Method for predicting screening efficiency of vibrating screen

Abstract

The invention relates to a method for predicting the screening efficiency of a vibrating screen, belongs to the technical field of screening efficiency prediction, and solves the problems of low prediction precision, large calculation amount, poor real-time performance and poor universality of a method for predicting the screening efficiency of the vibrating screen in the prior art. Constructing a vibrating screen discrete element model and a particle factory space model; constructing a screening efficiency sample data set based on the vibrating screen discrete element model and the particle factory space model; for the particles of each size, training a plurality of machine learning models based on the screening efficiency sample data set, and taking the machine learning model with the highest screening efficiency prediction precision as a screening efficiency prediction model of the particles; a sample to be detected is obtained, the corresponding screening efficiency prediction model is selected based on the particle size, the percentage of the fed particles and the total amount of the fed particles serve as input parameters, and the screening efficiency corresponding to the particle size can be obtained. The method for predicting the screening efficiency of the vibrating screen is high in prediction precision, high in calculation efficiency, high in real-time performance and capable of adapting to various complex working conditions.

Description

Technical Field

[0001] The invention relates to the technical field of screening efficiency prediction, and in particular to a method for predicting the screening efficiency of a vibrating screen. Background Art

[0002] Vibrating screen is a kind of screening equipment widely used in coal, mining, metallurgy, building materials, chemical industry and other industries. Its working principle is to use the exciting force generated by the vibration motor or exciter to make the screen box vibrate periodically, and the material is thrown up and moved forward on the screen surface, thereby achieving the separation of materials of different particle sizes. Screening efficiency is an important indicator to measure the working effect of the vibrating screen, which is affected by many factors such as the operating state of the screening equipment (amplitude, vibration frequency, vibration intensity, speed, etc.), structural parameters (screen, screen surface opening rate, aperture size, inclination, etc.) and material characteristics (water content, feed rate, particle size distribution, etc.).

[0003] In the prior art, there are two main methods for calculating screening efficiency: (1) constructing a screening efficiency prediction model based on dynamics and probability theory, and predicting the screening efficiency of the vibrating screen through theoretical deduction and experimental data fitting. Although the established model can explain the basic phenomena in the screening process to a certain extent, its prediction accuracy is often unsatisfactory when faced with complex working conditions and diversified materials; (2) discrete element method, which numerically solves the particle motion equation and simulates the interaction between particles and between particles and the screen surface, and can reproduce the screening process more realistically. For example, DEM simulation has been used to evaluate the screening efficiency of circular vibrating screens. The results show that different vibration parameters have a significant effect on the screening efficiency. However, each simulation using the discrete element method is for specific screening parameters, and can only predict the screening efficiency under specific parameters. If the screening efficiency of the vibrating screen under multiple conditions is predicted, a large amount of computing resources and storage resources will be consumed, which makes large-scale industrial applications or real-time simulations difficult and expensive.

[0004] Therefore, it is necessary to provide a vibrating screen screening efficiency prediction method with high prediction accuracy, high calculation efficiency, strong real-time performance and adaptability to various complex working conditions. Summary of the invention

[0005] In view of the above analysis, an embodiment of the present invention aims to provide a method for predicting the screening efficiency of a vibrating screen, so as to solve the problems of low prediction accuracy, large amount of calculation, poor real-time performance and poor versatility of the existing method for predicting the screening efficiency of a vibrating screen.

[0006] An embodiment of the present invention provides a method for predicting screening efficiency of a vibrating screen, comprising:

[0007] Construct discrete element model of vibrating screen and spatial model of pellet plant;

[0008] Constructing a screening efficiency sample data set based on the vibration screen discrete element model and the particle factory space model, wherein the screening efficiency sample data set includes: particle size, feed particle percentage, feed particle total amount, and particle screening efficiency;

[0009] For each size of particles, multiple machine learning models are trained based on the screening efficiency sample data set, and the machine learning model with the highest screening efficiency prediction accuracy is used as the screening efficiency prediction model for this type of particles;

[0010] Obtain the sample to be tested, select the corresponding screening efficiency prediction model based on the particle size, and use the percentage of feed particles and the total amount of feed particles as input parameters to obtain the screening efficiency corresponding to the particle size.

