Screening efficiency prediction method

Abstract

The invention relates to a screening efficiency prediction method, belongs to the technical field of screening efficiency prediction, and solves the problem that accurate and effective screening efficiency cannot be obtained in time in the prior art. Acquiring historical screening sample data of the vibrating screen, processing the vibration signal to obtain a second feature component, replacing the vibration signal in the historical sample data to form a sample data set, and performing data expansion on the sample data set to obtain a training sample data set; constructing a plurality of first training sample subsets with different parameter combinations based on the training sample data set, and establishing screening efficiency prediction models corresponding to the first training sample subsets; and training a corresponding screening efficiency prediction model based on each first training sample subset, taking the screening efficiency prediction model with the highest prediction precision as a target screening efficiency prediction model, obtaining an input parameter combination corresponding to the current to-be-tested vibrating screen, and inputting the input parameter combination into the target screening efficiency prediction model to obtain the screening efficiency of the to-be-tested vibrating screen. The screening efficiency corresponding to the current control parameter is rapidly and efficiently obtained.

Description

Technical Field

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

[0002] Vibrating screen is a key equipment widely used in mining, chemical, pharmaceutical, food and other industries. It uses the vibration force generated by a vibration motor or exciter to screen the material on the screen surface, thereby achieving particle grading and screening.

[0003] The level of screening efficiency directly affects the efficiency of the production process and the quality of the product. For example, in the mining industry, efficient screening can ensure the uniform particle size of the ore, which is helpful for subsequent crushing and enrichment operations, and improves smelting efficiency and metal recovery rate; in the pharmaceutical industry, the screening process determines the particle size distribution of the pharmaceutical powder, thereby affecting the dissolution rate and bioavailability of the drug. By continuously optimizing the design and operating parameters of the vibrating screen, the screening efficiency can be further improved to meet the growing production needs and promote the continuous progress and development of related industries.

[0004] However, the design of operating parameters in the prior art is mainly divided into: (1) design by technicians based on experience, but this method is difficult to obtain screening efficiency in a timely manner by manually setting parameters in a complex production environment, such as processing a variety of materials with different properties or under harsh working conditions; (2) simulation prediction using a finite element model, but the model for predicting screening efficiency usually needs to consider multiple variables, such as material properties, vibration parameters, screen structure, etc., which makes the model complex. For example, discrete element method (DEM) simulation requires a large amount of computing resources to process the interactions and motion trajectories between particles. This highly complex model may be limited by computing power in practical applications, especially when processing large-scale data sets or real-time predictions, and thus cannot obtain effective screening efficiency. Summary of the invention

[0005] In view of the above analysis, an embodiment of the present invention aims to provide a screening efficiency prediction method to solve the problem in the prior art that accurate and effective screening efficiency cannot be obtained in a timely manner.

[0006] An embodiment of the present invention provides a screening efficiency prediction method, comprising:

[0007] S1: Obtain historical screening sample data of the vibrating screen. Each piece of historical screening sample data includes the vibration signal, vibration frequency, exciting force, screen hole size, screen surface inclination, as well as material particle size, material moisture, feed speed, and screening efficiency of the vibrating screen;

[0008] S2: using an empirical mode decomposition (EMD) algorithm and a common coupling point (PCC) algorithm to process the vibration signal to obtain an intrinsic mode function (IMF) that meets the requirements, and reconstructing all intrinsic mode functions (IMF) that meet the requirements to obtain a first eigencomponent;

[0009] S3: performing Fourier transform on the first characteristic component to obtain a second characteristic component, replacing the vibration signal in the historical sample data with the second characteristic component to form a sample data set, and performing data expansion on the sample data set to obtain a training sample data set;

[0010] S4: constructing a plurality of first training sample subsets with different parameter combinations based on the training sample data set, and establishing a screening efficiency prediction model corresponding to the first training sample subsets;

[0011] S5: training the corresponding screening efficiency prediction models based on the first training sample subsets, taking the screening efficiency prediction model with the highest prediction accuracy as the target screening efficiency prediction model, and taking the parameter combination of the first training sample subset corresponding to the target screening efficiency prediction model as the input parameter combination;

[0012] S6: Obtain an input parameter combination corresponding to the current vibrating screen to be tested, and input it into a target screening efficiency prediction model to obtain the screening efficiency of the vibrating screen to be tested.

