Moving bed high-temperature particle flow and heat transfer characteristic simulation prediction method and system based on deep learning, and electronic equipment

Abstract

The invention relates to the technical field of heat storage and energy conservation, and provides a moving bed high-temperature particle flow and heat transfer characteristic simulation and prediction method and system based on deep learning and electronic equipment. The method comprises the steps that high-temperature particle flow characteristic numerical simulation is conducted, and a three-dimensional moving bed high-temperature particle flow numerical model is established; simulating association of a bed structure, particle properties and particle accumulation and flow pattern rules, and establishing a particle distribution regulation and control strategy; performing numerical simulation on heat transfer characteristics of high-temperature particles, establishing association between particle flow characteristics and a temperature field, simulating a gas-solid two-phase composite heat transfer process of the particles, and establishing a temperature field control strategy; intelligent prediction of particle distribution and a temperature field is carried out, intelligent prediction of particle distribution and the temperature field at a specific moment is realized based on deep learning, and a regulation strategy based on numerical simulation is optimized. The particle flow and heat exchange characteristics of the moving bed are clarified, a theoretical support and a technical solution are provided for improving the waste heat recovery efficiency of the blast furnace slag, and the energy utilization rate can be increased.

Description

Technical Field

[0001] The present invention relates to the field of heat storage and energy-saving technology, and in particular to a simulation prediction method, system and electronic equipment for high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning. Background Art

[0002] Blast furnace slag is a by-product of the steel production process. The discharge temperature is as high as 1500℃. It is one of the few high-temperature waste heat resources that cannot be recovered in the steel industry. The existing technology mainly uses water quenching to rapidly cool blast furnace slag, which has serious problems such as waste heat loss, water resource waste, and environmental pollution.

[0003] The moving bed waste heat recovery technology for blast furnace slag is highly anticipated for its advantages of low remixing, high waste heat recovery rate, and resource utilization of blast furnace slag after waste heat recovery. The flow and heat transfer of high-temperature particles in the moving bed are affected by many factors, and face problems such as long experimental cycles, complex research methods, and difficulty in replicating real working conditions. Currently, numerical simulation based on the Euler-Lagrange method is the main research method. However, since this method involves a large amount of calculation scale and simulation time, and the influence mechanism of the operating parameters and structural parameters of the moving bed on the particle flow and heat transfer characteristics is still unclear, the further improvement of waste heat recovery efficiency is constrained. Summary of the invention

[0004] The purpose of the present invention is to provide a simulation prediction method, system and electronic equipment for the flow and heat transfer characteristics of high-temperature particles in a moving bed based on deep learning, which can clarify the flow and heat transfer characteristics of particles in a moving bed, summarize the design criteria and operation methods of a moving bed for improving the waste heat recovery efficiency, and improve the recovery rate of waste heat and waste energy of blast furnace slag.

[0005] To solve the above technical problems, according to one aspect of the present invention, the present invention provides the following technical solutions: A simulation and prediction method for high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning includes the following steps: S1. Numerical simulation of high-temperature particle flow characteristics: Establish a three-dimensional moving bed high-temperature particle flow numerical model, simulate the relationship between bed structure, particle properties and particle accumulation, flow pattern, and establish a particle distribution control strategy; S2. Numerical simulation of heat transfer characteristics of high-temperature particles: Establish the correlation between particle flow characteristics and temperature field, simulate the gas-solid two-phase composite heat transfer process of particles, and establish the temperature field control strategy; S3. Intelligent prediction of particle distribution and temperature field: Based on deep learning, intelligent prediction of particle distribution and temperature field at a specific moment is achieved, and control strategies based on numerical simulation are optimized.

[0006] As a preferred scheme of the simulation and prediction method of the moving bed high-temperature particle flow and heat transfer characteristics based on deep learning described in the present invention, wherein: in the step S1, a three-dimensional moving bed high-temperature particle flow numerical model is established to simulate the bottom structure design and outlet diameter of the bed structure and the particle characteristics of the particle diameter, stacking density, multi-particle size mixing, initial temperature, etc., the influence on the particle stacking characteristics, flow type, porosity, stacking density, average residence time, stagnation zone size, and the flow characteristics of the particles including stacking uniformity, velocity uniformity, flow trajectory, etc., to establish a particle distribution control strategy and measures to suppress cavities and stagnation phenomena.

