High-carbon enterprise carbon emission monitoring method and device based on electric carbon data fusion

Abstract

The invention relates to the technical field of carbon emission monitoring, and discloses a high-carbon enterprise carbon emission monitoring method and device based on electric carbon data fusion, and the key point of the technical scheme is that the method comprises the following steps: S1, obtaining the characteristics of a target enterprise, and determining the carbon emission source of the target enterprise; s2, deploying a data acquisition system according to the carbon emission source, accessing a full-dimensional data source, and acquiring target data; s3, performing data processing and integration on the acquired target data, and constructing a standardized data set; s4, constructing an electricity and carbon calculation model, training through the standardized data set, and performing carbon emission accounting through the trained electricity and carbon calculation model; and S5, correcting the trained model, outputting a monitoring result after correction is completed, and performing real-time calculation and analysis by using an advanced mathematical model through deep integration of enterprise power consumption data and multi-source production and energy data so as to realize refined and dynamic monitoring of carbon emission of high-carbon enterprises.

Description

Technical Field

[0001] This invention relates to the field of carbon emission monitoring technology, and more specifically, to a method and apparatus for monitoring carbon emissions of high-carbon enterprises based on the fusion of electrical carbon data. Background Technology

[0002] As global climate change becomes increasingly severe, reducing greenhouse gas emissions and achieving the goals of "carbon peaking and carbon neutrality" has become a global consensus. For high-carbon enterprises, which are major sources of energy consumption and carbon emissions, accurate monitoring, reporting, and verification (MRV) of their carbon emission data is fundamental to implementing effective carbon management and emission reduction measures.

[0003] "Electricity-carbon data fusion" refers to the systematic use of technical means to deeply integrate, correlate, and collaboratively calculate the electricity consumption data of high-carbon enterprises with multi-source production, operation, and energy consumption data. This establishes a precise mapping relationship between electricity consumption and carbon emissions, providing core technical support for carbon emission monitoring. Currently, carbon emission accounting for high-carbon enterprises mainly relies on traditional methods based on material balance and emission factors, failing to fully leverage the technical advantages of electricity-carbon data fusion. These methods have the following shortcomings: The use of monthly, quarterly, or even annual statistical data makes it difficult to achieve real-time or near-real-time monitoring, which is not conducive to timely detection and handling of abnormal emissions.

[0004] Relying on empirical emission factors and failing to adequately consider individual differences in enterprise production processes, equipment efficiency, and energy structure leads to low accuracy in accounting results.

[0005] Data on electricity consumption, production operations, and energy procurement within enterprises are scattered across different systems, lacking effective integration and correlation analysis, thus failing to fully tap the value of the data.

[0006] Traditional methods involve a large amount of manual recording and calculation, which is prone to errors and carries the risk of data falsification.

[0007] Therefore, it is necessary to develop a carbon emission monitoring solution that is real-time, accurate, and highly automated. Summary of the Invention

[0008] The purpose of this invention is to provide a method and device for monitoring carbon emissions of high-carbon enterprises based on the fusion of electricity and carbon data. By deeply integrating enterprise electricity consumption data with multi-source production and energy data, and using advanced mathematical models for real-time calculation and analysis, it can achieve refined and dynamic monitoring of carbon emissions of high-carbon enterprises.

[0009] The above-mentioned technical objective of the present invention is achieved through the following technical solution: a method for monitoring carbon emissions of high-carbon enterprises based on electrical carbon data fusion, comprising the following steps: S1. Obtain the characteristics of the target company and determine its carbon emission sources; S2. Based on carbon emission sources, deploy a data acquisition system, connect to data sources from all dimensions, and collect target data; S3. Process and integrate the collected target data to construct a standardized dataset; S4. Construct an electric carbon calculation model and train it using a standardized dataset. Then, use the trained electric carbon calculation model to calculate carbon emissions. S5. Correct the trained model and output the monitoring results after correction.

[0010] As a preferred embodiment of the present invention, S1 includes: S11. Obtain the complete production process of the target company's high-carbon industry and identify the types of carbon emission sources in conjunction with industry standards; S12. Based on the physical distribution, emission intensity, and monitoring priority of carbon emission sources, the plant area is divided into grids, emission intensity is quantified and graded, and monitoring priority is assigned weights. Combined with the principles of production process correlation and data measurability, independent monitoring units with clear boundaries and no overlap are formed. Based on the emission source type, accounting method requirements, and monitoring accuracy requirements, the core monitoring parameters of each independent monitoring unit are matched and determined. S13. Based on the enterprise's production scale, equipment configuration, and historical production data, the Apriori algorithm is used to filter frequent itemsets by minimum support and strong association rules by minimum confidence to mine the association rules between production processes and carbon emissions. Combined with K-means clustering analysis, the enterprise is profiled and classified.

[0011] As a preferred embodiment of the present invention, S2 includes: S21. Configure multiple types of intelligent sensor terminals; S22. Connect to the enterprise's existing systems, and access the production management system, energy management system, and distributed control system through industrial protocols to obtain corresponding data; obtain power data in real time by connecting to the power system. S23. Establish a hierarchical communication transmission architecture to store the collected data.

