A big data management and accounting system for the carbon footprint of a heat treatment process full life cycle

Abstract

This invention discloses a big data management and accounting system for the entire lifecycle of carbon footprint in heat treatment processes, belonging to the interdisciplinary field of industrial internet, environmental big data, and intelligent manufacturing technologies. The system adopts a layered distributed architecture, including a sensing and acquisition layer, a data fusion layer, a core algorithm layer, an application service layer, and a user interaction layer. The sensing and acquisition layer collects data in real time from multiple heterogeneous data sources throughout the entire lifecycle of heat treatment. The data fusion layer cleans, verifies, correlates, and stores the multi-source data, constructing a unified, high-quality carbon footprint database. The core algorithm layer achieves high-precision accounting, dynamic tracking, and intelligent optimization of the carbon footprint of heat treatment. This invention solves the problems of incomplete boundaries, insufficient accuracy, data silos, and low levels of intelligence in traditional carbon footprint accounting, effectively supporting the refined carbon management and green, low-carbon transformation of heat treatment enterprises.

Description

Technical Field

[0001] This invention relates to the intersection of computer data processing and industrial low-carbon technology, and in particular to a big data management and accounting system for the entire life cycle of carbon footprint of heat treatment processes. Background Technology

[0002] Heat treatment is a key fundamental process in the equipment manufacturing industry chain, encompassing core steps such as quenching, tempering, annealing, and carburizing. The production process involves significant energy consumption and greenhouse gas emissions, making it a key area for carbon emission reduction control in the industrial sector. With the advancement of the "dual carbon" goals, the need for accurate accounting, full life-cycle traceability, and refined management of the carbon footprint of heat treatment processes is becoming increasingly urgent, but existing technologies have significant limitations.

[0003] Traditional carbon footprint accounting for heat treatment relies heavily on manual statistics, with limited data sources covering only direct energy data such as electricity and natural gas consumption. It fails to incorporate the entire lifecycle, including upstream carbon emissions from raw materials, auxiliary material consumption, equipment operation and maintenance, waste disposal, product transportation, and end-of-life recycling. This incomplete accounting boundary leads to significant biases in carbon footprint results. Furthermore, heat treatment process parameters are complex, with significant differences in energy consumption and emission characteristics across different furnace types, process curves, furnace loads, and workpiece materials. Traditional static accounting, using fixed emission factors, cannot match real-time process conditions and dynamic production scenarios, resulting in insufficient accuracy.

[0004] At the data management level, heat treatment production data is scattered across multiple systems, including PLC control systems, energy metering instruments, ERP, MES, and supply chain management. The data formats are heterogeneous, storage is independent, and there is a lack of unified standard interfaces and efficient integration mechanisms, creating data silos that make it difficult to support correlation analysis of the entire lifecycle carbon footprint. Some existing systems only perform simple carbon data aggregation, lacking the capabilities for big data cleaning, noise reduction, correlation, and quality verification. Missing or abnormal data cannot be intelligently corrected, affecting the reliability of accounting.

[0005] At the algorithmic level, current carbon accounting mostly relies on basic formulas for calculation, without building dedicated models that incorporate the characteristics of heat treatment processes. This makes it impossible to handle multivariate coupling, nonlinear emission relationships, or achieve dynamic prediction of carbon footprint, intelligent optimization of emission reduction pathways, and uncertainty analysis. Furthermore, existing systems lack full lifecycle visualization and management, carbon data traceability, emission reduction effect evaluation, and decision support functions, making it difficult to meet the needs of heat treatment enterprises for refined carbon management and green production.

[0006] In summary, existing technologies cannot achieve high-precision, dynamic, and intelligent big data management and accounting of the carbon footprint of thermal treatment throughout its entire lifecycle. A system solution that integrates multi-source data, specialized algorithms, and full-process control is needed. Summary of the Invention

[0007] This invention discloses a big data management and accounting system for the entire lifecycle of carbon footprint in heat treatment processes. It can effectively solve the problems in the background technology. By constructing a multi-source heterogeneous data fusion system, a carbon footprint accounting algorithm model specifically for heat treatment, a big data management engine for the entire lifecycle, and an intelligent decision-making module, this invention achieves accurate accounting, dynamic tracking, visual management, and intelligent optimization of the carbon footprint of the entire heat treatment process, from raw material acquisition, production and manufacturing, auxiliary material consumption, equipment use, transportation and distribution to recycling and waste. It solves the problems of incomplete accounting boundaries, insufficient accuracy, data silos, and low level of intelligence in traditional accounting methods.

