Method and system for dynamic calculation of fuel end carbon emission factor based on computational fluid dynamics

Abstract

The application discloses a fuel end carbon emission factor dynamic calculation method and system based on computational fluid dynamics, and belongs to the technical field of low-carbon energy-saving and environmental protection. The method comprises the following steps: obtaining geometric parameters of a combustion equipment, equipment operation condition data, fuel component data and field measurement data; obtaining standard parameters from the equipment operation condition data, constructing a CFD simulation model based on the geometric and fuel data, simulating with the standard parameters and verifying with field data deviation to obtain a standard simulation model; based on the model, performing multi-condition simulation with the operation condition data and the fuel data as boundaries, obtaining simulation data and calculating a carbon emission factor; selecting a data fitting, machine learning or deep learning model according to the data magnitude, and constructing a dynamic carbon emission factor calculation model; comparing the outlet temperature with the model output in real time, and when the deviation exceeds a threshold value, calling the standard simulation model again to obtain new data, and updating and optimizing the dynamic model after fusion. The application realizes high-precision dynamic calculation of the carbon emission factor.

Description

Technical Field

[0001] This invention belongs to the field of low-carbon, energy-saving and environmental protection technology, and in particular relates to a dynamic calculation method and system for fuel-end carbon emission factors based on computational fluid dynamics. Background Technology

[0002] In the carbon market system, the carbon emission factor is a key fundamental parameter characterizing the carbon emission level of a unit of activity or unit of fuel, and is an important basis for carbon emission accounting and quota allocation. Its accurate acquisition and dynamic updating directly affect the scientific validity and fairness of carbon emission accounting results, and are an important technical support for the effective operation of the carbon market. A precise emission factor accounting system helps the market rationally plan the quantity and price of carbon emission quotas, promoting the achievement of carbon peaking and carbon neutrality (dual-carbon) goals; at the same time, it also helps enterprises achieve refined carbon management, formulate reasonable carbon quota plans, and optimize carbon asset allocation.

[0003] Currently, fossil fuel combustion remains a major source of carbon emissions, and its emissions are typically calculated using the traditional emission factor method. This method mainly references the accounting guidelines issued by the Intergovernmental Panel on Climate Change (IPCC) and generally uses uniform default values ​​for calculation. It fails to adequately consider the differences in fuel composition and combustion equipment among different companies, leading to significant deviations in the calculation results and making it difficult to reflect the true emission levels. Therefore, there is an urgent need for a dynamic calculation method for fuel-side carbon emission factors that can adapt to various fuel characteristics and operating conditions of different combustion equipment. This would enable high-precision calculation and dynamic updating of emission factors, thereby improving the accuracy and reliability of carbon accounting results and supporting the scientific operation of the carbon market. Summary of the Invention

[0004] To address the above problems, this invention proposes a method and system for dynamically calculating fuel-end carbon emission factors based on computational fluid dynamics.

[0005] To achieve the above objectives, the present invention employs the following technical solution:

[0006] Dynamic calculation methods for fuel-end carbon emission factors based on computational fluid dynamics include:

[0007] Acquire and preprocess relevant data from emission control enterprises to obtain preprocessed combustion equipment geometric parameters, equipment operating condition data, fuel composition data, and on-site measurement data;

[0008] Standard operating condition parameters are determined based on equipment operating condition data; a computational fluid dynamics simulation model is constructed based on the geometric parameters of the combustion equipment and fuel composition data; the computational fluid dynamics simulation model is simulated based on the standard operating condition parameters; the simulation results are verified for deviation based on field measurement data; and a standard simulation model is obtained when preset conditions are met.

[0009] Based on the standard simulation model, and using equipment operating condition data and fuel composition data, simulations under various combustion conditions are performed to obtain simulation data under different combustion conditions, and the corresponding carbon emission factors are calculated based on the simulation data.

[0010] The modeling method is selected according to the data volume of the simulation data, and a dynamic carbon emission factor calculation model is constructed with carbon emission factor and combustion equipment outlet temperature as outputs.

[0011] The temperature deviation is obtained by comparing the real-time collected combustion equipment outlet temperature with the combustion equipment outlet temperature output by the dynamic carbon emission factor calculation model. When the temperature deviation exceeds the preset temperature threshold, the corresponding equipment operating condition data and fuel composition data are recorded, and the standard simulation model is re-simulated under the corresponding combustion conditions. The dynamic carbon emission factor calculation model is then updated and optimized based on the simulation data obtained from the simulation.

[0012] As a preferred embodiment of the present invention, the relevant data of the controlled emission enterprise includes the geometric parameters of the combustion equipment, the equipment operating condition data, the fuel composition data, and the on-site measurement data; the geometric parameters of the combustion equipment are obtained through the design drawings of the combustion equipment, including the geometric dimensions of the burner, the shape and size of the air distribution port, the geometric dimensions of the fuel nozzle, and the shape and size of the flue gas outlet;

[0013] The equipment operating condition data is collected through combustion equipment sensors and enterprise production records, including load, fuel consumption, fuel mass flow rate and fuel velocity of each burner, total air volume, and air volume, air velocity and air temperature of each air outlet.

[0014] The fuel composition data includes elemental analysis data and industrial analysis data. The elemental analysis data includes the content of carbon, hydrogen, oxygen, nitrogen, and sulfur, while the industrial analysis data includes the content of moisture, ash, volatile matter, and fixed carbon.

[0015] The on-site measurement data is acquired by sensors arranged inside and at the outlet of the combustion equipment, including temperature and velocity at the sensor location, outlet flue gas temperature, outlet flue gas velocity, and at least one of carbon dioxide concentration, carbon monoxide concentration, and oxygen concentration in the outlet flue gas.

[0016] As a preferred embodiment of the present invention, the preprocessing includes:

[0017] Unit standardization and format conversion are performed on the geometric parameters of the combustion equipment; outlier removal, missing value filling, and unit standardization are performed on the equipment operating condition data; data verification and normalization are performed on the fuel composition data; and filtering, noise reduction, and dimensional alignment are performed on the field measurement data.

[0018] As a preferred embodiment of the present invention, the step of determining standard operating condition parameters based on equipment operating condition data and constructing a computational fluid dynamics simulation model based on combustion equipment geometric parameters and fuel composition data includes:

[0019] The data segment with load fluctuation less than the preset fluctuation threshold and key parameter change rate lower than the preset change rate threshold is selected from the equipment operation condition data. Cluster analysis is performed on the equipment operation condition data of the data segment. One or more typical operating conditions are selected from the cluster results as standard operating conditions, and the corresponding standard operating condition parameters are extracted.

[0020] A three-dimensional model is constructed based on the geometric parameters of the combustion device, and the structure of the combustion device in the three-dimensional model is simplified, specifically: wall chamfers, fillets, bolt holes, and nozzles with diameters smaller than the preset diameter value are ignored, and adjacent channels with a spacing smaller than the preset spacing value are merged.

[0021] When the simplified combustion equipment structure is regular, a structured mesh is used for mesh generation; otherwise, an unstructured mesh is used. After selecting the turbulence model and combustion model according to the type of combustion equipment structure, the physical property parameters and chemical reaction parameters of the fuel are set according to the fuel composition data, and finally the computational fluid dynamics simulation model is obtained.

[0022] As a preferred embodiment of the present invention, the computational fluid dynamics simulation model is simulated based on the standard operating condition parameters, and the simulation results are verified for deviation based on field measurement data. When preset conditions are met, a standard simulation model is obtained, including:

[0023] After setting boundary conditions for the computational fluid dynamics simulation model based on the aforementioned standard operating condition parameters, the simulation calculation is performed to obtain the simulation results; and the deviation between the simulation results and the field measurement data is calculated, expressed as:

[0024]

[0025] In the formula, To account for the discrepancy between the simulation results and the field measurement data, For simulation results, This is data measured on-site;

[0026] If the deviation between the simulation results and the field measurement data does not exceed the preset deviation threshold, the computational fluid dynamics simulation model is determined as the standard simulation model; if the deviation between the simulation results and the field measurement data exceeds the preset deviation threshold, the model parameters and mesh of the computational fluid dynamics simulation model are adjusted, and the simulation calculation is performed again until the deviation does not exceed the preset deviation threshold, and the standard simulation model is output.

[0027] When there are multiple sets of standard operating condition parameters, the deviation between the simulation results and the field measurement data corresponding to each set of standard operating condition parameters must not exceed the preset deviation threshold.

