Sorting strategy determination method and device

Abstract

The invention relates to a method and a device for determining a sorting strategy. The method comprises the following steps: acquiring a target condition parameter corresponding to a target sorting task; the target condition parameters comprise sorting action parameters of the sorting equipment, coal quality characteristics of the to-be-sorted coal and real-time sorting data; the real-time sorting data comprises coal flow and particle size distribution of the to-be-sorted coal; according to the target condition parameters, a target sorting strategy is determined, so that the sorting efficiency and the product quality corresponding to the target sorting task are optimal; the target sorting strategy at least comprises adjustment of screening granularity and change of dense medium density. According to the method, the sorting strategy can be dynamically adjusted, it is ensured that the sorting process is always kept in the optimal state, and therefore the sorting efficiency is improved, the invalid sorting and repeated processing time is shortened, the adaptability of the sorting process is enhanced, and the technical problems that the sorting condition is complex and the optimal sorting effect is difficult to achieve in the coal sorting process are solved.

Description

Technical Field

[0001] The present application relates to the field of coal technology, and in particular to a method and device for determining a sorting strategy. Background Art

[0002] Coal sorting refers to the process of processing coal to remove impurities, reduce ash and sulfur content, and thus improve coal quality. This is the application of mineral processing technology in the coal industry. The main basis for coal sorting is the differences in various physical properties of minerals, surface physical and chemical properties, and chemical properties.

[0003] At present, due to the diversity of coal quality characteristics (such as ash content, moisture content, sulfur content, particle size distribution, etc.) and the sorting action parameters of the sorting equipment (such as the sieve hole size and vibration frequency of the screening machine, the medium density and flow rate of the heavy medium separator, etc.), the existing coal sorting process faces complex sorting conditions and it is difficult to achieve the best sorting effect. Therefore, it is urgently needed to be improved. Summary of the invention

[0004] Based on this, it is necessary to provide a method and device for determining a sorting strategy that can optimize the sorting process in response to the above technical problems.

[0005] In a first aspect, the present application provides a method for determining a sorting strategy, the method comprising: Obtain target condition parameters corresponding to the target sorting task; the target condition parameters include sorting action parameters of the sorting equipment, coal quality characteristics of the coal to be sorted, and real-time sorting data; the real-time sorting data includes coal flow and particle size distribution of the coal to be sorted; According to the target condition parameters, the target sorting strategy is determined so that the sorting efficiency and product quality corresponding to the target sorting task are optimized; the target sorting strategy at least includes adjusting the screening particle size and changing the density of the heavy medium.

[0006] In one embodiment, determining a target sorting strategy based on target condition parameters includes: Matching the target condition parameters with the simulation condition parameters in the simulation test; The simulation sorting strategy corresponding to the successfully matched simulation condition parameters is determined as the target sorting strategy.

[0007] In one embodiment, the simulation sorting strategy corresponding to each simulation condition parameter is determined in the following manner: Construct a Markov model of the sorting process; the state space of the Markov model includes simulation condition parameters corresponding to the simulation sorting task, the action space of the Markov model includes adjusting the screening particle size and changing the density of the heavy medium, and the reward function of the Markov model is determined according to the sorting efficiency and product quality; For each simulation condition parameter, the action space in the Markov model is adjusted to obtain a simulation sorting strategy corresponding to the simulation condition parameter, so that the sorting efficiency and product quality corresponding to the simulation sorting strategy corresponding to the simulation condition parameter are optimized.

[0008] In one embodiment, constructing a Markov model of the sorting process includes: Construct a 3D model of the sorting equipment; In the state space of the Markov model, add the state parameters of the three-dimensional model; In the action space of the Markov model, control parameters of the moving parts in the 3D model are added.

[0009] In one of the embodiments, the state parameters of the three-dimensional model include geometric shape parameters, texture coordinate parameters, material property parameters, structural property parameters and physical property parameters; Correspondingly, in the state space of the Markov model, the state parameters of the three-dimensional model are added, including: In the state space of the Markov model, the state representation corresponding to the state parameters of the three-dimensional model and the state transfer matrix corresponding to the three-dimensional model are added; the state representation is used to describe the position, speed and shape of the sorting equipment at different times; In one of the embodiments, the control parameters of the moving component in the three-dimensional model include position parameters, rotation parameters, speed parameters, acceleration parameters and switch parameters of the moving component; In the action space of the Markov model, the control parameters of the moving parts in the 3D model are added, including: In the action space of the Markov model, control actions corresponding to the control parameters of the moving parts in the three-dimensional model are added.

