Digital twinning method and system for staged combustion of coal-biomass fluidized bed

Abstract

The invention discloses a digital twinning method and system for staged combustion of a coal-biomass fluidized bed. The method comprises the following steps: establishing a comprehensive database comprising three-dimensional simulation data and operation data of a coal-biomass fluidized bed boiler; modeling and coupling the gas phase and the particle phase by using a neural network to obtain a coal-biomass combustion model which is used for predicting a temperature field, a pressure field and a velocity field of the coal-biomass fluidized bed boiler; and finally, integrating the coal-biomass combustion model with a boiler PLC system of the coal-biomass fluidized bed to realize data transmission, so as to obtain a coal-biomass combustion digital twinborn model and realize real-time control and dynamic optimization. On-line real-time prediction of the boiler is achieved, and long-term updating of the model and continuous improvement of the intelligent level are achieved through continuous accumulation and dynamic training of operation data.

Description

Technical Field

[0001] The present invention relates to boiler combustion and digital twin technologies, and in particular to a digital twin method and system for staged combustion of coal-biomass fluidized beds. Background Art

[0002] The co-combustion of coal and biomass can reduce the dependence on fossil fuels and effectively utilize the characteristics of biomass to reduce the emissions of greenhouse gases such as carbon dioxide (CO 2 ) etc. Different from the staged combustion of traditional single fuels, the staged combustion strategy in coal-biomass combustion needs to be optimized according to the different combustion characteristics of the two fuels. By dividing the combustion process into multiple stages and separately controlling the oxygen supply, temperature, and fuel distribution in each stage, the staged combustion technology can better adapt to the different volatilities and combustion speeds of coal and biomass during the combustion process.

[0003] Traditional numerical simulation methods such as CFD (Computational Fluid Dynamics) technology can accurately describe the multiphase flow and chemical reactions in the staged combustion process of coal-biomass in fluidized bed boilers. However, such simulation methods have a large computational amount, are difficult to meet the real-time response requirements of industrial sites, thereby reducing the fuel utilization efficiency, and are also prone to slagging, corrosion, and ash accumulation. Pollutants such as nitrogen oxides (NOx), sulfur dioxide (SO 2 ) and particulate matter cannot be effectively emitted. Summary of the Invention

[0004] Object of the Invention: The object of the present invention is to provide a digital twin method and system for staged combustion of coal-biomass fluidized beds, which can realize the online real-time prediction of coal-biomass fluidized bed boilers.

[0005] Technical Solution: The digital twin method for staged combustion of coal-biomass fluidized beds according to the present invention includes the following steps:

[0006] Define the target parameters to be predicted by the digital twin, determine the input parameters and evaluation criteria;

[0007] Utilize the three-dimensional simulation data of computational fluid dynamics and the operation data of coal-biomass fluidized bed boilers to construct a comprehensive database for staged combustion of coal-biomass fluidized beds;

[0008] Based on the comprehensive database, use neural networks to separately model the gas phase and the particle phase, and couple the neural network models of the gas phase and the particle phase through the drag force function and the heat transfer function between the gas and particle phases to obtain a coal-biomass combustion model for predicting the temperature field, pressure field, and velocity field of coal-biomass fluidized bed boilers;

[0009] Integrate the coal-biomass combustion model with the boiler PLC system of the coal-biomass fluidized bed to achieve data transmission, and obtain a coal-biomass combustion digital twin model.

[0010] Furthermore, the target parameters that need to be predicted for the digital twin are determined, and the input parameter range and evaluation criteria include:

[0011] The target parameters to be predicted include the boiler temperature field distribution, pressure field distribution, velocity field distribution, pollutant emission concentration, combustion efficiency, and thermal efficiency;

[0012] The input parameters include the structural parameters of the coal-biomass fluidized bed boiler and the physical and chemical property parameters of coal and biomass particles;

[0013] The evaluation criteria include the allowable deviations of combustion efficiency, pollutant emission concentration, and thermal efficiency.

[0014] Furthermore, in the comprehensive database, standardize the three-dimensional simulation data and real-time operation data, remove outliers, and perform time series alignment.