[0013] Based on the further improvement of the above method, the data expansion of the sample data set to obtain the training sample data set includes:

[0014] A generative adversarial network is used to expand the sample data set; wherein the loss function of the generative adversarial network is:

[0015] L c =λ1δ1+λ2δ2+λ3δ3,

[0016] Among them, λ1, λ2, and λ3 are the weight coefficients of adversarial loss δ1, content loss δ2, and diversity loss δ3, respectively. Diversity loss δ3 is:

[0017]

[0018] Among them, z i 、z j is a random noise vector, G c () is the generator function, and N is the number of samples in the sample data set.

[0019] Based on the further improvement of the above method, the data expansion of the sample data set to obtain the training sample data set includes: screening the sample data set with a screening efficiency greater than a preset screening threshold to form a first sample set;

[0020] For any sample in the first sample set, perform the following operations: for each sample, calculate the distance between the sample and other samples, and obtain k nearest neighbor samples of the sample based on the calculated distance; generate a fitting sample corresponding to the sample based on the sample and the corresponding k nearest neighbor samples; add the generated fitting sample to the sample data set to obtain a training sample data set.

[0021] Based on the further improvement of the above method, generating a fitting sample corresponding to the sample according to the sample and the corresponding k nearest neighbor samples includes:

[0022]

[0023] Among them, SampleNew i is the newly generated fitting sample corresponding to the i-th sample, Sample i is the i-th sample, j is the jth neighboring sample corresponding to the ith sample, δ is a random vector, j = 1, 2, 3, ..., k.

[0024] Based on the further improvement of the above method, the empirical mode decomposition (EMD) algorithm and the common coupling point (PCC) algorithm are used to process the vibration signal to obtain an intrinsic mode function (IMF) that meets the requirements, and all intrinsic mode functions (IMF) that meet the requirements are reconstructed to obtain a first characteristic component, including: executing the empirical mode decomposition (EMD) algorithm on the acquired vibration signal to generate IMF components; calculating the Pearson correlation coefficient (PCC) between each IMF component and its corresponding vibration signal, screening the components whose Pearson correlation coefficient (PCC) in the IMF components is greater than a first preset threshold to form an IMF component set; sorting the components in the IMF component set in descending order according to the Pearson correlation coefficient (PCC), assigning a larger weight to the IMF component with a higher order, and assigning a smaller weight to the component with a lower order, the sum of the weights corresponding to the components is 1, and taking the weighted sum of the components and their corresponding weights as the first characteristic component.

[0025] Based on further improvement of the above method, the sample data of each first training sample subset contains different numbers or types of parameters, and the parameters of the sample data in each first training sample subset include: the second characteristic component, material particle size, and material moisture.

[0026] Based on the further improvement of the above method, the screening efficiency prediction model is a neural network model, and its loss function is:

[0027]

[0028] in, is the root mean square error, Smoothing loss, α and β are constants obtained based on trial and error; wherein the smoothing loss is obtained as follows:

[0029]

[0030] Among them, y i is the screening efficiency output of the i-th iteration, y i-1 is the screening efficiency output of the i-1th iteration, y i-2 is the screening efficiency output of the i-2th iteration, μ is the smoothing control coefficient, 0<μ<1, i is the number of iterations, i=3,4,5,…,M.

[0031] Based on further improvement of the above method, the method further includes: after obtaining the target screening efficiency prediction model, performing a pruning operation on the target screening efficiency prediction model.

[0032] Based on the further improvement of the above method, the pruning operation includes: calculating the importance score of each hidden layer in the current target screening efficiency prediction model, for each hidden layer, calculating the weight threshold of each neuron based on the importance score and the custom weight, subtracting the neurons with weights lower than the weight threshold, and fine-tuning the pruned network to ensure that the prediction performance is not affected.

[0033] Based on the further improvement of the above method, the threshold of each neuron is evaluated based on the importance score and the custom weight, including:

[0034]

[0035] in,

[0036]

[0037] 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.

[0038] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects:

[0039] 1. The present invention provides a screening efficiency prediction method, which is implemented based on a machine learning model. It can quickly and efficiently obtain the screening efficiency corresponding to the input parameter combination according to the existing input parameter combination, so that technicians can adjust the input parameters in time according to the screening efficiency to obtain satisfactory screening efficiency.

[0040] 2. The present invention provides a screening efficiency prediction method, which adopts two data sample expansion methods, including generative adversarial network data expansion and K-nearest neighbor data expansion, to improve the generalization ability of the screening efficiency prediction model, reduce overfitting, enhance robustness, and simplify the data acquisition cost.

