold and dynamic threshold models, establish a whitelist of third-party audit institutions for enterprises to choose from, and store the hash value of the audit report on the chain.

[0096] In terms of data integration and estimation, multi-source heterogeneous data are integrated to build a dynamic compensation model, a lag compensation algorithm is developed and emission factors are adjusted to achieve more accurate real-time carbon emission estimation, and full-chain data standardization and correlation are achieved through middleware; in terms of privacy protection, blockchain and zero-knowledge proof technology are combined to ensure the authenticity of enterprise data without leaking original information, carbon emission results are stored on the chain, original data is encrypted and stored, and cross-enterprise federated learning ensures that data does not leave the domain; in the audit and supervision link, AI audit robots are deployed to automatically compare data, issue graded warnings, integrate multi-threshold models, provide third-party audit whitelists, and store audit reports on the chain, which improves the accuracy of logistics carbon emission data monitoring, privacy protection and the level of intelligent audit and supervision.

[0097] The second embodiment differs from the first embodiment mainly in that the fusion and dynamic compensation module collects and pre-processes the maritime / air transport satellite trajectory and port AIS data, including:

[0098] A11. Satellite trajectory data acquisition: Access Sentinel-1 synthetic aperture radar satellite images to obtain the real-time position, speed and heading data of the ship's latitude and longitude, with a time resolution of Δt s minute;

[0099] A12. Data cleaning: Eliminate outliers. The speed outlier elimination condition is V>V max or V <V min , fill in missing values (using linear interpolation);

[0100] A13, AIS data collection: Obtain AIS messages from the MarineTraffic platform and parse fields such as the vessel's MMSI number, draft, and load status (full / empty);

[0101] A14. Time alignment: Match satellite data with AIS data according to UTC timestamps. Time alignment error: ∈ t <Δt a minute.

[0102] In this embodiment, the design correction of the lag compensation algorithm by the fusion and dynamic compensation module includes:

[0103] A2.1. Historical data fitting: Collect carbon emission data (unit: kg CO2 / nautical mile) of ships on the same route over the past 12 months and construct a regression model:

[0104] Carbon emissions = α·V 2 +β·W+γ·C+∈;

[0105] Among them, α, β, γ are regression coefficients, V is the ship speed, W is the deadweight tonnage, C is the sea state coefficient, and ε is the error term;

[0106] A2.2 Calculation of sea state coefficient:

[0107]

[0108] Among them, H represents wave height, U represents wind speed, k1 and k2 are weight coefficients, H ref 、U ref are the reference thresholds for wave height and wind speed, respectively;

[0109] A2.3. Calculation of 72-hour lag compensation: Input current satellite / AIS data (e.g., speed 15 knots, load 50,000 tons, sea state 3), substitute it into the regression model to predict real-time carbon emissions. The compensation formula is:

[0110]

[0111] Among them, E comp represents the carbon emissions after compensation, represents the model prediction value, represents the average value of lagged data;

[0112] A2.4. Error control: Validate the model through cross-validation (leave one out method) and adjust the parameters to control the error MAE<∈ m (∈ m =8%);

[0113] A2.5. For compensated carbon emissions, Kalman filtering is used for smoothing: a state-space model is established, integrating satellite predictions with port loading and unloading records (such as changes in container weights) to dynamically adjust carbon emission estimates.

[0114] A2.6. Port anchor point calibration: When the ship approaches the port (distance < 10 nautical miles), the regression model is calibrated based on the port loading and unloading data to eliminate the accumulated error; Calibration trigger condition: d port <D th , where d port Indicates the distance from the port, D th Indicates the distance threshold.

[0115] In this embodiment, the steps of dynamically adjusting the emission factor for the air transport scenario by the fusion and dynamic compensation module include:

[0116] A3.1 Data Collection:

[0117] Aircraft model data: Build a database of aircraft models (e.g., Boeing 747-8, Airbus A350), and associate parameters such as engine type (e.g., GE90-115B), fuel efficiency η (L / 100km), etc.

[0118] Route weather data: access FlightRadar24API to obtain real-time wind speed Vwind (m / s), wind direction (°), air temperature (°C) and turbulence intensity (mild / moderate / severe);

[0119] A3.2. Feature conversion: Convert wind direction to angle θ with heading and calculate effective wind speed V eff =V wind ·cosθ.