[0015] Furthermore, before integrating the coal-biomass combustion model with the boiler PLC system of the coal-biomass fluidized bed to achieve data transmission, it also includes:

[0016] Input the sampled sample parameters into the coal-biomass combustion model to obtain the probability distribution of the prediction results, calculate the relative contribution degree of the input parameters to the prediction results through sensitivity analysis, and optimize the parameters of the coal-biomass combustion model according to the relative contribution degree to obtain an optimized coal-biomass combustion model.

[0017] Furthermore, the process of inputting the sampled sample parameters into the coal-biomass combustion model to obtain the probability distribution of the prediction results and calculating the relative contribution degree of the input parameters to the prediction results through sensitivity analysis includes:

[0018] Use Monte Carlo sampling to construct the uncertainty distribution of key input parameters to obtain sampled samples, and input the sampled sample parameters into the coal-biomass combustion model to obtain the probability distribution of the prediction results;

[0019] Use the Sobol method to perform sensitivity analysis on the input parameters and calculate the relative contribution degree of the input parameters to the prediction results.

[0020] Furthermore, the coupling of the gas-phase and particle-phase neural network models through the drag force function and the heat transfer function between the gas and particle phases includes:

[0021]

[0022] S u= ∑ p F d ;

[0023] S q = hA(T g - T p );

[0024] Wherein, F d is the drag force, C d is the drag coefficient, A p is the projected area of the particle, ∑ p is the summation over all particles, ρ g is the gas density, u g is the gas velocity, u p is the particle velocity, S u is the momentum source term affected by the drag force, S q is the energy source term of heat transfer between the gas phase and the solid phase, h is the convective heat transfer coefficient, A is the surface area of the particle, T g is the gas temperature, T p is the particle temperature;

[0025] Iteratively calculate the drag force F d and then calculate S u . Input S u into the gas-phase neural network model and iteratively calculate S q . Input S q into the gas-phase neural network model and the particle-phase neural network model for constraining the models.

[0026] Furthermore, it also includes embedding the continuity equation, the energy conservation equation, and the momentum conservation equation into the neural network structures of the gas phase and the particles for constraint, and optimizing the neural network model in the form of a loss function;

[0027] The loss function of the gas-phase neural network model is as follows:

[0028]

[0029] In the formula, is the continuity loss function, is the momentum loss function, is the energy loss function of the gas-phase neural network model, is the loss function based on experimental data, w 1 , w 2 , w 3 , w 4 are the weights;

[0030] In the formula, ρ gis the gas density, is the predicted velocity component of the gas in the x i 、x j direction, is the predicted gas pressure, μ t is the turbulent viscosity, S u is the momentum source term affected by the drag force, C p,g is the specific heat capacity of the gas, is the predicted temperature of the gas, t is the time, is the velocity vector of the gas, is the temperature gradient of the gas, k g is the thermal conductivity of the gas, S q is the heat transfer energy source term between the gas phase and the solid phase, N is the total number of actual data points, α 1 、α 2 、α 3 are the weights

[0031] The loss function of the particle phase neural network model is as follows:

[0032]

[0033] In the formula, is the energy loss function of the particle phase neural network model, is the loss function based on experimental data, w 5 and w 6 are the weights;

[0034]

[0035] In the formula, m p is the particle mass, C p,p is the specific heat capacity of the particle, is the predicted temperature of the particle, h is the convective heat transfer coefficient, A is the surface area of the particle, σ is the Stefan-Boltzmann constant, ∈ is the radiation emissivity of the particle, T s is the ambient temperature, is the chemical reaction heat release rate, is the position coordinate of the particle, α 4 、α 5 are the weights;

[0036] The position coordinate of the particle is updated by the following formula:

[0037]

[0038] In the formula, u pis the predicted particle velocity vector, k represents the particle number, m represents the current moment, m* is the intermediate calculation moment, m + 1 represents the next moment of position update, Δt is the time step of iteration, and Δx k is the difference in particle positions predicted by the particle-phase neural network model, a ext includes the effects of gravity and drag force.