[0028] As a preferred embodiment of the present invention, the step of performing simulations under various combustion conditions based on a standard simulation model, equipment operating condition data, and fuel composition data, to obtain simulation data under different combustion conditions, and calculating the corresponding carbon emission factor based on the simulation data, includes:

[0029] Multiple sets of different equipment operating condition parameter combinations are selected from the equipment operating condition data. Each set of equipment operating condition parameters corresponds to a set of combustion conditions. For each set of combustion conditions, the corresponding equipment operating condition parameters and fuel composition data are used as the entry boundary conditions of the standard simulation model for simulation calculation to obtain the simulation data under that combustion condition.

[0030] The simulation data includes: the carbon content of the fuel, the sum of the fuel mass flow rates of each burner, the outlet carbon dioxide mass flow rate, the temperature field, and the velocity field.

[0031] Based on the carbon content of the fuel, the sum of the fuel mass flow rates of each burner, and the outlet carbon dioxide mass flow rate, the corresponding carbon emission factor is calculated, expressed as:

[0032]

[0033]

[0034] In the formula, As a carbon emission factor, The carbon content of the fuel, For carbon oxidation rate, This represents the mass flow rate of carbon dioxide at the outlet. This is the sum of the fuel mass flow rates of each burner.

[0035] As a preferred embodiment of the present invention, the step of selecting a corresponding modeling method based on the data volume of the simulation data and constructing a dynamic carbon emission factor calculation model with carbon emission factor and combustion equipment outlet temperature as outputs includes:

[0036] If the data volume of the simulation data does not exceed the preset first threshold, then the data fitting model is used as the dynamic carbon emission factor calculation model; the data fitting model uses curve fitting or regression analysis methods, with the carbon content of the load and fuel as input, to establish a functional expression for the carbon emission factor and the outlet temperature of the combustion equipment.

[0037] If the volume of the simulated data exceeds a preset first threshold but does not exceed a preset second threshold, a machine learning model is used as the dynamic carbon emission factor calculation model. The machine learning model divides the simulated data into a training set, a validation set, and a test set. It uses the minimum prediction error of the carbon emission factor and the outlet temperature of the combustion equipment as the objective loss function, employs a gradient descent algorithm for training, and uses a Bayesian optimization algorithm to optimize the model's hyperparameters. The inputs to the machine learning model include load, elemental analysis data, industrial analysis data, and the air volume, wind speed, and wind temperature of each air outlet.

[0038] If the volume of the simulated data exceeds a preset second threshold, a deep learning model is used as the dynamic carbon emission factor calculation model. The deep learning model divides the simulated data into a training set, a validation set, and a test set. The target loss function is to minimize the prediction error of the carbon emission factor and the outlet temperature of the combustion equipment. An adaptive learning rate optimization algorithm is used for training, and an early stopping method is used to prevent overfitting. The inputs of the deep learning model include load, elemental analysis data, industrial analysis data, air volume, wind speed and wind temperature at each air outlet, temperature field and component characteristics. The component characteristics include at least the concentration fields of carbon dioxide, carbon monoxide and oxygen.

[0039] The outputs of the three models are the carbon emission factor and the combustion equipment outlet temperature.

[0040] As a preferred embodiment of the present invention, the step of comparing the real-time collected combustion equipment outlet temperature with the combustion equipment outlet temperature output by the dynamic carbon emission factor calculation model to obtain the temperature deviation; when the temperature deviation exceeds a preset temperature threshold, recording the corresponding equipment operating condition data and fuel composition data, and re-simulating the standard simulation model under the corresponding combustion condition, and updating and optimizing the dynamic carbon emission factor calculation model based on the simulation data obtained from the simulation, including:

[0041] The temperature at the outlet of the combustion equipment is collected in real time by a temperature sensor placed at the outlet of the combustion equipment, and compared with the temperature at the outlet of the combustion equipment output by the dynamic carbon emission factor calculation model. The temperature deviation is calculated. When the temperature deviation does not exceed the preset temperature deviation threshold, the operation of the dynamic carbon emission factor calculation model is maintained. When the temperature deviation exceeds the preset temperature deviation threshold, the equipment operating condition data and fuel composition data at the time of the deviation are recorded.

[0042] The standard simulation model is re-invoked, and the recorded equipment operating condition data and fuel composition data are used as boundary conditions. After optimizing the solution parameters of the standard simulation model, simulation is performed under the corresponding combustion conditions to obtain new simulation data.

[0043] The new simulation data is fused with the original simulation data to obtain updated simulation data; and the carbon emission factor calculation model is retrained and optimized based on the updated simulation data.

[0044] A dynamic calculation system for fuel-side carbon emission factors based on computational fluid dynamics includes:

[0045] The data acquisition module is used to acquire and preprocess relevant data from controlled emission enterprises to obtain preprocessed combustion equipment geometric parameters, equipment operating condition data, fuel composition data, and on-site measurement data.

[0046] The simulation module is used to determine standard operating condition parameters based on equipment operating condition data; to construct a computational fluid dynamics simulation model based on the geometric parameters of the combustion equipment and fuel composition data; to perform simulation calculations on the computational fluid dynamics simulation model based on the standard operating condition parameters; to verify the deviation of the simulation results based on field measurement data; and to obtain a standard simulation model when preset conditions are met.

[0047] The emission factor calculation module is used to perform simulations under various combustion conditions based on equipment operating condition data and fuel composition data according to the standard simulation model, obtain simulation data under different combustion conditions, and calculate the corresponding carbon emission factor based on the simulation data.

[0048] The emission factor modeling module is used to select the corresponding modeling method according to the data volume of the simulation data, and to construct a dynamic carbon emission factor calculation model with carbon emission factor and combustion equipment outlet temperature as outputs.

[0049] The model feedback optimization module is used to compare the real-time collected combustion equipment outlet temperature with the combustion equipment outlet temperature output by the dynamic carbon emission factor calculation model to obtain the temperature deviation. When the temperature deviation exceeds the preset temperature threshold, the corresponding equipment operating condition data and fuel composition data are recorded, and the standard simulation model is re-simulated under the corresponding combustion conditions. Based on the simulation data obtained from the simulation, the dynamic carbon emission factor calculation model is updated and optimized.

[0050] The beneficial effects of this invention are as follows: This invention abandons the traditional IPCC default value method and, for the first time, deeply integrates computational fluid dynamics (CFD) simulation with data-driven modeling. By constructing a high-fidelity standard simulation model and using field measurement data for multiple rounds of deviation verification and parameter adjustment, it ensures that the flow field, temperature field, and component field of the combustion process are highly consistent with reality. On this basis, a large amount of high-confidence simulation data is obtained through batch simulation under multiple operating conditions, thereby training a dynamic carbon emission factor calculation model with fuel carbon content, air distribution parameters, etc. as inputs. Compared with static factors, this method can accurately reflect the real-time impact of fuel composition fluctuations, equipment structural differences, and changes in operating conditions on carbon emissions, significantly reducing calculation errors and providing accurate data support for carbon accounting. This invention clusters and extracts various typical load and air distribution combinations from historical equipment operating data and simultaneously introduces fuel element / industrial analysis data as boundary conditions, performing simulations covering the entire operating range on a standard model. The dynamically trained model maintains high prediction accuracy across different loads, fuel batches, and air distribution schemes, overcoming the limitations of traditional methods that are only applicable to design conditions. This represents a leap from "single-point static" to "full-domain dynamic" carbon emission factors. This invention employs a real-time feedback mechanism based on the combustion equipment outlet temperature. When the deviation between the model's output temperature and the sensor's measured temperature exceeds a threshold, the system automatically records the complete operating parameters and fuel data for the triggered condition, re-invokes the standard simulation model for high-precision verification, and integrates the newly generated simulation data into the original dataset, triggering incremental learning or complete retraining of the dynamic model. This closed-loop strategy enables the carbon emission factor calculation model to continuously evolve with equipment aging, changes in fuel sources, and adjustments to operating strategies, always maintaining optimal prediction performance. The dynamic carbon emission factor provided by this invention can be directly used for enterprise-level carbon emission accounting and quota compliance. The high-precision, traceable, and dynamically updated factor system effectively suppresses accounting disputes, enhances the credibility of carbon market data, assists enterprises in refined carbon asset management, and provides key technical support for the scientific advancement of carbon peaking and carbon neutrality goals. In summary, this invention has outstanding advantages in terms of accuracy, adaptability, robustness, and market application. Attached Figure Description

[0051] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a flowchart of the method in an embodiment of the present invention; Figure 2 This is a system structure diagram in an embodiment of the present invention. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention are within the scope of protection of the present invention.