[0010] In one embodiment, the action space in the Markov model is adjusted to obtain a simulation sorting strategy corresponding to the simulation condition parameters, including: The action space in the Markov model is adjusted to obtain candidate sorting strategies corresponding to the simulation condition parameters, so that the sorting efficiency and product quality corresponding to the candidate sorting strategies corresponding to the simulation condition parameters are optimized; Simulate the sorting process corresponding to the candidate sorting strategy based on the three-dimensional model to obtain the sorting result; If the sorting result passes the verification after being verified by the expert review model, the candidate sorting strategy corresponding to the simulation condition parameter is used as the simulation sorting strategy corresponding to the simulation condition parameter.

[0011] In one embodiment, obtaining the coal quality characteristics of the coal to be sorted includes: According to the coal image of the coal to be sorted, the associated equipment corresponding to the coal to be sorted is determined; the associated equipment corresponding to the coal to be sorted includes at least one of a conveyor belt, a crusher, a screening machine, a heavy medium separator, and a flotation machine; Acquire device operation characteristics of the associated device; the device operation characteristics of the associated device include device sound characteristics and device vibration characteristics of the associated device; The coal quality characteristics of the coal to be sorted are determined based on the coal image and equipment operation characteristics.

[0012] In one embodiment, the coal quality characteristics of the coal to be sorted are determined based on the coal image and the equipment operation characteristics, including: Construct a three-dimensional coal pile model corresponding to the coal to be sorted; The coal quality characteristics of the coal to be sorted are determined based on the three-dimensional coal pile model, coal image and equipment operation characteristics.

[0013] In a second aspect, the present application also provides a sorting strategy determination device, comprising: An acquisition module is used to acquire target condition parameters corresponding to the target sorting task; the target condition parameters include sorting action parameters of the sorting equipment, coal quality characteristics of the coal to be sorted, and real-time sorting data; the real-time sorting data includes coal flow and particle size distribution of the coal to be sorted; The strategy determination module is used to determine the target sorting strategy according to the target condition parameters so as to optimize the sorting efficiency and product quality corresponding to the target sorting task; the target sorting strategy at least includes adjusting the screening particle size and changing the density of the heavy medium.

[0014] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented: Obtain target condition parameters corresponding to the target sorting task; the target condition parameters include sorting action parameters of the sorting equipment, coal quality characteristics of the coal to be sorted, and real-time sorting data; the real-time sorting data includes coal flow and particle size distribution of the coal to be sorted; According to the target condition parameters, the target sorting strategy is determined so that the sorting efficiency and product quality corresponding to the target sorting task are optimized; the target sorting strategy at least includes adjusting the screening particle size and changing the density of the heavy medium.

[0015] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the following steps are implemented: Obtain target condition parameters corresponding to the target sorting task; the target condition parameters include sorting action parameters of the sorting equipment, coal quality characteristics of the coal to be sorted, and real-time sorting data; the real-time sorting data includes coal flow and particle size distribution of the coal to be sorted; According to the target condition parameters, the target sorting strategy is determined so that the sorting efficiency and product quality corresponding to the target sorting task are optimized; the target sorting strategy at least includes adjusting the screening particle size and changing the density of the heavy medium.

[0016] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor: Obtain target condition parameters corresponding to the target sorting task; the target condition parameters include sorting action parameters of the sorting equipment, coal quality characteristics of the coal to be sorted, and real-time sorting data; the real-time sorting data includes coal flow and particle size distribution of the coal to be sorted; According to the target condition parameters, the target sorting strategy is determined so that the sorting efficiency and product quality corresponding to the target sorting task are optimized; the target sorting strategy at least includes adjusting the screening particle size and changing the density of the heavy medium.

[0017] The above-mentioned sorting strategy determination method and device, the present application obtains the coal quality characteristics and sorting data of the coal to be sorted, as well as the parameter status of the sorting equipment in real time, and can dynamically adjust the sorting strategy. This dynamic adjustment can ensure that the sorting process is always kept in the optimal state, thereby improving the sorting efficiency and reducing the time of invalid sorting and repeated processing. Due to the diversity of coal quality characteristics and sorting action parameters, this method can enhance the adaptability of the sorting process through real-time data monitoring and strategy adjustment. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 A schematic diagram of a flow chart of a method for determining a sorting strategy in one embodiment; Figure 2 A schematic diagram of a process for determining the coal quality characteristics of the coal to be sorted in one embodiment; Figure 3 A schematic diagram of a flow chart of steps for constructing a Markov model of a sorting process in one embodiment; Figure 4 It is a schematic flow chart of adjusting the action space in the Markov model to obtain the simulation sorting strategy steps corresponding to the simulation condition parameters in one embodiment; Figure 5 It is a structural block diagram of a sorting strategy determination device in one embodiment. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0021] In an exemplary embodiment, a method for determining a sorting strategy is provided, such as Figure 1 As shown, the method includes: S101, obtaining target condition parameters corresponding to the target sorting task.

