A dynamic multi-factor coupling prediction method for coke oven gas

Abstract

The application provides a coke oven gas dynamic multi-factor coupling prediction method, relates to the coking technology field, and comprises the following steps: S1, data acquisition and preprocessing; S2, dynamic reaction model establishment: simulating the dynamic change of gas precipitation in the coking process, considering the interaction between various factors, and establishing a dynamic mathematical model; S3, model verification and correction; S4, real-time calculation and prediction. The coke oven gas dynamic multi-factor coupling prediction method comprehensively considers various factors such as coal quality characteristics, coke oven heating system and operation parameters, and embeds a multi-factor coupling mechanism, effectively overcomes the problem of large prediction error of the traditional coal quality element balance method for high volatile matter coal, and avoids the problem of poor generalization ability caused by the dependence of the empirical formula method on specific coal data. The method does not depend on the data of specific coal, can accurately predict atypical coal, and significantly improves the prediction accuracy and generalization ability.

Description

Technical Field

[0001] This application relates to the field of coke oven dry distillation technology, and in particular to a dynamic multi-factor coupled prediction method for coke oven gas. Background Technology

[0002] Coke oven gas is a byproduct of coal distillation in coke ovens. Its main components are hydrogen, methane, carbon monoxide, and small amounts of nitrogen and hydrocarbons. It has a medium-to-high calorific value and is an important chemical raw material and fuel. Its composition and yield directly reflect the coke oven heating regime and coal quality characteristics. Coke oven gas forecasting can optimize coke oven thermal operation, predict gas yield and quality, and provide a design basis for subsequent purification processes. This reduces energy consumption, increases chemical product recovery rates, and simultaneously reduces emissions of carcinogens such as benzo[a]pyrene, achieving both economic and environmental benefits.

[0003] Existing methods for predicting coke oven gas mainly include the coal quality element balance method, empirical formula method, and numerical simulation. Among them, the coal quality element balance method is based on industrial / elemental analysis data of coal and calculates the theoretical gas production through the conservation of carbon, hydrogen, and oxygen elements. Although it is fast in calculation, it assumes that all carbon and hydrogen elements are completely converted into gas and ignores the secondary reactions of tar and semi-coke. Therefore, the prediction error for high-volatile coal types can reach 15% to 20%, and it cannot reflect the dynamic changes in gas evolution during the coking process.

[0004] Empirical formulas rely on data specific to certain coal types, have poor generalization ability, and exhibit significant prediction biases for atypical coals (such as high-sulfur coal or blended coal). Furthermore, they cannot provide detailed information on coal gas composition, limiting their application in optimization control. In addition, numerical simulations, such as CPFD or CFD methods, are computationally extremely expensive, requiring high-performance computing resources. Their models rely on numerous assumptions, such as tar cracking kinetic parameters, leading to discrepancies between the results and actual production. Currently, they are primarily used for theoretical research rather than real-time control. Summary of the Invention

[0005] This application is made in view of the above-mentioned problems, and its purpose is to provide a dynamic multi-factor coupled prediction method for coke oven gas to solve the problems mentioned in the background art. To achieve the above objectives, the present invention provides the following technical solution: a method for dynamic multi-factor coupled prediction of coke oven gas, comprising the following steps: S1. Data Acquisition and Preprocessing: Collect comprehensive and accurate basic data to provide a reliable basis for subsequent calculations, and organize and pre-process the data to improve data quality; S2. Establish a dynamic reaction model: Simulate the dynamic changes in gas evolution during coke oven dry distillation, consider the interaction between multiple factors, and establish a dynamic mathematical model; S3. Model Validation and Correction: Validate and correct the model based on actual production data to make the model's prediction results more consistent with the actual data and improve the model's predictive ability. S4. Real-time calculation and prediction: Based on the real-time collected coke oven operating parameters, the validated model is used to perform real-time calculation and prediction of gas yield and composition, providing timely guidance for coke oven thermal operation.