[0011] Based on the further improvement of the above method, the construction of the vibration screen discrete element model and the particle factory space model includes: constructing a vibration screen discrete element model, the vibration screen discrete element model includes: a screen and a material baffle, the material baffle is located on the feed side, and based on the preset screen length, the screen discrete element model is divided to obtain multiple sub-screen discrete element models with the same screen length; constructing a particle factory space model, the particle factory space model is arranged above the screen discrete element model and close to the material baffle, and the particle generation method and particle parameters of the particle factory space model are set based on business needs.

[0012] Based on the further improvement of the above method, the screening efficiency sample data set is constructed based on the vibration screen discrete element model and the particle factory space model, including:

[0013] A1: Based on the discrete element model of the vibration screen and the spatial model of the particle factory, a vibration screen screening process simulation is implemented, and a simulation result is obtained, wherein the simulation result includes the spatial position of each particle corresponding to each time step in the simulation process, the percentage of input particles, and the total amount of input particles;

[0014] A2: Calculating the particle screening efficiency of different particle sizes corresponding to each sub-sieve discrete element model based on the spatial position of each particle;

[0015] A3: The percentage of input particles of each type of particle, particle screening efficiency and total amount of input particles at each time step corresponding to each sub-sieve discrete element model are taken as a sample in the screening efficiency sample data set corresponding to the sieve length;

[0016] A4: Determine whether the amount of sample data currently generated meets the preset quantity requirement. If so, stop generating the sample data. If not, adjust the current particle generation method and particle parameters based on business requirements and return to A1.

[0017] Based on the further improvement of the above method, the simulation result refers to the percentage of feed particles, the total amount of feed particles and the spatial position of each particle corresponding to each sub-sieve discrete element model in each time step after the screening process reaches a steady state; the steady state means that the sum of the screening particle rate of the sieve discrete element model and the particle outflow rate at the discharge end per unit time is equal to the particle feed rate.

[0018] Based on the further improvement of the above method, the particle screening efficiency of different particle sizes corresponding to each sub-sieve discrete element model is calculated based on the spatial position of each particle, including:

[0019] For the first sub-sieve discrete element model near the feed end, the screening efficiency of each particle size is calculated based on the particle size generated in the particle plant spatial model, the feed particle percentage, the total feed particle amount, and the number of particles of various sizes below the first sub-sieve discrete element model;

[0020] For other sub-sieve discrete element models, the particle size, particle percentage and total particle amount at the discharge end of the previous sub-sieve discrete element model are used as the particle size, feed particle percentage and total feed particle amount of this sub-sieve discrete element model. The screening efficiency of each particle size is calculated based on the particle size, feed particle percentage and total feed particle amount of this sub-sieve discrete element model and the number of particles of various sizes below this sub-sieve discrete element model.

[0021] Based on the further improvement of the above method, the method of training multiple machine learning models based on the screening efficiency sample data set includes:

[0022] Performing data preprocessing on the screening efficiency sample data set to obtain a training sample data set;

[0023] The hyperparameters of each machine learning model are optimized based on the random grid search method, and the optimization results are used as the initial hyperparameters of each machine learning model;

[0024] K-fold cross validation is used to train the optimized machine learning models.

[0025] Based on the further improvement of the above method, the particle generation method adopts a dynamic method, and the particle parameters include particle size distribution, which are 3mm, 5mm, 7mm, and 10mm respectively.

[0026] Based on the further improvement of the above method, the machine learning model with the highest predicted screening efficiency is used as the screening efficiency prediction model of the particles, including:

[0027] For 3 mm particles, the screening efficiency prediction model is the K nearest regression model;

[0028] For 5 mm particles, the screening efficiency prediction model is a multi-layer perceptron regression model.