[0007] As a preferred scheme of the simulation and prediction method of the flow and heat transfer characteristics of high-temperature particles in a moving bed based on deep learning described in the present invention, wherein: in the step S1, the Hertz-Mindin no-slip contact model is used based on the DEM method to simulate the flow law of particles in the moving bed; the hard sphere model is used to describe the collision contact between particles, assuming that no deformation occurs during the collision, and the collision is a process in which different unit bodies are superimposed on each other, and the contact model is converted into a vibration model to simplify the calculation of contact force.

[0008] As a preferred solution of the simulation prediction method of high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning described in the present invention, wherein: in the step S1, the particle flow characteristics are quantitatively analyzed through uniformity evaluation indicators and stagnation evaluation indicators.

[0009] As a preferred scheme of the simulation and prediction method of the flow and heat transfer characteristics of high-temperature particles in a moving bed based on deep learning described in the present invention, wherein: in the step S2, for the particle heat exchange coupling behavior in the solid-solid heat exchange process of the particles flowing around the heating surface and the particle cooling behavior in the gas-solid two-phase heat exchange process, based on the above-mentioned particle flow characteristics and stacking characteristics, the key parameters of the particles including particle size, mixing degree, coordination number, movement speed, initial temperature, heating surface structure, etc. are simulated to affect the heat transfer coefficient between the particles and the heating surface, the particle temperature distribution and the temperature drop rate, the relationship between the particle flow characteristics and the temperature field is established, and a control strategy for the temperature field is proposed.

[0010] As a preferred scheme of the simulation and prediction method of the high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning described in the present invention, wherein: in the step S2, it is assumed that the moving bed is composed of a discrete particle phase and a continuous fluid phase, and the fluid volume fraction phase is introduced into the traditional control equation to characterize the influence of the presence of particles on the fluid motion, and the distribution law of the moving bed including the gas phase flow field, particle flow field, gas-solid temperature field, etc. is obtained.

[0011] As a preferred embodiment of the simulation and prediction method of the flow and heat transfer characteristics of high-temperature particles in a moving bed based on deep learning described in the present invention, wherein: in the step S2, the CFD-DEM coupling method is used to simulate the gas-solid heat exchange process of the high-temperature moving bed particles. During the coupling process, the solid particles obtain the force of the flow field on the particles, the transferred heat, and the momentum and energy source terms of the solid phase acting on the grid within the grid unit. Specifically, the CFD-DEM coupling method includes: DEM solution, calculation of the force and heat flux between the colliding particles, and tracking at the particle scale; CFD solution, solution of the flow field, obtaining the fluid phase volume fraction, fluid resistance and momentum, and tracking the fluid motion; when the calculation time reaches the same, the parameters are transferred to each other, the velocity, position and force of the particles are updated, and the CFD-DEM coupling is completed.

[0012] As a preferred scheme of the simulation and prediction method of the high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning described in the present invention, wherein: in the step S3, based on numerical simulation, the particle flow field and temperature field image data are provided by simulating solid-solid and gas-solid two-phase dense particle flow, the deep learning model is trained to predict the particle distribution and temperature distribution of the specified working conditions at the specified time, the control strategy based on the simulation data is optimized, and the moving bed design criteria and operation methods for improving the waste heat recovery efficiency are summarized.

[0013] As a preferred solution of the simulation prediction method of the high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning described in the present invention, wherein: in the step S3, a progressively growing multi-scale image generation network is used to generate a gas phase / particle temperature distribution map, including: A generative model is constructed and attention weights are used to obtain regions of interest at different scales. Low-resolution data has a lower dimension and is easy to achieve good distribution matching; high-resolution images focus on learning image details.

[0014] The temporal cue features are embedded, and word order encoding and fusion are implemented in the latent space to keep the underlying microstructure consistent and generate the temperature distribution map at the corresponding moment.