[0006] Furthermore, in step S1, the basic data includes coal quality characteristic data (elemental composition, volatile matter content, sulfur content, etc.) and coke oven heating regime parameters (heating temperature, heating time, excess air coefficient, etc.). The coal quality characteristic data is obtained through laboratory testing, and the coke oven heating regime parameters are collected using sensors and monitoring systems at the coke oven site.

[0007] Furthermore, in step S1, preliminary processing is performed: using data processing software (such as Excel, Python's pandas library, etc.), data cleaning algorithms are used to clean the collected data, remove outliers and erroneous data, and statistical methods are used to standardize the data, such as Z-score standardization, so that data of different dimensions are comparable.

[0008] Furthermore, in step S2, chemical kinetics software (such as ChemKin) is used to construct and solve the dynamic reaction model. The construction process includes the following sub-steps: S21. Define the model objectives and boundary conditions: Define the physical boundaries (such as carbonization chamber and combustion chamber) and time range (coking cycle) of the coke oven dry distillation process, and define the input / output variables; S22. Selection of Modeling Method and Mathematical Framework: A set of differential equations based on mass conservation (used to describe the change of mass of each component in the system over time, ensuring that the total mass of substances in the model remains constant, i.e., in the process of coke oven dry distillation, the total mass of elements such as carbon, hydrogen, and oxygen in coal should be conserved during the conversion into coal gas, tar, and semi-coke), energy conservation (used to describe the input, output, and conversion of energy in the system, i.e., the energy change in the process of coke oven dry distillation), and chemical kinetics (used to describe the chemical reaction rate and its influencing factors, i.e., the influence of various factors such as temperature, pressure, and catalysts on the reaction rate of coal pyrolysis and gasification in coke oven dry distillation) is used to describe the dynamic process of coal gas evolution, define key reaction paths, and introduce reaction rate equations; S23. Introduce a multi-factor coupling mechanism: embed coal quality characteristics (such as volatile matter, ash, and sulfur), heating regime (temperature field and heating rate), pressure and atmosphere effects, etc. into the model. Based on the embedded multi-factor information, adjust the reaction rate equation in the model so that it can more accurately reflect the reaction rate in the actual dry distillation process. At the same time, based on the interaction of multiple factors, modify the product distribution model to predict the specific composition and yield of coal gas under different conditions. S24. Dynamic process phased modeling: Based on the changing characteristics of coal quality, temperature range and the timing of key reactions, the dry distillation process is divided into stages such as drying period, pyrolysis period and semi-coke shrinkage period. Sub-models are established for each stage, and dynamic transfer between stages is realized through state variables. S25. Solution and Numerical Implementation: Discretization: The partial differential equations are discretized using the finite difference method or the finite element method, dividing the continuous spatial and temporal domains into finite grid points or elements; Numerical solver selection: Based on the characteristics of the discretized algebraic equation system, an explicit or implicit numerical solver is selected for solving the problem. Explicit solvers are suitable for simple problems and have a fast calculation speed but poor stability, while implicit solvers are suitable for complex problems and have a slower calculation speed but good stability. Software platform implementation: Numerical solutions to the model are achieved using MATLAB, COMSOL, or multiphysics simulation platforms.

[0009] Furthermore, the specific input / output variables are as follows: Input variables: coal properties (volatile matter, ash content, sulfur content, etc.), heating regime (temperature field distribution, heating rate), operating parameters (pressure, atmosphere); Output variables: gas yield, gas composition (H2, CH4, CO, etc.), and the amount of tar and semi-coke produced.