[0120] In this embodiment, the dynamic adjustment of emission factors includes:

[0121] Calculation of baseline emission factor: Calculate the baseline emission factor EF according to the ICAO standard formula:

[0122]

[0123] Where k' is a constant used to adjust the baseline emission factor. Indicates fuel flow, F thrust Indicates thrust;

[0124] Meteorological correction factors include:

[0125] Headwind correction factor Where c1 is the adjustment constant of EF for meteorological conditions;

[0126] Temperature correction factor Where c2 is the adjustment constant of EF with respect to temperature;

[0127] Turbulence Correction: It means that moderate turbulence increases EF by c3, and severe turbulence increases it by c4;

[0128] The calculation formula for the dynamic emission factor based on the meteorological correction factor is:

[0129] EF dyn =EF base ×(1+Δ wind )×(1+Δ temp )×(1+Δ turb ).

[0130] In this embodiment, the calculation method of total carbon emissions is further adjusted to take into account the different flight phases:

[0131] Flight phase division: The flight is divided into four phases: taxiing, climbing, cruising, and descending. Assume that the total carbon emissions of the flight are TCE, the dynamic emission factor is EF, the fuel consumption is expressed as FC, and the correction factor is expressed as CF. Calculate the carbon emissions separately:

[0132] Sliding phase: FC 滑行 = taxi time × taxi fuel flow;

[0133] Climbing phase: FC 爬升 = Climb time × Climb fuel flow × (1 + CF 爬升 );

[0134] Cruise phase: FC 巡航 =cruise time × taxi fuel flow;

[0135] Descending stage: FC 下降 = descent time × descent fuel flow × (1 + CF 下降 );

[0136] The formula for calculating total carbon emissions is:

[0137] TCE=∑(EF 动态 ×FC)=EF 动态,滑行 ×FC 滑行 +EF 动态,爬升 ×FC 爬升 +EF 动态,巡航 ×FC 巡航 +EF

[0138] 动态,下降 ×FC 下降 .

[0139] In this embodiment, the steps of converting electricity and fuel consumption into carbon footprints in the fusion and dynamic compensation module include:

[0140] A4.1. Data Preprocessing: Clean outliers (e.g., negative battery charge, zero fuel consumption), unify time granularity (e.g., monthly / quarterly aggregation), and correlate warehousing and transportation data (e.g., matching by order ID or logistics node).

[0141] A4.2. Parameter design:

[0142] E wh : Annual electricity consumption of storage facilities;

[0143] F trans : Annual fuel consumption of transportation vehicles (such as diesel, gasoline, natural gas);

[0144] EF elec : Electricity carbon emission factor, i.e., the average value of the regional power grid;

[0145] EF fuel : fuel carbon emission factor;

[0146] D: Total annual transport mileage, used to allocate fuel consumption;

[0147] η: Transport load utilization rate = actual load / maximum load;

[0148] A4.3. Dynamic configuration of parameters:

[0149] Select electricity carbon emission factor EF according to regional power grid type elec (e.g. China Southern Power Grid = 0.52kgCO2e / kWh);

[0150] Select the fuel carbon emission factor EF according to the fuel type fuel (e.g. biodiesel = 2.5kgCO2e / L);

[0151] Monitor load data through IoT devices and update the load utilization rate η of transport vehicles in real time;

[0152] A4.4. Convert electricity and fuel consumption into carbon footprint (CO2 equivalent). The calculation process includes:

[0153] A4.4.1. Calculate the carbon footprint of storage electricity: CF wh =E wh ×EF elec ;

[0154] Example: If E wh =100,000 kWh / year, EF elec =0.5kgCO2e / kWh, then:

[0155] CF wh =100,000×0.5=50,000kgCO2e / year;

[0156] A4.4.2. Calculate the carbon footprint of transportation fuels:

[0157] Example: If Ftrans = 50,000 L / year, η = 0.8, and EFfuel = 3.15 kgCO2e / L, then:

[0158]

[0159] A4.4.3. Calculate the total carbon footprint: CF total =CF wh +CF trans ;

[0160] A4.5. Model validation and optimization: Verify model accuracy by comparing historical data (error rate should be <5%); introduce machine learning to predict future carbon footprint (e.g. based on transportation volume growth trends).

[0161] In maritime and air transport scenarios, satellite trajectory, AIS, and meteorological data are collected through multiple channels. Data quality is ensured through pre-processing such as cleaning and alignment, laying the foundation for accurate analysis. For lagged maritime transport data, a regression model is constructed, taking into account sea condition coefficients. Kalman filtering and port calibration are combined to effectively compensate for lag errors and improve real-time estimation accuracy. In air transport scenarios, emission factors are dynamically adjusted, taking into account flight phases and meteorological factors, making carbon emission calculations more realistic. Warehousing electricity and transportation fuel consumption are uniformly converted into carbon footprints. Through dynamic parameter configuration, taking into account regional power grids, fuel types, and load utilization, scientific calculations are achieved, and the optimization model is verified to ensure accurate and reliable results.