[0008] A big data management and accounting system for the entire lifecycle of carbon footprint of heat treatment processes is proposed. It adopts a layered distributed architecture, consisting of a sensing and acquisition layer, a data fusion layer, a core algorithm layer, an application service layer, and a user interaction layer. Each layer works together through standardized interfaces.

[0009] Preferably, the sensing and acquisition layer interfaces with various data sources throughout the entire lifecycle of heat treatment, including real-time process parameters and energy metering data collected by workshop PLCs and DCS systems, equipment operating status and auxiliary material consumption data collected by IoT sensors, production plans, workpiece materials, furnace loading quantities, and batch information synchronized by ERP and MES systems, upstream carbon emission data of raw materials uploaded by the supply chain system, transportation mileage and transportation mode data from the logistics management system, and end-stage data from waste disposal ledgers and recycling records. The acquisition layer supports multi-protocol access and uses an edge computing gateway to perform preliminary data cleaning and standardization preprocessing, reducing the transmission pressure on the cloud.

[0010] Preferably, the data fusion layer is built on the Hadoop and Spark ecosystem and includes structured and unstructured databases such as a raw carbon footprint database, a process parameter database, an emission factor database, a life cycle basic database, and a quality verification database. The core of the data fusion layer is to achieve multi-source heterogeneous data fusion. It uses ETL tools to complete data extraction, transformation, and loading, establishes unified data encoding rules and association mapping relationships, and solves the problems of differences in data formats, dimensions, and definitions across different systems. It also includes a data quality verification module, which sets up algorithms for outlier detection, missing value imputation, duplicate data removal, and logical consistency verification.

[0011] Preferably, the core algorithm layer serves as the system's central hub, integrating five core algorithms: a carbon footprint accounting algorithm for the entire lifecycle of heat treatment, a dynamic emission factor optimization algorithm, a multivariate coupled carbon analysis algorithm, a carbon footprint prediction algorithm, and an emission reduction path optimization algorithm. These algorithms are custom-built to suit the characteristics of heat treatment processes.

[0012] Preferably, the application service layer provides modular services based on the core algorithm output results, including full life cycle carbon footprint accounting, carbon data traceability, process carbon emission analysis, emission reduction effect evaluation, automatic carbon report generation, and carbon early warning and control, covering the entire process of accounting, management, analysis, and decision-making.

[0013] The user interaction layer provides a visual operating interface for enterprise managers, process engineers, and environmental protection personnel through web terminals, mobile terminals, and workshop large screens, supporting data query, chart display, parameter configuration, and process approval functions.

[0014] Preferably, the carbon footprint accounting algorithm for the entire life cycle of heat treatment is combined with the boundary of the entire heat treatment process to construct a hierarchical accounting model, which divides the carbon footprint into four categories: direct emissions, energy indirect emissions, raw material indirect emissions, and other indirect emissions, so as to achieve full life cycle coverage accounting.

[0015] The core formula of the algorithm is: ; in, For the total carbon footprint over the entire life cycle, Direct carbon emissions include emissions from fuel combustion and process chemical reactions. Indirect emissions due to purchased electricity and heat This is due to hidden emissions from upstream raw materials and auxiliary materials. Emissions from transportation, waste disposal, equipment operation and maintenance, and recycling processes.

[0016] Direct carbon emission accounting focuses on fuel combustion and process reactions in heat treatment furnaces, using a combination of measured data and industry coefficients. ; Fuel consumption is collected in real time by energy meters. To correspond to fuel emission factors and distinguish between lower heating value and measured calorific value, These are parameters related to emissions from the process (carburizing carbon potential, decomposition rate). The emission coefficient is determined based on different materials and process types.