[0039] The digital twin system for staged combustion of coal-biomass fluidized bed according to the present invention includes:

[0040] A parameter setting unit for clarifying the target parameters to be predicted by the digital twin, determining the input parameters and evaluation criteria;

[0041] A database establishment unit for constructing a comprehensive database for staged combustion of coal-biomass fluidized bed by using three-dimensional simulation data of computational fluid dynamics and operation data of a coal-biomass fluidized bed boiler;

[0042] A coal-biomass combustion model establishment unit for respectively modeling the gas phase and the particle phase based on the comprehensive database by using a neural network, and coupling the neural network models of the gas phase and the particle phase through a drag force function and a heat transfer function between the gas and the particle phases to obtain a coal-biomass combustion model for predicting the temperature field, pressure field and velocity field of a coal-biomass fluidized bed boiler;

[0043] A coal-biomass combustion digital twin model establishment unit for integrating the coal-biomass combustion model with the boiler PLC system of the coal-biomass fluidized bed to realize data transmission and obtain a coal-biomass combustion digital twin model.

[0044] The electronic device according to the present invention includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the computer program is loaded into the processor, it implements the digital twin method for staged combustion of coal-biomass fluidized bed.

[0045] The computer-readable storage medium according to the present invention stores a computer program, and when the computer program is executed by a processor, it implements the digital twin method for staged combustion of coal-biomass fluidized bed.

[0046] Beneficial effects: Compared with the prior art, the advantages of the present invention are as follows: (1) By establishing a comprehensive database with multi-source data, the present invention ensures the accuracy and stability of the digital twin model under different working conditions and adapts to the dynamic changes of boiler operating conditions; (2) Through the continuous accumulation and dynamic training of operation data, the digital twin model of the present invention realizes the long-term update of the model and the continuous improvement of the intelligent level; (3) The present invention realizes the online real-time prediction of the boiler and the performance test of the operating condition combination, facilitating the operators to timely obtain the feedback of the boiler combustion state for optimization and the formulation of subsequent energy conservation and emission reduction plans; (4) The present invention supports fault prediction, performance test and optimization analysis, and provides reliable control suggestions and operating condition optimization plans without disturbing the actual equipment operation; (5) The present invention realizes the real-time and accurate prediction of data such as speed, pressure, and temperature in the coal-biomass fluidized bed boiler, dynamically feedbacks the combustion state, and adjusts the fuel flow and air supply according to the load demand and emission targets to optimize the combustion efficiency, avoid CO accumulation, and reduce the generation of NOx, SO 2 and particulate matter and reduce slagging and fouling corrosion; (6) The present invention adopts multi-scale and multi-physical field full-chain verification to improve the accuracy and reliability of model prediction and meet the high-precision prediction requirements under complex working conditions. Brief Description of the Drawings

[0047] Figure 1 is a flow chart of the digital twin method for the staged combustion of coal-biomass fluidized bed in the present invention;

[0048] Figure 2 is a schematic diagram of the physical mechanism and coupling principle of the coal-biomass combustion model in the embodiment of the present invention;

[0049] Figure 3 is a schematic diagram of the establishment and application of the coal-biomass combustion digital twin model in the embodiment of the present invention. Detailed Embodiment

[0050] The technical solution of the present invention will be further described below with reference to the drawings.

[0051] As Figure 1 shown, the digital twin method for the staged combustion of coal-biomass fluidized bed includes the following steps.

[0052] (1) Define the target parameters to be predicted and optimized by the digital twin, and determine the input parameter range and evaluation criteria.

[0053] The target parameters to be predicted and optimized include the boiler temperature field distribution, pressure field distribution, velocity field distribution, pollutant emission concentration, combustion efficiency, and thermal efficiency;

[0054] The input parameters include the structural dimension parameters of the coal-biomass fluidized bed boiler and the physical and chemical property parameters of coal and biomass particles. The structural dimension parameters of the coal-biomass fluidized bed boiler include the boiler geometric dimensions, bed height, and heat transfer area. The physical and chemical property parameters of coal and biomass particles include the density, particle size distribution, calorific value, particle shape factor, moisture content, elemental composition, volatile content, ash content, and fixed carbon content of coal and biomass;

[0055] Establish evaluation criteria. The performance indicators include combustion efficiency, pollutant emission concentration, and thermal efficiency. Set the allowable deviation for the target parameters to be ±15%.