[0015] A discriminant model is constructed to constrain the data distribution of the generated image at multiple scales to be consistent with the input time, to further ensure that the generated temperature distribution map has the consistency of the underlying microstructure information with the input time.

[0016] To solve the above technical problem, according to another aspect of the present invention, the present invention provides the following technical solution: A simulation and prediction system for high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning, including: High-temperature particle flow characteristics numerical simulation module: establish a three-dimensional moving bed high-temperature particle flow numerical model, simulate the bed structure, particle properties and particle accumulation, flow pattern law, and establish particle distribution control strategy; Numerical simulation module of high-temperature particle heat transfer characteristics: establish the correlation between particle flow characteristics and temperature field, simulate the gas-solid two-phase composite heat transfer process of particles, and establish temperature field control strategy; Intelligent prediction module for particle distribution and temperature field: Based on deep learning, it realizes intelligent prediction of particle distribution and temperature field at a specific moment, and optimizes the control strategy based on numerical simulation.

[0017] To solve the above technical problem, according to another aspect of the present invention, the present invention provides the following technical solution: An electronic device comprises: a processor, a memory, an internal bus, and a program or instruction stored in the memory and executable on the processor. When the program or instruction is executed by the processor, the simulation prediction method of the high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning is implemented.

[0018] A readable storage medium stores a program or instruction, which, when executed by a processor, implements the simulation prediction method of moving bed high-temperature particle flow and heat transfer characteristics based on deep learning.

[0019] The beneficial effects of the present invention are as follows: The present invention proposes a simulation prediction method, system and electronic equipment for high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning, which at least includes the following beneficial effects: (1) By analyzing and predicting the mechanism of particle flow and heat transfer characteristics of blast furnace slag moving bed, it is helpful to break through the current situation where the waste heat recovery efficiency is hindered by unclear mechanism, and provide theoretical support and technical solutions for improving the waste heat recovery efficiency of blast furnace slag.

[0020] (2) Combining numerical simulation and deep learning methods, providing particle flow field and temperature field image data through solid-solid and gas-solid two-phase dense particle flow simulation, and training deep learning models to predict data under specified conditions and at specified times at a low cost can effectively improve accuracy and reduce calculation time, while improving computing efficiency to achieve effective parameter design, and promote the application of deep learning technology in particle flow research and steel industry prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying creative work.

[0022] Figure 1 A schematic diagram of the steps of a simulation and prediction method for high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning; Figure 2 Schematic diagram of the CFD-DEM coupling method for numerical simulation of heat transfer of high-temperature particles; Figure 3 This is a structural block diagram of the simulation and prediction system for high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning; Figure 4 It is a structural block diagram of an electronic device according to an embodiment of the present application.

[0023] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0024] The following will be described clearly and completely in conjunction with the technical solutions in the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] The simulation and prediction method for the flow and heat transfer characteristics of high-temperature particles in a moving bed based on deep learning provided by the present invention can overcome the shortcomings of time-consuming and low-efficiency experimental and numerical methods. By analyzing and predicting the mechanism of the flow and heat transfer characteristics of particles in a moving bed of blast furnace slag, a large amount of image data can be obtained at a relatively low price and cost. The deep learning method is combined to extract the image features of the particle distribution and temperature cloud map to realize the prediction of the future state of the particle distribution and temperature field. It can effectively improve the accuracy and reduce the calculation time, while improving the calculation efficiency to realize effective parameter design, improve the waste heat recovery efficiency, and promote energy conservation and emission reduction.

[0026] Figure 1 A flowchart of a method for simulating and predicting the flow and heat transfer characteristics of high-temperature particles in a moving bed based on deep learning provided by an embodiment of the present invention. Figure 1The simulation prediction method of high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning can be performed by a simulation prediction device of high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning, and the device can be set on a server. For example, the server can be a physical server including an independent host, a virtual server carried by a host cluster, a cloud server, etc., and the embodiment of the present invention is not limited to this. The simulation prediction method of high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning includes the following steps: Step S1, numerical simulation of high-temperature particle flow characteristics, establishing a three-dimensional moving bed high-temperature particle flow numerical model, simulating the relationship between bed structure, particle properties and particle accumulation, flow pattern rules, and establishing a particle distribution control strategy.