[0012] As a preferred embodiment of the present invention, S3 includes: S31. Use the sliding window replacement method to fill in missing data, filter high-frequency noise through fast Fourier transform, and identify and remove outliers based on the 3σ principle or the isolated forest algorithm. S32. Establish mapping relationships between power data, production data, and energy data based on key fields; uniformly convert data from different units to a standard model to form a structured data matrix; S33. An adaptive Kalman filter algorithm is used, combined with the sensor's factory calibration parameters and the on-site environmental compensation factor, to dynamically calibrate the collected data.

[0013] As a preferred embodiment of the present invention, S4 includes: S41. Collect historical data on high-carbon industries in the region where the target enterprise is located, form a time series training dataset according to the processing standard in S3, and divide it into training set and test set; S42. Construct and train the ARDL-ECM electrocarbon model; S43. Carbon emission accounting is performed using the trained ARDL-ECM electric carbon model.

[0014] As a preferred embodiment of the present invention, S42 includes: S421. Using electricity consumption data and energy consumption data as input variables and carbon emissions as output variables, determine the objective function of the ARDL model; S422. The least squares method is used to estimate the coefficients to be fitted in the model, and the lag order of the variables is determined by the AIC information criterion to obtain the basic ARDL model. S423. Analyze the residual sequence of the ARDL model, construct an error correction model, combine the two to form a complete ARDL-ECM electric carbon calculation model, use the test set to verify the model accuracy, and determine if the prediction error rate is less than the preset threshold. If it does not meet the threshold, adjust the lag order and retrain.

[0015] As a preferred embodiment of the present invention, S43 includes: S431. Input the current standardized dataset obtained in S3 into the trained ARDL-ECM model to calculate the energy consumption of each monitoring unit of the target enterprise. S432. Combining the carbon emission accounting standard library for high-carbon industries, the emission factor method is used to calculate direct carbon emissions, and the material balance method is used to calculate carbon emissions during the production process. S433. Obtain enterprise net purchased electricity data and regional electricity carbon emission factors, and calculate indirect implicit carbon emissions. S434. Subtract the carbon emissions implied by carbon sequestration products to obtain the company's final carbon emissions.

[0016] As a preferred embodiment of the present invention, S5 includes: S51. Based on the measured data from the continuous emission monitoring system and the verification data from a third party, the least squares method is used to dynamically correct the model output results to compensate for the deviations caused by fluctuations in production conditions. S52. Generate carbon emission monitoring reports according to the time dimension and upload the data to the monitoring management platform for display through heat maps and trend curves.

[0017] A carbon emission monitoring device for high-carbon enterprises based on electrical carbon data fusion includes: a processor and a memory, wherein the memory stores a computer program executable by the processor, and the processor implements the above-mentioned method when executing the computer program.

[0018] In summary, the present invention has the following beneficial effects: by deeply integrating with the power system and the enterprise's real-time production system, and combining high-frequency data acquisition and rapid model calculation, it achieves near real-time monitoring and dynamic tracking of enterprise carbon emissions, overcoming the problem of data lag in traditional methods.

[0019] The ARDL-ECM model was adopted, which can effectively process time series data, capture the long-term equilibrium and short-term dynamic relationship between variables, and improve the modeling accuracy of the relationship between electricity consumption and carbon emissions.

[0020] By integrating data from multiple sources, the limitations of a single data source are avoided, resulting in more comprehensive and accurate accounting results.

[0021] A dynamic emission factor library and model calibration mechanism have been established, which can adapt to changes in enterprise production conditions and the external environment, and continuously ensure the accuracy of accounting.

[0022] From data collection, processing, model calculation to result output and anomaly warning, the entire process is highly automated, reducing human intervention, lowering the risk of human error and data falsification, and improving the efficiency and reliability of monitoring work.

[0023] It breaks down data silos between various systems within an enterprise, enabling effective integration and correlation analysis of multi-source data. It is not only used for carbon emission accounting, but also provides data support and decision-making basis for enterprises to optimize production processes, save energy and reduce consumption, and improve overall operational efficiency. Attached Figure Description

[0024] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0025] It is readily understood that, based on the technical solution of this invention, various embodiments of the invention can be conceived by those skilled in the art without altering the essential spirit of the invention. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this invention and should not be considered as the entirety of the invention or as limitations or restrictions on the technical solution of this invention. Rather, these embodiments are provided to enable those skilled in the art to gain a more thorough understanding of the invention. Preferred embodiments of the invention are described below in conjunction with the accompanying drawings, which form part of this application and, together with the embodiments of the invention, serve to illustrate the innovative concept of the invention.

[0026] like Figure 1 As shown, this invention provides a method for monitoring carbon emissions of high-carbon enterprises based on electrical carbon data fusion, comprising the following steps: S1. Obtain the characteristics of the target enterprise and accurately identify carbon emission sources.