[0029] Based on the further improvement of the above method, the method further includes: after obtaining the multi-layer perceptron regression model corresponding to 5 mm, performing a pruning operation on it.

[0030] Based on the further improvement of the above method, the pruning operation includes: calculating the importance score of each hidden layer in the current multi-layer perceptron regression model, for each hidden layer, calculating the weight threshold of each neuron based on the importance score and the custom weight, subtracting the neurons whose weight is lower than the weight threshold, and fine-tuning the pruned network to ensure that the prediction performance is not affected;

[0031] The threshold for evaluating each neuron based on importance score and custom weights includes:

[0032]

[0033] in,

[0034]

[0035] w ij is the weight of the jth neuron in the ith hidden layer, σ i is the importance score of the i-th hidden layer, σ is a constant, and w s is the standard weight threshold of the i-th hidden layer, w avg is the average weight of all neurons in the i-th hidden layer, 0<σ i <1,0<σ<1.

[0036] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects:

[0037] 1. The present invention provides a method for predicting the screening efficiency of a vibrating screen. The discrete element model of the vibrating screen is divided based on a preset length to obtain multiple sub-screen discrete element models of the same length. For each operating condition simulation of the particle factory space model, after steady state, each time step can generate multiple groups of simulation data corresponding to the operating condition and the preset length of the screen. Compared with the prior art that each time step can only generate a group of simulation data corresponding to the operating condition and the preset length, the present invention is more efficient in constructing a screening efficiency sample data set; in addition, the discrete element models of each sub-screen are arranged continuously, and the output of the discrete element model of the previous sub-screen is used as the input of the discrete element model of the next sub-screen. Compared with the prior art that only obtains sample data for a screen model of a fixed length, the screening sample data obtained by simulating each sub-screen discrete element model in the present invention is closer to the actual vibrating screen screening process.

[0038] 2. The present invention provides a method for predicting the screening efficiency of a vibrating screen. A screening efficiency prediction model for different particle sizes is constructed based on a machine learning model, which can more flexibly meet business needs. In addition, compared with the method of obtaining screening efficiency by using dynamics, probability theory and discrete element method in the prior art, the method proposed by the present invention adopts a machine learning model to obtain screening efficiency. The method has higher prediction accuracy and calculation efficiency, strong real-time performance and can be adapted to various complex working conditions.

[0039] 3. The present invention provides a method for predicting the screening efficiency of a vibrating screen. After obtaining the screening efficiency prediction model, a pruning operation is performed on the target screening efficiency prediction model, which can reduce the amount of calculation and memory requirements, improve model efficiency and generalization ability, and reduce energy consumption; further, the process of obtaining the weight threshold of each neuron based on importance score and custom weight calculation for neuron pruning is a dynamic process, which realizes efficient model compression and can quickly realize a compact model.

[0040] In the present invention, the above-mentioned technical solutions can also be combined with each other to achieve more preferred combination solutions. Other features and advantages of the present invention will be described in the subsequent description, and some advantages can become obvious from the description, or can be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained through the contents particularly pointed out in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings are only used for the purpose of illustrating specific embodiments and are not to be considered as limiting the present invention. In the entire drawings, the same reference symbols represent the same components;

[0042] Figure 1 This is an example diagram of a method for predicting screening efficiency of a vibrating screen in an embodiment of the present invention;

[0043] Figure 2 This is an example diagram of a discrete element model of a vibrating screen in an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0045] A specific embodiment of the present invention discloses a method for predicting the screening efficiency of a vibrating screen, such as Figure 1 As shown, including:

[0046] S1: Construct the discrete element model of the vibrating screen and the spatial model of the pellet plant, including:

[0047] S11: constructing a discrete element model of a vibrating screen, wherein the discrete element model of the vibrating screen comprises a screen and a material baffle, wherein the material baffle is located at the feeding side, and dividing the screen discrete element model based on a preset screen length to obtain a plurality of sub-screen discrete element models with the same screen length.