[0056] Design different operating condition parameters, including fuel flow rate, fuel ratio, total air flow rate, primary air flow rate, secondary air flow rate, fluidization velocity, and environmental boundary conditions including inlet temperature and pressure; and design test schemes under different operating loads of 30%, 50%, and 100% to verify the system stability and adaptability.

[0057] (2) Based on computational fluid dynamics (CFD) simulation and boiler operation data, construct a comprehensive database for staged combustion of coal-biomass fluidized bed.

[0058] Use CFD simulation technology to construct a three-dimensional simulation model of the boiler. Set the geometric model based on the boiler structure parameters, and combine the characteristics of coal-biomass fuel and operating conditions to set boundary conditions. Simulate the gas-solid two-phase flow, staged combustion reaction, heat transfer, and pollutant generation processes, and output the spatial coordinates of each calculation node and the temperature, pressure, velocity, density, and component concentration at the corresponding points.

[0059] Collect the actual operation data of the boiler, and obtain the actual operation information of the boiler through DCS and sensors, including boiler temperature, boiler pressure, fuel flow rate, fuel ratio, air supply temperature and pressure, air staging ratio, and flue gas emission concentration.

[0060] Integrate the simulation data and operation data to construct a comprehensive database. Adopt the unified data storage format HDF5, perform standardized processing, outlier removal, and time series alignment on the data to complete the integration of multi-source data.

[0061] (3) Use a neural network constrained by physical mechanisms to construct a coal-biomass combustion model.

[0062] Such as Figure 2As shown in the figure, based on the comprehensive database, boiler structure parameters, fuel characteristics, and simulation result data are extracted. For the multiphase flow characteristics, neural networks constrained by physical mechanisms are used to model the gas and particle phases respectively. Considering the differences in the chemical composition, density, particle size, and volatile content of coal and biomass fuels, as well as the interaction between the gas and solid phases, including drag force and heat transfer, the gas and particles are coupled to obtain a coal-biomass combustion model. The inputs of this model are boiler structure parameters, fuel properties, initial conditions, and operating conditions, and it can be used to predict the distribution of the boiler temperature field, pressure field, and velocity field. At the same time, physical mechanisms are incorporated, and the continuity equation, energy conservation equation, and momentum conservation equation are embedded into the neural network structure for constraint, and the coal-biomass combustion model is optimized in the form of a loss function.

[0063] Loss function of the gas-phase neural network model is as follows:

[0064]

[0065] In the formula, is the continuity loss function, is the momentum loss function, is the energy loss function of the gas-phase neural network model, is the loss function based on experimental data, w 1 、w 2 、w 3 、w 4 are weights;

[0066]

[0067] In the formula, ρ g is the gas density, is the predicted velocity component of the gas in the x i 、x j direction, is the predicted gas pressure, μ t is the turbulent viscosity, S u is the momentum source term affected by the drag force, C p,g is the specific heat capacity of the gas, is the predicted temperature of the gas, t is time, is the velocity vector of the gas, is the temperature gradient of the gas, k g is the thermal conductivity of the gas, S q is the energy source term of heat transfer between the gas phase and the solid phase, N is the total number of actual data points, a 1 、a 2 、a 3 are weights.

[0068] Loss function of the particle-phase neural network model As follows:

[0069]

[0070] In the formula, is the energy loss function of the particle-phase neural network model, is the loss function based on experimental data, and w 5 、w 6 are weights;

[0071]

[0072] In the formula, m p is the particle mass, C p,p is the specific heat capacity of the particle, is the predicted temperature of the particle, h is the convective heat transfer coefficient, A is the surface area of the particle, σ is the Stefan-Boltzmann constant, ∈ is the radiation emissivity of the particle, T s is the ambient temperature, is the chemical reaction heat release rate, is the position coordinate of the particle, and α 4 、a 5 are weights.