[0010] Furthermore, in step S23, the embedding methods for coal quality characteristics, heating regime, pressure, and atmosphere effects are as follows: Coal quality characteristics embedded: including volatile matter content, ash content, sulfur content, etc. These factors directly affect the pyrolysis characteristics and product distribution of coal. Specific values ​​are obtained through experimental measurement or by referring to coal quality analysis reports. Coal quality characteristic parameters are used as input variables of the model. The influence of different coal qualities on coal gas evolution is reflected by adjusting the parameters in the reaction rate equation. Heating regime embedding: This involves temperature field distribution and heating rate. Temperature field changes affect the reaction rate of each component in coal, while heating rate affects the speed of the dry distillation process. Temperature field data and heating rate are recorded by temperature sensors and heating control system. Temperature field data and heating rate are used as time-dependent input variables of the model. The gas evolution process under different heating regimes is simulated by dynamically adjusting the reaction rate. Pressure and atmosphere effects are embedded: The pressure and atmosphere (such as oxygen concentration, carbon dioxide concentration, etc.) in the coke oven affect the type and quantity of gas released. Real-time monitoring is carried out through pressure sensors and gas analyzers. The pressure and atmosphere data are used as boundary conditions or input variables of the model. The influence of these factors on gas release is reflected by adjusting the reaction path and reaction rate equation.

[0011] Furthermore, in step S24, the specific sub-models for each stage are as follows: Drying period sub-model: Since the main process during the drying period is the evaporation of moisture in the coal, the coal quality remains basically unchanged, but the temperature gradually increases to prepare for the subsequent pyrolysis reaction. This sub-model mainly simulates the evaporation process of moisture, taking into account the effects of temperature, humidity and ventilation conditions on the evaporation rate. Pyrolysis sub-model: As the temperature further increases, volatile matter in coal begins to be released in large quantities, undergoing complex pyrolysis reactions to generate coal gas, tar, and semi-coke. The pyrolysis stage is the main period for coal gas generation and is also the part that the model needs to focus on simulating. This sub-model focuses on simulating the release of volatile matter and pyrolysis reactions, including primary pyrolysis and possible secondary pyrolysis reactions. It is necessary to define key reaction pathways, introduce reaction rate equations, and consider the influence of coal quality characteristics (such as volatile matter content, ash content, etc.) on the reaction rate. Semi-coke shrinkage period sub-model: In the later stage of pyrolysis, semi-coke begins to shrink, and the internal pore structure changes. At the same time, there may be slight secondary reactions. The semi-coke shrinkage period also has a certain impact on the final composition and yield of coal gas. This sub-model simulates the shrinkage process of semi-coke, the changes in the internal pore structure, and the possible slight secondary reactions.

[0012] Furthermore, in step S24, dynamic transfer between stages is achieved through state variables (including key parameters such as temperature, pressure, gas composition, and semi-coke quality) to realize dynamic transfer between sub-models of each stage (drying period sub-model, pyrolysis period sub-model, and semi-coke shrinkage period sub-model). That is, at the end of each stage, the state variables are updated according to the simulation results of that stage and used as the initial conditions for the next stage to input into the sub-model of the next stage.

[0013] Furthermore, the specific operations of step S3 are as follows: collect actual coal gas yield and composition data as a verification dataset, and compare the model prediction results with the verification dataset point by point to calculate the prediction error. Indicators such as absolute error, relative error, and mean square error can be used to quantify the prediction accuracy. Analyze the prediction error to identify the causes of deviations, such as unreasonable model assumptions, improper parameter settings, or unconsidered key factors. Adjust the model parameters or structure accordingly. In actual operation, prioritize parameter adjustment, such as correcting parameters in the reaction rate equation, adjusting coal quality characteristic parameters, and optimizing heating regime parameters. If parameter adjustment cannot significantly improve prediction accuracy, consider optimizing the model structure, such as introducing more complex chemical reaction pathways, a more refined model considering multi-factor coupling effects, and adding stage divisions to the dry distillation process. After adjusting the model parameters or structure, re-perform model predictions and comparative verifications to evaluate the improvement effect until the model prediction results highly match the actual data.

[0014] Furthermore, in step S4, a data interface is established to input the real-time coke oven operating parameters into the dynamic reaction model, the model is run, the gas yield and composition under the current conditions are calculated, and the results are output in the form of numbers or charts. The output results include the gas yield, the proportion of each gas component, and other relevant indicators (such as temperature distribution, pressure change, etc.).