[0017] Indirect energy emissions focus on purchased electricity and heat: ; , Electricity and heat consumption Using the regional power grid's annual dynamic emission factor, A heat supply benchmark factor is adopted to replace the traditional fixed value, thereby improving the accuracy of the calculation.

[0018] Indirect emissions of raw materials cover the entire upstream stage of the life cycle, through List association: ; For the first Raw material usage For the first dosage of similar excipients , To address the hidden emission factors upstream of the materials, we integrate emission data from the entire process of raw material mining, processing, and transportation.

[0019] Other indirect emissions are accounted for by category: transportation is calculated based on transportation mode, mileage, and load; waste treatment is calculated based on the type of waste, treatment volume, and treatment method (landfill, incineration, recycling) corresponding to the emission factors; and equipment operation and maintenance is calculated based on maintenance consumables and repair energy consumption, so as to achieve full life cycle accounting without omission.

[0020] Preferably, the dynamic emission factor optimization algorithm combines real-time process data with machine learning to achieve dynamic optimization of emission factors.

[0021] The algorithm uses heat treatment furnace type, process type, workpiece material, furnace loading, temperature curve, running time, and equipment aging degree as input features to construct a random forest regression model and correct the basic emission factors.

[0022] First, a massive amount of historical data is collected, including actual energy consumption, emission monitoring values ​​and corresponding process parameters under different operating conditions. Key influencing factors are extracted through feature engineering, redundant features are removed, and then a nonlinear mapping relationship between process parameters and emission coefficients is established through model training to obtain the correction coefficient k under different scenarios.

[0023] Formula for calculating dynamic emission factor: ; As the industry benchmark emission factor, This is a dynamic correction factor, determined by the process type. Material Furnace loading capacity Temperature curve runtime Equipment aging degree The decision is made jointly. The system collects process parameters in real time and substitutes them into the model for calculation. Value, updated in real time This ensures that the accounting results match the real-time production status.

[0024] Preferably, the multivariate coupled carbon analysis algorithm is based on the combination of the analytic hierarchy process and the partial least squares method to construct a multivariate attribution model.

[0025] First, the weights of each influencing factor are determined through hierarchical analysis. The target layer is then decomposed into a criterion layer and an indicator layer. The weights of each indicator are determined by combining expert scoring and data statistics. The target layer is the total carbon footprint, the criterion layer includes process parameters, equipment performance, material properties, energy structure, and management level, and the indicator layer includes temperature, holding time, furnace efficiency, carbon content of materials, energy type, and furnace loading rate.

[0026] Then, partial least squares analysis was used to analyze the correlation and contribution of each variable to the carbon footprint, and multicollinearity interference was eliminated to obtain the influence coefficient of each factor on the carbon footprint. Calculation formula: ; For each influencing variable, This represents random error. The algorithm automatically outputs the ranking of the contributions of each variable, identifies key influencing factors in carbon footprint, and provides precise direction for emission reduction.

[0027] Preferably, the carbon footprint prediction algorithm is based on an LSTM neural network to construct a prediction model and process the nonlinearity and hysteresis characteristics of the thermal treatment carbon footprint time series data.

[0028] Using historical carbon footprint time-series data, production plans, preset process parameters, energy prices, and emission policies as inputs, the model captures long-term dependencies through the memory units of an LSTM network to uncover the inherent patterns in the time-series data. During model training, the model is input with carbon footprint data from previous periods and related variables, and outputs the predicted carbon footprint value for the next period. An attention mechanism is also introduced to strengthen the weight allocation of key process nodes, thereby improving prediction accuracy.

[0029] The prediction formula simplifies to: ; To predict carbon footprint for the next moment, For the historical carbon footprint of the previous m moments, These are the relevant variables for the process and production at the corresponding time point. The system supports both short-term and long-term forecasting.

[0030] Preferably, the emission reduction path optimization algorithm is based on carbon footprint accounting and attribution results. This algorithm constructs a multi-objective optimization model with the objectives of maximizing carbon emission reduction, minimizing production costs, and optimizing production efficiency. Combined with the constraints of heat treatment process, the optimal emission reduction path is solved by a genetic algorithm.

