Method and system for preparing ultrafine barium sulfate by airflow pulverization based on reinforcement learning
Through the airflow crushing technology based on reinforcement learning, the problems of high production costs and large environmental pollution in the preparation of ultrafine barium sulfate are solved, and precise control of particle size distribution and morphology is achieved, which improves production efficiency and economic benefits.
Patent Information
- Application Number
- CN202411099975.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-12
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-08-12
AI Technical Summary
The existing ultrafine barium sulfate preparation technology has problems such as high production costs, large environmental pollution and complex process control, which is difficult to meet the market's strict demand for particle size distribution and morphology regulation.
The airflow crushing method based on reinforcement learning is adopted, and the initial parameters of the airflow crusher are set through simulation calculations, and the reinforcement learning algorithm is used to optimize the parameters such as airflow velocity, temperature and pressure during the crushing process in real time to achieve accurate crushing of barium sulfate.
This method can more accurately control the particle size distribution and morphology of powder, reduce energy consumption and waste of raw materials, improve production efficiency, reduce production costs and improve environmental benefits.
Smart Images

Figure CN118634973B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of artificial intelligence, and in particular to a method and system for preparing ultrafine barium sulfate by airflow pulverization based on reinforcement learning. Background Art
[0002] As an important functional inorganic material, ultrafine barium sulfate plays an indispensable role in many industrial fields, especially in the manufacturing industries of plastics, rubber, coatings and inks. Ultrafine barium sulfate is widely used as a filler or pigment due to its high density, high whiteness and good chemical stability. Since these application fields have strict requirements on the particle size and morphology of barium sulfate, the technical research on the preparation of ultrafine barium sulfate is particularly important. The existing ultrafine barium sulfate preparation technologies mainly include physical and chemical methods, such as ball milling, microwave-assisted synthesis, coprecipitation, etc. These methods have been proven in industry to effectively prepare barium sulfate powders with uniform particle size and good dispersibility. However, these technologies often have disadvantages such as high energy consumption, low production efficiency, difficulty in particle size control and morphology regulation, and are difficult to meet the increasingly stringent market demand.
[0003] In order to address the shortcomings of the existing ultrafine barium sulfate preparation technology, the industry has tried to optimize it through various methods. For example, the improved ball milling technology achieves better crushing effect by adjusting the size and material of the grinding balls and optimizing the operating parameters of the mill; in the chemical synthesis method, different surfactants are added or microwave-assisted technology is used to improve the uniformity of particle size and crystal morphology. Although these improvements have improved product quality to a certain extent, there are still problems such as high production costs, high environmental pollution, and complex process control. In addition, these traditional methods usually rely on experience to set parameters and lack the ability to flexibly regulate specific batches and material properties, which limits their application in the large-scale production of high-quality ultrafine barium sulfate powders.
[0004] Therefore, there is an urgent need for a technical solution that can more accurately control the particle size distribution and morphology of the powder, reduce energy consumption and raw material waste, and improve production efficiency. Summary of the invention
[0005] In order to solve the deficiencies of the prior art, the embodiment of the present application provides a method and system for preparing ultrafine barium sulfate by airflow pulverization based on reinforcement learning. The present application solves the technical problems of the prior art such as high production cost, severe environmental pollution and complex process control.
[0006] The embodiment of the present application provides a method for preparing ultrafine barium sulfate by airflow pulverization based on reinforcement learning, comprising: separating qualified barium sulfate raw materials into particle sizes suitable for airflow pulverization; setting the initial working parameters of the airflow pulverizer based on simulation calculations to perform airflow pulverization on the separated barium sulfate raw materials; collecting the operating data of the airflow pulverizer in real time, and using a reinforcement learning algorithm to calculate the operating data of the airflow pulverizer to optimize the working parameters in real time.
[0007] In a possible implementation, wherein the separation of qualified barium sulfate raw materials into particle sizes suitable for airflow milling comprises: selecting qualified barium sulfate raw materials for screening and cleaning to remove surface impurities and moisture; using a ball mill to coarsely grind the barium sulfate raw materials into a target particle size range; and passing the coarsely ground barium sulfate raw materials through a vibrating screen to screen the particle size to separate the barium sulfate raw materials with a particle size suitable for airflow milling.