[0027] Specifically, S1 includes: S11. First, conduct in-depth research on the target company to identify its specific high-carbon industry segment, such as the processes of coking, sintering, blast furnace ironmaking, and converter steelmaking in the steel industry, and thoroughly analyze its complete production process. Based on relevant standards such as the "Guidelines for Greenhouse Gas Emission Accounting Methods and Reporting for Key Industry Enterprises" issued by the National Development and Reform Commission, identify the main carbon emission sources generated by the company during its production process.

[0028] Emission source classification: The identified carbon emission sources are specifically classified as follows: Sources of emissions from fossil fuel combustion include coal-fired boilers used for power supply and gas-fired kilns used for heating.

[0029] Emissions from industrial production processes include emissions from non-energy-related processes such as carbonate decomposition in the sintering process of the steel industry and limestone calcination in cement production.

[0030] Net purchased electricity for production implies emissions sources, which refers to the carbon emissions from upstream power generation implied by a company's purchase of electricity from the grid.

[0031] S12. Based on the physical distribution, emission intensity, and monitoring priority of carbon emission sources, a combination of methods including grid-based division of the plant area, quantitative grading of emission intensity, and weighting of monitoring priority, combined with the principles of production process correlation and data measurability, is used to divide the plant into independent monitoring units with clear boundaries and no overlap. Based on the emission source type, accounting method requirements, and monitoring accuracy requirements, the core monitoring parameters of each independent monitoring unit are matched and determined. For example, for combustion emission sources, fuel consumption, flue gas flow rate, and CO2 concentration need to be monitored; for production process emission sources, raw material input and product output need to be monitored.

[0032] The specific implementation method is as follows: First, based on the plant's layout, the area is divided into grids according to the locations of production workshops, equipment clusters, and emission outlets to ensure that carbon emission sources within the same grid are spatially close and related to production processes. Then, based on historical emission data and equipment rated capacity, emission sources are divided into three intensity levels: high, medium, and low, using emission per unit time and emission per unit product as indicators. Subsequently, the analytic hierarchy process (AHP) is used to calculate monitoring priority scores based on emission intensity, regulatory control requirements, and relevance to emission reduction targets, i.e., preset weight parameters, and classify them into mandatory monitoring, key monitoring, and routine monitoring levels. Finally, monitoring units are constructed according to the rule of setting up independent units for high-priority sources, merging units for medium-priority sources based on relevance, and integrating units for low-priority sources based on processes. For different types of emission sources such as combustion, production process, and implicit power generation, core monitoring parameters such as fuel consumption, flue gas parameters, carbon content of raw materials / products, and electricity consumption are matched respectively, and the accuracy standards for parameter measurement are clarified.

[0033] S13. Based on the target company's production scale, main production equipment models and configurations, energy structure, and historical production operation data and carbon emission data for the past 1 to 3 years (including the proportion of coal, gas, and electricity), data mining techniques are employed. For example, the Apriori algorithm is used to filter frequent itemsets by minimum support and strong association rules by minimum confidence to uncover the association rules between production processes and carbon emissions. Combined with clustering analysis methods, such as the K-means algorithm, the company's production operation characteristics are classified to construct a company profile, such as stable, fluctuating, intermittent, or frequently anomalous types. This profile provides an important basis for the subsequent optimization design of data collection schemes and the dynamic adjustment of monitoring model parameters.

[0034] Regarding the Apriori algorithm, historical production data is first discretized according to dimensions such as production process, equipment operating parameters, energy consumption, and carbon emissions. Minimum support and minimum confidence are set as screening thresholds. The Apriori algorithm iteratively scans the data, filtering out frequent 1-itemsets, frequent 2-itemsets, and so on, that meet the minimum support, until the highest-order frequent itemset is generated. Then, rules that meet the minimum confidence are extracted from the frequent itemsets. For example, blast furnace ironmaking process + coke consumption ≥ 50t / h → carbon emissions ≥ 100t / h, forming a strong correlation rule between production process and carbon emissions. Finally, K-means clustering analysis is combined, using the core features in the association rules as clustering indicators, such as the proportion of high-emission processes and energy consumption fluctuation coefficient, to classify enterprises into different profile categories such as stable, fluctuating, and intermittent.

[0035] S2. Deploy a data acquisition system to achieve access to data sources across all dimensions.

[0036] Specifically, S2 includes: S21. Deploy corresponding smart sensor terminals for each monitoring unit and carbon emission source identified in S1. These devices include, but are not limited to: The power data acquisition equipment is equipped with smart meters that have high-precision metering functions and support power communication protocols such as DL / T645. It can collect total power consumption, power consumption of workshops / plants, power consumption of key production equipment, and power consumption data for different time periods.

[0037] Energy consumption monitoring equipment includes flow meters and gas meters used to monitor the consumption of fossil fuels such as coal, natural gas, and heavy oil, as well as metering instruments used to monitor the consumption of secondary energy sources such as steam.