[0048] For example, SolidWorks can be used to construct a discrete element model of a vibrating screen. Furthermore, taking into account factors such as computational efficiency and resource consumption, the present invention removes unnecessary details and complex structures in the discrete element model of the vibrating screen, and based on business needs, ultimately simplifies the discrete element model of the vibrating screen into a discrete element model of the screen mesh and a discrete element model of the material baffle, such as Figure 2 As shown, the left side is a material baffle and the lower side is a screen. In the present invention, the left side of the screen is defined as the feed end and the right side is the discharge end. It can be understood that the material baffle is used to limit the range of movement of particles, and the upper surface of the screen discrete element model is used to filter materials of different particle sizes.

[0049] After obtaining the sieve discrete element model, the sieve discrete element model can be divided based on the preset sieve length specified by business needs to obtain multiple sub-sieve discrete element models with the same sieve length. The present invention does not limit the preset sieve length, and the value of the preset sieve length can be set according to business needs or actual needs. For example, the preset sieve length can be set to 250 mm.

[0050] S12: constructing a particle factory space model, wherein the particle factory space model is arranged above the screen discrete element model and close to the material baffle, and setting the particle generation method and particle parameters of the particle factory space model based on business requirements.

[0051] Figure 2 The location of the particle factory space model is not shown in the figure, but it is generally placed at the upper left of the screen (i.e. above the feed end) so that the particles it produces can fall into the screen area.

[0052] The present invention does not limit the software for constructing the discrete element model, and discrete element software such as EDEM, YADE, LIGGGHTS, DEMms, SDEM, etc. can be used. For example, EDEM can be used to define the particle factory space, including:

[0053] Particle definition: Users can define the physical properties of particles, such as shape (spherical, blocky, fibrous, etc.), size, mass, density, etc.

[0054] Particle generation: Through the API source code, the particle generation process in the simulation space can be realized, including the position, speed, rotation angle, etc. of particle generation and the particle generation method. Users can write code to generate a single particle, or batch generate particle groups with different properties (for example, generate particle groups with different particle sizes according to the different particle sizes defined in the particle definition step);

[0055] Particle behavior control: The particle factory API provides methods to control the movement and interaction of particles. Users can specify collisions, friction, and bonding between particles, as well as interactions between particles and simulation devices.

[0056] Customized simulation: By modifying and extending the API source files, users can create customized simulation scenarios to achieve specific particle flow patterns, particle loading and unloading processes, particle-device interactions, etc.

[0057] Integration of external data: API supports importing initial conditions and property parameters of particles from external data sources, such as importing particle shapes from CAD models or obtaining physical properties of particles from experimental data;

[0058] Performance optimization: By writing API code, users can optimize simulation performance, such as reducing simulation calculation time through algorithm optimization, or adjusting the particle management strategy during the simulation process;

[0059] Visualization: During the simulation, the API also provides a visualization output of the particle state. Users can programmatically control the output of particle position, velocity and other information for result analysis and display.

[0060] User interaction and automation: The API allows users to interactively control the simulation process, such as starting, pausing, and stopping the simulation, as well as dynamically adjusting particle parameters during the simulation. At the same time, users can automate a series of simulation tasks through the API to improve work efficiency.

[0061] The above contents are common operations of the particle factory space. The purpose of the particle factory space used in the present invention is to generate particles that meet business needs. The present invention does not specifically limit the specific operations of generating particles, such as the particle generation method and the particle size, which is subject to meeting business needs. For example, the particle generation method can be dynamic, and the particle size can be 3mm, 5mm, 7mm, or 10mm.

[0062] In addition, periodic boundaries are adopted on both sides of the area between the screen inlet and outlet ends to limit the activity range of particles generated in the particle factory space.

[0063] S2: constructing a screening efficiency sample data set based on the vibrating screen finite element model and the particle factory space model, wherein the screening efficiency sample data set includes: particle size, feed particle percentage, feed particle total amount and particle screening efficiency.