[0041] 3. The present invention provides a screening efficiency prediction method. After obtaining the target screening efficiency prediction model, the target screening efficiency prediction model is pruned, 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 the importance score and custom weight calculation for pruning neurons is a dynamic process, which realizes efficient model compression and can quickly realize a compact model.

[0042] 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

[0043] 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;

[0044] Figure 1 The figure is a flowchart of a screening efficiency prediction method in an embodiment of the present invention. DETAILED DESCRIPTION

[0045] 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.

[0046] A specific embodiment of the present invention discloses a screening efficiency prediction method, such as Figure 1 As shown, including:

[0047] S1: Obtain historical screening sample data of the vibrating screen. Each piece of historical screening sample data includes the vibration signal, vibration frequency, exciting force, screen hole size, screen surface inclination, as well as material particle size, material moisture, feed speed, and screening efficiency of the vibrating screen.

[0048] In actual application scenarios, there are many parameters that affect the screening efficiency of the vibrating screen, such as the type of vibrating screen, the model of the vibration motor, etc., and some parameters are difficult to obtain, and some parameters have little effect on the screening efficiency. Therefore, after statistically sorting out the parameters that affect the screening efficiency, the present invention selects one of the parameter screening methods such as principal component analysis PCA, factor analysis, feature selection algorithm or decision tree, and finally uses the vibration signal, vibration frequency, exciting force, sieve hole size, screen surface inclination, material particle size, material humidity, and feed speed of the vibrating screen as the parameters that affect the screening efficiency.

[0049] The vibration signal obtained in the present invention is collected based on the sensor installed on the vibrating screen, and the sensor can be installed in the screen box, screen, bracket and other positions of the vibrating screen. The present invention does not limit the number and collection position of the sensors for collecting vibration signals, as long as they meet the actual needs. It can be understood that the vibration frequency and exciting force are directly determined by the type of vibrating screen. Parameters such as screen hole size, screen surface inclination, material particle size, material humidity, feed speed, feed amount, opening rate, bulk density, etc. can be determined according to the actual scenario.

[0050] In addition, the screening efficiency of the vibrating screen is closely related to the vibration signal. Therefore, the more vibration signals collected, the more accurate the prediction result of the final screening efficiency.

[0051] Preferably, the present invention collects three vibration signals, including one screen box, one screen mesh, and one bracket.

[0052] S2: using an empirical mode decomposition (EMD) algorithm and a common coupling point (PCC) algorithm to process the vibration signal to obtain an intrinsic mode function (IMF) that meets the requirements, and reconstructing all intrinsic mode functions (IMF) that meet the requirements to obtain a first characteristic component.

[0053] Specifically, an empirical mode decomposition (EMD) algorithm is performed on the acquired vibration signal to generate IMF components; the Pearson correlation coefficient (PCC) between each IMF component and its corresponding vibration signal is calculated, and components whose Pearson correlation coefficient (PCC) is greater than a first preset threshold in the IMF components are screened to form an IMF component set; each component in the IMF component set is sorted in descending order according to the Pearson correlation coefficient (PCC), and a larger weight is assigned to the front-ranked IMF component, and a smaller weight is assigned to the rear-ranked component, the sum of the weights corresponding to each component is 1, and the weighted sum of each component and its corresponding weight is taken as the first characteristic component.

[0054] For example, the first preset threshold value may be 0.7, and the first characteristic component IMf total for:

[0055] IMF total =W1*IMf1+W2*IMf2+W3*IMf3+.....+W N *LMf N ,

[0056] in,

[0057] W1+W2+W3+...+W N =1,

[0058] IMf1>IMf2>IMf3>...>IMf N ,

[0059] W1>W2>W3>…>W N ,

[0060] Among them, W N corresponds to IMf N The weight of IMf N is the component in the IMF component set, and N is the number of components in the IMF component set.

[0061] S3: Perform Fourier transform on the first characteristic component to obtain a second characteristic component, replace the vibration signal in the historical sample data with the second characteristic component to form a sample data set, and perform data expansion on the sample data set to obtain a training sample data set.

[0062] Among them, performing Fourier transform on the first characteristic component to obtain the second characteristic component includes: performing Fourier transform on the first characteristic component to obtain a frequency domain signal, extracting the mean, peak value, and skewness of the frequency domain signal, and using the mean, peak value, and skewness as the second characteristic component.