[0022] Among them, the target condition parameters include the sorting action parameters of the sorting equipment, the coal quality characteristics of the coal to be sorted and the real-time sorting data.

[0023] Specifically, a communication connection with the sorting equipment is established, and a parameter query request is sent to the sorting equipment to obtain the sorting action parameters of the equipment. The sorting action parameters may include the equipment's operating speed, vibration frequency, sorting accuracy, screen size, etc. The acquired equipment parameters are verified to ensure that the parameter values ​​are within a reasonable range and match the current task requirements.

[0024] Optionally, the coal quality characteristics of the coal to be sorted may include the ash content, moisture content, sulfur content, calorific value, etc. of the coal.

[0025] Among them, the real-time sorting data includes the coal flow and particle size distribution of the coal to be sorted.

[0026] Specifically, the data monitoring system is started to collect key data in the sorting process in real time, such as equipment operating status, sorting efficiency, coal flow, etc. By analyzing the real-time sorting data, we can understand the impact of different coal quality characteristics on the sorting effect.

[0027] S102, determining a target sorting strategy according to target condition parameters so as to optimize the sorting efficiency and product quality corresponding to the target sorting task.

[0028] Among them, the target sorting strategy at least includes adjusting the screening particle size and changing the density of the heavy medium.

[0029] In one possible implementation, the target sorting strategy can be determined based on a rule-based approach: a sorting strategy is selected or generated based on pre-set rules and thresholds. For example, if the coal ash content exceeds a certain threshold, the screening particle size is increased or the heavy medium density is adjusted.

[0030] In another possible implementation, the target sorting strategy can also be determined based on machine learning methods: supervised learning, reinforcement learning and other algorithms are used to train models to predict the optimal sorting strategy based on historical sorting data and target condition parameters. If the historical data contains condition parameters and the corresponding optimal strategy, a classification or regression model can be used to predict the target strategy.

[0031] Then, the determined sorting strategy is sent to the control system of the sorting equipment to perform corresponding adjustment operations.

[0032] The above-mentioned sorting strategy determination method, the present application obtains the coal quality characteristics and sorting data of the coal to be sorted, as well as the parameter status of the sorting equipment in real time, and can dynamically adjust the sorting strategy. This dynamic adjustment can ensure that the sorting process is always kept in the optimal state, thereby improving the sorting efficiency and reducing the time of invalid sorting and repeated processing. Due to the diversity of coal quality characteristics and sorting action parameters, this method can enhance the adaptability of the sorting process through real-time data monitoring and strategy adjustment.

[0033] In an exemplary embodiment, a target sorting strategy is determined according to target condition parameters, including: matching the target condition parameters with various simulation condition parameters in a simulation experiment; and determining the simulation sorting strategy corresponding to the successfully matched simulation condition parameters as the target sorting strategy.

[0034] Among them, the simulation condition parameters are real historical experiments, and each simulation condition parameter should cover different coal quality characteristics, real-time sorting data (such as coal flow, particle size distribution) and sorting action parameters of the sorting equipment.

[0035] Specifically, for each combination of condition parameters, the actual sorting process is simulated through the simulation model, and the corresponding optimal sorting strategy and its effect (such as sorting efficiency, product quality, etc.) are recorded. These data and strategies are organized into a database for subsequent matching and use.

[0036] Exemplarily, the acquired target condition parameters are compared one by one with each simulation condition parameter in the simulation test database or matched using a more advanced matching algorithm (such as nearest neighbor algorithm, fuzzy matching, etc.). The goal is to find the simulation condition parameter combination that is most similar or closest to the target condition. Once a successfully matched simulation condition parameter combination is found, the corresponding simulation sorting strategy is directly used as the target sorting strategy.

[0037] In an exemplary embodiment, Figure 2 As shown, the simulation sorting strategy corresponding to each simulation condition parameter is determined in the following way: S201, construct a Markov model of the sorting process.