[0031] The algorithm encoding transforms decision variables such as process parameters, equipment parameters, energy structure, and furnace loading method into chromosomes. It iterative optimization is achieved through selection, crossover, and mutation operations. The fitness function comprehensively considers carbon emission reduction benefits, cost changes, and process compliance.

[0032] Fitness function: ; For the combination of decision variables, For carbon emission reductions, Due to cost changes, The production efficiency coefficient. , , These are weighting coefficients, which can be adjusted according to the company's needs. The algorithm outputs multiple optimization schemes, including process curve optimization, equipment modification, energy structure adjustment, and furnace loading method optimization, and evaluates the emission reduction effect, cost input, and implementation difficulty of each scheme.

[0033] The advantages of this invention compared to the prior art are: 1. It has full coverage of accounting boundaries and high accuracy, breaking through traditional limitations. It covers the entire life cycle of heat treatment and combines dynamic emission factors with dedicated algorithms.

[0034] 2. Break down data silos across multiple systems to achieve automatic data collection, cleaning, correlation, and traceability, reducing manual intervention, enabling efficient data fusion and intelligent management, and lowering management costs.

[0035] 3. By adopting the LSTM model to achieve accurate prediction and combining it with the early warning mechanism, the process shifts from post-event accounting to pre-event prediction and in-event control, dynamic prediction and proactive control, helping enterprises to reduce emissions in advance.

[0036] 4. Employing a multi-objective optimization algorithm, it outputs emission reduction solutions that fit actual production conditions, taking into account emission reduction, cost, and efficiency. It can directly guide process optimization and equipment transformation, making emission reduction decisions scientific and highly implementable.

[0037] 5. The entire process is visualized and easy to use. The visual interface intuitively displays carbon footprint data, is easy to operate, and is suitable for use by personnel in different positions, supporting enterprises' green production decisions. Attached Figure Description

[0038] Figure 1 This is a schematic diagram of the overall system architecture of the present invention.

[0039] Figure 2 This is a flowchart of the carbon footprint accounting algorithm for the entire life cycle of heat treatment in this invention.

[0040] Figure 3 This is a structural diagram of the dynamic emission factor optimization algorithm model of the present invention.

[0041] Figure 4 This is a flowchart illustrating the execution of the emission reduction path optimization algorithm of this invention. Detailed Implementation

[0042] The technical solution of the present invention will be further described in detail below through embodiments and in conjunction with the accompanying drawings. Many specific details are set forth in the following description to provide a full understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of the invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0043] Implementation, for example Figures 1-4 As shown, the system adopts a layered distributed architecture, consisting of a perception and acquisition layer, a data fusion layer, a core algorithm layer, an application service layer, and a user interaction layer. Each layer works together through standardized interfaces.

[0044] In one optional embodiment of this invention, the sensing and acquisition layer interfaces with various data sources throughout the entire lifecycle of heat treatment, including real-time process parameters (temperature, holding time, heating rate, furnace pressure) and energy metering data (electricity consumption, natural gas, propane, and other fuel consumption) collected by workshop PLCs and DCS systems; equipment operating status and auxiliary material (quenching oil, protective gas) consumption data collected by IoT sensors; production plans, workpiece materials, furnace loading quantities, and batch information synchronized by ERP and MES systems; upstream carbon emission data for raw materials such as steel and alloys uploaded by the supply chain system; transportation mileage and transportation mode data from the logistics management system; and end-stage data such as waste disposal ledgers and recycling records. The acquisition layer supports multi-protocol access and performs preliminary data cleaning and standardization preprocessing through an edge computing gateway, reducing the transmission pressure on the cloud.

[0045] In one optional embodiment of this invention, the data fusion layer constructs a distributed big data storage and processing platform based on the Hadoop and Spark ecosystem. It includes structured and unstructured databases such as a raw carbon footprint database, a process parameter database, an emission factor database, a lifecycle database, and a quality verification database. The core of this layer is the fusion of multi-source heterogeneous data. It uses ETL tools to extract, transform, and load data, establishing unified data encoding rules and associated mapping relationships to address differences in data formats, dimensions, and definitions across different systems. Simultaneously, it incorporates a data quality verification module that uses algorithms for outlier detection, missing value imputation, duplicate data removal, and logical consistency verification to ensure the accuracy and integrity of carbon footprint data, providing a high-quality data foundation for subsequent accounting.