[0008] In a possible implementation, the initial working parameters of the airflow pulverizer are set based on simulation calculation to perform airflow pulverization on the separated barium sulfate raw material, including: constructing a pulverization kinetic equation, a crushing frequency equation and a generation rate equation to obtain a target simulation model; bringing the zero-order working parameters into the target simulation model to perform numerical simulation and obtain a particle size distribution; iteratively simulating and calculating the zero-order working parameters according to the particle size distribution and the target particle size threshold to adjust the zero-order working parameters; when the numerical simulation value of the zero-order working parameter corresponding to the particle size distribution reaches the target particle size threshold, the current zero-order working parameter is used as the initial working parameter of the airflow pulverizer to perform airflow pulverization on the separated barium sulfate raw material.
[0009] In a possible implementation, the pulverization dynamics equation includes:
[0010] ,in, express The first-order derivative at time t is represents the particle number concentration of the particle size L at time t, It represents the rate at which coarse particles are broken into particles of size L at time t. Represents the breakage frequency of particles with particle size L at time t.
[0011] In a possible implementation, the crushing frequency equation includes: ,in, represents the breakage frequency of particles with particle size L at time t, represents the primary coefficient, represents the particle density, represents the relative speed, represents a function containing the geometric parameters of the system, represents the first coefficient, represents the second coefficient, represents the third coefficient, Represents the fourth coefficient; wherein the primary coefficient is 0.5, the first coefficient is 0.9, the second coefficient is -0.3, the third coefficient is 1.2, and the fourth coefficient is 0.7.
[0012] In a possible implementation, the generation rate equation includes: ,in, It represents the rate at which coarse particles are broken into particles of size L at time t. represents the particle number concentration of the particle size L at time t, represents the scaling factor of the probability of breakage, represents the influence index of particle size on the probability of breakage; wherein the scaling factor is , the value range of the impact index is 1.2 to 1.5.
[0013] In a possible implementation, the real-time collection of airflow pulverizer operating data and the use of a reinforcement learning algorithm to calculate the airflow pulverizer operating data to optimize working parameters in real time include: constructing and initializing a policy network and a value network, wherein the policy network is used to generate actions, and the value network is used to evaluate the value of the current state; the intelligent agent interacts with the environment and obtains rewards by performing actions to collect data including states, actions, and rewards; based on the collected data, a loss function is constructed using a PPO algorithm to update the parameters of the policy network; based on the collected data, the parameters of the value network are updated using the loss function of the value network.
[0014] .The embodiment of the present application also provides a system for preparing ultrafine barium sulfate by airflow pulverization based on reinforcement learning, comprising: a pre-processing unit, a pulverizing unit and a control unit; wherein the pre-processing unit is used to separate qualified barium sulfate raw materials into particle sizes suitable for airflow pulverization; the pulverizing unit is used to set the initial working parameters of the airflow pulverizer based on simulation calculations to perform airflow pulverization on the separated barium sulfate raw materials; the control unit is used to collect the operating data of the airflow pulverizer including particle size, temperature and pressure in real time, and calculate the operating data of the airflow pulverizer using a reinforcement learning algorithm to optimize the working parameters in real time.
[0015] An embodiment of the present application also provides a device for preparing ultrafine barium sulfate by airflow pulverization based on reinforcement learning, comprising: a processor, a memory, and a system bus; wherein the processor and the memory are connected via the system bus; the memory is used to store one or more programs, and the one or more programs include instructions, which, when executed by the processor, enable the processor to execute the method described in the above embodiment.