[0038] Among them, the state space of the Markov model includes the coal quality characteristics of the coal to be sorted and the sorting action parameters of the sorting equipment. The action space of the Markov model includes adjusting the screening particle size and changing the density of the heavy medium. The reward function of the Markov model is determined according to the sorting efficiency and product quality.

[0039] It can be understood that the construction process of the Markov model is: Define the state space: The state space should include the coal quality characteristics of the coal to be sorted (such as ash content, moisture content, sulfur content, particle size distribution, etc.) and the sorting action parameters of the sorting equipment (such as the sieve hole size and vibration frequency of the screening machine, the medium density and flow rate of the heavy medium separator, etc.).

[0040] Define the action space: The action space includes the operating parameters that can be adjusted, such as adjusting the screening particle size and changing the density of the dense medium.

[0041] Define the reward function: The reward function should be defined based on sorting efficiency and product quality. For example, it can be set as a positive reward for improved sorting efficiency and improved product quality, and a negative penalty for reduced sorting efficiency and reduced product quality.

[0042] S202, for each simulation condition parameter, adjusting the action space in the Markov model to obtain a simulation sorting strategy corresponding to the simulation condition parameter, so that the sorting efficiency and product quality corresponding to the simulation sorting strategy corresponding to the simulation condition parameter are optimized.

[0043] Optionally, for each set of simulation condition parameters, initialize the state of the Markov model. Use a deep reinforcement learning algorithm to train a policy network that can select the best action based on the target state. During the training process, the algorithm simulates the sorting process, evaluates the quality of each action based on the reward function, and adjusts the parameters of the policy network through the back-propagation algorithm to maximize the long-term reward. After the training is completed, the obtained policy network is the simulation sorting strategy under the simulation condition parameters.

[0044] Furthermore, based on deep reinforcement learning, we can further combine dynamic programming algorithms to optimize long-term strategies. Dynamic programming algorithms can consider rewards for multiple time steps in the future, thereby making more globally optimal decisions. At the same time, using the historical sorting strategies and historical sorting conditions of historical sorting tasks, we can perform transfer learning, accelerate the learning process of new strategies, and improve the generalization ability of strategies.

[0045] For example, taking Deep Q-Network (DQN) as an example, the specific implementation steps are as follows: Initialize the Q network (a deep neural network) to estimate the Q value (i.e., expected future reward) for each state-action pair. Initialize the experience replay buffer to store state transition samples during training. For each set of simulation condition parameters, reset the state of the Markov model. At each time step, select an action based on the target state and the Q network (such as adjusting the sieving granularity or changing the density of the medium). Execute the action, observe the new state and reward, and store this state transition sample in the experience replay buffer.

[0046] A batch of samples are randomly drawn from the experience replay buffer. For each sample, the target Q value is calculated (based on the target Q network and the reward function). The loss is calculated using the target Q value and the predicted value of the target Q network, and the parameters of the Q network are updated through the back-propagation algorithm. Repeat the simulation sorting process and train the Q network until the Q network converges or reaches the preset number of training rounds. The converged Q network is the simulation sorting strategy under the simulation condition parameters, and the optimal action can be selected according to the target state. During the training process, a dynamic programming algorithm can be used to optimize the long-term strategy, considering the rewards of multiple time steps in the future. At the same time, the data of historical sorting tasks can be used for transfer learning to accelerate the learning process of the Q network.

[0047] In an exemplary embodiment, Figure 3 As shown, a Markov model of the sorting process is constructed, including: S301, constructing a three-dimensional model of the sorting equipment.

[0048] Optionally, design a 3D model of the sorting equipment using 3D modeling software (such as Blender, 3ds Max, SolidWorks, etc.). Create an accurate geometric model based on the actual size, shape, and structure of the equipment. Add necessary details to the model, such as texture, color, material, etc., to improve the realism of the simulation.

[0049] S302, adding the state parameters of the three-dimensional model to the state space of the Markov model.

[0050] Among them, the state parameters of the three-dimensional model include geometric shape parameters, texture coordinate parameters, material property parameters, structural property parameters and physical property parameters.

[0051] Exemplarily, the geometric shape parameters include: Vertex position: the three-dimensional coordinates of each vertex in the three-dimensional model. Edge length or area: the length or area of ​​the edge or face in the model, which can reflect the size and shape changes of the model. Normal direction: the normal direction of the vertex or face, used to describe the orientation of the surface.

[0052] Texture and material parameters, including: Texture coordinates: describe the mapping position of the texture image on the model surface. Material properties: such as color, glossiness, reflectivity, transparency, etc., which can affect the appearance of the model when rendered.