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention, by comprehensively considering multiple factors such as coal quality characteristics, coke oven heating regime, and operating parameters, and embedding a multi-factor coupling mechanism, effectively overcomes the problem of large prediction errors for high-volatile coal types in traditional coal element balance methods. It also avoids the poor generalization ability caused by the reliance on specific coal type data in empirical formula methods. This method does not depend on specific coal type data and can accurately predict atypical coal types, including high-sulfur coal and blended coals, significantly improving prediction accuracy and generalization ability.

[0016] 2. This invention establishes a dynamic reaction model to simulate the dynamic changes in gas evolution during the dry distillation process of a coke oven in stages, including key stages such as the drying period, pyrolysis period, and semi-coke shrinkage period. This accurately reflects the real-time changes in gas composition and yield during coking. Furthermore, this method requires no high-performance computing resources; through a reasonable mathematical framework and numerical solution strategy, it reduces computational costs, making it particularly suitable for real-time calculation and prediction. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this drawing or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this drawing. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the steps of a dynamic multi-factor coupling prediction method for coke oven gas according to the present invention.

[0019] The purpose, features, and advantages of this accompanying drawing will be further explained in conjunction with the embodiments and with reference to the accompanying drawing. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this application clearer, the following description and illustration are provided in conjunction with embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.

[0021] Obviously, the following description is merely some examples or embodiments of this application. Those skilled in the art can apply this application to other similar scenarios without any inventive effort. Furthermore, it is understood that although the effort involved in such development may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, any changes to design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as insufficient disclosure of the content of this application.

[0022] Unless otherwise specified, the terms "comprising" and "including" as used in this application can be open-ended or closed-ended. For example, "comprising" and "including" can mean that other components not listed may also be included, or that only the listed components may be included.

[0023] Unless otherwise specified, the term "or" is inclusive in this application. For example, the phrase "A or B" means "A, B, or both A and B". More specifically, the condition "A or B" is satisfied by any of the following conditions: A is true (or exists) and B is false (or does not exist); A is false (or does not exist) and B is true (or exists); or both A and B are true (or exist).

[0024] Terminology Explanation: Example 1: Prediction of Coke Oven Gas Yield and Composition for High Volatile Coal Types Step S1, Data Acquisition and Preprocessing: Coal quality characteristics data: A high volatile coal type was selected, and its elemental composition (C: 85%, H: 5%, O: 8%, N: 1%, S: 1%) was obtained through laboratory testing. The volatile matter content was 35%, the ash content was 10%, and the sulfur content was 1%.

[0025] Coke oven heating parameters: Data collected using on-site sensors and monitoring systems in the coke oven, heating temperature is 1100℃, heating time is 18 hours, and excess air coefficient is 1.2.

[0026] Data preprocessing: Data cleaning and standardization were performed using Python's pandas library. After removing outliers, the data standardization (Z-score standardization) results are as follows: volatile matter content standardized value is 1.2, ash content standardized value is -0.5, and sulfur content standardized value is 0.8.

[0027] Step S2: Establish a dynamic reaction model: Model Objectives and Boundary Conditions: The carbonization chamber is defined as the physical boundary, and the coking cycle is 20 hours. Input variables include coal properties (volatile matter 35%, ash 10%, sulfur 1%), heating regime (temperature 1100℃, uniform heating rate), and operating parameters (pressure 1 atm, atmosphere: air). Output variables are coal gas yield, coal gas composition (H2, CH4, CO, etc.), and the amount of tar and semi-coke produced.

[0028] Modeling methods and mathematical framework: A set of differential equations based on mass conservation, energy conservation and chemical kinetics is used to describe the dynamic process of coal gas evolution.

[0029] Multi-factor coupling mechanism: Coal quality characteristics embedded: high volatile content, adjusting parameters in the reaction rate equation to reflect the release of more volatile substances.

[0030] Heating regime embedding: uniform temperature field, constant heating rate, simulating a stable pyrolysis environment.