[0038] Flue gas analysis equipment, based on non-dispersive infrared spectroscopy or other advanced technologies, is used for real-time online monitoring of CO2, SO2, and NO in chimneys or exhaust stacks. x The concentration of the gas and parameters such as flue gas flow rate, temperature, and pressure.

[0039] Production process parameter monitoring equipment, such as online component analyzers for analyzing the composition of raw materials or products, X-ray fluorescence spectrometers, sensors for monitoring equipment operating load, timers for recording production time, etc.

[0040] S22. Integration with existing enterprise systems: Through industrial communication protocols such as Modbus, OPCUA, and IEC61850, the deployed intelligent sensing devices will be seamlessly integrated with the enterprise's existing production management system, energy management system, distributed control system, data acquisition and monitoring system. This will enable the automatic collection of information such as production plans, equipment operating status, raw material procurement and consumption records, product output data, and fuel lower heating value.

[0041] Deep integration with the power system enables secure and reliable data exchange with national or local power dispatch centers and power supply companies via dedicated lines or public wireless networks, such as 4G, 5G, and NB-IoT. This allows for real-time acquisition of detailed power data, including total electricity consumption, electricity consumption of each line, power factor, voltage, and current, providing high-precision, high-frequency power consumption data support for subsequent carbon model calculations.

[0042] Specifically, the data obtained from the production management system includes: raw material input, product output, production batches, process start / stop status, and equipment operating load rate. The data obtained from the energy management system includes: consumption of various fossil fuels, lower heating value of fuels, supply and consumption of secondary energy, including steam, compressed air, etc. The data acquired by the distributed control system includes: key process parameters and process gas components. Key process parameters include reaction temperature, pressure, material ratio, and valve opening. Process gas components include oxygen, carbon dioxide, etc. Connect to the power system to obtain real-time power data, including: total power consumption, power consumption by workshop / equipment, active power / reactive power, and power consumption time distribution; The core data input into the subsequent ARDL-ECM electrocarbon model includes: raw material input, product output, equipment operating load rate, consumption of various fossil fuels, lower heating value of fuels, secondary energy consumption, key process temperature / pressure, total electricity consumption, electricity consumption of each workshop / equipment, and active power.

[0043] S23. Data transmission network construction: Establish a layered and redundant data communication transmission architecture. At the equipment layer, wired or wireless communication methods such as RS-485, M-BUS, and LoRa can be used between metering equipment and field acquisition terminals. At the aggregation layer, high-speed power line carrier, fiber optic, and 4G / 5G communication methods can be used between field acquisition terminals and enterprise data centers or cloud system master stations to ensure that all-dimensional data such as power data, production data, and energy consumption data can be transmitted to the data processing center in real time, stably, and securely.

[0044] S3. Data processing and integration: building standardized, high-quality datasets.

[0045] S3 includes: S31. Data cleaning: Preprocessing the collected multi-source raw data to remove noise and outliers. Specifically, this includes: Missing values ​​can be handled by methods such as sliding window mean replacement, imputation based on time series prediction models such as ARIMA and LSTM, or reasoning completion using knowledge graphs.

[0046] Outlier detection and removal involves identifying and removing abnormal data caused by equipment failure, communication interference, human error, or other reasons by setting thresholds, based on statistical distributions, or using machine learning algorithms.

[0047] For noise filtering, for sensor data containing random noise, such as flue gas flow and concentration data, fast Fourier transform can be used for frequency domain analysis to filter out noise components at specific frequencies, or methods such as moving average and Kalman filtering can be used for smoothing.

[0048] S32, Data Association and Standardization.

[0049] Data association uses key fields such as unified timestamps, equipment numbers, and monitoring unit IDs as links to connect and match scattered data from different data sources and different devices, establishing mapping relationships between data. For example, it can link data such as blast furnace electricity consumption, iron ore consumption, coke consumption, gas production, and CO2 emissions within a certain time period.

[0050] Data standardization involves converting data in different formats and units into a unified standard data model. For example, standardizing the unit for gas concentration to mg / m³. 3 Alternatively, ppm can be used to standardize energy consumption to kWh or tons of standard coal, temperature to ℃, and pressure to MPa, forming a structured data matrix.

[0051] S33. Data calibration: To ensure the accuracy and reliability of the data, the cleaned and standardized data are further calibrated based on the requirements of the MRV system.

[0052] Equipment calibration involves establishing an initial calibration model using the factory calibration parameters of the sensor equipment and periodic field calibration data, and making preliminary corrections to the collected data.

[0053] Cross-calibration involves comparing data from the same source (such as CO2 analyzer readings from different locations on the same flue) using different measurement principles or locations, and employing an adaptive weighted fusion algorithm for mutual calibration to improve measurement accuracy.

[0054] Model calibration utilizes the error distribution patterns between historical monitoring data and authoritative third-party verification data or high-precision instrument measured data to establish a dynamic error compensation model, and performs final precise calibration of the data to ensure that the data error rate is controlled within a preset range, such as ≤2%.