[0073] The position coordinate of the particle is updated by the following formula:

[0074]

[0075] In the formula, u p is the predicted particle velocity vector, k represents the particle number, m represents the current moment, m* is the intermediate calculation moment, m + 1 represents the next moment of position update, Δt is the time step of iteration, and Δx k is the predicted particle position difference of the particle-phase neural network model, and a ext includes the effects of gravity and drag force.

[0076] The gas-solid two-phase neural network model is coupled through drag force and heat transfer:

[0077]

[0078] S u =∑ p F d ;

[0079] S q =hA(T g -T p );

[0080] Among them, F d is the drag force, C d is the drag coefficient, Ap is the projected area of the particle, Σ p is the sum over all particles, ρ g is the gas density, u g is the gas velocity, u p is the particle velocity, S u is the momentum source term affected by the drag force, S q is the energy source term for heat transfer between the gas phase and the solid phase, h is the convective heat transfer coefficient, A is the surface area of the particle, T g is the gas temperature T p is the particle temperature.

[0081] Iteratively calculate the drag force F d and then calculate the momentum source term S u Put it into the momentum loss function of the gas-phase neural network model, and iteratively calculate the heat transfer source term S q Put them into the energy loss functions of the gas-phase neural network and the particle-phase neural network respectively to constrain the model. At the same time, F d is used to calculate a when updating the particle position ext , and then iteratively update the particle position information.

[0082] (4) Optimize and verify the model based on the uncertainty optimization method and consistency analysis.

[0083] Define the uncertainty distribution of the input parameters based on prior knowledge; based on the comprehensive database, use the Monte Carlo method for input sampling, input the samples into the coal-biomass combustion model, obtain the probability distribution of the output results, and preliminarily quantify the prediction uncertainty of the target parameters; use the Sobol method for sensitivity analysis, calculate its relative contribution and interaction effects on the prediction uncertainty of the target parameters, identify the main sources and propagation paths affecting the output uncertainty, and on the basis of the above uncertainty quantification and sensitivity analysis, introduce the uncertainty optimization method to tune the model hyperparameters. Through integrating the coal-biomass combustion model, CFD simulation, experimental bench and industrial boiler tests, construct a multi-scale verification process. During the verification process, use the uncertainty quantification method to gradually verify and optimize, accumulate uncertainty, achieve full-chain verification, and conduct consistency analysis through cross-validation to evaluate the prediction accuracy of the model and the feasibility of industrial actual application, and optimize the model adaptability.

[0084] (5) Under the framework of advanced process control (APC), integrate the optimized coal-biomass combustion model into the boiler PLC system coupled with the existing DCS to establish a coal-biomass combustion digital twin model, and conduct real-time control and optimization of the boiler.

[0085] Such as Figure 3As shown in the figure, the DCS system of the boiler is connected through the OPC UA protocol to obtain real-time operation data. The PLC system is connected through the Modbus protocol to realize the real-time execution of boiler control instructions. Kafka is used to build an efficient data transmission channel to transmit the real-time collected data to a SQL-based integrated database for storage. The RESTful API is used for the real-time call of the vertical coal-biomass combustion digital twin model to complete the generation of prediction and regulation instructions.

[0086] The user interface (GUI) is developed based on the Qt framework. In the interface, the temperature field, pressure field, flow velocity field, pollutant emissions, combustion performance indicators, and operating conditions input are displayed in real time, supporting 2D / 3D visualization. A manual input module is provided for testing specific operating conditions or simulating optimization schemes. The abnormal alarm function is integrated. Alarm events are triggered by combining real-time data monitoring, and a fault diagnosis report is generated through logical inference. The WebSocket technology is used to achieve real-time data synchronization between the interface and the model, ensuring the instant feedback and visual display of regulation suggestions. The regulation instructions are executed through the PLC, and historical data is stored for subsequent performance analysis and fault backtracking.