[0064] The constructing of a screening efficiency sample data set based on the vibrating screen discrete element model and the particle factory spatial model includes:

[0065] A1: Based on the discrete element model of the vibration screen and the spatial model of the particle factory, the vibration screen screening process simulation is implemented, and the simulation results are obtained. The simulation results include the spatial position of each particle corresponding to each time step in the simulation process, the percentage of input particles, and the total amount of input particles. The total amount of input particles in the present invention refers to the total number of particles of different sizes generated by the particle factory spatial model according to business needs.

[0066] The simulation results refer to the percentage of feed particles, the total amount of feed particles, and the spatial position of each particle corresponding to each sub-sieve discrete element model in each time step after the screening process reaches a steady state.

[0067] It is understandable that before the simulation process is performed, simulation parameters need to be set, including the screen surface movement mode, screen surface properties, screen surface movement state, screened material shape and properties, feed amount, and feed speed. Each simulation parameter can be set according to business needs. For example, the screen surface movement mode can be simple harmonic vibration.

[0068] After setting the simulation parameters, iterative simulation can be performed until the sum of the mass of particles passing through the screen per unit time and the particle discharge rate at the discharge end is equal to the particle input rate, that is, the screening process enters a steady state. Continue to perform discrete element iterative solution, simulate the steady state for at least 10 seconds, and save the discrete element simulation data.

[0069] The discrete element simulation data mainly includes the position information (3D coordinates), motion information (force, velocity, acceleration, angular velocity, torque, etc.) and attribute information (mass, density, Young's modulus, sliding friction coefficient, rolling friction coefficient, etc.) of particles and geometric bodies (screens). Taking EDEM as an example, the information of particles and geometric models at each time step is stored in h5 format files, and the naming method is "time step + .h5".

[0070] In the present invention, according to business requirements, the data required by the present invention are the spatial position of each particle corresponding to each time step, the percentage of input particles, and the total amount of input particles.

[0071] The present invention divides the vibration screen discrete element model based on a preset length to obtain multiple sub-screen discrete element models of the same length, each sub-screen discrete element model is arranged continuously, and the output of the previous sub-screen discrete element model is used as the input of the next sub-screen discrete element model.

[0072] A2: Calculating the particle screening efficiency of different particle sizes corresponding to each sub-sieve discrete element model based on the spatial position of each particle, including:

[0073] For the first sub-sieve discrete element model near the feed end, the screening efficiency of each particle size is calculated based on the particle size generated in the particle factory spatial model, the feed particle percentage, the total feed particle amount, and the number of particles of various sizes below the first sub-sieve discrete element model.

[0074] For other sub-sieve discrete element models, the particle size, particle percentage and total particle amount at the discharge end of the previous sub-sieve discrete element model are used as the particle size, feed particle percentage and total feed particle amount of this sub-sieve discrete element model. The screening efficiency of each particle size is calculated based on the particle size, feed particle percentage and total feed particle amount of this sub-sieve discrete element model and the number of particles of various sizes below this sub-sieve discrete element model.

[0075] The screening efficiency for each particle size in each sub-sieve DEM is calculated as follows:

[0076] B1: Count the number Q of particles of this size located below the sieve;

[0077] B2: Calculate the product of the percentage of the input particles of this particle size and the total amount of the input particles as the total amount M of this particle size;

[0078] B3: Screening efficiency S = Q / M.

[0079] A3: The percentage of feed particles of each type of particle, the particle screening efficiency and the total amount of feed particles at each time step corresponding to each sub-sieve discrete element model are taken as a sample in the screening efficiency sample data set corresponding to the sieve length.

[0080] For each working condition simulation of the particle factory space model of the present invention, after steady state, each time step can generate multiple sets of simulation data corresponding to the working condition and the preset length of the screen. For example, if the current screen length is 1000mm and the preset screen length is 250mm, the current screen can be divided into 4 250mm sub-screens. Therefore, for the same working condition parameters, within one time step, the method proposed by the present invention can simultaneously obtain 4 sample data corresponding to the 250mm screen length.