[0053] Structural attribute parameters are deformation and animation parameters, including: Joint angles and positions in skeletal animation: used to describe the skeletal animation state of the model. Vertex weights: In skeletal animation, each vertex may be affected by multiple bones, and vertex weights describe the degree of such influence. Deformation parameters: parameters for deformation operations such as stretching, bending, and twisting.

[0054] Physical property parameters are physical and dynamic parameters, including: mass, center of gravity and inertia tensor: used for physical simulation and collision detection. Velocity and acceleration: describe the motion state of the model in physical space. Torque and force: the force and torque acting on the model can affect the movement and deformation of the model.

[0055] Specifically, in the state space of the Markov model, the state parameters of the three-dimensional model are added, including: In the state space of the Markov model, the state representation corresponding to the state parameters of the three-dimensional model and the state transfer matrix corresponding to the three-dimensional model are added.

[0056] Among them, the state representation is used to describe the position, speed and shape of the sorting equipment at different times.

[0057] Exemplarily, the state space of the Markov model is determined, which should be able to represent all possible states of the sorting device. The state may include the position, speed, shape of the device and any other parameters related to the sorting process. The state parameters (geometric shape parameters, texture coordinate parameters, material property parameters, structural property parameters, physical property parameters) of the three-dimensional model are mapped to the state space of the Markov model. For each state parameter, a corresponding state representation is defined, which should be able to clearly describe the state of the sorting device at different times.

[0058] According to the operation logic of the sorting equipment and the possible state transition, a state transition matrix is ​​constructed. Each element in the matrix represents the probability of transitioning from one state to another. Make sure that the state transition matrix conforms to the properties of the Markov chain, that is, the current state is only related to the previous state.

[0059] S303, adding control parameters of the moving parts in the three-dimensional model to the action space of the Markov model.

[0060] The control parameters of the moving parts in the three-dimensional model include position parameters, rotation parameters, speed parameters, acceleration parameters and switch parameters of the moving parts.

[0061] Specifically, adding the control parameters of the moving parts in the three-dimensional model in the action space of the Markov model includes: adding the control actions corresponding to the control parameters of the moving parts in the three-dimensional model in the action space of the Markov model.

[0062] Exemplarily, determine the action space of the Markov model, which should contain all possible control actions. The control action should be able to change the state of the sorting equipment, such as moving, rotating, accelerating, decelerating, switching, etc. Map the control parameters (position parameters, rotation parameters, speed parameters, acceleration parameters, switching parameters) of the moving parts in the three-dimensional model to the action space of the Markov model. Define corresponding control actions for each control parameter, and these actions should be able to be directly applied to the moving parts in the three-dimensional model.

[0063] Write control logic to calculate and apply necessary control parameters based on the state of the Markov model and the selected action. Ensure that the control logic can respond to state changes in real time and accurately control the movement of the moving parts.

[0064] For example, position parameters describe the position of a moving part in three-dimensional space, such as X, Y, and Z coordinates. Rotation parameters are used to describe the rotation state of a moving part, such as the rotation angle around the X, Y, and Z axes. Speed ​​parameters are used to describe the translation or rotation speed of a moving part. Acceleration parameters are used to describe the rate of change of the speed of a moving part. Switching parameters are used to control the switch state of a switchable moving part (such as a joint of a robotic arm).

[0065] The move action is used to specify that the moving part moves a certain distance in a certain direction. The rotate action is used to specify that the moving part rotates a certain angle around a certain axis. The speed adjustment action is used to increase or decrease the translation or rotation speed of the moving part. The acceleration adjustment action is used to change the acceleration of the moving part. The switch action is used to switch the on / off state of the moving part.

[0066] In this embodiment, by constructing a three-dimensional model of the sorting equipment and combining it with the Markov model, accurate simulation and optimization of the sorting process can be achieved.

[0067] In an exemplary embodiment, Figure 4 As shown, the action space in the Markov model is adjusted to obtain the simulation sorting strategy corresponding to the simulation condition parameters, including: S401, adjusting the action space in the Markov model to obtain candidate sorting strategies corresponding to the simulation condition parameters, so that the sorting efficiency and product quality corresponding to the candidate sorting strategies corresponding to the simulation condition parameters are optimized.