[0162] The third embodiment differs from the first embodiment mainly in that the privacy protection module specifically includes:

[0163] B1. Blockchain + Zero-Knowledge Proof Storage: Enterprise data is first processed through homomorphic encryption and then generated into a ZKP, which supports selective disclosure (e.g., only proving that "carbon emissions meet a certain standard") without exposing the entire data (e.g., load parameters only prove compliance without revealing the specific value). The hash value of the carbon emissions calculation result is uploaded to the blockchain, and the original data is stored in the IPFS distributed network, requiring the enterprise's private key to authorize access.

[0164] B2. Federated Learning Privacy Computing: Build a cross-enterprise federated learning model where each enterprise only uploads encrypted gradient parameters to collaboratively optimize carbon emission prediction models (e.g., jointly train transportation route optimization algorithms), without data leaving the domain.

[0165] B3. NFTization of carbon credits: Convert emission reductions into NFTs, record metadata such as emission reduction behavior, time, and location, and automatically trade through smart contracts to improve the liquidity of carbon assets; the NFT hash value is linked to blockchain evidence to ensure that carbon credits cannot be tampered with and are traceable.

[0166] The specific operations for enterprises to upload encrypted gradient parameters include:

[0167] B2.1. Local training: Each enterprise uses its own data to train the model locally and obtain gradient parameters;

[0168] Encryption: Gradient parameters are encrypted to ensure data security during transmission. Noise is added to the gradient parameters to prevent the original data from being inferred through the model while maintaining model accuracy (e.g., adding Laplace noise).

[0169] Parameter upload: Upload the encrypted gradient parameters to the federated learning server.

[0170] Leveraging blockchain and zero-knowledge proof technology, enterprise data is homomorphically encrypted and ZKPs are generated, enabling selective disclosure and preventing raw data leaks. Carbon emission results are stored on-chain, and raw data is encrypted and stored on IPFS, ensuring data security. In a cross-enterprise federated learning model, enterprises only upload encrypted gradient parameters, and data does not leave the domain. Encryption and noise processing ensure transmission security, prevent data back-propagation, maintain model accuracy, and achieve a balance between privacy and efficiency. Carbon credits are NFT-ified to enhance carbon asset liquidity. Linking NFTs to blockchain evidence ensures that carbon credits are tamper-proof and traceable.

[0171] The fourth implementation mode mainly differs from the first implementation mode in that the work content of the intelligent audit and supervision module specifically includes:

[0172] C1. AI Anomaly Detection: Deploy AI audit robots to automatically compare multi-dimensional data such as fuel consumption, load, and GPS trajectory, flagging abnormal fluctuations (e.g., a 50% drop in load but unchanged fuel consumption), triggering manual review. Manual review then provides graded warnings based on the severity of the anomaly (e.g., Level 1 anomalies require review within 2 hours, Level 2 anomalies within 24 hours). Audit rules include static and dynamic rules. Static rules set specific thresholds for audit items, while dynamic rules create thresholds using threshold models trained with historical data.

[0173] C2. Third-party audit pool: Establish a whitelist of third-party audit institutions. Enterprises can independently select institutions for verification. The hash value of the audit report is uploaded to the chain, and the entire audit process is recorded.

[0174] The AI audit robot automatically compares multi-dimensional data, can accurately mark abnormal fluctuations and issue graded warnings, and combines static and dynamic audit rules to improve the comprehensiveness and accuracy of anomaly detection; a whitelist of third-party audit institutions is established for companies to choose from, and audit reports are uploaded to the chain and the process is recorded, which not only ensures the fairness and professionalism of the audit, but also enhances the credibility and traceability of the results, and optimizes the logistics carbon emissions review and supervision process.

[0175] At the same time, the contents not described in detail in this specification belong to the existing technology well known to those skilled in the art.

[0176] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0177] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. The logistics carbon emission data monitoring system based on blockchain and edge computing is characterized by: include: The fusion and dynamic compensation module integrates multi-source heterogeneous data on maritime / air transport satellite trajectories, port AIS data, and warehousing power and transportation fuel to build a dynamic compensation model for cross-border transport. By cross-validating satellite trajectory predictions with AIS data, a lag compensation algorithm is developed, combining historical data with fitting to estimate carbon emissions in real time. For air transport scenarios, the emission factor is dynamically adjusted by integrating aircraft model and route meteorological data. Develop logistics carbon data middleware to standardize data formats across the entire supply chain, and deploy an intelligent mapping engine to link storage electricity and transportation fuel data to the same carbon footprint model; The privacy protection module is based on blockchain and zero-knowledge proof technology. When enterprises submit data, they generate ZKP to verify the authenticity without leaking the original information. The hash value of the carbon emission calculation result is stored on the chain, and the original data is encrypted and stored in the IPFS distributed network, requiring enterprise private key authorization to access. A cross-enterprise federated learning model is built, and each enterprise only uploads encrypted gradient parameters to ensure data privacy and security. The intelligent audit and supervision module deploys AI audit robots to automatically compare multi-dimensional data, mark abnormal fluctuations, and issue warnings based on severity. The audit rules integrate static threshold and dynamic threshold models, establish a whitelist of third-party audit institutions for enterprises to choose from, and store the hash value of the audit report on the chain.