[0016] In the method, system and equipment for preparing ultrafine barium sulfate by airflow pulverization based on reinforcement learning provided above, the embodiment of the present application opens up new possibilities by introducing airflow pulverization technology based on reinforcement learning. This technology uses reinforcement learning algorithms to intelligently optimize the pulverization process, and ensures the optimal pulverization effect in a dynamically changing production environment by adjusting operating parameters such as airflow velocity, temperature, pressure, etc. in real time. Compared with traditional airflow pulverization technology, the method based on reinforcement learning can more accurately control the particle size distribution and morphology of the powder, reduce energy consumption and raw material waste, and improve production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] In order to more clearly illustrate the embodiments of the present application 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 some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0018] Figure 1 A schematic diagram of a process for preparing ultrafine barium sulfate by airflow pulverization based on reinforcement learning provided in an embodiment of the present application;
[0019] Figure 2 A schematic flow chart of a method for setting initial working parameters of an airflow pulverizer based on simulation calculation provided in an embodiment of the present application;
[0020] Figure 3 A schematic block diagram of a system for preparing ultrafine barium sulfate by airflow pulverization based on reinforcement learning provided in an embodiment of the present application. DETAILED DESCRIPTION
[0021] Various exemplary embodiments of the present application will now be described in detail with reference to the accompanying drawings. It should be noted that unless otherwise specifically stated, the relative arrangement of components and steps, numerical expressions and numerical values set forth in these embodiments do not limit the scope of the present application.
[0022] Those skilled in the art will appreciate that the terms "first", "second" and the like in the embodiments of the present application are only used to distinguish different steps, devices or modules, etc., and do not represent any specific technical meaning, nor do they represent the necessary logical order between them. It should also be understood that in the embodiments of the present application, "multiple" may refer to two or more, and "at least one" may refer to one, two or more. It should also be understood that for any component, data or structure mentioned in the embodiments of the present application, in the absence of explicit limitation or contrary revelation given in the context, it can generally be understood as one or more. In addition, the term "and / or" in the present application is only a kind of association relationship describing the associated objects, indicating that there can be three kinds of relationships, for example, A and / or B, which can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in the present application generally indicates that the associated objects before and after are an "or" relationship. It should also be understood that the description of each embodiment in the present application emphasizes the differences between the embodiments, and the same or similar parts can refer to each other. For the sake of brevity, they will not be repeated one by one.
[0023] At the same time, it should be understood that, for ease of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship. The following description of at least one exemplary embodiment is actually only illustrative and is by no means intended to limit the present application and its application or use. The techniques, methods and devices known to ordinary technicians in the relevant fields may not be discussed in detail, but where appropriate, the techniques, methods and devices should be considered as part of the specification. It should be noted that similar numbers and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further discussed in subsequent drawings.
[0024] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0025] Figure 1 The schematic diagram of the process of preparing ultrafine barium sulfate by airflow pulverization based on reinforcement learning provided in the embodiment of the present application. Figure 1 As shown, at step S101, the barium sulfate raw material of qualified quality is separated into particle sizes suitable for airflow milling. It should be understood that before starting airflow milling, the barium sulfate raw material must be screened and pretreated. The selected raw material must meet certain chemical purity and physical properties to ensure the quality of the final product.
[0026] First, select qualified barium sulfate raw materials for screening and cleaning to remove surface impurities and moisture. Screen the barium sulfate raw materials to exclude those with oversized particles or irregular shapes, which may affect the crushing efficiency and product uniformity. Cleaning removes impurities and attached tiny particles on the surface of the raw materials, which is critical to reducing contamination during the crushing process. The cleaned raw materials need to be dried to remove surface moisture to prevent moisture from affecting the crushing effect and powder fluidity during the airflow crushing process.
[0027] Secondly, use a ball mill to coarsely grind the barium sulfate raw material to the target particle size range. After the raw material pretreatment is completed, the particle size of the barium sulfate needs to be adjusted by mechanical grinding to reach a particle size range suitable for airflow milling. Use a ball mill to coarsely grind the barium sulfate raw material to initially break up large particles and grind them to a predetermined particle size range. Continue to use finer abrasives and higher grinding speeds for fine grinding to further even out the size and shape of the particles and reduce the width of the particle size distribution.