[0063] Optionally, preprocessing operations such as data cleaning, data conversion, data dimension reduction, and data reconstruction are performed on the historical screening sample data to improve data quality and enhance the accuracy of screening efficiency prediction. It should be noted that the data preprocessing operation can be determined according to business needs, and the present invention is not limited here, and is subject to the ability to achieve business needs. Preferably, the material particle size and material humidity can be normalized to obtain a first material particle size and a first material humidity. At this point, the vibration signal, material particle size, and material humidity in the historical sample data can be replaced with a second characteristic component, a first material particle size, and a first material humidity. By normalizing part of the data, the accuracy of the screening efficiency prediction model can be further improved.

[0064] In addition, the present invention proposes two methods for data sample expansion, including:

[0065] Method 1: Generative adversarial network is used for data expansion. The loss function of the generative adversarial network is:

[0066] L c =λ1δ1+λ2δ2+λ3δ3,

[0067] Among them, λ1, λ2, λ3 are the weight coefficients of adversarial loss δ1, content loss δ2 and diversity loss δ3 respectively. Adversarial loss is used to improve the quality of generated samples, and the calculation method is expressed as:

[0068] δ1=E x～pdata [logD c (x)]+E z～pz [log(1-D c (G c (z)))],

[0069] Among them, pdata is the distribution of real data, z is random noise, pz is the distribution of random noise, D c () is the discriminator function, G c () is the generator function, E represents the expectation, and x represents the real sample. The calculation method of content loss is expressed as:

[0070]

[0071] Among them, ||||2 is the L2 norm, Nc is the number of samples, x i is the i-th true sample, z i is the i-th random noise vector. The diversity loss is:

[0072]

[0073] Among them, z i 、z j is a random noise vector, G c () is the generator function, and N is the number of samples in the sample data set.

[0074] Method 2: Sample data with a screening efficiency greater than a preset screening threshold in the sample data set are screened to form a first sample set. For any sample in the first sample set, the following operations are performed:

[0075] A1: For each sample, calculate the distance between the sample and other samples, and obtain the k nearest neighbor samples of the sample based on the calculated distance.

[0076] A2: Generate a fitting sample corresponding to the sample based on the sample and its corresponding k nearest neighbor samples, including:

[0077]

[0078] Among them, SampleNew i is the newly generated fitting sample corresponding to the i-th sample, Sample i is the i-th sample, Sample j is the jth neighboring sample corresponding to the ith sample, δ is a random vector, j = 1, 2, 3, ..., k.

[0079] A3: Add the generated fitting samples to the sample data set to obtain a training sample data set.

[0080] The present invention does not limit which method is used to expand the sample. Only one method may be used, or two methods may be used at the same time, depending on the actual number of samples that can meet the needs of the user.

[0081] After the samples are expanded by method one or method two, the expanded samples are merged with the sample data set to obtain the training sample data set.

[0082] S4: constructing a plurality of first training sample subsets with different parameter combinations based on the training sample data set, and establishing a screening efficiency prediction model corresponding to the first training sample subsets.

[0083] The number or type of parameters contained in the sample data of each first training sample subset is different, and the parameters of the sample data in each first training sample subset include: the second characteristic component, the material particle size, and the material moisture. Establishing a screening efficiency prediction model corresponding to the first training sample subset means that the input parameters of the screening efficiency prediction model correspond to the number of parameters in the first training sample subset.

[0084] S5: Based on the screening efficiency prediction models corresponding to the first training sample subsets, the screening efficiency prediction model with the highest prediction accuracy is used as the target screening efficiency prediction model, and the parameter combination of the first training sample subset corresponding to the target screening efficiency prediction model is used as the input parameter combination.

[0085] The screening efficiency in the first training sample subset is used as the true value, and the remaining parameters are used as input data to train the screening efficiency prediction model. When the change value of the loss function value of the screening efficiency prediction model in multiple consecutive iterations is less than the preset threshold or reaches the upper limit of the number of iterations, the training of the screening efficiency prediction model is completed. Specifically, after obtaining the first training sample subset, it is divided into a training set, a test set, and a validation set according to a certain ratio, and then the screening efficiency prediction model is trained. Preferably, it is divided into a training set, a test set, and a validation set in a ratio of 8:1:1; the preset threshold is 0.001, and the upper limit of the number of iterations is 1000.