[0053] like Figure 1 As shown, this is an embodiment of the present invention, which provides a dynamic calculation method for fuel-end carbon emission factors based on computational fluid dynamics, including:

[0054] S1. Obtain and preprocess relevant data from emission control enterprises to obtain preprocessed combustion equipment geometric parameters, equipment operating condition data, fuel composition data, and on-site measurement data.

[0055] In this embodiment, the relevant data of the controlled emission enterprise includes the geometric parameters of the combustion equipment, the equipment operating condition data, the fuel composition data, and the on-site measurement data. The geometric parameters of the combustion equipment are obtained through the design drawings of the combustion equipment, including the geometric dimensions of the burner, the shape and size of the air distribution port, the geometric dimensions of the fuel nozzle, and the shape and size of the flue gas outlet. The equipment operating condition data are obtained through the combustion equipment sensors and the enterprise's production records, including load, fuel consumption, fuel mass flow rate and fuel velocity of each burner, total air volume, and air volume, velocity, and temperature of each air outlet. The fuel composition data includes elemental analysis data and industrial analysis data. The elemental analysis data includes the content of carbon, hydrogen, oxygen, nitrogen, and sulfur, while the industrial analysis data includes the content of moisture, ash, volatile matter, and fixed carbon. The on-site measurement data are obtained through sensors arranged inside and at the outlet of the combustion equipment, including the temperature and velocity at the sensor location, the outlet flue gas temperature, the outlet flue gas velocity, and at least one of the following concentrations in the outlet flue gas: carbon dioxide concentration, carbon monoxide concentration, and oxygen concentration.

[0056] Unit standardization and format conversion are performed on the geometric parameters of the combustion equipment; outlier removal, missing value filling, and unit standardization are performed on the equipment operating condition data; data verification and normalization are performed on the fuel composition data; and filtering, noise reduction, and dimensional alignment are performed on the field measurement data.

[0057] Specifically, complete geometric parameters of the combustion equipment are obtained through design drawings or 3D model files. For areas with complex structures, such as multiple similar structures in the design drawings, the positional relationships and spatial coordinates are determined one by one. Units of all geometric parameters of the combustion equipment are standardized, such as converting them to millimeters or meters, and formats are converted, such as converting CAD drawings to standard formats suitable for mesh generation software, such as STEP or IGES.

[0058] For the operating condition data of the combustion equipment collected by the sensors, outliers are removed, such as by sampling 3σ principle and box plot method. Data points that are obviously outside the physical range due to combustion equipment sensor failure are deleted. Missing values ​​are filled, such as by linear interpolation or the mean of adjacent time periods. Units are also standardized, such as unifying all temperatures to degrees Celsius and flow rates to kilograms per hour, to ensure that the data sequence is continuous and the physical units are consistent.

[0059] Fuel composition data is obtained through fuel analysis reports. First, the data is validated to check if the sum of all components is 100% or within a reasonable fluctuation range. If significant deviations are found, the data must be returned to the analysis stage for reconfirmation. The fuel composition data is then normalized, for example, by converting the component content to small values ​​between 0 and 1.

[0060] For on-site measurement data, low-pass filters or moving averages are used to remove high-frequency noise; and the same physical quantities collected by different sensors are standardized to the same unit, such as temperature in degrees Celsius and speed in meters per second.

[0061] S2. Determine standard operating condition parameters based on equipment operating condition data; construct a computational fluid dynamics simulation model based on combustion equipment geometric parameters and fuel composition data, and perform simulation calculations on the computational fluid dynamics simulation model based on the standard operating condition parameters. Verify the deviation of the simulation results based on field measurement data. When the preset conditions are met, a standard simulation model is obtained.

[0062] In this embodiment, data segments with load fluctuations less than a preset fluctuation threshold and key parameter change rates lower than a preset change rate threshold are selected from the equipment operating condition data. Cluster analysis is performed on the equipment operating condition data of the data segments. One or more typical operating conditions are selected from the clustering results as standard operating conditions, and the corresponding standard operating condition parameters are extracted.

[0063] A three-dimensional model is constructed based on the geometric parameters of the combustion device, and the structure of the combustion device in the three-dimensional model is simplified. Specifically, wall chamfers, fillets, bolt holes, and nozzles with diameters smaller than the preset diameter value are ignored, and adjacent channels with a spacing smaller than the preset spacing value are merged. When the simplified combustion device structure is regular, a structured mesh is used for mesh generation; otherwise, an unstructured mesh is used. After selecting the turbulence model and combustion model according to the type of combustion device structure, the physical property parameters and chemical reaction parameters of the fuel are set according to the fuel composition data, and finally, a computational fluid dynamics simulation model is obtained.

[0064] Specifically, from the acquired equipment operating condition data, the first step is to filter out data segments that are operating stably. Specific filtering criteria include:

[0065] The load fluctuation range within a continuous 30-minute period, i.e., the maximum relative deviation from the average load during that period, shall not exceed ±5%;

[0066] Other key operating parameters, including fuel mass flow rate of each burner, total air volume, air volume of each air outlet, and air velocity, have a relative change rate of no more than ±2% per minute.

[0067] Within the aforementioned stable data segment, a complete sequence of operating parameters is extracted. Cluster analysis methods, such as the K-means clustering algorithm, are then employed. The K value is set to 3-5 based on the complexity of the equipment's operating conditions; in this embodiment, it is 4. After clustering, the operating point closest to the cluster center is selected as a typical operating point from each cluster. Depending on actual needs, one or more typical operating points can be selected as standard operating conditions, and the corresponding standard operating condition parameters are extracted. These standard operating condition parameters include at least: load, fuel mass flow rate of each burner, total air volume, air volume at each air outlet, and air temperature.

[0068] Based on the acquired geometric parameters of the combustion equipment, a 3D model of the combustion equipment is created in 3D modeling software such as SolidWorks, SpaceClaim, or ANSYS DesignModeler. To ensure the feasibility of subsequent mesh generation and simulation calculations, the structure of the combustion equipment in the 3D model is simplified through engineering procedures, specifically including:

[0069] Ignore minor structural details that have little impact on flow and combustion, including: wall chamfers, fillets, bolt holes, etc.

[0070] For nozzles or holes with a diameter smaller than the preset diameter value, such as less than 5 mm, if they are isolated and do not affect the main field, they will be deleted directly; if they are adjacent to the main channel, they will be merged into the adjacent main channel.

[0071] For adjacent channels with a spacing smaller than the preset spacing value, such as less than 10 mm, they are merged into an equivalent channel to reduce geometric complexity.

[0072] The simplified 3D model should retain the key structural features of the burner, the air distribution channel, the fuel nozzle, and the flue gas outlet, etc., to ensure that the physical nature of flow and combustion remains unchanged.

[0073] Import the simplified 3D model into a meshing tool, such as ANSYS Meshing, Pointwise, or the snappyHexMesh / blockMesh tool in OpenFOAM. Select a meshing strategy based on the geometric complexity of the 3D model, specifically:

[0074] If the simplified combustion device structure is generally regular, i.e., mainly composed of rectangular, cylindrical, or hexahedral channels, without twisting or complex free-form surfaces (e.g., surfaces with a radius of curvature greater than 10 times the grid scale can be considered regular), then a structured grid is used. Orthogonalizing the structured grid along the flow direction can improve computational efficiency and convergence.

[0075] If the combustion equipment structure contains irregular areas, such as irregularly shaped burners, deflecting air vents, asymmetric outlets, or complex curved surfaces, then unstructured meshes, such as tetrahedral or hexahedral core meshes, should be used. Unstructured meshes can better adapt to complex geometric boundaries, but mesh quality needs to be controlled, such as keeping the skewness below 0.85, the aspect ratio below 5:1, which can be relaxed to 10:1 in the boundary layer region, and the orthogonality quality greater than 0.15.

[0076] Local mesh refinement should be performed in key areas, including the combustion reaction zone, the vicinity of fuel nozzles, the wall boundary layer region, and areas with significant temperature, velocity, and composition gradients. During refinement, the ratio of adjacent mesh sizes (i.e., the ratio of coarse to fine mesh sizes) should not exceed 1.2–1.5 to ensure numerical stability. After mesh generation, output the mesh file, such as in .msh, .cas, or OpenFOAM's polyMesh format.