[0055] After the above processing, all high-quality, standardized data are organized according to dimensions such as time series or monitoring units to construct a unified, standardized dataset that can be used by the model.

[0056] S4. Construction of the electric carbon calculation model and carbon emission accounting.

[0057] S4 includes: S41. Construct a standard library for carbon emission accounting in high-carbon industries and prepare for model training. Standards integration involves systematically reviewing and integrating internationally accepted accounting standards such as the ISO 14064 series of standards from the International Organization for Standardization, the IPCC National Greenhouse Gas Inventory Guidelines, the GHG Protocol from the World Resources Institute and the World Business Council for Sustainable Development, and the MRV regulations of the European Union Emissions Trading System.

[0058] The system integrates domestic standards, deeply incorporating the "Guidelines for Greenhouse Gas Emission Accounting Methods and Reporting for Enterprises in 24 Industries" issued by the National Development and Reform Commission, carbon emission verification technical specifications issued by various localities, and special accounting guidelines issued by relevant industry associations, among other domestic accounting requirements.

[0059] A dynamic emission factor database is established and maintained, containing emission factors from the combustion of different types of fossil fuels, process emission factors from different industrial production processes, and electricity emission factors from different regions and types of power grids. This database can be updated regularly or in real time based on changes in regional energy structure, technological advancements, and policy updates.

[0060] The accounting boundaries are defined to clearly define the accounting boundaries for various types of carbon emissions, such as the boundaries for emissions from fuel combustion, emissions from industrial production processes, emissions from net purchased electricity and heat, and the deduction boundaries for emissions implied by carbon sequestration products, so as to ensure the accuracy of the accounting scope.

[0061] S42. Train the electric carbon calculation model, namely the ARDL-ECM model; Historical data preparation involves collecting historical data from multiple high-carbon enterprises of the same type within the target company's region for more than three years. The core data includes: electricity consumption data (total electricity consumption, electricity consumption by process, active power of key equipment, and distribution of electricity consumption during different time periods), energy consumption data (consumption of fossil fuels such as coal, natural gas, and heavy oil, lower heating value of fuels, and consumption of secondary energy such as steam / compressed air), key production parameters (raw material input, product output, and core process temperature / pressure), and corresponding comprehensive carbon emission data (direct carbon emissions, carbon emissions during production processes, and indirect carbon emissions) verified by a third party or recognized by the industry. This ensures that the data covers the input and output dimensions required for model training.

[0062] The data preprocessing process involves rigorous S3 data cleaning, association standardization, and dynamic calibration of collected historical data. This includes filling in missing data using a sliding window replacement method, filtering noise using Fast Fourier Transform, removing outliers using the 3σ principle / isolation forest algorithm, establishing multi-source data mapping relationships based on key fields, unifying data units to form a structured matrix, and finally performing dynamic calibration using an adaptive Kalman filter algorithm to create a high-quality time series dataset. The training and test sets are then divided in a 7:3 or 8:2 ratio.

[0063] The model structure was determined, and the Autoregressive Distributed Lag-Error Correction Model (ARDL-ECM) was selected as the core electric carbon calculation model. This model can effectively handle non-stationary time series data, and accurately capture the long-term equilibrium relationship between electricity consumption, energy consumption, production parameters and carbon emissions, as well as the dynamic adjustment effect brought about by short-term operating condition fluctuations, thus adapting to the carbon emission modeling needs of high-carbon enterprises in complex production scenarios.

[0064] Model parameter estimation and lag order determination: The objective function of the ARDL model is defined, and the pre-processed electricity consumption, energy consumption, and key production parameters are used as input variables, while the total carbon emissions are used as the output variable. The objective function of the ARDL model is constructed to quantify the relationship between multi-dimensional inputs and carbon emissions. Coefficient estimation: Based on the training set data, the least squares method is used to solve the coefficients to be fitted for each input variable in the model, and the weights of different variables on carbon emissions are clarified. Lag order optimization involves iterating through a preset range of lag orders using the Akaike information criterion or the Schwarz information criterion, selecting the combination that minimizes the criterion value as the optimal lag order, balancing the model fitting accuracy and complexity, and obtaining the basic ARDL model.

[0065] The error correction term is integrated, and the stationarity of the residual sequence of the basic ARDL model is tested. If the residual sequence is non-stationary, it indicates that the model has not fully captured the short-term dynamic relationships of the variables. An error correction term is constructed based on this residual sequence to quantify the short-term adjustment strength when the variables deviate from the long-term equilibrium state. The basic ARDL model and the error correction term are integrated to form a complete ARDL-ECM electric carbon calculation model that has both the ability to characterize long-term correlations and adapt to short-term dynamics.

[0066] Model validation and optimization involve inputting test set data into the complete ARDL-ECM model and calculating the error rate between predicted and actual carbon emissions. Root mean square error (RMSE) and mean absolute percentage error (MAPE) are used as core evaluation indicators. If the error rate exceeds a preset threshold, the process returns to adjusting the input variable selection range or lag order, and parameter estimation and model building are repeated until the model's prediction accuracy meets monitoring requirements.