[0028] Finally, the coarsely ground barium sulfate raw material is screened by a vibrating screen to separate the barium sulfate raw material with a particle size suitable for airflow crushing. The ground barium sulfate is screened by a vibrating screen or air separation equipment, and only particles that reach the target particle size can enter the next airflow crushing process.
[0029] like Figure 1 As shown, at step S102, the initial working parameters of the airflow mill are set based on the simulation calculation to perform airflow milling on the separated barium sulfate raw material (such as Figure 2 The specific method will be shown in Figure 2 ). In one embodiment, the method includes: constructing a pulverization kinetic equation, a crushing frequency equation, and a generation rate equation to obtain a target simulation model; bringing the zero-order working parameter into the target simulation model to perform numerical simulation and obtain a particle size distribution; performing iterative simulation calculation on the zero-order working parameter according to the particle size distribution and the target particle size threshold to adjust the zero-order working parameter; when the numerical simulation value of the particle size distribution corresponding to the zero-order working parameter reaches the target particle size threshold, the current zero-order working parameter is used as the initial working parameter of the airflow pulverizer to perform airflow pulverization on the separated barium sulfate raw material.
[0030] like Figure 1As shown, at step S103, the airflow pulverizer operation data is collected in real time, and the airflow pulverizer operation data is calculated using a reinforcement learning algorithm to optimize the working parameters in real time. Specifically, it includes: constructing and initializing a policy network and a value network, wherein the policy network is used to generate actions, and the value network is used to evaluate the value of the current state; the intelligent agent interacts with the environment and obtains rewards by performing actions to collect data including states, actions, and rewards; based on the collected data, a loss function is constructed using the PPO algorithm to update the parameters of the policy network; based on the collected data, the parameters of the value network are updated using the loss function of the value network.
[0031] In this process, first of all, two main networks need to be built and initialized: the policy network and the value network. The main function of the policy network is to generate decision actions, while the value network is used to evaluate the expected return of a given state. Through these two networks, the operating parameters of the airflow mill, such as wind speed, temperature, particle feed rate, etc., can be finely controlled to ensure the optimization of crushing efficiency and crushing quality.
[0032] First, the construction of the policy network is based on a deep learning model, usually a convolutional neural network (CNN) or a recurrent neural network (RNN). For example, when processing the dynamic data of an airflow mill, an RNN can be used to capture the characteristics of time series data. The initialization of the policy network involves setting the network weights, which are usually set by a random initialization method, such as Xavier initialization or He initialization, to ensure the stability of the network in the early stages of training.
[0033] The design of the value network focuses on evaluating the value of each state, that is, the total reward that may be obtained in the future based on the current state. The value network can also be built using a deep learning framework, and its training depends on feedback signals obtained from the environment. For example, if the airflow crushing efficiency is improved, the value of the state increases, and the network learns this feedback by adjusting parameters.
[0034] The interaction between the agent and the environment is carried out by constantly trying different operating parameters in actual operation. After each operation, the system will evaluate the results of the operation, such as the uniformity of the crushing particle size and production efficiency, and then feed these results back to the policy network and the value network. The actions performed by the agent include adjusting the feed rate, changing the rotation speed of the grinding disc, etc. After each action is performed, a reward value related to the result will be obtained.
[0035] The PPO algorithm plays a key role in this process. The algorithm optimizes network parameters by maximizing the expected rewards of actions generated by the policy network. Specifically, the PPO algorithm ensures the stability of the learning process by comparing the difference in rewards generated by the new and old strategies and limiting the difference to a small range. The mathematical formula is expressed as follows:
[0036] ;
[0037] in, is a ratio function, which represents the ratio of the new strategy to the old strategy; is the advantage function, which measures how much better it is to take an action than the average; and A very small number, usually set to 0.1 or 0.2, used to limit the magnitude of policy updates.
[0038] The parameter update of the value network is based on minimizing its loss function, which usually defines the mean squared error between the predicted value and the actual return: , here, Is the value network for the state The value prediction of is the actual reward obtained from the state.