[0077] Based on the type of combustion equipment structure and flow characteristics, the turbulence model and combustion model are selected as follows:

[0078] Turbulence model: For combustion equipment structures with high Reynolds number, strong swirling flow, or complex geometry, the k-ε model or shear stress transport model that can be realized should be selected first; for combustion equipment structures with significant near-wall effects, a low Reynolds number turbulence model or enhanced wall treatment can be used.

[0079] Combustion Model: For gaseous fuels or volatile solid or liquid fuels, eddy dissipation models or eddy dissipation conceptual models can be used; for scenarios requiring detailed chemical reactions, laminar flame surface models or reaction progress variable models are selected. The selection is based on fuel type, combustion chamber temperature distribution, and NOx formation prediction requirements.

[0080] Based on the obtained fuel composition data, the physical property parameters and chemical reaction parameters of the fuel are set, specifically as follows:

[0081] Physical properties include fuel density, specific heat capacity, thermal conductivity, and dynamic viscosity. For gaseous fuels, density can be calculated using the ideal gas law; for liquid or solid fuels, a property curve showing the change with temperature is required.

[0082] Chemical reaction parameters: Based on the carbon, hydrogen, oxygen, nitrogen, and sulfur content in the elemental analysis data, determine the stoichiometric reaction equation for the fuel and set parameters such as the reaction rate constant, activation energy, and pre-exponential factor. If a simplified chemical reaction mechanism is used, such as the 2-step or 4-step mechanism of methane combustion, the chemical reaction parameters need to be adjusted accordingly.

[0083] All physical property parameters and chemical reaction parameters are input into the simulation solver in SI form, such as ANSYS Fluent, OpenFOAM, CFX, etc. After completing the above settings, a preliminary computational fluid dynamics simulation model is formed.

[0084] After setting boundary conditions for the computational fluid dynamics simulation model based on the aforementioned standard operating condition parameters, the simulation calculation is performed to obtain the simulation results; and the deviation between the simulation results and the field measurement data is calculated, expressed as:

[0085]

[0086] In the formula, To account for the discrepancy between the simulation results and the field measurement data, For simulation results, This is data measured on-site;

[0087] If the deviation between the simulation results and the field measurement data does not exceed the preset deviation threshold, the computational fluid dynamics simulation model is determined as the standard simulation model; if the deviation between the simulation results and the field measurement data exceeds the preset deviation threshold, the model parameters and mesh of the computational fluid dynamics simulation model are adjusted, and the simulation calculation is performed again until the deviation does not exceed the preset deviation threshold, and the standard simulation model is output.

[0088] When there are multiple sets of standard operating condition parameters, the deviation between the simulation results and the field measurement data corresponding to each set of standard operating condition parameters must not exceed the preset deviation threshold.

[0089] Specifically, based on the fuel mass flow rate and fuel velocity of each burner, and the air volume and velocity of each air inlet in the standard operating condition parameters, the boundary conditions for the fuel inlet and air inlet are set respectively. The details are as follows:

[0090] For the fuel inlet, the mass flow rate inlet boundary is preferred, and the mass flow rate value is set according to the fuel mass flow rate of each burner; if the fuel velocity data is more reliable, the velocity inlet boundary can also be used. For the air inlet, either the velocity inlet boundary or the mass flow rate inlet boundary is selected according to the air volume and velocity of each air outlet.

[0091] Simultaneously, based on the elemental analysis data in the fuel composition data, the component mass fraction at the fuel inlet is set, including the mass percentages of elements such as carbon, hydrogen, oxygen, nitrogen, and sulfur; the temperature of each inlet is set according to the air temperature in the standard operating condition parameters. The flue gas outlet is set as the pressure outlet boundary, and an outlet static pressure value is given, typically one standard atmosphere, or set according to the actual back pressure of the equipment. For the combustion equipment wall, boundary conditions are selected based on the actual heat exchange situation, specifically:

[0092] If the outer wall of the equipment has an insulation layer with good insulation effect, it can be approximated as an insulated wall surface; if there is no insulation or heat dissipation needs to be considered, it can be set as a given heat flux density or convective heat transfer boundary. For computational fluid dynamics simulation models with multiple burners or multiple air outlets, it should be ensured that each inlet boundary is correctly connected to the interior of the computational domain to ensure accurate flow and component transport.

[0093] After setting the boundary conditions, check the physical rationality of all boundary conditions, such as the inlet flow velocity should be in the subsonic range and the temperature should not change abruptly.

[0094] Submit the computational fluid dynamics simulation model with pre-defined boundary conditions to the simulation solver for steady-state or transient simulation calculations. The simulation solver settings include:

[0095] A pressure-based solver is employed, suitable for low-velocity compressible or incompressible combustion flows. The pressure-velocity coupling method utilizes either the SIMPLE or PISO algorithm. A second-order upwind scheme is used for the convection term, and a central difference scheme is employed for the diffusion term to improve computational accuracy.

[0096] The convergence criterion is as follows: Set an absolute residual threshold for each physical equation: the energy equation must be below 1 × 10⁻⁻⁻⁶. 6 The continuity equation, momentum equation, turbulence equation, and component equation are all below 1×10⁻ 4 Simultaneously, key physical quantities, including the combustion equipment outlet temperature and the outlet carbon dioxide mass flow rate, are monitored as a function of iteration steps. When the relative rate of change of the above key physical quantities is less than 0.1% in 100 consecutive iterations, the calculation is determined to be converged by combining the absolute residual threshold condition.

[0097] After completing the simulation calculation, extract the key output parameters from the simulation results, including at least: temperature and velocity at the sensor location, outlet flue gas temperature, outlet flue gas velocity, carbon dioxide concentration, carbon monoxide concentration, and oxygen concentration in the outlet flue gas.

[0098] The output values ​​at the corresponding measurement points in the simulation results are compared with the field measurement data, and the deviation between the two is calculated. Specifically:

[0099] For each set of corresponding data points, such as the temperature at a certain sensor location, the deviation is obtained by calculating the expression for the difference between the simulation result and the field measurement data. For multiple sets of data, such as multiple measuring points or multiple sets of standard operating conditions, the deviation of each set is calculated separately, and the maximum or average value is taken as the basis for judging the overall deviation.

[0100] A preset deviation threshold, such as 5% or 10%, can be determined based on the actual accuracy requirements of the controlled emission enterprise. If the deviation of all corresponding data points does not exceed this threshold, the simulation accuracy of the current computational fluid dynamics simulation model is considered to meet the requirements, and the computational fluid dynamics simulation model is determined as the standard simulation model.

[0101] If the deviation at any data point exceeds the deviation threshold, it indicates that the current computational fluid dynamics simulation model fails to accurately reflect the actual combustion process; and the computational fluid dynamics simulation model will be adjusted, including but not limited to:

[0102] Adjust turbulence model parameters, such as turbulence intensity and turbulence viscosity ratio, within a range that is usually no more than ±20% of the default value; adjust combustion model parameters, such as reaction rate constant and activation energy; and adjust radiation model parameters, such as absorption coefficient and scattering coefficient. Adjustments can be based on sensitivity analysis or engineering experience.

[0103] For areas with large deviations, such as the combustion reaction zone and the vicinity of the combustion equipment outlet measuring point, local mesh refinement is carried out, reducing the mesh size to one-half to one-quarter of the original size.

[0104] After the adjustment is completed, re-execute the simulation calculation, deviation calculation and deviation threshold judgment until all deviations meet the preset requirements.

[0105] S3. Based on the standard simulation model, and using equipment operating condition data and fuel composition data, perform simulations under various combustion conditions to obtain simulation data under different combustion conditions, and calculate the corresponding carbon emission factors based on the simulation data.