[0046] In one optional embodiment of the present invention, the core algorithm layer is the system hub, integrating five core algorithms: a full life cycle carbon footprint accounting algorithm for heat treatment, a dynamic emission factor optimization algorithm, a multivariate coupled carbon analysis algorithm, a carbon footprint prediction algorithm, and an emission reduction path optimization algorithm. It is customized for the characteristics of heat treatment processes, and is different from general carbon accounting models, achieving high-precision and dynamic accounting and analysis.

[0047] In one optional embodiment of this invention, the application service layer provides modular services based on the core algorithm output results, including full lifecycle carbon footprint accounting, carbon data traceability, process carbon emission analysis, emission reduction effect evaluation, automatic carbon report generation, and carbon early warning and control, covering the entire process of accounting, management, analysis, and decision-making. The user interaction layer provides a visual operation interface for enterprise managers, process engineers, and environmental protection personnel through terminals such as web terminals, mobile terminals, and workshop large screens, supporting functions such as data query, chart display, parameter configuration, and process approval.

[0048] Implementation, for example Figure 2 As shown, the carbon footprint accounting algorithm for the entire life cycle of heat treatment combines the boundaries of the entire heat treatment process to construct a hierarchical accounting model, which divides the carbon footprint into four categories: direct emissions, energy indirect emissions, raw material indirect emissions, and other indirect emissions, thereby achieving full life cycle coverage accounting.

[0049] The core formula of the algorithm is: ; in, For the total carbon footprint over the entire life cycle, Direct carbon emissions include emissions from fuel combustion and process chemical reactions. Indirect emissions due to purchased electricity and heat This is due to hidden emissions from upstream raw materials and auxiliary materials. Emissions from transportation, waste disposal, equipment operation and maintenance, and recycling processes.

[0050] Direct carbon emission accounting focuses on fuel combustion and process reactions in heat treatment furnaces, using a combination of measured data and industry coefficients. ; The consumption of fuels such as natural gas and propane is collected in real time by energy meters. To correspond to fuel emission factors and distinguish between lower heating value and measured calorific value, The carbon potential and decomposition rate of carburized emissions during the process are used to determine the emissions. The emission coefficient for the process reaction is customized based on different materials for carburizing and nitriding.

[0051] Indirect energy emissions focus on purchased electricity and heat: ; , Electricity and heat consumption Using the regional power grid's annual dynamic emission factor, A heat supply benchmark factor is adopted to replace the traditional fixed value, thereby improving the accuracy of the calculation.

[0052] Indirect emissions of raw materials cover the entire upstream stage of the life cycle, through List association: ; For the first Raw material usage For the first dosage of similar excipients , To address the hidden emission factors upstream of the materials, we integrate emission data from the entire process of raw material mining, processing, and transportation.

[0053] Other indirect emissions are accounted for by category. For example, transportation emissions are calculated based on transportation mode, mileage, and load; waste treatment emissions are calculated based on waste type, treatment volume, landfill, incineration, recycling, and other treatment methods and corresponding emission factors; and equipment operation and maintenance emissions are calculated based on maintenance consumables and repair energy consumption, so as to achieve complete accounting throughout the entire life cycle.

[0054] Implementation, for example Figure 3 As shown, the dynamic emission factor optimization algorithm combines real-time process data with machine learning to achieve dynamic optimization of emission factors. The algorithm uses heat treatment furnace type, process type, workpiece material, furnace loading, temperature curve, running time, and equipment aging degree as input features to construct a random forest regression model to correct the basic emission factors.

[0055] First, a massive amount of historical data is collected, including actual energy consumption, emission monitoring values ​​and corresponding process parameters under different operating conditions. Key influencing factors are extracted through feature engineering, redundant features are removed, and then a nonlinear mapping relationship between process parameters and emission coefficients is established through model training to obtain the correction coefficient k under different scenarios.

[0056] Formula for calculating dynamic emission factor: ; As the industry benchmark emission factor, This is a dynamic correction factor, determined by the process type. Material Furnace loading capacity Temperature curve runtime Equipment aging degree The decision is made jointly. The system collects process parameters in real time and substitutes them into the model for calculation. Value, updated in real time This ensures that the accounting results match the real-time production status.