[0087] The digital twin system for the staged combustion of coal-biomass fluidized bed according to the present invention includes:

[0088] A parameter setting unit for clarifying the target parameters to be predicted by the digital twin, determining the input parameters and evaluation criteria;

[0089] A database establishment unit for constructing an integrated database for the staged combustion of coal-biomass fluidized bed by using the three-dimensional simulation data of computational fluid dynamics and the operation data of the coal-biomass fluidized bed boiler;

[0090] A coal-biomass combustion model establishment unit for respectively modeling the gas phase and the particle phase based on the integrated database by using neural networks, and coupling the neural network models of the gas phase and the particle phase through a drag force function and a heat transfer function between the gas and the particle phases to obtain a coal-biomass combustion model for predicting the temperature field, pressure field, and velocity field of the coal-biomass fluidized bed boiler;

[0091] A coal-biomass combustion digital twin model establishment unit for integrating the coal-biomass combustion model with the boiler PLC system of the coal-biomass fluidized bed to realize data transmission, and obtaining a coal-biomass combustion digital twin model.

[0092] The electronic device according to the present invention includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the computer program is loaded into the processor, the digital twin method for the staged combustion of coal-biomass fluidized bed is implemented.

[0093] The computer-readable storage medium according to the present invention stores a computer program, and when the computer program is executed by a processor, it implements the digital twin method for staged combustion of coal-biomass fluidized bed.

[0094] The computer-readable storage medium may include RAM, ROM, EEPROM, CD-ROM or other optical disc storage devices, magnetic disk storage devices or other magnetic storage devices, flash memory, or any other medium that can be used to store program code in the form of instructions or data structures and can be accessed by a computer.

[0095] The processor is used to execute the computer program stored in the memory to implement each step in the method involved in the above embodiments.

Claims

1. A digital twin method for coal-biomass fluidized bed staged combustion, characterized in that: The steps include: Clarify the target parameters that the digital twin needs to predict, and determine the input parameters and evaluation criteria; Using the 3D simulation data of computational fluid dynamics and the operation data of coal-biomass fluidized bed boilers, a comprehensive database of coal-biomass fluidized bed staged combustion is constructed; Based on the comprehensive database, the gas phase and the particle phase are modeled by using a neural network respectively, and the neural network models of the gas phase and the particle phase are coupled through a drag function and a heat transfer function between the gas phase and the particle phase to obtain a coal-biomass combustion model, which is used to predict the temperature field, pressure field and velocity field of the coal-biomass fluidized bed boiler; The coal-biomass combustion model is integrated with the boiler PLC system of the coal-biomass fluidized bed to realize data transmission, and a coal-biomass combustion digital twin model is obtained.

2. The digital twin method for coal-biomass fluidized bed staged combustion according to claim 1, characterized in that: The target parameters that need to be predicted by the digital twin are clarified, and the input parameter range and evaluation criteria are determined, including: The target parameters that need to be predicted include boiler temperature field distribution, pressure field distribution, velocity field distribution, pollutant emission concentration, combustion efficiency and thermal efficiency; The input parameters include the structural parameters of the coal-biomass fluidized bed boiler, and the physical and chemical property parameters of coal and biomass particles; The evaluation criteria include combustion efficiency, pollutant emission concentration and allowable deviation of thermal efficiency.

3. The digital twin method for coal-biomass fluidized bed staged combustion according to claim 1, characterized in that: In the comprehensive database, the three-dimensional simulation data and the real-time operation data are standardized, outliers are removed and time series are aligned.

4. The digital twin method for coal-biomass fluidized bed staged combustion according to claim 1, characterized in that: Before the coal-biomass combustion model is integrated with the coal-biomass fluidized bed boiler PLC system to realize data transmission, it also includes: The sample parameters are input into the coal-biomass combustion model to obtain the probability distribution of the prediction results. The relative contribution of the input parameters to the prediction results is calculated through sensitivity analysis. The parameters of the coal-biomass combustion model are tuned according to the relative contribution to obtain an optimized coal-biomass combustion model.

5. The digital twin method for coal-biomass fluidized bed staged combustion according to claim 4, characterized in that: The sampling parameters are input into the coal-biomass combustion model to obtain the probability distribution of the prediction results, and the relative contribution of the input parameters to the prediction results is calculated through sensitivity analysis, including: Monte Carlo sampling is used to construct the uncertainty distribution of key input parameters, and the sampled samples are obtained. The sampled sample parameters are input into the coal-biomass combustion model to obtain the probability distribution of the prediction results; The Sobol method is used to perform sensitivity analysis on the input parameters and calculate the relative contribution of the input parameters to the prediction results.