[0068] Optionally, the action space in the Markov model is adjusted based on the simulation condition parameters. This may involve modifying certain actions in the action space, or adjusting the transition probabilities between actions, so that the sorting efficiency and product quality are optimized under given simulation conditions. Optimization algorithms (such as genetic algorithms, particle swarm optimization algorithms, etc.) can be used to search for the optimal action space configuration. Based on the adjusted action space, a set of candidate sorting strategies are generated. Each strategy corresponds to a specific action space configuration, that is, a specific set of control actions and decision rules.

[0069] S402, simulating a sorting process corresponding to a candidate sorting strategy based on the three-dimensional model to obtain a sorting result.

[0070] Optionally, load the previously constructed three-dimensional model of the sorting equipment into the simulation environment. According to the candidate sorting strategy, set the relevant parameters in the simulation environment, including the operating parameters of the sorting equipment, the physical property parameters of the coal, etc. Based on the three-dimensional model and the set simulation parameters, perform the simulation of the sorting process. This usually involves simulating the flow, screening, and discharge of coal in the sorting equipment. After the simulation is completed, obtain the sorting results from the simulation environment, including indicators such as sorting efficiency, product quality, and energy consumption.

[0071] S403, if the result of the sorting result after verification by the expert review model is that the verification is passed, the candidate sorting strategy corresponding to the simulation condition parameter is used as the simulation sorting strategy corresponding to the simulation condition parameter.

[0072] Optionally, an expert review model is established, which can objectively and comprehensively evaluate candidate sorting strategies based on sorting results and preset evaluation criteria. The sorting results obtained by simulation are input into the expert review model for verification. The verification process may include comparing simulation results with experimental data, evaluating the improvement of sorting efficiency and product quality, analyzing the reduction of energy consumption and cost, etc.

[0073] If the result of the expert review model verification is that the verification is passed, that is, the candidate sorting strategy performs well under the simulation conditions and meets the expected sorting efficiency and product quality requirements, then the candidate sorting strategy is determined as the simulation sorting strategy corresponding to the simulation condition parameters. If the verification fails, it is necessary to return to step S401, continue to adjust the action space in the Markov model, and generate a new candidate sorting strategy for simulation and verification until a simulation sorting strategy that meets the requirements is found.

[0074] In an exemplary embodiment, coal quality characteristics of the coal to be sorted are obtained, including: determining associated equipment corresponding to the coal to be sorted based on a coal image of the coal to be sorted; obtaining equipment operation characteristics of the associated equipment; and determining the coal quality characteristics of the coal to be sorted based on the coal image and the equipment operation characteristics.

[0075] Among them, the associated equipment corresponding to the coal to be sorted includes at least one of a conveyor belt, a crusher, a screening machine, a heavy medium separator, and a flotation machine.

[0076] The device operation characteristics of the associated device include device sound characteristics and device vibration characteristics of the associated device.

[0077] Specifically, use a camera or image sensor to shoot the coal to be sorted to obtain a clear coal image. The image should contain the appearance characteristics of the coal, such as color, texture, shape, etc., which can be related to the coal quality characteristics (such as ash, moisture, sulfur, etc.). According to the coal sorting process, determine the equipment directly related to the coal to be sorted, such as conveyor belts, crushers, screens, heavy medium separators, and flotation machines. These equipment will treat the coal differently during the coal sorting process, so their operating parameters can contain information about the coal quality characteristics. For each associated equipment, install corresponding sensors to monitor its operating parameters. In particular, pay attention to equipment sound parameters (such as noise level, sound frequency, etc.) and equipment vibration parameters (such as vibration amplitude, vibration frequency, etc.). These parameters can reflect the flow state, collision or sorting effect of the coal inside the equipment, thereby indirectly reflecting the coal quality characteristics.

[0078] Then, the acquired coal images and equipment operation characteristics are input into the pre-trained machine learning model. The model should be able to infer the coal quality characteristics of the coal to be sorted based on these input data and the prior knowledge of coal sorting. Possible models include convolutional neural networks (CNN) for image processing, and regression models or classification models for correlation analysis between equipment operation characteristics and coal quality characteristics. The output results should include key coal quality indicators such as ash, moisture, sulfur, calorific value, etc. Ensure that the acquisition of coal images and equipment operation characteristics is synchronized so that they can be accurately associated. Machine learning models need to be trained with a large amount of labeled data, which should include coal images of different coal quality characteristics and corresponding equipment operation characteristics. During the model training process, the image features and equipment operation characteristics that are most helpful for predicting coal quality characteristics should be carefully selected. As the coal sorting process changes and new data accumulates, the machine learning model should be updated regularly to improve the prediction accuracy.