[0027] This step is a step for numerically simulating the flow characteristics of high-temperature particles. An optional method for numerically simulating the flow characteristics may be: establishing a three-dimensional moving bed high-temperature particle flow numerical model, simulating the bottom structure design and outlet diameter of the bed structure and the particle characteristics of the particle diameter, stacking density, multi-particle size mixing, initial temperature, etc., the influence on the particle stacking characteristics, flow type, porosity, stacking density, average residence time, stagnation zone size, as well as the flow characteristics such as particle stacking uniformity, velocity uniformity, flow trajectory, etc., establishing a particle distribution control strategy and measures to suppress cavities and stagnation phenomena.

[0028] An optional numerical simulation method for flow characteristics can be: based on the DEM method, the Hertz-Mindin no-slip contact model is used to simulate the flow law of particles in a moving bed. The hard sphere model is used to describe the collision contact between particles. It is assumed that no deformation occurs during the collision. The collision is a process in which different units are superimposed on each other. The contact model is converted into a vibration model to simplify the calculation of contact force.

[0029] Specifically, the analysis object is regarded as a sufficiently large number of discrete units, each particle is an independent unit, the contact force is calculated according to the interaction force between the particles at any time, and the motion parameters of the unit are solved according to Newton's laws of motion. This is repeated and finally a macroscopic prediction of the overall motion is achieved. The motion equation of particle i at time t is: in, and They are the normal contact force and the tangential contact force: in, and are the normal and tangential velocity components, respectively; and Particles Relative to particles The normal deformation and tangential deformation of and are the normal and tangential contact elastic coefficients, respectively; and are the normal and tangential contact viscoelastic damping coefficients, respectively.

[0030] Based on the above implementation of moving bed particle flow simulation, the relationship between key parameters such as moving bed structure and particle properties and particle accumulation and flow pattern is simulated.

[0031] An optional evaluation index includes: a uniformity evaluation index and a stagnation evaluation index, which are used to quantitatively analyze the particle flow characteristics.

[0032] Step S2, numerical simulation of the heat transfer characteristics of high-temperature particles, establishing the correlation between the particle flow characteristics and the temperature field, simulating the gas-solid two-phase composite heat exchange process of the particles, and establishing a temperature field control strategy.

[0033] This step is a numerical simulation step of the heat transfer characteristics of high-temperature particles. An optional method for numerical simulation of heat transfer characteristics may be: for the particle heat transfer coupling behavior of the solid-solid heat transfer process of particles flowing around the heating surface and the particle cooling behavior of the gas-solid two-phase heat transfer process, based on the above-mentioned particle flow characteristics and stacking characteristics, simulate the influence of key parameters such as particle size, mixing degree, coordination number, movement speed, initial temperature, heating surface structure, etc. on the heat transfer coefficient between particles and the heating surface, particle temperature distribution and temperature drop rate, establish the relationship between particle flow characteristics and temperature field, and propose a control strategy for the temperature field.

[0034] Optionally, this step assumes that the moving bed consists of a discrete particle phase and a continuous fluid phase, and introduces the fluid volume fraction phase into the traditional control equation to characterize the influence of the presence of particles on the fluid motion, so as to obtain the distribution laws of the gas phase flow field, particle flow field, gas-solid temperature field, etc. in the moving bed.

[0035] An optional numerical simulation method for heat transfer characteristics may be: using the CFD-DEM coupling method to simulate the gas-solid heat exchange process of high-temperature moving bed particles. During the coupling process, the solid particles obtain the force of the flow field on the particles, the transferred heat, and the momentum and energy source terms of the solid phase acting on the grid within the grid unit.

[0036] like Figure 2 Shown is an optional CFD-DEM coupling method of the present invention, comprising: DEM solution is used to calculate the forces and heat flux between colliding particles and track them at the particle scale.

[0037] CFD solves the flow field, obtains the fluid phase volume fraction, fluid resistance and momentum, and tracks the fluid motion.

[0038] When the calculation time reaches the same value, the parameters are transferred to each other to update the velocity, position and force of the particles, thus completing the CFD-DEM coupling.