[0031] Pressure and atmosphere effects are embedded: the pressure is 1 atm and the atmosphere is air. The effect of oxygen on the coal oxidation reaction is considered.

[0032] Dynamic process phased modeling: divided into drying period (first 2 hours), pyrolysis period (2-18 hours), and semi-coke shrinkage period (18-20 hours).

[0033] The partial differential equations are discretized using the finite difference method, and an implicit numerical solver is selected for solving them. The numerical solution of the model is realized using the MATLAB platform.

[0034] Step S3, Model Validation and Correction: Actual production data showed a coal gas yield of 18.5%, with H2 at 55%, CH4 at 25%, and CO at 15%. The model predicted a coal gas yield of 18.2%, with H2 at 54%, CH4 at 26%, and CO at 16%. The calculated prediction error was 0.03%, which is extremely small.

[0035] Step S4, Real-time Calculation and Prediction: A data interface was established to obtain the coke oven operating parameters in real time, input them into the dynamic reaction model, and calculate that the gas yield under the current conditions is 18.3%, with H2 at 54.5%, CH4 at 25.5%, and CO at 15.5%.

[0036] in conclusion: 1. The model prediction results are highly consistent with the actual data. The prediction error of coal gas yield is reduced to 0.3%, and the prediction error of composition is within 1%, which significantly improves the prediction accuracy and can predict the composition of coal gas in detail.

[0037] 2. It can dynamically reflect the changes in gas evolution during the coking process, providing strong support for optimized control.

[0038] 3. Low computational cost, no need for high-performance computing resources, suitable for real-time control in actual production.

[0039] Example 2: Prediction and Optimization of Coke Oven Gas Yield and Composition in Coal Blending Step S1, Data Acquisition and Preprocessing: Coal quality characteristics data: Two types of coal were selected and blended in a certain proportion: Coal A (volatile matter 25%, ash 12%, sulfur 0.8%) and Coal B (volatile matter 15%, ash 8%, sulfur 1.2%), with a blending ratio of A:B=6:4.

[0040] Coke oven heating parameters: heating temperature 1050℃, heating time 18 hours, excess air coefficient 1.15.

[0041] Data preprocessing: Excel was used for data cleaning and standardization. The standardized value for volatile matter content was 0.5, the standardized value for ash content was -0.2, and the standardized value for sulfur content was 2.1.

[0042] Step S2: Establish a dynamic reaction model: Similar to Example 1, but the coal quality characteristics in the input variables are the comprehensive characteristics after blending (volatile matter 21%, ash 10.4%, sulfur 0.96%).

[0043] It also employs a set of differential equations based on the principles of mass conservation, energy conservation, and chemical kinetics.

[0044] Multi-factor coupling mechanism: Coal quality characteristics embedded: Considering the coal blending ratio, the reaction rate equation is adjusted.

[0045] Heating regime embedding: The temperature field and heating rate are set according to the actual heating regime.

[0046] Pressure and atmosphere effects are embedded: the pressure is 1 atm and the atmosphere is the adjusted air ratio.

[0047] Dynamic process phased modeling: Similar to Example 1, the focus is on simulating the reaction characteristics of coal blending during the pyrolysis period.

[0048] Solution and numerical implementation: The finite element method is used for discretization, and COMSOL is used for numerical solution.

[0049] Step S3, Model Validation and Correction: Actual production data was collected, including: actual coal gas yield of 160 m³ / t coal, H₂ content of 50%, CH₄ content of 20%, CO content of 20%, and sulfur content of 1.2%. Comparing the model's predictions with the actual data, it was found that, except for sulfur content, the model's prediction error was less than 1%, indicating accurate predictions. However, the predicted sulfur content of the coal gas was 1.8%, resulting in a prediction deviation of +0.6%, meaning an overestimation of 50%.

[0050] The reason for the bias is that the model did not fully consider the conversion pathways of sulfur in high-sulfur coal, such as the competitive reactions in which sulfur is released in the form of H2S or SO2, which led to the distortion of sulfur distribution prediction.