[0079] Furthermore, the coal quality characteristics of the coal to be sorted are determined based on the coal image and equipment operation characteristics, including: constructing a three-dimensional coal pile model corresponding to the coal to be sorted; and determining the coal quality characteristics of the coal to be sorted based on the three-dimensional coal pile model, the coal image and equipment operation characteristics.

[0080] For example, a 3D scanning technology or a stereoscopic vision technology is used to perform an all-round scan on the coal pile to be sorted, and 3D point cloud data of its surface is obtained. Based on the obtained 3D point cloud data, a 3D coal pile model of the coal to be sorted is constructed using 3D modeling software.

[0081] Specifically, use image processing technology (such as edge detection, texture analysis, etc.) to extract key features in coal images, such as coal particle size distribution, color, gloss, etc. Analyze the operating characteristics of the sorting equipment, including the vibration frequency of the equipment, screen inclination, feeding speed, etc.

[0082] Comprehensively analyze the three-dimensional coal pile model, coal image features and equipment operation characteristics. Use the three-dimensional coal pile model to understand the overall structure and distribution of the coal pile. Combine the coal image features to analyze the surface characteristics of the coal, such as particle size, color, gloss, etc.; consider the impact of equipment operation characteristics on coal sorting, such as vibration frequency may affect the stratification effect of coal, and the screen inclination may affect the transmittance of coal of different particle sizes.

[0083] Based on the above comprehensive analysis, the coal quality characteristics of the coal to be sorted are determined, including key indicators such as coal particle size distribution, ash content, sulfur content, calorific value, etc.

[0084] It is to be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0085] Based on the same inventive concept, the embodiment of the present application also provides a sorting strategy determination device for implementing the sorting strategy determination method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more embodiments of the sorting strategy determination device provided below can refer to the limitations of the sorting strategy determination method above, and will not be repeated here.

[0086] In an exemplary embodiment, Figure 5 As shown, a sorting strategy determination device is provided, comprising: The acquisition module 11 is used to acquire the target condition parameters corresponding to the target sorting task; the target condition parameters include the sorting action parameters of the sorting equipment, the coal quality characteristics of the coal to be sorted and the real-time sorting data; the real-time sorting data includes the coal flow and particle size distribution of the coal to be sorted; The strategy determination module 12 is used to determine the target sorting strategy according to the target condition parameters so as to optimize the sorting efficiency and product quality corresponding to the target sorting task; the target sorting strategy at least includes adjusting the screening particle size and changing the density of the heavy medium.

[0087] Each module in the above-mentioned sorting strategy determination device can be implemented in whole or in part by software, hardware and a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0088] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented: Obtain target condition parameters corresponding to the target sorting task; the target condition parameters include sorting action parameters of the sorting equipment, coal quality characteristics of the coal to be sorted, and real-time sorting data; the real-time sorting data includes coal flow and particle size distribution of the coal to be sorted; According to the target condition parameters, the target sorting strategy is determined so that the sorting efficiency and product quality corresponding to the target sorting task are optimized; the target sorting strategy at least includes adjusting the screening particle size and changing the density of the heavy medium.

[0089] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: Obtain target condition parameters corresponding to the target sorting task; the target condition parameters include sorting action parameters of the sorting equipment, coal quality characteristics of the coal to be sorted, and real-time sorting data; the real-time sorting data includes coal flow and particle size distribution of the coal to be sorted; According to the target condition parameters, the target sorting strategy is determined so that the sorting efficiency and product quality corresponding to the target sorting task are optimized; the target sorting strategy at least includes adjusting the screening particle size and changing the density of the heavy medium.

[0090] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps: Obtain target condition parameters corresponding to the target sorting task; the target condition parameters include sorting action parameters of the sorting equipment, coal quality characteristics of the coal to be sorted, and real-time sorting data; the real-time sorting data includes coal flow and particle size distribution of the coal to be sorted; According to the target condition parameters, the target sorting strategy is determined so that the sorting efficiency and product quality corresponding to the target sorting task are optimized; the target sorting strategy at least includes adjusting the screening particle size and changing the density of the heavy medium.

[0091] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0092] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0093] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for determining a sorting strategy, characterized in that: The method comprises: Obtaining target condition parameters corresponding to the target sorting task; the target condition parameters include sorting action parameters of the sorting equipment, coal quality characteristics of the coal to be sorted, and real-time sorting data; the real-time sorting data includes coal flow and particle size distribution of the coal to be sorted; According to the target condition parameters, a target sorting strategy is determined to optimize the sorting efficiency and product quality corresponding to the target sorting task; the target sorting strategy at least includes adjusting the screening particle size and changing the density of the heavy medium.