2. The logistics carbon emission data monitoring system based on blockchain and edge computing according to claim 1 is characterized by: The fusion and dynamic compensation module includes the following steps for collecting and preprocessing maritime / air transport satellite trajectories and port AIS data: A11. Satellite trajectory data acquisition: Access Sentinel-1 synthetic aperture radar satellite images to obtain the real-time position, speed and heading data of the ship's latitude and longitude, with a time resolution of Δt s minute; A12. Data cleaning: Eliminate outliers. The speed outlier elimination condition is V>V max or V <V min , fill missing values; A13, AIS data collection: Obtain AIS messages from the MarineTraffic platform and parse the vessel's MMSI number, draft, and load status fields; A14. Time alignment: Match satellite data with AIS data according to UTC timestamps. Time alignment error: ∈ t <Δt a minute.

3. The logistics carbon emission data monitoring system based on blockchain and edge computing according to claim 1 is characterized by: The design corrections of the hysteresis compensation algorithm by the fusion and dynamic compensation module include: A2.

1. Historical data fitting: Collect carbon emission data of ships on the same route over the past 12 months and build a regression model: Carbon emissions = α·V 2 +β·W+γ·C+∈; Among them, α, β, γ are regression coefficients, V is the ship speed, W is the deadweight tonnage, C is the sea state coefficient, and ε is the error term; A2.2 Calculation of sea state coefficient: Among them, H represents wave height, U represents wind speed, k1 and k2 are weight coefficients, H ref 、U ref are the reference thresholds for wave height and wind speed, respectively; A2.

3. Calculation of 72-hour lag compensation: Input current satellite / AIS data and substitute it into the regression model to predict real-time carbon emissions. The compensation formula is: Among them, E comp represents the carbon emissions after compensation, represents the model prediction value, represents the average value of lagged data; A2.

4. Error control: Validate the regression model through cross-validation algorithm and adjust the parameters to control the error MAE<∈ m ; A2.

5. For post-compensation carbon emissions, Kalman filtering is used for smoothing: a state-space model is established, integrating satellite predictions with port loading and unloading records to dynamically adjust carbon emission estimates; A2.

6. Port anchor point calibration: When the ship approaches the port, the regression model is calibrated based on the port loading and unloading data to eliminate the accumulated error; Calibration trigger condition: d port <D th , where d port Indicates the distance from the port, D th Indicates the distance threshold.

4. The logistics carbon emission data monitoring system based on blockchain and edge computing according to claim 1 is characterized by: The steps of dynamically adjusting the emission factor for the air transport scenario by the fusion and dynamic compensation module include: A3.1 Data Collection: Aircraft model data: Establish an aircraft model database, associate engine type, and fuel efficiency η parameters; Route weather data: access FlightRadar24API to obtain real-time wind speed V wind , wind direction, temperature and turbulence intensity; A3.

2. Feature conversion: Convert wind direction to angle θ with heading and calculate effective wind speed V eff =V wind ·cosθ.