[0106] In this embodiment, multiple different combinations of equipment operating condition parameters are selected from the equipment operating condition data, with each set of parameters corresponding to a combustion condition. For each combustion condition, the corresponding equipment operating condition parameters and fuel composition data are used as the inlet boundary conditions of a standard simulation model for simulation calculation, yielding simulation data for that combustion condition. The simulation data includes: fuel carbon content, the sum of fuel mass flow rates of each burner, outlet carbon dioxide mass flow rate, temperature field, and velocity field. Based on the fuel carbon content, the sum of fuel mass flow rates of each burner, and the outlet carbon dioxide mass flow rate, the corresponding carbon emission factor is calculated, expressed as:

[0107]

[0108]

[0109] In the formula, As a carbon emission factor, The carbon content of the fuel, For carbon oxidation rate, This represents the mass flow rate of carbon dioxide at the outlet. The carbon oxidation rate is the sum of the fuel mass flow rates of each burner. It is calculated based on the elemental analysis data corresponding to the current combustion condition and the sum of the fuel mass flow rates of each burner, representing the theoretical maximum carbon dioxide generation corresponding to the complete combustion of total carbon in the fuel. A deviation-verified standard simulation model based on field measurement data is used to simulate the combustion process under this combustion condition, extracting the outlet carbon dioxide mass flow rate of the combustion equipment from the simulation results. The carbon oxidation rate under this combustion condition is determined based on the ratio of the total carbon generated to the total carbon input for combustion. Furthermore, the carbon oxidation rate is calculated independently for each combustion condition under different loads, air distribution parameters, and fuel batches, achieving dynamic acquisition of the carbon oxidation rate.

[0110] Specifically, after obtaining the standard simulation model, based on the acquired equipment operating condition data and fuel composition data, multiple combinations of different equipment operating condition parameters covering the actual operating range of the equipment are selected. The selection rules include:

[0111] The common load range of the covered equipment is selected from the minimum stable load to the rated load at equal intervals, such as every 10% of the rated load, or critical load points, such as 30%, 50%, 75%, and 100% of the rated load.

[0112] It covers fuel composition data from different batches, including different combinations of elemental analysis data and industrial analysis data;

[0113] It includes various air distribution schemes, such as combinations of different total air supply volumes, air volume, air velocity, and air temperature at each air outlet.

[0114] Each set of equipment operating parameters corresponds to an independent set of combustion conditions. For each set of combustion conditions, the following parameters are extracted as the input boundary conditions for the standard simulation model: load, fuel consumption, fuel mass flow rate and fuel velocity of each burner, total air volume, air volume, air velocity and air temperature of each air outlet; and carbon and hydrogen content of the fuel from the elemental analysis data; and moisture and ash content from the industrial analysis data.

[0115] For each selected combustion condition, the corresponding equipment operating parameters and fuel composition data are used as inlet boundary conditions and substituted into the established standard simulation model. The specific operation is as follows:

[0116] In the standard simulation model, modify the parameter values ​​in the inlet boundary conditions one by one, including: mass flow rate, component mass fraction, temperature, etc., to make them consistent with the equipment operating condition parameters and fuel composition data corresponding to the current combustion conditions;

[0117] Keep the calibrated turbulence model, combustion model, radiation model, and solution parameters in the standard simulation model unchanged; submit the simulation task for each combustion condition sequentially until all combustion conditions meet the convergence condition. After the simulation of each combustion condition is completed, extract the following simulation data from the solution results:

[0118] The carbon content of the fuel is read directly from the inlet boundary condition setting value for this combustion condition, in units of mass percentage (%) or mass fraction (between 0 and 1).

[0119] The sum of the fuel mass flow rates of each burner, that is, the total mass flow rate of all fuel inlets, in kilograms per second or kilograms per hour;

[0120] The mass flow rate of carbon dioxide at the outlet is obtained by applying the area integral of the density and velocity of carbon dioxide at the cross-section of the flue gas outlet, and the unit is kilograms per second or kilograms per hour.

[0121] Temperature field refers to the temperature distribution data of the entire area inside and outside the combustion equipment. It can be output in the form of structured grid points, such as the temperature value of each grid node or extract typical cross-sections, such as the temperature cloud map data of the central symmetry plane and the outlet cross-section.

[0122] The velocity field is the velocity vector distribution data of the entire area inside and outside the combustion equipment. It can also output three-dimensional velocity components or extract the velocity distribution of key sections by grid nodes.

[0123] S4. Select the corresponding modeling method according to the data volume of the simulation data, and construct a dynamic carbon emission factor calculation model with carbon emission factor and combustion equipment outlet temperature as outputs.

[0124] Specifically, after obtaining the simulation data, the data volume of the simulation data is counted, that is, the number of simulation data sets for different combustion conditions. In this embodiment, the first threshold is preset to 100 sets; the second threshold is preset to 5000 sets, and the corresponding modeling method is selected according to the range of the data volume.

[0125] If the number of data sets does not exceed 100, a data fitting model is used as the dynamic carbon emission factor calculation model; if the number of data sets exceeds 100 but does not exceed 5000, a machine learning model is used as the dynamic carbon emission factor calculation model; if the number of data sets exceeds 5000, a deep learning model is used as the dynamic carbon emission factor calculation model.

[0126] The above thresholds can be adjusted according to the complexity of the actual operating conditions of the equipment and the enterprise's computing resources. For example, for small and medium-sized combustion equipment, the preset first threshold can be set to 50 groups and the preset second threshold can be set to 2000 groups; for large multi-burner systems, they can be appropriately increased.

[0127] In this embodiment, if the data volume of the simulated data does not exceed a preset first threshold, a data fitting model is used as the dynamic carbon emission factor calculation model. The data fitting model uses curve fitting or regression analysis methods, with the carbon content of the load and fuel as inputs, to establish a functional expression for the carbon emission factor and the outlet temperature of the combustion equipment.

[0128] Specifically, only the two parameters that have the most significant impact on carbon emission factors and combustion equipment outlet temperature are selected as inputs: load and fuel carbon content. Load can be obtained from equipment operating condition data, and fuel carbon content is derived from elemental analysis data of fuel composition data.

[0129] Curve fitting or regression analysis methods are used to establish functional expressions for carbon emission factors and combustion equipment outlet temperature, respectively, regarding load and fuel carbon content. For example, polynomial fitting, such as quadratic, cubic, or exponential polynomial fitting functions, can be employed. The fitting process aims to minimize the error between predicted and simulated values, and the function coefficients are determined using the least squares method.

[0130] The data fitting model takes the load and fuel carbon content as inputs and outputs the carbon emission factor under the current operating conditions in real time, in kilograms of carbon dioxide per kilogram of fuel and the flue gas temperature at the combustion equipment outlet, in degrees Celsius.

[0131] Use a small amount of simulated data that was not used in the fitting process, such as reserving 10% of the samples to verify the fitting accuracy. If the error is within an acceptable range, such as a deviation of less than 5%, then deploy the data fitting model as a dynamic carbon emission factor calculation model.

[0132] If the data volume of the simulated data exceeds a preset first threshold but does not exceed a preset second threshold, a machine learning model is used as the dynamic carbon emission factor calculation model. The machine learning model divides the simulated data into a training set, a validation set, and a test set. It uses the minimum prediction error of the carbon emission factor and the outlet temperature of the combustion equipment as the objective loss function, employs a gradient descent algorithm for training, and uses a Bayesian optimization algorithm to optimize the model's hyperparameters. The inputs of the machine learning model include load, elemental analysis data, industrial analysis data, and the air volume, wind speed, and wind temperature of each air outlet.

[0133] Specifically, all simulation data is divided proportionally into training, validation, and test sets. For example, 60% of the simulation data is used for training, 20% for validating hyperparameters, and 20% for final testing. The division ensures that different operating conditions are similarly distributed across the sets.

[0134] The input to the machine learning model includes the following dimensions: load, elemental analysis data, industrial analysis data, air volume, air speed and air temperature of each air outlet. If the equipment has multiple independent air outlets, the three parameters of each air outlet are used as independent input features.

[0135] Common machine learning regression algorithms such as random forest, gradient boosting tree, or support vector regression can be used. The initial machine learning model structure can be set empirically, for example, the number of decision trees in a random forest can be set to 100, and the maximum depth can be set to 10.

[0136] The objective loss function is to minimize the prediction errors of carbon emission factor and combustion equipment outlet temperature. The objective loss function can be either mean squared error or mean absolute error. When combining the errors of the two output variables, a weighted sum is used, such as a weight of 0.5 for each.

[0137] Gradient descent algorithms, such as Adam, RMSprop, or node splitting algorithms for tree models, are used to train machine learning models.

[0138] Bayesian optimization algorithms are used to optimize the hyperparameters of the machine learning model, such as learning rate, tree depth, and regularization coefficient. The loss function values ​​under different combinations of hyperparameters are evaluated on the validation set, and the hyperparameter configuration with the minimum loss is selected.