[0051] To correct the prediction bias of sulfur content, the model needs to be optimized by introducing the sulfur conversion reaction pathway: Operation 1: Add a new reaction pathway, supplementing the sulfur conversion reaction in the dynamic reaction model, including: Pyrolysis stage: Organic sulfur in coal decomposes into H2S and CS2. ; Semi-coke shrinkage stage: Residual sulfur reacts with oxygen to produce SO2. ; Gas-phase reaction: H2S reacts with CO to produce COS: ; Operation 2: Correction of the reaction rate equation. Define kinetic parameters for the newly added reaction pathway. The rate equation for the H2S formation reaction can be expressed as: ; in, The reaction rate constant is... For activation energy, The gas constant is For temperature.

[0052] In addition, the standardized sulfur content of the original coal is 2.1 (corresponding to an actual sulfur content of 0.96%), but the sulfur conversion characteristics of high-sulfur coals (such as coal type B with a sulfur content of 1.2%) are not fully captured. Therefore, the standardized value needs to be recalculated. Assuming the sulfur content range of high-sulfur coal expands to 1.5%, the standardized value is adjusted as follows: ; ( (These are the mean and standard deviation of the sulfur content distribution; they need to be updated based on actual coal quality data.) Step S4, Real-time Calculation and Prediction: A data interface was established to obtain coke oven operating parameters in real time, which were then input into the dynamic reaction model to calculate the gas yield under the current conditions: 162 m³ / t coal; gas composition: H2: 50.5%; CH4: 20.2%; CO: 19.8%; sulfur content: 1.1% (after model correction, the sulfur content prediction is more accurate and closer to the actual value).

[0053] in conclusion: 1. It can accurately predict atypical coal types such as blended coal without relying on data for specific coal types, demonstrating strong generalization ability.

[0054] 2. It can provide detailed information on the composition of coal gas, providing strong support for optimized control.

[0055] 3. It dynamically reflects the changes in gas evolution during the coking process, improving the stability and efficiency of the production process.

[0056] 4. Low computational cost, suitable for real-time control in actual production, reducing energy consumption and chemical product loss.

[0057] It should be noted that this application is not limited to the above-described embodiments. The above embodiments are merely examples, and any embodiments with the same structure and effect as the technical concept within the scope of this application are included in the technical scope of this application. Furthermore, various modifications that can be conceived by those skilled in the art to the embodiments, and other ways of constructing by combining some of the constituent elements of the embodiments, without departing from the spirit of this application, are also included in the scope of this application.

Claims

1. A method for dynamic multi-factor coupled prediction of coke oven gas, characterized in that, Includes the following steps: S1. Data Acquisition and Preprocessing: Collect comprehensive and accurate basic data, and organize and pre-process the data. S2. Establish a dynamic reaction model: Simulate the dynamic changes in gas evolution during coke oven dry distillation, consider the interaction between multiple factors, and establish a dynamic mathematical model; S3. Model Validation and Correction: Validate and correct the model based on actual production data to make the model prediction results more consistent with the actual data; S4. Real-time calculation and prediction: Based on the real-time collected coke oven operating parameters, the gas yield and composition are calculated and predicted in real time using the validated model.

2. The method for dynamic multi-factor coupled prediction of coke oven gas according to claim 1, characterized in that, In step S1, the basic data includes coal quality characteristic data and coke oven heating system parameters. The coal quality characteristic data is obtained through laboratory testing and includes, but is not limited to, elemental composition, volatile matter content, and sulfur content. The coke oven heating system parameters are collected using sensors and monitoring systems at the coke oven site and include, but are not limited to, heating temperature, heating time, and excess air coefficient.

3. The method for dynamic multi-factor coupled prediction of coke oven gas according to claim 1, characterized in that, In step S1, the preliminary processing involves using data processing software to clean the collected data, remove outliers and erroneous data, and then using statistical methods to standardize the data.