5. The logistics carbon emission data monitoring system based on blockchain and edge computing according to claim 4 is characterized by: Dynamic adjustment of emission factors includes: Calculation of baseline emission factor: Calculate the baseline emission factor EF according to the ICAO standard formula: Where k' is a constant used to adjust the baseline emission factor. Indicates fuel flow, F thrust Indicates thrust; Meteorological correction factors include: Headwind correction factor Where c1 is the adjustment constant of EF for meteorological conditions; Temperature correction factor Where c2 is the adjustment constant of EF with respect to temperature; Turbulence Correction: It means that moderate turbulence increases EF by c3, and severe turbulence increases it by c4; The calculation formula for the dynamic emission factor based on the meteorological correction factor is: EF dyn =EF base ×(1+D) wind )×(1+D temp )×(1+D turb )。 6. The blockchain and edge computing-based logistics carbon emission data monitoring system according to claim 4 is characterized by: Taking into account the different flight stages, the calculation method of total carbon emissions is further adjusted: Flight phase division: The flight is divided into four phases: taxiing, climbing, cruising, and descending. Assume that the total carbon emissions of the flight are TCE, the dynamic emission factor is EF, the fuel consumption is expressed as FC, and the correction factor is expressed as CF. Calculate the carbon emissions separately: Sliding phase: FC 滑行 = taxi time × taxi fuel flow; Climbing phase: FC 爬升 = Climb time × Climb fuel flow × (1 + CF 爬升 ); Cruise phase: FC 巡航 =cruise time × taxi fuel flow; Descending stage: FC 下降 = descent time × descent fuel flow × (1 + CF 下降 ); The formula for calculating total carbon emissions is: TCE=∑(EF 动态 ×FC)=EF 动态,滑行 ×FC 滑行 +EF 动态,爬升 ×FC 爬升 +EF 动态,巡航 ×FC 巡航 +EF 动态,下降 ×FC 下降 。 7. The logistics carbon emission data monitoring system based on blockchain and edge computing according to claim 1 is characterized by: The steps of converting electricity and fuel consumption into carbon footprint in the fusion and dynamic compensation module include: A4.

1. Data preprocessing: Clean outliers, unify time granularity, and correlate warehousing and transportation data. A4.

2. Parameter design: E wh : Annual electricity consumption of storage facilities; F trans : Annual fuel consumption of transportation vehicles; EF elec : Electricity carbon emission factor, i.e., the average value of the regional power grid; EF fuel : fuel carbon emission factor; D: Total annual transport mileage, used to allocate fuel consumption; η: Transport load utilization rate = actual load / maximum load; A4.

3. Dynamic configuration of parameters: Select electricity carbon emission factor EF according to regional power grid type elec ; Select the fuel carbon emission factor EF according to the fuel type fuel ; Monitor load data through IoT devices and update the load utilization rate η of transport vehicles in real time; A4.

4. Convert electricity and fuel consumption into carbon footprints. The calculation process includes: A4.4.

1. Calculate the carbon footprint of storage electricity: CF wh =E wh ×EF elec ; A4.4.

2. Calculate the carbon footprint of transportation fuels: A4.4.

3. Calculate the total carbon footprint: CF total =CF wh +CF trans ; A4.

5. Model validation and optimization: Verify model accuracy by comparing historical data; introduce machine learning to predict future carbon footprint.

8. The logistics carbon emission data monitoring system based on blockchain and edge computing according to claim 1 is characterized by: The privacy protection module specifically includes: B1. Blockchain + Zero-Knowledge Proof Storage: Enterprise data is first processed through homomorphic encryption and then generated into a ZKP, which supports selective disclosure without exposing the entire data. The hash value of the carbon emission calculation result is uploaded to the blockchain, and the original data is stored in the IPFS distributed network, requiring the enterprise's private key to authorize access. B2. Federated Learning Privacy Computing: Build a cross-enterprise federated learning model where each enterprise only uploads encrypted gradient parameters to collaboratively optimize the carbon emissions prediction model without data leaving the domain. B3. NFTization of carbon credits: Convert emission reductions into NFTs, record metadata of emission reduction behavior, time, and location, and automatically trade through smart contracts to improve the liquidity of carbon assets; the NFT hash value is linked to blockchain evidence to ensure that carbon credits cannot be tampered with and are traceable.

9. The logistics carbon emission data monitoring system based on blockchain and edge computing according to claim 8 is characterized by: The specific operations for enterprises to upload encrypted gradient parameters include: B2.

1. Local training: Each enterprise uses its own data to train the model locally and obtain gradient parameters; Encryption: Gradient parameters are encrypted to ensure data security during transmission. Noise is added to the gradient parameters to prevent the original data from being inferred through the model while maintaining model accuracy. Parameter upload: Upload the encrypted gradient parameters to the federated learning server.

10. The logistics carbon emission data monitoring system based on blockchain and edge computing according to claim 1 is characterized by: The work content of the intelligent audit and supervision module specifically includes: C1. AI Anomaly Detection: Deploy AI audit robots to automatically compare multi-dimensional data on fuel consumption, load, and GPS trajectory, flag abnormal fluctuations, and trigger manual review. Manual review then provides graded warnings based on the severity of the anomaly. Audit rules include static and dynamic rules. Static rules set specific thresholds for audit items, while dynamic rules create thresholds using threshold models trained with historical data. C2. Third-party audit pool: Establish a whitelist of third-party audit institutions. Enterprises can independently select institutions for verification. The hash value of the audit report is uploaded to the chain, and the entire audit process is recorded.

Images