[0139] Training stops when the target loss function on the validation set no longer decreases significantly or reaches the preset number of iterations, and the model is finally confirmed as a dynamic carbon emission factor calculation model.

[0140] If the volume of the simulated data exceeds a preset second threshold, a deep learning model is used as the dynamic carbon emission factor calculation model. The deep learning model divides the simulated data into a training set, a validation set, and a test set. The target loss function is to minimize the prediction error of the carbon emission factor and the outlet temperature of the combustion equipment. An adaptive learning rate optimization algorithm is used for training, and an early stopping method is used to prevent overfitting. The input of the deep learning model includes load, elemental analysis data, industrial analysis data, air volume, wind speed and wind temperature of each air outlet, temperature field and component characteristics. The component characteristics include at least the concentration fields of carbon dioxide, carbon monoxide and oxygen. Among them, the key section extraction method is used to extract the feature values ​​of the central symmetry plane and the outlet section of the combustion equipment as the input of the deep learning model for the temperature field and component characteristics.

[0141] The outputs of the three models are the carbon emission factor and the combustion equipment outlet temperature.

[0142] Specifically, the dataset is also divided into training, validation, and test sets proportionally, with a recommended ratio of 70%, 15%, and 15%, respectively. For temperature and concentration field data, spatial downsampling or extraction of key cross-sectional features is required to reduce the input dimensionality.

[0143] The inputs to the deep learning model include: load, elemental analysis data, industrial analysis data, air volume, wind speed and temperature at each air outlet, and temperature field. It extracts the temperature distribution of several representative sections or measuring points within the combustion equipment from the simulation data, which can be processed into a two-dimensional grid or a one-dimensional sequence. The component characteristics include at least the concentration fields of carbon dioxide, carbon monoxide and oxygen, and the concentration values ​​at key locations can also be extracted.

[0144] Employing multi-layer fully connected neural networks and convolutional neural networks, this system is suitable for inputting gridded temperature or concentration fields or using long short-term memory networks, and is applicable to time-series operating conditions. The number of nodes in the input layer is determined based on the feature dimension, and the hidden layers can be designed to have 2-5 layers, with each layer containing 64-512 neurons. The output layer contains two nodes, corresponding to the carbon emission factor and the combustion equipment outlet temperature, respectively.

[0145] The objective loss function is to minimize the prediction error of carbon emission factor and combustion equipment outlet temperature, and the mean square error or mean absolute error is commonly used.

[0146] An adaptive learning rate optimization algorithm, such as Adam, is used for training, with the initial learning rate set to 0.001. Early stopping is used to prevent overfitting, including: calculating the loss on the validation set after each training cycle; if the validation set loss does not decrease for 10 consecutive cycles, training is terminated early, and the model parameters with the minimum validation set loss are restored.

[0147] After training, the deep learning model is saved in a standard format, such as ONNX or TensorFlowSavedModel; it is ultimately confirmed as a dynamic carbon emission factor calculation model.

[0148] S5. The temperature deviation is obtained by comparing the real-time collected combustion equipment outlet temperature with the combustion equipment outlet temperature output by the dynamic carbon emission factor calculation model. When the temperature deviation exceeds the preset temperature threshold, the corresponding equipment operating condition data and fuel composition data are recorded, and the standard simulation model is re-simulated under the corresponding combustion condition. The dynamic carbon emission factor calculation model is updated and optimized based on the simulation data obtained from the simulation.

[0149] In this embodiment, the outlet temperature of the combustion equipment is collected in real time by a temperature sensor placed at the outlet of the combustion equipment, and compared with the outlet temperature of the combustion equipment output by the dynamic carbon emission factor calculation model to calculate the temperature deviation. When the temperature deviation does not exceed the preset temperature deviation threshold, the operation of the dynamic carbon emission factor calculation model is maintained. When the temperature deviation exceeds the preset temperature deviation threshold, the equipment operating condition data and fuel composition data at the time of the deviation are recorded. The standard simulation model is called again, and the recorded equipment operating condition data and fuel composition data are used as boundary conditions to optimize the solution parameters of the standard simulation model. Then, simulation is performed under the corresponding combustion condition to obtain new simulation data. The new simulation data is fused with the original simulation data to obtain updated simulation data. The carbon emission factor calculation model is retrained and optimized based on the updated simulation data.

[0150] Specifically, a temperature sensor with an accuracy of no less than ±0.5% is installed at the outlet of the combustion equipment to collect the outlet temperature of the combustion equipment in real time.

[0151] At the same time, the current equipment operating condition data and fuel composition data are input into the dynamic carbon emission factor calculation model in real time, and the model immediately outputs the corresponding combustion equipment outlet temperature.

[0152] The temperature deviation is calculated by comparing the combustion equipment outlet temperature measured by the sensor with the outlet temperature predicted by the dynamic carbon emission factor calculation model. The temperature deviation can be expressed as an absolute deviation (measured value minus predicted value) or a relative deviation (absolute deviation divided by measured value). In practical engineering, the absolute deviation is preferred.

[0153] A preset temperature deviation threshold is established. This threshold can be determined based on the actual calculation accuracy requirements of the emission control enterprise and the stable operating range of the combustion equipment, such as ±5 degrees Celsius or ±3 degrees Celsius. When the calculated temperature deviation does not exceed this threshold, it indicates that the prediction results of the current dynamic carbon emission factor calculation model are in good agreement with the actual operating conditions and are within the effective range. No intervention is required, and the current operating state of the dynamic carbon emission factor calculation model should be maintained.

[0154] When the temperature deviation exceeds the temperature deviation threshold, it is determined that the current dynamic carbon emission factor calculation model's prediction for this operating condition is inaccurate, and an update process needs to be initiated. Furthermore, the equipment operating condition data and fuel composition data at the time the deviation is triggered are automatically recorded.

[0155] The system automatically calls upon a calibrated and stored standard simulation model, using the recorded equipment operating condition data and fuel composition data at the time of triggering the deviation as the entry boundary conditions for this model. During the setting process, the solution parameters of the standard simulation model are locally optimized according to the specific circumstances of the deviation. Optimization measures include, but are not limited to:

[0156] Appropriately refine the mesh in the combustion reaction zone and the combustion equipment outlet area, such as reducing the local mesh size to half to a quarter of its original size, to improve local calculation accuracy; fine-tune the empirical constants in the turbulence model or combustion model to make the simulation results closer to the current actual operating conditions. Specifically, this includes:

[0157] For the turbulence model, the turbulence model constants Cμ (default 0.09, adjustment range 0.08~0.10), C1ε, C2ε, etc. can be adjusted;

[0158] For the combustion model, the empirical constants A (default 4.0, adjustment range 2.0~6.0) and B (default 0.5, adjustment range 0.3~0.7) can be adjusted.

[0159] If a laminar flame surface model is used, the strain rate parameter can be adjusted.

[0160] Reduce the convergence residual threshold of the simulation solver, for example, by reducing the absolute residual threshold of the energy equation from... Tighten to Other equations from Tighten to In order to obtain more accurate simulation results.

[0161] After setting up, submit the simulation calculation and execute the simulation under the corresponding combustion conditions. After the calculation converges, extract new simulation data from the simulation results and calculate the corresponding carbon emission factor.

[0162] The new simulation data is then merged with the existing simulation data. The merging method involves appending the new simulation data to the end of the existing simulation data to form expanded and updated simulation data. Based on the updated simulation data, step S4 is re-executed. Specifically, this includes:

[0163] If the current dynamic carbon emission factor calculation model is a data fitting model, then the curve fitting or regression analysis is performed again using the updated simulation data, and the coefficients in the function expression are updated.

[0164] If the current dynamic carbon emission factor calculation model is a machine learning model or a deep learning model, incremental learning is performed based on the original model parameters. For example, the old model parameters can be used as initial values, and the model can be trained for several more rounds using updated simulation data. Alternatively, the training set, validation set, and test set can be completely re-divided, and a new dynamic carbon emission factor calculation model can be trained from scratch. This embodiment adopts an incremental learning approach to retain knowledge of historical operating conditions while quickly adapting to new operating conditions.