4. The method for dynamic multi-factor coupled prediction of coke oven gas according to claim 1, characterized in that, Step S2 includes the following sub-steps: S21. Define the model objectives and boundary conditions: Define the physical boundaries and time range of the coke oven dry distillation process, and define the input / output variables; S22. Selection of modeling method and mathematical framework: A set of differential equations based on mass conservation, energy conservation and chemical kinetics is used to describe the dynamic process of coal gas evolution, key reaction paths are defined and reaction rate equations are introduced. S23. Introduce a multi-factor coupling mechanism: embed coal quality characteristics, heating regime, pressure and atmosphere effects into the model to correct reaction rate and product distribution; S24. Dynamic process phased modeling: The dry distillation process is divided into drying period, pyrolysis period and semi-coke shrinkage period. Sub-models are established for each, and dynamic transfer between stages is realized through state variables. S25. Solution and numerical implementation: After discretizing the partial differential equation, a numerical solver is selected for solution, and the model solution is implemented with the help of a software platform.

5. The method for dynamic multi-factor coupled prediction of coke oven gas according to claim 4, characterized in that, The specific input / output variables are as follows: Input variables: coal quality characteristics, heating regime, operating parameters; Output variables: gas yield, gas composition, and the amount of tar and semi-coke produced.

6. The method for dynamic multi-factor coupled prediction of coke oven gas according to claim 4, characterized in that, In step S23, the embedding methods for coal quality characteristics, heating regime, pressure, and atmosphere effects are as follows: Coal quality characteristic embedding: Specific values ​​are obtained through experimental measurement or by referring to coal quality analysis reports. Coal quality characteristic parameters are used as input variables of the model. The influence of different coal qualities on coal gas evolution is reflected by adjusting the parameters in the reaction rate equation. Heating regime embedding: Temperature field data and heating rate are recorded by temperature sensors and heating control system. Temperature field data and heating rate are used as time-dependent input variables of the model. The gas evolution process under different heating regimes is simulated by dynamically adjusting the reaction rate. Pressure and atmosphere effects embedded: Real-time monitoring is performed using pressure sensors and gas analyzers. Pressure and atmosphere data are used as boundary conditions or input variables of the model. The influence of these factors on gas evolution is reflected by adjusting the reaction path and reaction rate equation.

7. The method for dynamic multi-factor coupled prediction of coke oven gas according to claim 4, characterized in that, In step S24, the specific sub-models for each stage are as follows: Drying period sub-model: simulates the evaporation process of moisture, considering the effects of temperature, humidity and ventilation conditions on the evaporation rate; Pyrolysis period sub-model: simulates the release of volatiles and pyrolysis reaction, including primary pyrolysis and secondary pyrolysis reaction; Semi-coke shrinkage sub-model: simulates the shrinkage process of semi-coke and the changes in its internal pore structure, as well as the accompanying slight secondary reactions.

8. The method for dynamic multi-factor coupled prediction of coke oven gas according to claim 7, characterized in that, In step S24, dynamic transfer between stages is achieved through state variables. That is, at the end of each stage, the state variables are updated according to the simulation results of that stage and used as the initial conditions for the next stage to input into the sub-model of the next stage.

9. The method for dynamic multi-factor coupled prediction of coke oven gas according to claim 1, characterized in that, The specific operation of step S3 is as follows: collect the actual gas yield and composition data in production as a verification dataset, compare the model prediction results with the verification dataset point by point, calculate the prediction error, analyze the prediction error to identify the cause of the deviation, adjust the model parameters or structure accordingly, and after adjustment, re-perform model prediction and comparison verification, evaluate the improvement effect, until the model prediction results are highly consistent with the actual data.

10. The method for dynamic multi-factor coupled prediction of coke oven gas according to claim 1, characterized in that, In step S4, the real-time coke oven operating parameters are input into the dynamic reaction model, the model is run, the gas yield and composition under the current conditions are calculated, and the results are output in the form of numbers or charts.

Images