[0039] Specifically, the Di Felice model is used to describe the drag force of the fluid on the particles, and the expression of the drag model is: Among them, the parameters and They are The no-slip Hertz-Mindlin model is used to describe the heat transfer between particles through heat conduction: in, and denote the gravity and geometric mean radius of the particles, respectively. is the equivalent elastic modulus.

[0040] For particle-fluid-particle heat conduction, the heat flow is defined as follows: In the formula, the geometric parameters are defined as: The upper limit of points is: In the above formula, and Refers to the temperature of different particles, represents the fluid lens radius of two non-touching particles, and They refer to the thermal conductivity of different particles, The high-temperature particles enter the bed from the upper part of the moving bed and directly contact with the air introduced from the bottom to exchange heat.

[0041] Convective heat flux is defined as follows: in, For particles Convective flux with the surrounding fluid, is the logarithmic mean temperature difference between the two phases, It represents the ratio of the convection heat transfer resistance to the conduction heat transfer resistance between the gas-solid phases and is calculated using the Gunn model to suit particle flow.

[0042] The Stefan-Boltzmann formula is used to calculate the net heat energy radiated from the particle to the surrounding environment: in, is the Stefan-Boltzmann constant; is the particle surface area; is the emissivity of the spherical particle; is the average temperature of particles in a closed grid cell.

[0043] Based on the particle movement and fluid movement, combined with the above-mentioned particle contact heat conduction, fluid-particle heat exchange, and particle radiation heat exchange, an accurate simulation of the heat exchange process of the moving bed high-temperature particles is achieved.

[0044] Step S3, intelligent prediction of particle distribution and temperature field, realizes intelligent prediction of particle distribution and temperature field at a specific moment based on deep learning, and optimizes the control strategy based on numerical simulation.

[0045] This step is the intelligent prediction step of particle distribution and temperature field. An optional intelligent prediction method can be: provide particle flow field and temperature field image data through solid-solid and gas-solid two-phase dense particle flow simulation, train the deep learning model to predict the particle distribution and temperature distribution of the specified working conditions at the specified time, optimize the control strategy based on simulation data, and summarize the moving bed design criteria and operation methods for improving waste heat recovery efficiency.

[0046] An optional intelligent prediction method may be: using a progressively growing multi-scale image generation network for gas phase / particle temperature distribution map generation, including: A generative model is constructed and attention weights are used to obtain regions of interest at different scales. Low-resolution data has a lower dimension and is easy to achieve good distribution matching; high-resolution images focus on learning image details.

[0047] The temporal cue features are embedded, and word order encoding and fusion are implemented in the latent space to keep the underlying microstructure consistent and generate the temperature distribution map at the corresponding moment.

[0048] A discriminant model is constructed to constrain the data distribution of the generated image at multiple scales to be consistent with the input time, to further ensure that the generated temperature distribution map has the consistency of the underlying microstructure information with the input time.

[0049] Specifically, in order to add decoupling constraints and avoid the possibility that the gas phase / particle temperature distribution graph may contain redundant features due to repeated semantic distribution and coupling features of the abstract semantic space, we define is a low-dimensional feature vector, and are the number of model layers and indexes respectively. Specifically, to measure the degree of coupling between feature maps, the similarity metric is defined as: In order to constrain the learning ability of different feature channels and improve the semantic strength of a single feature map representation, the coupled separation training objective is constructed as follows: exist Under the constraints of We focus more on balancing the semantic distribution in different directions to ensure that the feature distribution in the latent space is denser, thereby generating more accurate gas / particle temperature distribution maps.

[0050] The present invention can produce the following effects by executing the simulation prediction method of high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning: (1) By analyzing and predicting the mechanism of particle flow and heat transfer characteristics of blast furnace slag moving bed, it is helpful to break through the current situation where the waste heat recovery efficiency is hindered by unclear mechanism, and provide theoretical support and technical solutions for improving the waste heat recovery efficiency of blast furnace slag.