[0067] S43, Carbon Emission Accounting.

[0068] For input data preparation, the target enterprise's current standardized dataset, obtained through S3 processing, including real-time or near-real-time electricity consumption data, production data, energy consumption data, etc., is organized according to the format required by the model and used as the input to the model.

[0069] The model computation involves inputting the prepared input data into the pre-trained and validated ARDL-ECM electricity carbon calculation model. Based on the learned complex relationship between electricity consumption, production activities, and carbon emissions, the model calculates the energy consumption or direct carbon emissions of each monitoring unit of the target enterprise within the corresponding time period.

[0070] Multidimensional emission calculation: Direct carbon emissions are calculated based on the energy consumption output from the model, combined with the corresponding fuel combustion emission factors in the carbon emission accounting standard library in S41, using the emission factor method to calculate the direct carbon emissions of each monitoring unit. The formula can be expressed as: Direct carbon emissions = Σ(consumption of the i-th fossil fuel × emission factor of the i-th fuel).

[0071] For industrial production processes where emissions cannot be accurately reflected directly by electricity data, such as limestone decomposition in cement production, the carbon emission calculation method is used. The formula can be expressed as: Carbon emissions from production process = (Raw material input × Raw material carbon content - Product output × Product carbon content - Waste output × Waste carbon content) × (44 / 12), where 44 / 12 is the molar mass ratio of CO2 to C.

[0072] Indirect carbon emissions are calculated by obtaining the target company's net electricity purchase data (electricity purchased from the grid minus the electricity supplied to the grid) and the corresponding regional grid average carbon emission factor. The formula can be expressed as: Indirect carbon emissions = Net electricity purchase × Regional grid carbon emission factor.

[0073] The implicit emissions of carbon sequestration products need to be deducted. If a company generates or consumes products with carbon sequestration properties during its production process (such as certain chemical products, biomass fuels, etc.), then the amount of carbon sequestration needs to be calculated and deducted. The formula can be expressed as: Implicit carbon emissions of carbon sequestration products = Σ(output of the s-th type of carbon sequestration product × carbon emission factor of the s-th type of carbon sequestration product).

[0074] Total carbon emissions are calculated by adding the direct carbon emissions, production process carbon emissions, and indirect carbon emissions obtained from the above calculations, and then subtracting the carbon emissions implied by carbon sequestration products. The final result is the target company's total carbon emissions during the accounting period. The formula can be expressed as: Total carbon emissions = Direct carbon emissions + Production process carbon emissions + Indirect carbon emissions - Carbon emissions implied by carbon sequestration products.

[0075] S5. Output of model calibration and monitoring results.

[0076] S5 includes: S51. Dynamic model calibration: In order to ensure the accuracy of long-term monitoring, the model needs to be continuously calibrated dynamically.

[0077] Real-time correction utilizes measured data from continuous emission monitoring systems deployed at key emission points, or periodic third-party on-site verification data, as the true values. Adaptive algorithms such as Kalman filtering and least squares are used to dynamically correct the output of the electrocarbon calculation model in real time or periodically to correct model deviations caused by factors such as equipment aging, operating condition drift, and environmental changes.

[0078] Parameter updates involve regularly collecting information on changes in the target company's production processes, equipment upgrades and modifications, and energy structure adjustments. Based on this information, the emission factors, model input variables, or model parameters in the carbon emission accounting standard library are updated in a timely manner to ensure the model's adaptability and calculation accuracy.

[0079] S52, Visualization of monitoring results and generation of reports.

[0080] Data visualization displays the final carbon emission accounting results, as well as key process data such as real-time electricity consumption, energy consumption of major equipment, and concentrations of major pollutants, through a carbon emission monitoring and management platform. Various formats can be used, including dashboards, trend charts, heat maps, carbon-electricity intensity analysis charts for different industries, and comparison charts of carbon-electricity intensity between enterprises and industries, to intuitively present the enterprise's carbon emission status.

[0081] The system generates multi-dimensional reports, automatically producing carbon emission monitoring reports based on different time and spatial dimensions. Report content should include carbon emission amounts, carbon emission intensity, carbon emission trend analysis, identification results of high-emission pathways, and carbon reduction potential recommendations based on data analysis.

[0082] S53. Anomaly Warning and Tracing; The warning threshold setting is based on the company's historical data, industry benchmarks, and the company profile built in S1. Dynamic carbon emission warning thresholds are set for different monitoring units and emission indicators, including instantaneous emission thresholds and cumulative emission thresholds.

[0083] Anomaly detection and alarm: The system monitors changes in carbon emission data in real time. When the monitored value exceeds the set warning threshold or abnormal emission fluctuations occur, the system can automatically identify the anomaly and issue warning notifications to relevant enterprise managers and government regulatory departments through various means.