[0057] In one optional embodiment of the present invention, the multivariate coupled carbon analysis algorithm is based on a combination of the analytic hierarchy process (AHP) and partial least squares method to construct a multivariate attribution model.

[0058] First, the weights of each influencing factor are determined through analytic hierarchy process (AHP). The target layer is then decomposed into a criterion layer and an indicator layer. The weights of each indicator are determined by combining expert scoring with statistical data. The target layer is the total carbon footprint, the criteria layer includes process parameters, equipment performance, material properties, energy structure, and management level, and the indicator layer includes temperature, holding time, furnace efficiency, carbon content of materials, energy type, and furnace loading rate.

[0059] Then, partial least squares analysis was used to analyze the correlation and contribution of each variable to the carbon footprint, and multicollinearity interference was eliminated to obtain the influence coefficient of each factor on the carbon footprint. Calculation formula: ; For each influencing variable, This is due to random error. The algorithm automatically outputs the contribution ranking of each variable, identifies key influencing factors of carbon footprint (high-temperature energy consumption, low furnace loading rate, and high-energy-consuming furnace type), and provides precise direction for emission reduction.

[0060] In one optional embodiment of the present invention, the carbon footprint prediction algorithm is based on an LSTM neural network to construct a prediction model and process the nonlinearity and hysteresis characteristics of the thermal treatment carbon footprint time series data.

[0061] Using historical carbon footprint time-series data, production plans, preset process parameters, energy prices, and emission policies as inputs, this model captures long-term dependencies through the memory units of an LSTM network to uncover inherent patterns in the time-series data. During model training, previous period carbon footprint data and associated variables are input, and the predicted carbon footprint value for the next period is output. An attention mechanism is also introduced to strengthen the weight allocation of key process nodes, improving prediction accuracy. The prediction formula is simplified to: ; To predict carbon footprint for the next moment, For the historical carbon footprint of the previous m moments, These are the relevant variables for the corresponding process and production at any given time. Furthermore, the system supports short-term (daily / weekly) and long-term (monthly / quarterly) forecasts.

[0062] Implementation, for example Figure 4 As shown, the emission reduction path optimization algorithm is based on carbon footprint accounting and attribution results. This algorithm constructs a multi-objective optimization model with the objectives of maximizing carbon emission reduction, minimizing production costs, and optimizing production efficiency. Combined with heat treatment process constraints (temperature range, holding time, material requirements), the optimal emission reduction path is solved by a genetic algorithm.

[0063] The algorithm encoding transforms decision variables such as process parameters, equipment parameters, energy structure, and furnace loading method into chromosomes. It iterative optimization is achieved through selection, crossover, and mutation operations. The fitness function comprehensively considers carbon emission reduction benefits, cost changes, and process compliance.

[0064] Fitness function: ; For the combination of decision variables, For carbon emission reductions, Due to cost changes, The production efficiency coefficient. , , These are weighting coefficients, which can be adjusted according to enterprise needs. The algorithm outputs multiple optimization solutions: process curve optimization (shortening holding time, reducing heating rate), equipment modification (replacing with high-efficiency furnace type, waste heat recovery), energy structure adjustment (increasing green electricity ratio), and furnace loading method optimization (improving furnace loading rate).

[0065] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A big data management and accounting system for the carbon footprint of a heat treatment process throughout its life cycle, characterized in that, It adopts a layered distributed architecture, consisting of a perception and acquisition layer, a data fusion layer, a core algorithm layer, an application service layer, and a user interaction layer. Each layer works together through standardized interfaces.

2. The big data management and accounting system for carbon footprint of heat treatment process life cycle according to claim 1, characterized in that, The aforementioned sensing and acquisition layer interfaces with various data sources throughout the entire lifecycle of heat treatment, including process parameters and energy metering data collected in real time by workshop PLCs and DCS systems, equipment operating status and auxiliary material consumption data collected by IoT sensors, production plans, workpiece materials, furnace loading quantities, and batch information synchronized by ERP and MES systems, upstream carbon emission data of raw materials uploaded by the supply chain system, transportation mileage and transportation mode data from the logistics management system, and end-stage data such as waste disposal ledgers and recycling records. The acquisition layer supports multi-protocol access and achieves preliminary data cleaning and standardization preprocessing through edge computing gateways, reducing the transmission pressure on the cloud.

