A control method and system for a powder wet vibrating sieve separation device based on image analysis

Abstract

The application provides a control method and system of a powder wet vibration screening device based on image analysis, and relates to the technical field of powder wet vibration screening device control. The method comprises the following steps: obtaining an original dry matter scanning image of powder to be screened, characteristic data of premixed slurry of the powder to be screened, and structure configuration information of the powder wet vibration screening device; generating a plurality of first wet vibration screening schemes based on screening operation state information under each initial test wet vibration screening scheme; determining an optimal wet vibration screening scheme based on operation video of the powder wet vibration screening device under each first wet vibration screening scheme for the premixed slurry of the powder to be screened; and controlling the powder wet vibration screening device to perform full-amount wet vibration screening on the premixed slurry of the powder to be screened based on the optimal wet vibration screening scheme. The method can efficiently and accurately determine the optimal vibration screening scheme of the powder wet vibration screening device.

Description

Technical Field

[0001] This invention relates to the field of control technology for wet powder vibrating sieving devices, and specifically to a control method and system for a wet powder vibrating sieving device based on image analysis. Background Technology

[0002] Powder wet vibrating sieving is a crucial step in the fine classification and processing of industrial materials, and the scientific control of its sieving parameters directly affects the powder's screen penetration efficiency and the quality of the finished product. Traditional powder wet vibrating sieving devices primarily rely on manual experience to set the vibration frequency and spray flow rate control parameters, or simply assess slurry characteristics and perform static adjustments through basic physicochemical sampling tests of the powder. However, this method has significant limitations. Manual trial and error and static adjustments have limited perceptual range and suffer from severe lag in dynamic feedback, making it difficult to fully reflect the transient changes in the rheological properties of powder slurries under complex excitation fields. In particular, it is severely inadequate in identifying and predicting the aggregation and accumulation of local microenvironments on the screen surface and the instantaneous clogging trend of the screen openings. This limitation not only leads to insufficient basis for selecting excitation and spray parameters, easily causing large-area screen clogging or material leakage at the discharge end, but also affects the overall stability of subsequent large-scale continuous sieving operations. Insufficient support from static physicochemical data further weakens the effectiveness of dynamic screening decisions, reduces the accuracy of anti-clogging and anti-material leakage control, and exacerbates the problems of excessive consumption of spray liquid resources and low overall screening efficiency. In addition, traditional methods rely on a large number of on-site personnel and repeated shutdowns for blind adjustments, resulting in low operating efficiency and an inability to meet the demands of high-throughput, intelligent, and precise modern wet powder screening industrial production.

[0003] Therefore, how to efficiently and accurately determine the optimal excitation screening scheme for a wet powder vibrating screen is an urgent problem to be solved. Summary of the Invention

[0004] The main technical problem solved by this invention is how to efficiently and accurately determine the optimal excitation and screening scheme for a wet vibrating sieve device for powder.

[0005] According to a first aspect, the present invention provides a control method for a wet vibrating sieving device for powder based on image analysis, comprising: acquiring a scanned image of the original dried material of the powder to be sieved, characteristic data of the premixed slurry of the powder to be sieved, and structural configuration information of the wet vibrating sieving device; determining multiple initial test wet excitation sieving schemes based on the scanned image of the original dried material of the powder to be sieved, the characteristic data of the premixed slurry of the powder to be sieved, and the structural configuration information of the wet vibrating sieving device, wherein the wet excitation sieving scheme includes vibration frequency and spray flow rate; acquiring a trial operation video of the wet vibrating sieving device under each initial test wet excitation sieving scheme for the premixed slurry of the powder to be sieved; and determining multiple initial test wet excitation sieving schemes based on the scanned image of the original dried material of the powder to be sieved, the characteristic data of the premixed slurry of the powder to be sieved, and the structural configuration information of the wet vibrating sieving device. The trial operation video of the powder wet vibrating sieving device under the initial test wet vibrating sieving scheme is used to determine the sieving operation status information under each initial test wet vibrating sieving scheme; based on the sieving operation status information under each initial test wet vibrating sieving scheme, multiple first wet vibrating sieving schemes are generated; the operation video of the powder wet vibrating sieving device under each first wet vibrating sieving scheme of the powder premixed slurry to be sieved is obtained; based on the operation video of the powder wet vibrating sieving device under each first wet vibrating sieving scheme of the powder premixed slurry to be sieved, the preferred wet vibrating sieving scheme is determined; based on the preferred wet vibrating sieving scheme, the powder wet vibrating sieving device is controlled to perform full wet vibrating sieving control of the powder premixed slurry to be sieved.

[0006] In one possible implementation, generating multiple first wet vibration screening schemes based on the screening operation status information under each initial test wet vibration screening scheme includes: clustering the screening operation status information under each initial test wet vibration screening scheme to obtain K mesh-through condition characterization clusters; determining the critical minimum vibration frequency for anti-clogging, the critical minimum spray flow rate for anti-clogging, the critical maximum vibration frequency for anti-material leakage, and the critical maximum spray flow rate for anti-material leakage based on the K mesh-through condition characterization clusters; and generating multiple first wet vibration screening schemes based on the screening operation status information under each initial test wet vibration screening scheme, the critical minimum vibration frequency for anti-clogging, the critical minimum spray flow rate for anti-clogging, the critical maximum vibration frequency for anti-material leakage, and the critical maximum spray flow rate for anti-material leakage.