[0039] Furthermore, the agent model is based on deep reinforcement learning algorithms (such as PPO, SAC, etc.), which mainly includes the following parts:
[0040] The definition of state space and action space, that is, the definition of state space and action space needs to be combined with the characteristics of the airflow pulverizing system. Specifically, the state space can include: medium gas pressure and temperature; nozzle geometry; target geometry and material; classifier working parameters; current product particle size distribution characteristic values (such as D90, D50, etc.). The action space can include: adjusting the medium gas pressure and temperature; adjusting the nozzle and target geometry; adjusting the classifier working parameters.
[0041] The construction of the reward function, that is, the construction of the reward function needs to comprehensively consider the product's particle size distribution, yield, energy consumption and other indicators. In order to simplify the problem, a comprehensive evaluation function can be constructed to weight the sum of various indicators. For example: ,in, is the weight coefficient, which can be determined through experiments and experience.
[0042] The particle size distribution score can be calculated based on the target values of D90 and D50, for example: ,in, and To adjust the parameters.
[0043] The yield score can be calculated based on the ratio of the actual yield to the target yield, for example: .
[0044] The energy consumption score can be calculated based on the ratio of unit product energy consumption to target energy consumption, for example: .
[0045] In the specific implementation, you can choose a deep reinforcement learning algorithm suitable for continuous action space, such as PPO (Proximal Policy Optimization) or SAC (Soft Actor Critic). Taking PPO as an example, its basic idea is to ensure the stability and efficiency of policy optimization by limiting the step size of policy update. The specific implementation steps are as follows:
[0046] 1. Initialization of the policy network and value network: Build and initialize the policy network and value network. The policy network is used to generate actions, and the value network is used to evaluate the value of the current state. 2. Interaction and data collection: The agent interacts with the environment, collects state, action, reward and other data by performing actions and obtaining rewards. 3. Policy update: Use the collected data to update the parameters of the policy network according to the loss function of the PPO algorithm. 4. Value update: Use the collected data to update the parameters of the value network according to the loss function of the value network.
[0047] During the training process of the reinforcement learning optimization framework, the agent gradually optimizes the key parameters of the airflow crushing system through continuous interaction with the simulated environment, so that the system performance is continuously improved. The specific steps are as follows: Initialize the environment model, agent model and reward function, set the initial system parameters and the initial strategy of the agent. In the initial state, the agent performs an action and obtains a new state and reward. Record the state, action, reward and new state to form an interaction round. According to the collected data, the parameters of the policy network and the value network are updated according to the loss function of the PPO algorithm. During the training process, the agent continuously adjusts the parameters of the airflow crushing system to gradually improve the system performance. After the training is completed, the optimized system parameters and agent strategy are obtained. Apply the optimized system parameters and agent strategy to the actual system to verify its effect. If the effect is not ideal, you can continue to adjust the reward function and training parameters, and retrain and optimize.
[0048] In one embodiment, the PPO algorithm is used to optimize a certain airflow pulverizing system. The specific process is as follows: 1. Initialization: Set the initial system parameters (such as medium gas pressure 6 bar, temperature 300K, etc.), build the environment model and agent model, and initialize the policy network and value network. 2. Interaction and data collection: In the initial state, the agent performs an action (such as adjusting the medium gas pressure to 7 bar) to obtain a new state (such as the change in the characteristic value of the product particle size distribution) and reward (calculated based on the particle size distribution, yield, and energy consumption). Record the state, action, reward, and new state. 3. Strategy and value update: According to the collected data, according to the loss function of the PPO algorithm, update the parameters of the policy network and value network. 4. Parameter optimization: During the training process, the agent continuously adjusts the parameters of the airflow pulverizing system, such as medium gas pressure, nozzle geometry, etc., to gradually improve the system performance. 5. Verification and testing: Apply the optimized system parameters (such as medium gas pressure 7.5 bar, nozzle diameter 2.5 mm, etc.) to the actual system to verify its effect. If the effect is not ideal, adjust the reward function and training parameters, and retrain and optimize.