[0086] The loss function of the screening efficiency prediction model is:

[0087]

[0088] in, is the root mean square error, Smoothing loss, α and β are constants obtained based on trial and error; where the root mean square error is:

[0089]

[0090] The smoothing loss is obtained as follows:

[0091]

[0092] Among them, y i is the screening efficiency output at the i-th iteration, is the screening efficiency recorded in the sample data set, Y i-1 is the screening efficiency output of the i-1th iteration, y i-2 is the screening efficiency output of the i-2th iteration, μ is the smoothing control coefficient, 0<μ<1, i is the number of iterations, i=3,4,5,…,M.

[0093] After obtaining various screening efficiency prediction models, the screening efficiency prediction model with the highest prediction accuracy is selected as the target screening efficiency prediction model.

[0094] The screening efficiency prediction model can be one of CNN, DNN or RNN models. The present invention does not limit the specific form of the screening efficiency model, such as the common BP neural network, SVM model, etc., and the ability to achieve screening efficiency prediction is used as the screening criterion.

[0095] After obtaining the target screening efficiency prediction model, in order to further reduce the complexity of the model, improve the generalization ability of the model, and reduce the risk of overfitting, the target screening efficiency prediction model can be pruned. The pruning operation includes:

[0096] The importance score of each hidden layer in the current target screening efficiency prediction model is calculated. For each hidden layer, the weight threshold of each neuron is calculated based on the importance score and the custom weight. 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.

[0097] The importance score is calculated based on the importance of the current hidden layer. The more important the layer is, the higher the importance score is, but the sum of all importance scores must be 1. The importance score can be calculated using methods such as the Garson method, gradient-based sensitivity analysis, deep Taylor decomposition (DTD), random permutation feature importance (PFI), SmoothGrad method, singular value decomposition (SVD), etc. The present invention does not limit this. After calculating the importance of each hidden layer, the weight is assigned based on the importance of the layer and the number of hidden layers in the target screening efficiency prediction model.

[0098] For example, the number of hidden layers in the current target screening efficiency prediction model is 5, and the importance is 34251, then the importance score of the third layer is 0.4, the importance score of the fourth layer is 0.2, the importance score of the second layer is 0.2, the importance score of the fifth layer is 0.1, and the importance score of the first layer is 0.1. According to the importance score, the first layer is pruned first, and then the fifth layer is pruned, and the layers are pruned in turn according to the importance score until the preset pruning rate is met. The pruning rate is set according to actual needs.

[0099] The threshold for evaluating each neuron based on importance score and custom weights includes:

[0100]

[0101] in,

[0102]

[0103] 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.

[0104] The weight of the current neuron is set based on the threshold of each neuron, the neurons with a weight value of 0 are deleted, and the pruned network is fine-tuned to ensure that the prediction performance is not affected. The pruned model is used as the final target screening efficiency prediction model.

[0105] Compared with the prior art, the present embodiment provides a screening efficiency prediction method, which is implemented based on a machine learning model and can quickly and efficiently obtain the screening efficiency corresponding to the input parameter combination according to the existing input parameter combination, so that technical personnel can adjust the input parameters in time according to the screening efficiency to obtain satisfactory screening efficiency; two data sample expansion methods are adopted, including generative adversarial network data expansion and K-nearest neighbor data expansion, which improves the generalization ability of the screening efficiency prediction model, reduces overfitting, enhances robustness, and simplifies data acquisition costs; after obtaining the target screening efficiency prediction model, the target screening efficiency prediction model is pruned, 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 the importance score and the custom weight calculation for pruning neurons is a dynamic process, which realizes efficient model compression and can quickly realize a compact model.

[0106] 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.

[0107] 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 screening efficiency prediction method, characterized in that: include: S1: Obtain historical screening sample data of the vibrating screen. Each piece of historical screening sample data includes the vibration signal, vibration frequency, exciting force, screen hole size, screen surface inclination, as well as material particle size, material moisture, feed speed, and screening efficiency of the vibrating screen; S2: using an empirical mode decomposition (EMD) algorithm and a common coupling point (PCC) algorithm to process the vibration signal to obtain an intrinsic mode function (IMF) that meets the requirements, and reconstructing all intrinsic mode functions (IMF) that meet the requirements to obtain a first eigencomponent; S3: performing Fourier transform on the first characteristic component to obtain a second characteristic component, replacing the vibration signal in the historical sample data with the second characteristic component to form a sample data set, and performing data expansion on the sample data set to obtain a training sample data set; S4: constructing a plurality of first training sample subsets with different parameter combinations based on the training sample data set, and establishing a screening efficiency prediction model corresponding to the first training sample subsets; S5: training the corresponding screening efficiency prediction models based on the first training sample subsets, taking the screening efficiency prediction model with the highest prediction accuracy as the target screening efficiency prediction model, and taking the parameter combination of the first training sample subset corresponding to the target screening efficiency prediction model as the input parameter combination; S6: Obtain an input parameter combination corresponding to the current vibrating screen to be tested, and input it into a target screening efficiency prediction model to obtain the screening efficiency of the vibrating screen to be tested.