[0081] A4: Determine whether the amount of sample data currently generated meets the preset quantity requirement. If so, stop generating the sample data. If not, adjust the current particle generation method and particle parameters based on business requirements and return to A1.

[0082] The present invention adopts a machine learning model as a screening efficiency prediction model, and the number of training samples is an important indicator affecting the prediction accuracy of the machine learning model. Therefore, it is necessary to set the number of training sample data here so that the prediction effect of the machine learning model meets the requirements.

[0083] If it is determined that the amount of sample data currently generated does not meet the requirements, the particle generation method and particle parameters in the particle factory space model can be adjusted to generate a training sample under a new working condition. For example, the amount of sample data can be set to 5000-10000.

[0084] S3: For each size of particles, multiple machine learning models are trained based on the screening efficiency sample data set, and the machine learning model with the highest screening efficiency prediction accuracy is used as the screening efficiency prediction model for this type of particles.

[0085] It can be understood that when constructing the screening efficiency sample data set, the present invention constructs a sample data set corresponding to each particle size for each particle size, as shown in step A3. Therefore, when obtaining the screening efficiency prediction model for the particle, the sample data set corresponding to the particle size is used to train the machine learning model, so as to obtain the screening efficiency prediction model for predicting the particle size.

[0086] The method of training multiple machine learning models based on the screening efficiency sample data set includes:

[0087] The screening efficiency sample data set is subjected to data preprocessing to obtain a training sample data set; the hyperparameters of each machine learning model are optimized based on a random grid search method, and the optimization results are used as the initial hyperparameters of each machine learning model; and the optimized machine learning models are trained using K-fold cross validation.

[0088] Among them, the machine learning model can be a support vector machine, a random forest, a Gaussian process regression, a BP neural network regression, a gradient enhancement regression, etc., which are not limited in the present invention. In addition, the random grid search method and the K-fold cross validation algorithm used in the present invention are well-known algorithms in the art, and no limitation is made on the specific implementation, but the prediction accuracy of each machine learning model is further improved through the two.

[0089] Furthermore, a multi-objective optimization algorithm (for example, particle swarm algorithm, genetic algorithm, Harris Eagle algorithm, etc.) can also be used to optimize the hyperparameters of each machine learning model, thereby constructing a machine learning prediction model with high prediction accuracy.

[0090] It can be understood that for different discrete element models of the vibrating screen, the corresponding prediction models of the screening efficiency of particles of various particle sizes are also different.

[0091] Exemplarily, for the vibrating screen with a screen length of 250 mm in the present invention, for 3 mm particles, the screening efficiency prediction model is a K-nearest nearest neighbor regression model; for 5 mm particles, the screening efficiency prediction model is a multi-layer perceptron regression model.

[0092] Exemplarily, if the final prediction model of the particle screening efficiency is a neural network model, a pruning algorithm can be used to further optimize the model to reduce the amount of calculation and memory requirements and improve the model efficiency and generalization ability. Specifically, the importance score of each hidden layer in the current neural network model is calculated, and for each hidden layer, the weight threshold of each neuron is calculated based on the importance score and the custom weight, and the neurons with weights lower than the weight threshold are removed, and the pruned network is fine-tuned to ensure that the prediction performance is not affected; the threshold of each neuron is evaluated based on the importance score and the custom weight, including:

[0093]

[0094] in,

[0095]

[0096] w ij is the weight of the jth neuron in the ith hidden layer, σ i is the importance score of the i-th hidden layer, σ is a constant, and w s is the standard weight threshold of the i-th hidden layer, w avg is the average weight of all neurons in the i-th hidden layer, 0<σ i <1,0<σ<1.

[0097] In addition, the method provided in the present invention can also be used to predict the screening efficiency of particles smaller than a fixed value, such as the screening efficiency of particles smaller than 6 mm. In essence, when constructing training samples, the corresponding percentage of input particles smaller than 6 mm, the total amount of input particles and the particle screening efficiency are used as training samples, and then different machine learning models are used to predict the screening efficiency, and the machine learning model with the highest accuracy in predicting screening efficiency is used as the screening efficiency prediction model for particles smaller than 6 mm.