[0049] Through the comprehensive application of the above methods, the real-time optimization of the working parameters of the airflow mill can be effectively achieved, thereby improving production efficiency and product quality. In addition, the implementation of this system can also reduce energy consumption and production costs, with significant economic and environmental benefits.
[0050] Figure 2 A flow chart of a method for setting the initial working parameters of a jet mill based on simulation calculation provided in an embodiment of the present application. Figure 2 As shown, at step S201, a pulverization kinetics equation, a crushing frequency equation and a generation rate equation are constructed to obtain a target simulation model.
[0051] Airflow pulverization is a process that uses high-speed airflow to accelerate particles and crush them through shear force and impact force when they collide with the annular target surface. The crushed fine particles and uncrushed coarse particles are separated in the classifier, the fine particles are collected as products, and the coarse particles are returned to the pulverizer for further crushing. Assuming that the airflow pulverization system is a steady-state continuous process, the following mathematical model can be constructed:
[0052] The kinetic equation of the pulverization process describes the particle number concentration of particle size L The rate of change of , wherein the pulverization kinetic equation includes: ,in, represents the particle number concentration of the particle size L at time t, It represents the rate at which coarse particles are broken into particles of size L at time t. Indicates the crushing frequency of particles with particle size L at time t. Crushing frequency The crushing frequency equation is related to the kinematic and kinetic parameters of the particles, including: ,in, represents the breakage frequency of particles with particle size L at time t, represents the primary coefficient, represents the particle density, represents the relative speed, represents a function containing the geometric parameters of the system, represents the first coefficient, represents the second coefficient, represents the third coefficient, Represents the fourth coefficient; wherein the primary coefficient is 0.5, the first coefficient is 0.9, the second coefficient is -0.3, the third coefficient is 1.2, and the fourth coefficient is 0.7.
[0053] Spawn Rate It can be expressed by selecting a suitable fracture function, and the specific form depends on the actual situation. In one implementation scenario, the generation rate equation includes: ,in, It represents the rate at which coarse particles are broken into particles of size L at time t. represents the particle number concentration of the particle size L at time t, represents the scaling factor of the probability of breakage, represents the influence index of particle size on the probability of breakage; wherein the scaling factor is , the value range of the impact index is 1.2 to 1.5.
[0054] like Figure 2 As shown, at step S202, the zero-level working parameters are brought into the target simulation model to perform numerical simulation and obtain the particle size distribution. In order to ensure the physical rationality of the model and the operability of the system, the following constraints need to be considered: material balance, the balance between the amount of input material and the amount of output material. Energy balance, the balance between the energy input and output of the system. Operating parameter range constraints, each operating parameter needs to change within a certain range to ensure the safe and stable operation of the system.
[0055] In order to concretize the mathematical model, this application takes a certain airflow crushing system as an example. Assume that the main parameters of the system are as follows: the medium gas is compressed air, the pressure is 6 bar, and the temperature is 300K; the nozzle diameter is 2mm, the annular target surface diameter is 100mm, and the material is stainless steel; the classifier is a cyclone classifier, and the working parameter is an inlet wind speed of 15m / s.
[0056] For a specific particle size L and time t, it can be determined by experimental measurement or numerical simulation The initial value of and by solving the kinetic equation, the Specifically, the crushing frequency is calculated based on the initial working parameters. Assuming the particle density is 4.5g / cm³, the medium gas density is 1.2kg / m³, and the relative velocity is calculated according to the above formula: , substituting into the crushing frequency formula, assuming the empirical constant , and assuming The function of the geometric parameters of the system takes the constant value 1, then: , after calculation, the specific crushing frequency is obtained The relationship with the particle size L.
[0057] According to the pulverization kinetic equation, assuming that the initial particle size distribution is known, it can be solved by numerical methods (such as finite difference method) Assuming that the initial particle size distribution is uniformly distributed between 10 μm and 100 μm, the numerical method is used to iteratively solve , the particle size distribution at different time points can be obtained. According to the crushing frequency and particle size distribution, an appropriate fracture function can be selected to represent the generation rate . Assume that the fracture function is: .in and is an empirical constant determined by fitting experimental data.