2. A screening efficiency prediction method according to claim 1, characterized in that: The step of expanding the sample data set to obtain a training sample data set includes: A generative adversarial network is used to expand the sample data set; wherein the loss function of the generative adversarial network is: L c =λ1δ1+λ2δ2+λ3δ3, Among them, λ1, λ2, and λ3 are the weight coefficients of adversarial loss δ1, content loss δ2, and diversity loss δ3, respectively. Diversity loss δ3 is: Among them, z i 、z j is a random noise vector, G c () is the generator function, and N is the number of samples in the sample data set.

3. A screening efficiency prediction method according to claim 1, characterized in that: The step of expanding the sample data set to obtain a training sample data set includes: The sample data whose screening efficiency is greater than a preset screening threshold value in the screening sample data set is formed into a first sample set; For any sample in the first sample set, perform the following operations: For each sample, the distance between the sample and other samples is calculated, and the k nearest neighbor samples of the sample are obtained according to the calculated distance; a fitting sample corresponding to the sample is generated according to the sample and the corresponding k nearest neighbor samples; the generated fitting sample is added to the sample data set to obtain a training sample data set.

4. The screening efficiency optimization method for a vibrating screen according to claim 3, characterized in that: The step of generating a fitting sample corresponding to the sample according to the sample and the corresponding k nearest neighbor samples includes: Among them, SampleNew i is the newly generated fitting sample corresponding to the i-th sample, Sample i is the i-th sample, Sample j is the jth neighboring sample corresponding to the ith sample, δ is a random vector, j = 1, 2, 3, ..., k.

5. A screening efficiency optimization method for a vibrating screen according to claim 2 or 4, characterized in that: The method of using the empirical mode decomposition (EMD) algorithm and the common coupling point (PCC) algorithm to process the vibration signal to obtain an intrinsic mode function (IMF) that meets the requirements, and reconstructing all intrinsic mode functions (IMF) that meet the requirements to obtain a first characteristic component includes: Performing the EMD algorithm on the acquired vibration signal to generate IMF components; Calculating the Pearson correlation coefficient PCC between each IMF component and its corresponding vibration signal, and screening the IMF components whose Pearson correlation coefficient PCC is greater than a first preset threshold to form an IMF component set; The components in the IMF component set are sorted in descending order according to the Pearson correlation coefficient PCC, and the IMF components with higher rankings are assigned larger weights, and the components with lower rankings are assigned smaller weights. The sum of the weights corresponding to each component is 1, and the weighted sum of each component and its corresponding weight is taken as the first characteristic component.

6. A screening efficiency prediction method according to claim 5, characterized in that: The sample data of each first training sample subset contains different numbers or types of parameters, and the parameters of the sample data in each first training sample subset include: the second characteristic component, material particle size, and material moisture.

7. A screening efficiency prediction method according to claim 6, characterized in that: The screening efficiency prediction model is a neural network model, and its loss function is: in, is the root mean square error, Smoothing loss, α and β are constants obtained based on trial and error; wherein the smoothing loss is obtained as follows: Among them, y i is the screening efficiency output of the i-th iteration, y i-1 is the screening efficiency output of the i-1th iteration, y i-2 is the screening efficiency output of the i-2th iteration, μ is the smoothing control coefficient, 0<μ<1, i is the number of iterations, i=3,4,5,…,M.

8. A screening efficiency prediction method according to claim 7, characterized in that: The method further includes: after obtaining the target screening efficiency prediction model, performing a pruning operation on the target screening efficiency prediction model.

9. A screening efficiency prediction method according to claim 8, characterized in that: The pruning operation includes: The importance score of each hidden layer in the current target screening efficiency prediction model is calculated. For each hidden layer, the weight threshold of each neuron is calculated based on the importance score and the custom weight. 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.

10. A screening efficiency prediction method according to claim 9, characterized in that: 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