[0165] After retraining, the new dynamic carbon emission factor calculation model is validated. Specifically, predictions are made using several sets of measured data before the trigger deviation, and the temperature deviation at the combustion equipment outlet temperature is compared to see if it falls back to within the preset temperature deviation threshold. If the requirement is met, the original dynamic carbon emission factor calculation model is replaced with the new dynamic carbon emission factor calculation model, and it continues to be put into real-time operation. If the temperature deviation threshold is still exceeded, the above steps are repeated until the temperature deviation meets the requirement.

[0166] The aforementioned feedback optimization mechanism enables the dynamic carbon emission factor calculation model to maintain high accuracy under conditions such as equipment aging, changes in fuel sources, and adjustments to operating strategies. The closed-loop logic of this method requires no frequent manual intervention and can run automatically and periodically, for example, performing a deviation check daily or weekly to ensure that the dynamic carbon emission factor calculation model remains synchronized with the actual operating status. Furthermore, the dynamic carbon emission factor calculation model is updated no more than once a day; if the error of the new model on the validation set exceeds 1.2 times that of the old model, it automatically rolls back to the old model.

[0167] like Figure 2 As shown, this is the second embodiment of the present invention. This embodiment provides a dynamic calculation system for fuel-end carbon emission factors based on computational fluid dynamics, including:

[0168] The data acquisition module is used to acquire and preprocess relevant data from controlled emission enterprises to obtain preprocessed combustion equipment geometric parameters, equipment operating condition data, fuel composition data, and on-site measurement data.

[0169] The simulation module is used to determine standard operating condition parameters based on equipment operating condition data; to construct a computational fluid dynamics simulation model based on the geometric parameters of the combustion equipment and fuel composition data; to perform simulation calculations on the computational fluid dynamics simulation model based on the standard operating condition parameters; to verify the deviation of the simulation results based on field measurement data; and to obtain a standard simulation model when preset conditions are met.

[0170] The emission factor calculation module is used to perform simulations under various combustion conditions based on equipment operating condition data and fuel composition data according to the standard simulation model, obtain simulation data under different combustion conditions, and calculate the corresponding carbon emission factor based on the simulation data.

[0171] The emission factor modeling module is used to select the corresponding modeling method according to the data volume of the simulation data, and to construct a dynamic carbon emission factor calculation model with carbon emission factor and combustion equipment outlet temperature as outputs.

[0172] The model feedback optimization module is used to compare the real-time collected combustion equipment outlet temperature with the combustion equipment outlet temperature output by the dynamic carbon emission factor calculation model to obtain the temperature deviation. When the temperature deviation exceeds the preset temperature threshold, the corresponding equipment operating condition data and fuel composition data are recorded, and the standard simulation model is re-simulated under the corresponding combustion conditions. Based on the simulation data obtained from the simulation, the dynamic carbon emission factor calculation model is updated and optimized.

[0173] In summary, this invention abandons the traditional IPCC default value method and, for the first time, deeply integrates computational fluid dynamics simulation with data-driven modeling. By constructing a high-fidelity standard simulation model and using field measurement data for multiple rounds of deviation verification, it ensures that the flow field, temperature field, and component field of the combustion process are highly consistent with reality, thereby obtaining high-confidence simulation data. The carbon emission factor calculation model trained based on this model can accurately reflect the real-time impact of fuel composition fluctuations, equipment structure differences, and changes in operating conditions on carbon emissions, significantly reducing calculation errors and providing accurate data support for carbon accounting. Based on the standard simulation model, this invention comprehensively covers combustion conditions under different loads, fuel batches, and air distribution schemes. It obtains simulation data across the entire operating range through batch simulation and adaptively selects data fitting, machine learning, or deep learning modeling methods according to the data volume. The constructed dynamic carbon emission factor calculation model maintains high prediction accuracy for various complex operating conditions, overcoming the limitation of traditional methods that are only applicable to design conditions, and realizing the leap from "single-point static" to "full-domain dynamic" carbon emission factors. This invention designs a real-time feedback mechanism based on the combustion equipment outlet temperature. When the deviation between the model's output temperature and the sensor's measured temperature exceeds a preset threshold, the system automatically records the complete operating parameters and fuel data for the triggered condition, re-invokes the standard simulation model for high-precision verification, and integrates the newly generated simulation data into the original dataset, triggering incremental learning or complete retraining of the dynamic model. This closed-loop strategy enables the carbon emission factor calculation model to continuously evolve with equipment aging, changes in fuel sources, and adjustments to operating strategies, always maintaining optimal predictive performance. The dynamic carbon emission factor provided by this invention can be directly used for enterprise-level carbon emission accounting and quota compliance, effectively suppressing accounting disputes, enhancing the credibility of carbon market data, and assisting enterprises in refined carbon asset management and the scientific advancement of carbon peaking and carbon neutrality goals.

[0174] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any other combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product, which includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions according to this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another.

[0175] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. This storage medium can be a read-only memory, a disk, or an optical disk, etc.

[0176] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in this application, and these should all be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A dynamic calculation method for fuel-end carbon emission factors based on computational fluid dynamics, characterized in that, include: Acquire and preprocess relevant data from emission-controlled enterprises to obtain preprocessed combustion equipment geometric parameters, equipment operating condition data, fuel composition data, and on-site measurement data; Standard operating condition parameters are determined based on equipment operating condition data; a computational fluid dynamics simulation model is constructed based on the geometric parameters of the combustion equipment and fuel composition data; the computational fluid dynamics simulation model is simulated based on the standard operating condition parameters; the simulation results are verified for deviation based on field measurement data; and a standard simulation model is obtained when preset conditions are met. Based on the standard simulation model, and using equipment operating condition data and fuel composition data, simulations under various combustion conditions are performed to obtain simulation data under different combustion conditions, and the corresponding carbon emission factors are calculated based on the simulation data. The corresponding modeling method is selected based on the data volume of the simulation data, and a dynamic carbon emission factor calculation model is constructed with carbon emission factor and combustion equipment outlet temperature as outputs. The temperature deviation is obtained by comparing the real-time collected combustion equipment outlet temperature with the combustion equipment outlet temperature output by the dynamic carbon emission factor calculation model. When the temperature deviation exceeds the preset temperature threshold, the corresponding equipment operating condition data and fuel composition data are recorded, and the standard simulation model is re-simulated under the corresponding combustion conditions. The dynamic carbon emission factor calculation model is then updated and optimized based on the simulation data obtained from the simulation.

2. The method for dynamic calculation of fuel-end carbon emission factors based on computational fluid dynamics according to claim 1, characterized in that, The data related to the controlled emission enterprises includes the geometric parameters of the combustion equipment, equipment operating condition data, fuel composition data, and on-site measurement data; the geometric parameters of the combustion equipment are obtained through the design drawings of the combustion equipment, including the geometric dimensions of the burner, the shape and size of the air distribution port, the geometric dimensions of the fuel nozzle, and the shape and size of the flue gas outlet; The equipment operating condition data is collected through combustion equipment sensors and enterprise production records, including load, fuel consumption, fuel mass flow rate and fuel velocity of each burner, total air volume, and air volume, air velocity and air temperature of each air outlet. The fuel composition data includes elemental analysis data and industrial analysis data. The elemental analysis data includes the content of carbon, hydrogen, oxygen, nitrogen, and sulfur, while the industrial analysis data includes the content of moisture, ash, volatile matter, and fixed carbon. The on-site measurement data is acquired by sensors arranged inside and at the outlet of the combustion equipment, including temperature and velocity at the sensor location, outlet flue gas temperature, outlet flue gas velocity, and at least one of carbon dioxide concentration, carbon monoxide concentration, and oxygen concentration in the outlet flue gas.

3. The method for dynamic calculation of fuel-end carbon emission factors based on computational fluid dynamics according to claim 1, characterized in that, The preprocessing includes: Unit standardization and format conversion are performed on the geometric parameters of the combustion equipment; outlier removal, missing value filling, and unit standardization are performed on the equipment operating condition data; data verification and normalization are performed on the fuel composition data; and filtering, noise reduction, and dimensional alignment are performed on the field measurement data.