[0051] (2) Combining numerical simulation and deep learning methods, providing particle flow field and temperature field image data through solid-solid and gas-solid two-phase dense particle flow simulation, and training deep learning models to predict data under specified conditions and at specified times at a low cost can effectively improve accuracy and reduce calculation time, while improving computing efficiency to achieve effective parameter design, and promote the application of deep learning technology in particle flow research and steel industry prediction.

[0052] The above is a simulation prediction method for the flow and heat transfer characteristics of high-temperature particles in a moving bed based on deep learning provided by an embodiment of the present invention. Based on the same inventive idea, an embodiment of the present invention also provides a corresponding simulation prediction system for the flow and heat transfer characteristics of high-temperature particles in a moving bed based on deep learning, such as Figure 3 As shown, including: The high-temperature particle flow characteristics numerical simulation module 201 establishes a three-dimensional moving bed high-temperature particle flow numerical model, simulates the relationship between bed structure, particle properties and particle accumulation, flow pattern rules, and establishes a particle distribution control strategy.

[0053] The high-temperature particle heat transfer characteristic numerical simulation module 202 establishes the correlation between the particle flow characteristic and the temperature field, simulates the gas-solid two-phase composite heat exchange process of the particles, and establishes the temperature field control strategy.

[0054] The particle distribution and temperature field intelligent prediction module 203 realizes the intelligent prediction of particle distribution and temperature field at a specific moment based on deep learning, and optimizes the control strategy based on numerical simulation.

[0055] The embodiment of the present invention also provides a corresponding electronic device for simulating and predicting the flow and heat transfer characteristics of high-temperature particles in a moving bed based on deep learning, such as Figure 4 As shown, including: At least one processor 301, memory 303 and internal bus 302; wherein the processor 301 and memory 303 are interconnected and communicated via the internal bus 302; the computer program or instructions stored in the memory 303 can be executed by the processor 301 to implement the steps of the above-mentioned embodiments.

[0056] The electronic device may be a computer chip or a physical implementation, or a product with a certain function. A typical implementation device is a computer, such as a personal computer, a smart phone, a tablet computer, a wearable device, or any combination of devices. The processor is a processor of the electronic device, and the memory includes but is not limited to a computer read-only memory, a random access memory, an optical disk, etc.

[0057] Based on the same inventive concept, an embodiment of the present invention further provides a non-volatile storage medium corresponding to the above method, on which a computer program or instruction is stored, which can be executed by the processor to implement the steps of the above embodiments.

[0058] The non-volatile storage medium includes permanent and non-permanent, removable and non-removable media, and information storage can be achieved by any method and technology. Examples include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes or any other non-transmission medium that can be used to store information that can be accessed by a computing device.

[0059] Each embodiment of the present invention is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device, electronic device and volatile storage medium, since they are basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the corresponding description of the method embodiment.

[0060] The device, electronic device and volatile storage medium provided in the embodiments of the present invention are one-to-one corresponding. Therefore, the device, electronic device and volatile storage medium also have similar beneficial technical effects as the corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the device, electronic device and volatile storage medium will not be repeated here.

[0061] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. All equivalent structural changes made by using the contents of the present invention specification under the inventive concept of the present invention, or directly / indirectly applied in other related technical fields are included in the patent protection scope of the present invention.

Claims

1. A simulation and prediction method for high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning, characterized in that: The steps include: S1. Numerical simulation of high-temperature particle flow characteristics: Establish a three-dimensional moving bed high-temperature particle flow numerical model, simulate the relationship between bed structure, particle properties and particle accumulation, flow pattern, and establish a particle distribution control strategy; S2. Numerical simulation of heat transfer characteristics of high-temperature particles: Establish the correlation between particle flow characteristics and temperature field, simulate the gas-solid two-phase composite heat transfer process of particles, and establish the temperature field control strategy; S3. Intelligent prediction of particle distribution and temperature field: Based on deep learning, intelligent prediction of particle distribution and temperature field at a specific moment is achieved, and control strategies based on numerical simulation are optimized.