[0084] Historical data tracing: The system possesses comprehensive data storage and tracing capabilities, capable of fully recording all raw data, processed data, model calculation processes and results, and alarm information. Users can easily query historical data for any time period, providing reliable data support for carbon emission verification, auditing, dispute resolution, and evaluation of the effectiveness of emission reduction measures.

[0085] Taking a large integrated steel enterprise as an example, the specific implementation process of the present invention is explained.

[0086] S1. Identify the target company's characteristics and accurately pinpoint carbon emission sources. This steel company comprises major production processes including coking, sintering, blast furnace ironmaking, converter steelmaking, hot rolling, and cold rolling. Through research and analysis, its main carbon emission sources were identified as: coke oven gas combustion, sintering machine fuel combustion and process emissions, blast furnace gas combustion, converter gas combustion, and purchased electricity. Based on these emission sources, the company was divided into multiple monitoring units, including a coking plant, a sintering plant, an ironmaking plant, and a steelmaking plant. Based on its historical production data, analysis showed that the company's production operations were relatively stable, but carbon emissions from the sintering and blast furnace processes fluctuated significantly, classifying it as a "basically stable but key process fluctuating" company.

[0087] S2. Deploy a data acquisition system to achieve access to all data sources. This involves deploying smart meters, gas meters, steam flow meters, and CEMS (Chemical Emission Monitoring System) on various branch plants, key production workshops, and equipment. Through the OPCUA protocol, connect to the company's MES and DCS systems to acquire data on molten iron production, molten steel production, sintered ore production, coke consumption, and various types of gas consumption. Connect via a dedicated line to the State Grid system to obtain real-time data on the company's total electricity consumption at different times and the electricity consumption of each branch plant.

[0088] S3. Data Processing and Integration: Construct a standardized, high-quality dataset. Clean the collected real-time data, removing abnormal fluctuations caused by sensor malfunctions. Utilize enterprise production plans and equipment operation logs to appropriately fill in missing production parameter data. Convert all data to standard units (such as electricity (kWh), gas volume (m³), temperature (°C), etc.), and perform time alignment and aggregation at 15-minute intervals to form a standardized dataset.

[0089] S4. Construction of the Electricity-Carbon Emission Calculation Model and Carbon Emission Accounting: Historical electricity consumption, energy consumption, output, and third-party verified carbon emission data for the past five years were collected from this steel company and several other steel companies of similar size in China. The training and test sets were divided in a 7:3 ratio. The ARDL-ECM model was trained, with input variables including: electricity consumption, coke consumption, blast furnace gas consumption, sinter output, and molten iron output for each branch plant; the output variable was the carbon emission of each branch plant / process. After model training, the prediction error rate was verified on the test set, reaching 3.8%, meeting the accuracy requirements. In actual monitoring, the current data processed in S3 was input into the model to calculate the direct carbon emissions of each branch plant. For carbonate decomposition emissions during the sintering process, supplementary calculations were performed using a material balance algorithm. Simultaneously, indirect carbon emissions were calculated based on purchased electricity and regional power grid emission factors. Finally, the total carbon emissions of the enterprise were obtained.

[0090] S5. Model Calibration and Monitoring Result Output: Utilizing measured CO2 concentration and flue gas flow data from the CEMS systems of each branch plant, the carbon emissions calculated by the model for each branch plant are compared and calibrated periodically (e.g., weekly), and model parameters are adjusted accordingly. The real-time carbon emissions, cumulative carbon emissions, and carbon emission intensity per unit product for each branch plant and major process are displayed on the large screen in the company's carbon emission monitoring center, showing historical trends in line graphs, bar charts, and other formats. When the carbon emissions of a certain process suddenly increase beyond the warning threshold, the system automatically sends SMS warnings to the environmental protection manager and workshop director, and marks the abnormal period and possible causes on the platform for rapid investigation and handling.

[0091] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this invention should be included within the protection scope of this invention.

[0092] It should be understood that, in order to simplify the present invention and help those skilled in the art understand its various aspects, in the above description of exemplary embodiments of the present invention, various features of the present invention are sometimes described in a single embodiment or with reference to a single figure. However, the present invention should not be construed as including all features included in the exemplary embodiments as essential technical features of the claims of this patent.

[0093] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0094] It should be understood that the modules, units, components, etc., included in the device of one embodiment of the present invention can be adaptively changed to be placed in a device different from that embodiment. Different modules, units, or components included in the device of the embodiment can be combined into a single module, unit, or component, or they can be divided into multiple sub-modules, sub-units, or sub-components.

[0095] The modules, units, or components in the embodiments of the present invention can be implemented in hardware, in software running on one or more processors, or in a combination thereof. Those skilled in the art should understand that... In practice, microprocessors or digital signal processors (DSPs) can be used to implement embodiments of the invention. The invention can also be implemented on computer program products or computer-readable media for performing some or all of the methods described herein.