2. The method according to claim 1, characterized in that Determining the target sorting strategy according to the target condition parameters includes: Matching the target condition parameters with the simulation condition parameters in the simulation test; The simulation sorting strategy corresponding to the successfully matched simulation condition parameters is determined as the target sorting strategy.

3. The method according to claim 2, characterized in that The simulation sorting strategy corresponding to each simulation condition parameter is determined in the following way: Constructing a Markov model of the sorting process; wherein the state space of the Markov model includes simulation condition parameters corresponding to the simulated sorting task, the action space of the Markov model includes adjusting the screening particle size and changing the density of the heavy medium, and the reward function of the Markov model is determined according to the sorting efficiency and product quality; For each simulation condition parameter, the action space in the Markov model is adjusted to obtain a simulation sorting strategy corresponding to the simulation condition parameter, so that the sorting efficiency and product quality corresponding to the simulation sorting strategy corresponding to the simulation condition parameter are optimized.

4. The method according to claim 3, characterized in that The Markov model of the sorting process is constructed, including: Constructing a three-dimensional model of the sorting equipment; In the state space of the Markov model, adding the state parameters of the three-dimensional model; In the action space of the Markov model, control parameters of the moving parts in the three-dimensional model are added.

5. The method according to claim 4, characterized in that The state parameters of the three-dimensional model include geometric shape parameters, texture coordinate parameters, material property parameters, structural property parameters and physical property parameters; Correspondingly, adding the state parameters of the three-dimensional model into the state space of the Markov model includes: In the state space of the Markov model, a state representation corresponding to the state parameters of the three-dimensional model and a state transfer matrix corresponding to the three-dimensional model are added; the state representation is used to describe the position, speed and shape of the sorting device at different times.

6. The method according to claim 4, characterized in that The control parameters of the moving parts in the three-dimensional model include position parameters, rotation parameters, speed parameters, acceleration parameters and switch parameters of the moving parts; Adding control parameters of the moving parts in the three-dimensional model to the action space of the Markov model includes: In the action space of the Markov model, a control action corresponding to the control parameter of the moving part in the three-dimensional model is added.

7. The method according to claim 4, characterized in that The step of adjusting the action space in the Markov model to obtain a simulation sorting strategy corresponding to the simulation condition parameters includes: The action space in the Markov model is adjusted to obtain a candidate sorting strategy corresponding to the simulation condition parameter, so that the sorting efficiency and product quality corresponding to the candidate sorting strategy corresponding to the simulation condition parameter are optimized; Simulating a sorting process corresponding to the candidate sorting strategy based on the three-dimensional model to obtain a sorting result; If the result of the sorting result after verification by the expert review model is that the verification is passed, the candidate sorting strategy corresponding to the simulation condition parameter is used as the simulation sorting strategy corresponding to the simulation condition parameter.

8. The method according to claim 1, characterized in that Obtaining the coal quality characteristics of the coal to be sorted, including: According to the coal image of the coal to be sorted, the associated equipment corresponding to the coal to be sorted is determined; the associated equipment corresponding to the coal to be sorted includes at least one of a conveyor belt, a crusher, a screening machine, a heavy medium separator, and a flotation machine; Acquire the device operation characteristics of the associated device; the device operation characteristics of the associated device include the device sound characteristics and device vibration characteristics of the associated device; The coal quality characteristics of the coal to be sorted are determined according to the coal image and the equipment operation characteristics.

9. The method according to claim 8, characterized in that Determining the coal quality characteristics of the coal to be sorted according to the coal image and the equipment operation characteristics includes: Constructing a three-dimensional coal pile model corresponding to the coal to be sorted; The coal quality characteristics of the coal to be sorted are determined according to the three-dimensional coal pile model, the coal image and the equipment operation characteristics.

10. A sorting strategy determination device, characterized in that: The device comprises: An acquisition module, used to acquire target condition parameters corresponding to the target sorting task; the target condition parameters include sorting action parameters of the sorting equipment, coal quality characteristics of the coal to be sorted, and real-time sorting data; the real-time sorting data includes coal flow and particle size distribution of the coal to be sorted; The strategy determination module is used to determine the target sorting strategy according to the target condition parameters so as to optimize the sorting efficiency and product quality corresponding to the target sorting task; the target sorting strategy at least includes adjusting the screening particle size and changing the density of the heavy medium.

Images