[0058] like Figure 2 As shown, at step S203, the zero-order working parameters are iteratively simulated and calculated according to the particle size distribution and the target particle size threshold to adjust the zero-order working parameters. After the initial parameter setting is completed, it is necessary to verify its effect through experiments. The specific steps are as follows: Install and debug the airflow crushing system according to the initial settings, and prepare the crude barium sulfate raw material. Start the system, conduct a crushing experiment, and monitor the product particle size distribution and yield in real time. Collect product samples, perform particle size analysis, and calculate particle size characteristic values (such as D90, D50, etc.). According to the experimental data, calibrate the empirical constants and fracture function parameters in the mathematical model. According to the corrected model, optimize the working parameters and improve the system performance.
[0059] In one embodiment, the product D90 is 7μm, D50 is 2μm, and the specific surface area is 8m² / g under the initial setting, which does not meet the expected target. According to the mathematical model and experimental data, the following adjustments can be made: increase the medium gas pressure and increase the compressed air pressure to 8bar to increase the relative speed , increase the frequency of crushing Adjust the nozzle diameter and reduce the nozzle diameter to 1.5mm to further increase the air flow velocity. Optimize the classifier parameters and increase the classifier inlet wind speed to 18m / s to improve the classification efficiency.
[0060] Through the above adjustments, the experiment is conducted again to verify the effect of the optimized parameters. Assume that the experimental results after adjustment show that the product D90 is 4μm, D50 is 0.8μm, and the specific surface area is 11m² / g, which meets the expected goal.
[0061] like Figure 2 As shown, at step S204, when the numerical simulation value of the particle size distribution corresponding to the zero-order working parameter reaches the target particle size threshold, the current zero-order working parameter is used as the initial working parameter of the airflow mill to perform airflow grinding on the separated barium sulfate raw material.
[0062] Figure 3 A schematic block diagram of a system for preparing ultrafine barium sulfate by airflow pulverization based on reinforcement learning provided in an embodiment of the present application. It should be understood that the system shown in the figure is exemplary and not restrictive. This means that the system architecture involved is not limited to a specific form or design, but is presented as an example. In other words, the architecture shown in the figure can be regarded as a way of expression to clearly describe related concepts and relationships, and does not exclude other forms of architecture. Therefore, when interpreting the architecture in the picture, it should be understood that the model is flexible and diverse, and its purpose is to provide an exemplary description rather than a restrictive provision on a specific form.
[0063] Specifically, the system includes a pre-processing unit 301, a crushing unit 302 and a control unit 303; wherein the pre-processing unit 301 is used to separate the barium sulfate raw material of qualified quality into particle size suitable for airflow crushing; the crushing unit 302 is used to set the initial working parameters of the airflow crusher based on simulation calculations to perform airflow crushing on the separated barium sulfate raw material; the control unit 303 is used to collect the airflow crusher operating data including particle size, temperature and pressure in real time, and use the reinforcement learning algorithm to calculate the airflow crusher operating data to optimize the working parameters in real time.
[0064] Furthermore, an embodiment of the present application also provides a device for preparing ultrafine barium sulfate by airflow pulverization based on reinforcement learning, comprising: a processor, a memory, and a system bus; the processor and the memory are connected via the system bus; the memory is used to store one or more programs, and the one or more programs include instructions, which, when executed by the processor, enable the processor to execute any of the above methods.
[0065] Furthermore, an embodiment of the present application also provides a computer program product, which, when executed on a terminal device, enables the terminal device to execute any of the above methods.
[0066] It can be known from the description of the above implementation methods that those skilled in the art can clearly understand that all or part of the steps in the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product can be stored in a storage medium such as ROM / RAM, a disk, an optical disk, etc., and includes a number of instructions for a computer device (which can be a personal computer, a server, or a network communication device such as a media gateway, etc.) to execute the methods described in the various embodiments of the present application or certain parts of the embodiments.
[0067] It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.
[0068] It should also be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.