3. The big data management and accounting system for the entire lifecycle carbon footprint of a heat treatment process according to claim 1, characterized in that, The data fusion layer is built on the Hadoop and Spark ecosystem and includes structured and unstructured databases such as the carbon footprint raw database, process parameter database, emission factor database, life cycle basic database, and quality verification database. The core of the data fusion layer is to achieve the fusion of multi-source heterogeneous data. It uses ETL tools to complete data extraction, transformation, and loading, establishes unified data encoding rules and associated mapping relationships, and solves the problem of differences in data format, dimension, and caliber between different systems. At the same time, it is equipped with a data quality verification module, which sets up algorithms for outlier detection, missing value imputation, duplicate data removal, and logical consistency verification.

4. The big data management and accounting system for the entire lifecycle carbon footprint of a heat treatment process according to claim 1, characterized in that, The core algorithm layer is the central hub of the system, integrating five core algorithms: carbon footprint accounting algorithm for the entire life cycle of heat treatment, dynamic emission factor optimization algorithm, multivariate coupled carbon analysis algorithm, carbon footprint prediction algorithm, and emission reduction path optimization algorithm. It is custom-built for the characteristics of heat treatment processes.

5. A big data management and accounting system for the entire lifecycle carbon footprint of a heat treatment process according to claim 1, characterized in that, The application service layer, based on the core algorithm output results, provides modular services for full lifecycle carbon footprint accounting, carbon data traceability, process carbon emission analysis, emission reduction effect evaluation, automatic carbon report generation, and carbon early warning and control, covering the entire process of accounting, management, analysis, and decision-making. The user interaction layer provides a visual operation interface for enterprise managers, process engineers, and environmental protection personnel through web terminals, mobile terminals, and workshop large screens, supporting data query, chart display, parameter configuration, and process approval functions.

6. A big data management and accounting system for the entire lifecycle carbon footprint of a heat treatment process according to claim 4, characterized in that, The carbon footprint accounting algorithm for the entire life cycle of heat treatment combines the boundaries of the entire heat treatment process to construct a hierarchical accounting model, which divides the carbon footprint into four categories: direct emissions, energy indirect emissions, raw material indirect emissions, and other indirect emissions, thereby achieving full life cycle coverage accounting. The core formula of the algorithm is: ； in, For the total carbon footprint over the entire life cycle, Direct carbon emissions include emissions from fuel combustion and process chemical reactions. Indirect emissions due to purchased electricity and heat This is due to hidden emissions from upstream raw materials and auxiliary materials. Emissions from transportation, waste disposal, equipment operation and maintenance, and recycling processes; Direct carbon emission accounting focuses on fuel combustion and process reactions in heat treatment furnaces, using a combination of measured data and industry coefficients. ； Fuel consumption is collected in real time by energy meters. To correspond to fuel emission factors and distinguish between lower heating value and measured calorific value, These are parameters related to emissions from the process, namely carburizing carbon potential and decomposition rate. The emission coefficient for process reaction is customized based on different materials and process types; Indirect energy emissions focus on purchased electricity and heat: ； , Electricity and heat consumption Using the regional power grid's annual dynamic emission factor, A heat supply benchmark factor is adopted to replace the traditional fixed value, thereby improving the accuracy of the calculation. Indirect emissions of raw materials cover the entire upstream stage of the life cycle, through List association: ； For the first Raw material usage For the first dosage of similar excipients , To address the hidden emission factors in the upstream of the materials, emission data from the entire process of raw material mining, processing, and transportation is integrated; Other indirect emissions are accounted for by category: transportation is calculated based on transportation mode, mileage, and load; waste treatment is calculated based on the emission factors corresponding to waste type, treatment volume, and treatment method; and equipment operation and maintenance is calculated based on maintenance consumables and repair energy consumption, so as to achieve full life cycle accounting without omission.