[0098] S4: Obtain the sample to be tested, select the corresponding screening efficiency prediction model based on the particle size, and use the percentage of feed particles and the total amount of feed particles as input parameters to obtain the screening efficiency corresponding to the particle size.

[0099] In practical applications, in order to obtain the screening efficiency of the current material, the types of particles contained in the current material (for example, size) can be identified, and the percentage of particles of each particle size can be calculated. The total amount of material and the percentage of this type of particle are used as input parameters and input into the screening efficiency prediction model of this particle size, thereby obtaining the screening efficiency of this type of particle in the current material.

[0100] Compared with the prior art, the present embodiment provides a method for predicting the screening efficiency of a vibrating screen, which divides the discrete element model of the vibrating screen based on a preset length to obtain multiple sub-sieve discrete element models of the same length. For each operating condition simulation of the particle factory space model, after steady state, each time step can generate multiple sets of simulation data corresponding to the operating condition and the preset length of the screen. Compared with the prior art in which each time step can only generate a set of simulation data corresponding to the operating condition and the preset length, the present invention is more efficient in constructing a screening efficiency sample data set; in addition, the discrete element models of each sub-sieve are arranged continuously, and the output of the discrete element model of the previous sub-sieve is used as the input of the discrete element model of the next sub-sieve. Compared with the prior art in which sample data is only obtained for a sieve model of a fixed length, the screening sample data obtained by simulating each sub-sieve discrete element model in the present invention is closer to the actual vibrating screen screening process. Based on the machine learning model, a screening efficiency prediction model for different particle sizes is constructed, which can meet business needs more flexibly; in addition, compared with the method of obtaining screening efficiency by dynamics, probability theory and discrete element method in the prior art, the method proposed by the present invention adopts a machine learning model to obtain screening efficiency. The method has higher prediction accuracy and calculation efficiency, strong real-time performance and can adapt to various complex working conditions. After obtaining the screening efficiency prediction model, pruning the target screening efficiency prediction model can reduce the amount of calculation and memory requirements, improve model efficiency and generalization ability, and reduce energy consumption; further, the process of obtaining the weight threshold of each neuron based on importance score and custom weight calculation for neuron deletion is a dynamic process, which realizes efficient model compression and can quickly realize a compact model.

[0101] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0102] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for predicting vibrating screen screening efficiency, characterized in that: include: Construct discrete element model of vibrating screen and spatial model of pellet plant; Constructing a screening efficiency sample data set based on the vibration screen discrete element model and the particle factory space model, wherein the screening efficiency sample data set includes: particle size, feed particle percentage, feed particle total amount, and particle screening efficiency; For each size of particles, multiple machine learning models are trained based on the screening efficiency sample data set, and the machine learning model with the highest screening efficiency prediction accuracy is used as the screening efficiency prediction model for this type of particles; Obtain the sample to be tested, select the corresponding screening efficiency prediction model based on the particle size, and use the percentage of feed particles and the total amount of feed particles as input parameters to obtain the screening efficiency corresponding to the particle size.

2. A method for predicting vibrating screen screening efficiency according to claim 1, characterized in that: The construction of the discrete element model of the vibrating screen and the particle plant space model includes: Constructing a discrete element model of a vibrating screen, the discrete element model of the vibrating screen comprising: a screen and a material baffle, the material baffle being located at the feeding side, and dividing the discrete element model of the screen based on a preset screen length to obtain a plurality of sub-screen discrete element models with the same screen length; A particle factory space model is constructed, wherein the particle factory space model is arranged above the screen discrete element model and close to the material baffle, and the particle generation method and particle parameters of the particle factory space model are set based on business requirements.