[0069] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for preparing ultrafine barium sulfate by airflow pulverization based on reinforcement learning. It is characterized in that include: Separate the qualified barium sulfate raw materials into particle sizes suitable for air flow milling; The initial working parameters of the airflow mill are set based on simulation calculations to perform airflow milling on the separated barium sulfate raw materials; The method includes: constructing a pulverization kinetic equation, a crushing frequency equation and a generation rate equation to obtain a target simulation model; bringing the zero-order working parameters into the target simulation model to perform numerical simulation and obtain the particle size distribution; iteratively simulating and calculating the zero-order working parameters according to the particle size distribution and the target particle size threshold to adjust the zero-order working parameters; when the numerical simulation value of the particle size distribution corresponding to the zero-order working parameter reaches the target particle size threshold, the current zero-order working parameter is used as the initial working parameter of the airflow mill to perform airflow pulverization on the separated barium sulfate raw material; Collecting the operating data of the airflow mill in real time, and calculating the operating data of the airflow mill using a reinforcement learning algorithm to optimize the working parameters in real time; Wherein, the pulverization dynamics equation includes: ,in, express About the moment The first derivative of express Particle size at the moment The particle number concentration, express The coarse particles are crushed to form particles of different sizes. The velocity of the particles, express Particle size at the moment The particle breakage frequency; Wherein, the crushing frequency equation includes: ,in, represents the primary coefficient, represents the particle density, represents the relative speed, represents a function containing the geometric parameters of the system, represents the first coefficient, represents the second coefficient, represents the third coefficient, represents the fourth coefficient; wherein the primary coefficient is 0.5, the first coefficient is 0.9, the second coefficient is -0.3, the third coefficient is 1.2, and the fourth coefficient is 0.7; Wherein, the generation rate equation includes: ,in, represents the scaling factor of the fragmentation probability, represents the influence index of particle size on the probability of breakage; wherein the scaling factor is , the value range of the impact index is 1.2 to 1.
5.
2. The method according to claim 1, It is characterized in that in, The method of separating the qualified barium sulfate raw material into particles with a particle size suitable for air flow pulverization comprises: Select qualified barium sulfate raw materials for screening and cleaning to remove surface impurities and moisture; The barium sulfate raw material is coarsely ground into a target particle size range using a ball mill; The coarsely ground barium sulfate raw material is screened by a vibrating screen to separate the barium sulfate raw material with a particle size suitable for air flow milling.
3. The method according to claim 1, It is characterized in that in, The real-time collection of airflow mill operating data and the use of a reinforcement learning algorithm to calculate the airflow mill operating data to optimize operating parameters in real time include: Construct and initialize a policy network and a value network, wherein the policy network is used to generate actions, and the value network is used to evaluate the value of the current state; The agent interacts with the environment and obtains rewards by performing actions to collect data including states, actions, and rewards; Based on the collected data, the PPO algorithm is used to construct a loss function to update the parameters of the policy network; Based on the collected data, the parameters of the value network are updated using the loss function of the value network.
4. A system for preparing ultrafine barium sulfate by airflow pulverization based on reinforcement learning using the method according to any one of claims 1 to 3, It is characterized in that It includes a pre-processing unit, a crushing unit and a control unit; wherein, The pre-treatment unit is used to separate the barium sulfate raw material of qualified quality into particles with a particle size suitable for air flow pulverization; The pulverizing unit is used to set the initial working parameters of the airflow pulverizer based on simulation calculations, so as to perform airflow pulverization on the separated barium sulfate raw material; The control unit is used to collect the air flow mill operating data including particle size, temperature and pressure in real time, and calculate the air flow mill operating data using a reinforcement learning algorithm to optimize the working parameters in real time.
5. Equipment for preparing ultrafine barium sulfate by airflow pulverization based on reinforcement learning, It is characterized in that include: A processor, a memory, and a system bus; wherein the processor and the memory are connected via the system bus; The memory is used to store one or more programs, wherein the one or more programs include instructions, and when the instructions are executed by the processor, the processor executes the method according to any one of claims 1 to 3.