4. The method for dynamic calculation of fuel-end carbon emission factors based on computational fluid dynamics according to claim 1, characterized in that, The standard operating condition parameters are determined based on the equipment operating condition data. A computational fluid dynamics simulation model is constructed based on the geometric parameters of the combustion equipment and fuel composition data, including: The data segment with load fluctuation less than the preset fluctuation threshold and key parameter change rate lower than the preset change rate threshold is selected from the equipment operation condition data. Cluster analysis is performed on the equipment operation condition data of the data segment. One or more typical operating conditions are selected from the cluster results as standard operating conditions, and the corresponding standard operating condition parameters are extracted. A three-dimensional model is constructed based on the geometric parameters of the combustion device, and the structure of the combustion device in the three-dimensional model is simplified, specifically: wall chamfers, fillets, bolt holes, and nozzles with diameters smaller than the preset diameter value are ignored, and adjacent channels with a spacing smaller than the preset spacing value are merged. When the simplified combustion equipment structure is regular, a structured mesh is used for mesh generation; otherwise, an unstructured mesh is used. After selecting the turbulence model and combustion model according to the type of combustion equipment structure, the physical property parameters and chemical reaction parameters of the fuel are set according to the fuel composition data, and finally the computational fluid dynamics simulation model is obtained.

5. The method for dynamic calculation of fuel-end carbon emission factors based on computational fluid dynamics according to claim 4, characterized in that, The computational fluid dynamics simulation model is simulated based on the standard operating condition parameters. Deviation verification of the simulation results is performed based on field measurement data. When preset conditions are met, a standard simulation model is obtained, including: After setting boundary conditions for the computational fluid dynamics simulation model based on the aforementioned standard operating condition parameters, the simulation calculation is performed to obtain the simulation results; and the deviation between the simulation results and the field measurement data is calculated, expressed as: ； In the formula, To account for the discrepancy between simulation results and field measurement data, For simulation results, This is data measured on-site; If the deviation between the simulation results and the field measurement data does not exceed the preset deviation threshold, the computational fluid dynamics simulation model is determined as the standard simulation model; if the deviation between the simulation results and the field measurement data exceeds the preset deviation threshold, the model parameters and mesh of the computational fluid dynamics simulation model are adjusted, and the simulation calculation is performed again until the deviation does not exceed the preset deviation threshold, and the standard simulation model is output. When there are multiple sets of standard operating condition parameters, the deviation between the simulation results and the field measurement data corresponding to each set of standard operating condition parameters must not exceed the preset deviation threshold.

6. The method for dynamically calculating fuel-end carbon emission factors based on computational fluid dynamics according to claim 1, characterized in that, The process involves performing simulations under various combustion conditions based on a standard simulation model, equipment operating condition data, and fuel composition data. Simulation data under different combustion conditions is then obtained, and the corresponding carbon emission factors are calculated based on the simulation data, including: Multiple sets of different equipment operating condition parameter combinations are selected from the equipment operating condition data. Each set of equipment operating condition parameters corresponds to a set of combustion conditions. For each set of combustion conditions, the corresponding equipment operating condition parameters and fuel composition data are used as the entry boundary conditions of the standard simulation model for simulation calculation to obtain the simulation data under that combustion condition. The simulation data includes: the carbon content of the fuel, the sum of the fuel mass flow rates of each burner, the outlet carbon dioxide mass flow rate, the temperature field, and the velocity field. Based on the carbon content of the fuel, the sum of the fuel mass flow rates of each burner, and the outlet carbon dioxide mass flow rate, the corresponding carbon emission factor is calculated, expressed as: ； ； In the formula, As a carbon emission factor, The carbon content of the fuel, For carbon oxidation rate, This represents the mass flow rate of carbon dioxide at the outlet. This is the sum of the fuel mass flow rates of each burner.

7. The method for dynamic calculation of fuel-end carbon emission factors based on computational fluid dynamics according to claim 6, characterized in that, The step of selecting a corresponding modeling method based on the data volume of the simulation data and constructing a dynamic carbon emission factor calculation model with carbon emission factor and combustion equipment outlet temperature as outputs includes: If the data volume of the simulation data does not exceed the preset first threshold, a data fitting model is used as the dynamic carbon emission factor calculation model. The data fitting model uses curve fitting or regression analysis methods, with the carbon content of the load and fuel as input, to establish a functional expression for the carbon emission factor and the outlet temperature of the combustion equipment. If the volume of the simulated data exceeds a preset first threshold but does not exceed a preset second threshold, a machine learning model is used as the dynamic carbon emission factor calculation model. The machine learning model divides the simulated data into a training set, a validation set, and a test set. It uses the minimum prediction error of the carbon emission factor and the outlet temperature of the combustion equipment as the objective loss function, employs a gradient descent algorithm for training, and uses a Bayesian optimization algorithm to optimize the model's hyperparameters. The inputs to the machine learning model include load, elemental analysis data, industrial analysis data, and the air volume, wind speed, and wind temperature of each air outlet. If the volume of the simulated data exceeds a preset second threshold, a deep learning model is used as the dynamic carbon emission factor calculation model. The deep learning model divides the simulated data into a training set, a validation set, and a test set. The target loss function is to minimize the prediction error of the carbon emission factor and the outlet temperature of the combustion equipment. An adaptive learning rate optimization algorithm is used for training, and an early stopping method is used to prevent overfitting. The inputs of the deep learning model include load, elemental analysis data, industrial analysis data, air volume, wind speed and wind temperature at each air outlet, temperature field and component characteristics. The component characteristics include at least the concentration fields of carbon dioxide, carbon monoxide and oxygen. The outputs of the three models are the carbon emission factor and the combustion equipment outlet temperature.

8. The method for dynamic calculation of fuel-end carbon emission factors based on computational fluid dynamics according to claim 5, characterized in that, The process involves comparing the real-time collected combustion equipment outlet temperature with the combustion equipment outlet temperature output by the dynamic carbon emission factor calculation model to obtain the temperature deviation. When the temperature deviation exceeds a preset temperature threshold, the corresponding equipment operating condition data and fuel composition data are recorded, and the standard simulation model is re-simulated under the corresponding combustion conditions. Based on the simulation data obtained from the simulation, the dynamic carbon emission factor calculation model is updated and optimized, including: The temperature at the outlet of the combustion equipment is collected in real time by a temperature sensor placed at the outlet of the combustion equipment, and compared with the temperature at the outlet of the combustion equipment output by the dynamic carbon emission factor calculation model. The temperature deviation is calculated. When the temperature deviation does not exceed the preset temperature deviation threshold, the operation of the dynamic carbon emission factor calculation model is maintained. When the temperature deviation exceeds the preset temperature deviation threshold, the equipment operating condition data and fuel composition data at the time of the deviation are recorded. The standard simulation model is re-invoked, and the recorded equipment operating condition data and fuel composition data are used as boundary conditions. After optimizing the solution parameters of the standard simulation model, simulation is performed under the corresponding combustion conditions to obtain new simulation data. The new simulation data is fused with the original simulation data to obtain updated simulation data; and the carbon emission factor calculation model is retrained and optimized based on the updated simulation data.

9. A dynamic calculation system for fuel-end carbon emission factors based on computational fluid dynamics, applied to the dynamic calculation method for fuel-end carbon emission factors based on computational fluid dynamics as described in any one of claims 1 to 8, characterized in that, include: The data acquisition module is used to acquire and preprocess relevant data from controlled emission enterprises to obtain preprocessed combustion equipment geometric parameters, equipment operating condition data, fuel composition data, and on-site measurement data. The simulation module is used to determine standard operating condition parameters based on equipment operating condition data; to construct a computational fluid dynamics simulation model based on the geometric parameters of the combustion equipment and fuel composition data; to perform simulation calculations on the computational fluid dynamics simulation model based on the standard operating condition parameters; to verify the deviation of the simulation results based on field measurement data; and to obtain a standard simulation model when preset conditions are met. The emission factor calculation module is used to perform simulations under various combustion conditions based on equipment operating condition data and fuel composition data according to the standard simulation model, obtain simulation data under different combustion conditions, and calculate the corresponding carbon emission factor based on the simulation data. The emission factor modeling module is used to select the corresponding modeling method according to the data volume of the simulation data, and to construct a dynamic carbon emission factor calculation model with carbon emission factor and combustion equipment outlet temperature as outputs. The model feedback optimization module is used to compare the real-time collected combustion equipment outlet temperature with the combustion equipment outlet temperature output by the dynamic carbon emission factor calculation model to obtain the temperature deviation. When the temperature deviation exceeds the preset temperature threshold, the corresponding equipment operating condition data and fuel composition data are recorded, and the standard simulation model is re-simulated under the corresponding combustion conditions. Based on the simulation data obtained from the simulation, the dynamic carbon emission factor calculation model is updated and optimized.

Images