2. The simulation prediction method for high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning according to claim 1 is characterized in that: In the step S1, a three-dimensional moving bed high-temperature particle flow numerical model is established to simulate the bottom structure design of the bed structure and the outlet diameter and particle characteristics of the particle diameter, packing density, multi-particle mixing, initial temperature, the influence on the particle packing characteristics, flow type, porosity, packing density, average residence time, stagnation zone size, and the flow characteristics of the particles including packing uniformity, velocity uniformity, and flow trajectory, and establish a particle distribution control strategy and measures to suppress cavities and stagnation phenomena.

3. The simulation prediction method for high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning according to claim 1 is characterized in that: In step S1, the Hertz-Mindin no-slip contact model is used based on the DEM method to simulate the flow law of particles in the moving bed; the hard sphere model is used to describe the collision contact between particles, assuming that no deformation occurs during the collision, and the collision is a process of superposition of different unit bodies. The contact model is converted into a vibration model to simplify the calculation of contact force; the uniformity evaluation index and the hysteresis evaluation index are used to quantitatively analyze the particle flow characteristics.

4. The simulation prediction method for high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning according to claim 1 is characterized in that: In the step S2, for the particle heat transfer coupling behavior in the solid-solid heat transfer process of the particles flowing around the heating surface and the particle cooling behavior in the gas-solid two-phase heat transfer process, based on the above-mentioned particle flow characteristics and stacking characteristics, the key parameters of the particles including particle size, mixing degree, coordination number, movement speed, initial temperature, and heating surface structure are simulated to affect the heat transfer coefficient between the particles and the heating surface, the particle temperature distribution, and the temperature drop rate, the relationship between the particle flow characteristics and the temperature field is established, and a control strategy for the temperature field is proposed.

5. The simulation prediction method for high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning according to claim 1 is characterized in that: In step S2, it is assumed that the moving bed is composed of a discrete particle phase and a continuous fluid phase, and the fluid volume fraction phase is introduced into the traditional control equation to characterize the influence of the presence of particles on the fluid movement, so as to obtain the distribution law of the gas phase flow field, particle flow field and gas-solid temperature field in the moving bed.

6. The simulation prediction method for high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning according to claim 1 is characterized in that: In step S2, the gas-solid heat exchange process of the high-temperature moving bed particles is simulated by using the CFD-DEM coupling method. During the coupling process, the solid particles obtain the force of the flow field on the particles, the transferred heat, and the momentum and energy source terms of the solid phase acting on the grid within the grid unit.

7. The simulation prediction method for high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning according to claim 1 is characterized in that: In step S3, based on numerical simulation, solid-solid and gas-solid two-phase dense particle flow simulation is used to provide particle flow field and temperature field image data, and a deep learning model is trained to predict the particle distribution and temperature distribution of a specified working condition at a specified time, and the control strategy based on the simulation data is optimized, and the moving bed design criteria and operation methods for improving waste heat recovery efficiency are summarized.

8. A simulation and prediction system for high-temperature particle flow and heat transfer characteristics of a moving bed based on deep learning, characterized in that: include: High-temperature particle flow characteristics numerical simulation module: establish a three-dimensional moving bed high-temperature particle flow numerical model, simulate the relationship between bed structure, particle properties and particle accumulation, flow pattern, and establish particle distribution control strategy; Numerical simulation module of high-temperature particle heat transfer characteristics: establish the correlation between particle flow characteristics and temperature field, simulate the gas-solid two-phase composite heat transfer process of particles, and establish temperature field control strategy; Intelligent prediction module for particle distribution and temperature field: Based on deep learning, it realizes intelligent prediction of particle distribution and temperature field at a specific moment, and optimizes the control strategy based on numerical simulation.

9. An electronic device, characterized in that: include: A processor, a memory, an internal bus, and a program or instruction stored in the memory and executable on the processor, which, when executed by the processor, implements the simulation and prediction method for the flow and heat transfer characteristics of high-temperature particles in a moving bed based on deep learning as described in any one of claims 1 to 7.

10. A readable storage medium, characterized in that: The readable storage medium stores a program or instruction, which, when executed by a processor, implements the simulation and prediction method of the flow and heat transfer characteristics of high-temperature particles in a moving bed based on deep learning as described in any one of claims 1 to 7.

Images