3. A method for predicting vibrating screen screening efficiency according to claim 2, characterized in that: The constructing of a screening efficiency sample data set based on the vibrating screen discrete element model and the particle factory spatial model includes: A1: Based on the discrete element model of the vibration screen and the spatial model of the particle factory, a vibration screen screening process simulation is implemented, and a simulation result is obtained, wherein the simulation result includes the spatial position of each particle corresponding to each time step in the simulation process, the percentage of input particles, and the total amount of input particles; A2: Calculating the particle screening efficiency of different particle sizes corresponding to each sub-sieve discrete element model based on the spatial position of each particle; A3: The percentage of input particles of each type of particle, particle screening efficiency and total amount of input particles at each time step corresponding to each sub-sieve discrete element model are taken as a sample in the screening efficiency sample data set corresponding to the sieve length; A4: Determine whether the amount of sample data currently generated meets the preset quantity requirement. If so, stop generating the sample data. If not, adjust the current particle generation method and particle parameters based on business requirements and return to A1.

4. A method for predicting vibrating screen screening efficiency according to claim 3, characterized in that: include: The simulation result refers to the percentage of input particles, the total amount of input particles, and the spatial position of each particle corresponding to each sub-sieve discrete element model in each time step after the screening process reaches a steady state; The steady state means that the sum of the particle screening rate of the screen discrete element model and the particle outflow rate at the discharge end per unit time is equal to the particle input rate.

5. A method for predicting vibrating screen screening efficiency according to claim 4, characterized in that: The calculating, based on the spatial position of each particle, the particle screening efficiency of different particle sizes corresponding to each sub-sieve discrete element model comprises: For the first sub-sieve discrete element model near the feed end, the screening efficiency of each particle size is calculated based on the particle size generated in the particle plant spatial model, the feed particle percentage, the total feed particle amount, and the number of particles of various sizes below the first sub-sieve discrete element model; For other sub-sieve discrete element models, the particle size, particle percentage and total particle amount at the discharge end of the previous sub-sieve discrete element model are used as the particle size, feed particle percentage and total feed particle amount of this sub-sieve discrete element model. The screening efficiency of each particle size is calculated based on the particle size, feed particle percentage and total feed particle amount of this sub-sieve discrete element model and the number of particles of various sizes below this sub-sieve discrete element model.

6. A method for predicting vibrating screen screening efficiency according to claim 5, characterized in that: The method of training multiple machine learning models based on the screening efficiency sample data set includes: Performing data preprocessing on the screening efficiency sample data set to obtain a training sample data set; The hyperparameters of each machine learning model are optimized based on the random grid search method, and the optimization results are used as the initial hyperparameters of each machine learning model; K-fold cross validation is used to train the optimized machine learning models.

7. A method for predicting vibrating screen screening efficiency according to claim 6, characterized in that: include: The particle generation method adopts a dynamic method, and the particle parameters include particle size distribution, which are 3mm, 5mm, 7mm, and 10mm respectively.

8. A method for predicting vibrating screen screening efficiency according to claim 7, characterized in that: The method of using the machine learning model with the highest predicted screening efficiency as the screening efficiency prediction model for the particles includes: For 3 mm particles, the screening efficiency prediction model is the K nearest regression model; For 5 mm particles, the screening efficiency prediction model is a multi-layer perceptron regression model.

9. A method for predicting vibrating screen screening efficiency according to claim 8, characterized in that: The method further comprises: After obtaining the multi-layer perceptron regression model corresponding to 5mm, a pruning operation is performed on it.

10. A method for predicting vibrating screen screening efficiency according to claim 9, characterized in that: The pruning operation includes: Calculate the importance score of each hidden layer in the current multi-layer perceptron regression model. For each hidden layer, calculate the weight threshold of each neuron based on the importance score and the custom weight. Remove neurons with weights lower than the weight threshold, and fine-tune the pruned network to ensure that the prediction performance is not affected. The threshold for evaluating each neuron based on importance score and custom weights includes: in, w ij is the weight of the jth neuron in the ith hidden layer, σ i is the importance score of the i-th hidden layer, σ is a constant, and w s is the standard weight threshold of the i-th hidden layer, w avg is the average weight of all neurons in the i-th hidden layer, 0<σ i <1,0<σ<1.

Images