6. The digital twin method for coal-biomass fluidized bed staged combustion according to claim 1, characterized in that: The coupling of the neural network models of the gas phase and the particle phase through the drag function and the heat transfer function between the gas phase and the particle phase includes: S u =∑ p F d ; S q =hA(T g -T p ); Among them, F d is the drag force, C d is the drag coefficient, A p is the projected area of ​​the particle, Σ p To sum over all particles, ρ g is the gas density, u g is the gas velocity, u p is the particle velocity, S u is the momentum source term affected by the drag force, S q is the energy source term for heat transfer between the gas phase and the solid phase, h is the convective heat transfer coefficient, A is the surface area of ​​the particle, T g is the gas temperature, T p is the particle temperature; Iterative calculation of drag force F d Then calculate S u , S u Input into the gas phase neural network model and iteratively calculate S q , S q Input into the gas phase neural network model and particle phase neural network model to constrain the model.

7. The digital twin method for coal-biomass fluidized bed staged combustion according to claim 1, characterized in that: It also includes embedding the continuity equation, energy conservation equation, and momentum conservation equation into the neural network structure of the gas phase and particles for constraints, and optimizing the neural network model in the form of a loss function; Loss function of the gas phase neural network model as follows: In the formula, is the continuity loss function, is the momentum loss function, is the energy loss function of the gas phase neural network model, is the loss function based on experimental data, w1, w2, w3, w4 are weights; In the formula, ρ g is the gas density, For gas at x i 、x j The predicted velocity component in the direction, To predict the gas pressure, μ t is the turbulent viscosity, S u is the momentum source term affected by the drag force, C p,g is the specific heat capacity of gas, is the predicted temperature of the gas, t is the time, is the velocity vector of the gas, is the temperature gradient of the gas, k g is the thermal conductivity of the gas, S q is the energy source term for heat transfer between gas phase and solid phase, N is the total number of actual data points, α1, α2, α3 are weights Loss function of the particle phase neural network model as follows: In the formula, is the energy loss function of the particle phase neural network model, is the loss function based on experimental data, w5 and w6 are weights; In the formula, m p is the particle mass, C p,p is the specific heat capacity of the particle, is the predicted temperature of the particle, h is the convective heat transfer coefficient, A is the surface area of ​​the particle, σ is the Stefan-Boltzmann constant, ∈ is the radiation emissivity of the particle, T s is the ambient temperature, is the heat release rate of the chemical reaction, is the position coordinate of the particle, α4 and α5 are weights; The position coordinates of the particles are updated by the following formula: In the formula, u p is the predicted particle velocity vector, k is the particle number, m is the current moment, m* is the intermediate calculation moment, m+1 is the next moment of position update, Δt is the iterative time step, Δx k is the particle position difference predicted by the particle phase neural network model, a ext Includes the effects of gravity and drag.

8. A digital twin system for coal-biomass fluidized bed staged combustion, characterized in that: include: The parameter setting unit is used to clarify the target parameters that the digital twin needs to predict, and determine the input parameters and evaluation criteria; A database building unit, used to build a comprehensive database of coal-biomass fluidized bed staged combustion by using three-dimensional simulation data of computational fluid dynamics and operation data of coal-biomass fluidized bed boiler; A coal-biomass combustion model establishment unit is used to respectively model the gas phase and the particle phase using a neural network based on the comprehensive database, and couple the neural network models of the gas phase and the particle phase through a drag function and a heat transfer function between the gas and particle phases to obtain a coal-biomass combustion model for predicting the temperature field, pressure field and velocity field of the coal-biomass fluidized bed boiler; The coal-biomass combustion digital twin model establishment unit is used to integrate the coal-biomass combustion model with the boiler PLC system of the coal-biomass fluidized bed to realize data transmission and obtain the coal-biomass combustion digital twin model.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the computer program is loaded into a processor, the digital twin method for coal-biomass fluidized bed staged combustion according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the digital twin method for coal-biomass fluidized bed staged combustion according to any one of claims 1 to 7 is implemented.

Images