Claims

1. A method for monitoring carbon emissions of high-carbon enterprises based on electrical carbon data fusion, characterized by: Includes the following steps: S1. Obtain the characteristics of the target company and determine its carbon emission sources; S2. Based on carbon emission sources, deploy a data acquisition system, connect to data sources from all dimensions, and collect target data; S3. Process and integrate the collected target data to construct a standardized dataset; S4. Construct an electric carbon calculation model and train it using a standardized dataset. Then, use the trained electric carbon calculation model to calculate carbon emissions. S5. Correct the trained model and output the monitoring results after correction.

2. The method for monitoring carbon emissions of high-carbon enterprises based on electrical carbon data fusion according to claim 1, characterized in that: S1 includes: S11. Obtain the complete production process of the target company's high-carbon industry and identify the types of carbon emission sources in conjunction with industry standards; S12. Based on the physical distribution, emission intensity, and monitoring priority of carbon emission sources, the plant area is divided into grids, emission intensity is quantified and graded, and monitoring priority is assigned weights. Combined with the principles of production process correlation and data measurability, independent monitoring units with clear boundaries and no overlap are formed. Based on the emission source type, accounting method requirements, and monitoring accuracy requirements, the core monitoring parameters of each independent monitoring unit are matched and determined. S13. Based on the enterprise's production scale, equipment configuration, and historical production data, the Apriori algorithm is used to filter frequent itemsets by minimum support and strong association rules by minimum confidence to mine the association rules between production processes and carbon emissions. Combined with K-means clustering analysis, the enterprise is profiled and classified.

3. The method for monitoring carbon emissions of high-carbon enterprises based on electrical carbon data fusion according to claim 2, characterized in that: S2 includes: S21. Configure multiple types of intelligent sensor terminals; S22. Connect to the enterprise's existing systems, and access the production management system, energy management system, and distributed control system through industrial protocols to obtain corresponding data; obtain power data in real time by connecting to the power system. S23. Establish a hierarchical communication transmission architecture to store the collected data.

4. The method for monitoring carbon emissions of high-carbon enterprises based on electrical carbon data fusion according to claim 3, characterized in that: S3 includes: S31. Use the sliding window replacement method to fill in missing data, filter high-frequency noise through fast Fourier transform, and identify and remove outliers based on the 3σ principle or the isolated forest algorithm. S32. Establish mapping relationships between power data, production data, and energy data based on key fields; uniformly convert data from different units to a standard model to form a structured data matrix; S33. An adaptive Kalman filter algorithm is used, combined with the sensor's factory calibration parameters and the on-site environmental compensation factor, to dynamically calibrate the collected data.

5. The method for monitoring carbon emissions of high-carbon enterprises based on electrical carbon data fusion according to claim 4, characterized in that: S4 includes: S41. Collect historical data on high-carbon industries in the region where the target enterprise is located, form a time series training dataset according to the processing standard of S3, and divide it into training set and test set; S42. Construct and train the ARDL-ECM electrocarbon model; S43. Carbon emission accounting is performed using the trained ARDL-ECM electric carbon model.

6. The method for monitoring carbon emissions of high-carbon enterprises based on electrical carbon data fusion according to claim 5, characterized in that: S42 includes: S421. Using electricity consumption data and energy consumption data as input variables and carbon emissions as output variables, determine the objective function of the ARDL model. S422. The least squares method is used to estimate the coefficients to be fitted in the model, and the lag order of the variables is determined by the AIC information criterion to obtain the basic ARDL model. S423. Analyze the residual sequence of the ARDL model, construct an error correction model, combine the two to form a complete ARDL-ECM electric carbon calculation model, use the test set to verify the model accuracy, and determine if the prediction error rate is less than the preset threshold. If it does not meet the threshold, adjust the lag order and retrain.

7. The method for monitoring carbon emissions of high-carbon enterprises based on electrical carbon data fusion according to claim 6, characterized in that: S43 includes: S431. Input the current standardized dataset obtained in S3 into the trained ARDL-ECM model to calculate the energy consumption of each monitoring unit of the target enterprise. S432. Combining the carbon emission accounting standard library for high-carbon industries, the emission factor method is used to calculate direct carbon emissions, and the material balance method is used to calculate carbon emissions during the production process. S433. Obtain enterprise net purchased electricity data and regional electricity carbon emission factors, and calculate indirect implicit carbon emissions. S434. Subtract the carbon emissions implied by carbon sequestration products to obtain the company's final carbon emissions.

8. The method for monitoring carbon emissions of high-carbon enterprises based on electrical carbon data fusion according to claim 7, characterized in that: S5 includes: S51. Based on the measured data from the continuous emission monitoring system and the verification data from a third party, the least squares method is used to dynamically correct the model output results to compensate for the deviations caused by fluctuations in production conditions. S52. Generate carbon emission monitoring reports according to the time dimension and upload the data to the monitoring management platform for display through heat maps and trend curves.

9. A carbon emission monitoring device for high-carbon enterprises based on electrical carbon data fusion, characterized in that: include: A processor and a memory, the memory storing a computer program executable by the processor, wherein the processor, when executing the computer program, implements the method of any one of claims 1-8.

Images