7. A big data management and accounting system for the entire lifecycle carbon footprint of a heat treatment process according to claim 4, characterized in that, The dynamic emission factor optimization algorithm achieves dynamic optimization of emission factors by combining real-time process data with machine learning. The algorithm uses heat treatment furnace type, process type, workpiece material, furnace loading, temperature curve, running time, and equipment aging degree as input features to construct a random forest regression model and correct the basic emission factors; First, a massive amount of historical data is collected, including actual energy consumption, emission monitoring values ​​and corresponding process parameters under different operating conditions. Key influencing factors are extracted through feature engineering, redundant features are removed, and then a nonlinear mapping relationship between process parameters and emission coefficients is established through model training to obtain the correction coefficient k under different scenarios. Formula for calculating dynamic emission factor: ； As the industry benchmark emission factor, This is a dynamic correction factor, determined by the process type. Material Furnace loading capacity Temperature curve runtime Equipment aging degree The decision is made jointly; the system collects process parameters in real time and substitutes them into the model for calculation. Value, updated in real time This ensures that the accounting results match the real-time production status.

8. A big data management and accounting system for the entire lifecycle carbon footprint of a heat treatment process according to claim 4, characterized in that, The multivariate coupled carbon analysis algorithm described above is based on a combination of the analytic hierarchy process (AHP) and partial least squares method to construct a multivariate attribution model. First, the weights of each influencing factor are determined through analytic hierarchy process (AHP). The target layer is then decomposed into a criterion layer and an indicator layer. Finally, the weights of each indicator are determined by combining expert scoring with statistical data. The target layer is the total carbon footprint; the criteria layer includes process parameters, equipment performance, material properties, energy structure, and management level; and the indicator layer includes temperature, holding time, furnace efficiency, carbon content of materials, energy type, and furnace loading rate. Then, partial least squares analysis was used to analyze the correlation and contribution of each variable to the carbon footprint, and multicollinearity interference was eliminated to obtain the influence coefficient of each factor on the carbon footprint. Calculation formula: ； For each influencing variable, To account for random errors, the algorithm automatically outputs the contribution ranking of each variable, identifies key influencing factors of carbon footprint, and provides precise direction for emission reduction.

9. A big data management and accounting system for the entire lifecycle carbon footprint of a heat treatment process according to claim 8, characterized in that, The carbon footprint prediction algorithm described above is based on an LSTM neural network to construct a prediction model and to handle the nonlinearity and hysteresis characteristics of thermal treatment carbon footprint time series data. Using historical carbon footprint time-series data, production plans, process parameter presets, energy prices, and emission policies as inputs, the model captures long-term dependencies through the memory units of an LSTM network to uncover the inherent patterns in the time-series data. During model training, the model inputs carbon footprint data from previous periods and related variables, and outputs the predicted carbon footprint value for the next period. Attention mechanisms are also introduced to strengthen the weight allocation of key process nodes and improve prediction accuracy. The prediction formula simplifies to: ; To predict carbon footprint for the next moment, For the front A moment in history's carbon footprint, These are the correlation variables for the corresponding time point in the process and production.

10. A big data management and accounting system for the entire lifecycle carbon footprint of a heat treatment process according to claim 9, characterized in that, The emission reduction path optimization algorithm described above is based on carbon footprint accounting and attribution results. This algorithm constructs a multi-objective optimization model with the objectives of maximizing carbon emission reduction, minimizing production costs, and optimizing production efficiency. Combined with the constraints of heat treatment process, the optimal emission reduction path is solved by a genetic algorithm. The algorithm encoding transforms decision variables such as process parameters, equipment parameters, energy structure, and furnace loading method into chromosomes. It iterative optimization is achieved through selection, crossover, and mutation operations. The fitness function comprehensively considers carbon emission reduction benefits, cost changes, and process compliance. Fitness function: ； For the combination of decision variables, For carbon emission reductions, Due to cost changes, The production efficiency coefficient. , , The weighting coefficients can be adjusted according to the needs of enterprises; the algorithm outputs multiple optimization schemes, including process curve optimization, equipment modification, energy structure adjustment, and furnace loading method optimization, and evaluates the emission reduction effect, cost input, and implementation difficulty of each scheme.

Images