[0007] In one possible implementation, determining the preferred wet vibration screening scheme based on the operating video of the wet vibration screening device for the powder premixed slurry to be screened under each first wet vibration screening scheme includes: constructing a wet vibration screening map, wherein the wet vibration screening map includes multiple nodes of the first wet vibration screening scheme and multiple edges between the nodes, and the node characteristics of each first wet vibration screening scheme node include the operating video of the wet vibration screening device for the powder premixed slurry to be screened under each first wet vibration screening scheme, and the edges are the vibration frequency difference and spray flow between the nodes of the first wet vibration screening scheme. The difference in quantity is calculated; the wet vibration screening spectrum is processed based on a graph neural network to determine multiple second wet vibration screening schemes; based on the trial operation video of the powder wet vibration screening device under each initial test wet vibration screening scheme, the operation video of the powder wet vibration screening device under each first wet vibration screening scheme, and the multiple second wet vibration screening schemes, an operation simulation video of the powder wet vibration screening device under each second wet vibration screening scheme is generated; based on the operation simulation video of the powder wet vibration screening device under each second wet vibration screening scheme, the preferred wet vibration screening scheme is determined.

[0008] In one possible implementation, the step of clustering the screening operation status information under each initial test wet vibration screening scheme to obtain K mesh-through condition characterization clusters includes: using the K-means clustering algorithm to cluster the screening operation status information under each initial test wet vibration screening scheme to obtain K mesh-through condition characterization clusters.

[0009] According to a second aspect, the present invention provides a control system for a wet vibrating sieving device for powder based on image analysis, comprising: an acquisition module for acquiring a scanned image of the original dried material of the powder to be sieved, characteristic data of the premixed slurry of the powder to be sieved, and structural configuration information of the wet vibrating sieving device; an initial scheme determination module for determining multiple initial test wet excitation screening schemes based on the scanned image of the original dried material of the powder to be sieved, the characteristic data of the premixed slurry of the powder to be sieved, and the structural configuration information of the wet vibrating sieving device, wherein the wet excitation screening scheme includes vibration frequency and spray flow rate; a vibration screening trial operation module for acquiring a trial operation video of the wet vibrating sieving device of the powder premixed slurry under each initial test wet excitation screening scheme; and a status information determination module for acquiring a trial operation video of the wet vibrating sieving device of the powder premixed slurry under each initial test wet excitation screening scheme; and a status information determination module for determining the status information based on the premixed slurry of the powder to be sieved under each initial test wet excitation screening scheme. The system includes: a trial operation video of the powder wet vibrating sieving device under each initial wet vibrating sieving scheme to determine the sieving operation status information under each initial wet vibrating sieving scheme; a first scheme generation module to generate multiple first wet vibrating sieving schemes based on the sieving operation status information under each initial wet vibrating sieving scheme; an operation video acquisition module to acquire operation videos of the powder premixed slurry to be sieved under each first wet vibrating sieving scheme; a preferred scheme determination module to determine a preferred wet vibrating sieving scheme based on the operation videos of the powder premixed slurry to be sieved under each first wet vibrating sieving scheme; and a sieving control module to control the powder wet vibrating sieving device to perform full wet vibrating sieving of the powder premixed slurry to be sieved based on the preferred wet vibrating sieving scheme.

[0010] In one possible implementation, the first scheme generation module is further configured to: cluster K mesh-through condition characterization clusters based on the screening operation status information under each initial test wet vibration screening scheme; determine the critical minimum vibration frequency for anti-clogging, the critical minimum spray flow rate for anti-clogging, the critical maximum vibration frequency for anti-material leakage, and the critical maximum spray flow rate for anti-material leakage based on the K mesh-through condition characterization clusters; and generate multiple first wet vibration screening schemes based on the screening operation status information under each initial test wet vibration screening scheme, the critical minimum vibration frequency for anti-clogging, the critical minimum spray flow rate for anti-clogging, the critical maximum vibration frequency for anti-material leakage, and the critical maximum spray flow rate for anti-material leakage.

[0011] In one possible implementation, the preferred solution determination module is further configured to: construct a wet-excitation screening map, the wet-excitation screening map including multiple first wet-excitation screening scheme nodes and multiple edges between the nodes, the node features of each first wet-excitation screening scheme node including the operation video of the powder premixed slurry to be screened under each first wet-excitation screening scheme, and the edges being the vibration frequency difference and spray flow rate difference between the first wet-excitation screening scheme nodes; process the wet-excitation screening map based on a graph neural network to determine multiple second wet-excitation screening schemes; generate an operation simulation video of the powder wet-excitation screening device under each second wet-excitation screening scheme based on the trial operation video of the powder wet-excitation screening device under each initial test wet-excitation screening scheme, the operation video of the powder wet-excitation screening device under each first wet-excitation screening scheme, and the multiple second wet-excitation screening schemes; and determine a preferred wet-excitation screening scheme based on the operation simulation video of the powder wet-excitation screening device under each second wet-excitation screening scheme.

[0012] In one possible implementation, the step of clustering the screening operation status information under each initial test wet vibration screening scheme to obtain K mesh-through condition characterization clusters includes: using the K-means clustering algorithm to cluster the screening operation status information under each initial test wet vibration screening scheme to obtain K mesh-through condition characterization clusters.

[0013] According to a third aspect, embodiments of the present invention provide an electronic device, including: a processor; a memory; and a computer program; wherein the computer program is stored in the memory and configured to be executed by the processor to implement the method as described above, the method including: acquiring an original dry scan image of the powder to be screened, premixed slurry characteristic data of the powder to be screened, and structural configuration information of a wet vibrating sieving device for the powder; determining multiple initial test wet vibrating sieving schemes based on the original dry scan image of the powder to be screened, the premixed slurry characteristic data of the powder to be screened, and the structural configuration information of the wet vibrating sieving device for the powder, the wet vibrating sieving scheme including vibration frequency and spray flow rate; acquiring the wet vibrating sieving of the premixed slurry of the powder to be screened under each initial test wet vibrating sieving scheme. The device undergoes trial operation video testing; based on the trial operation video of the wet vibrating sieving device for the powder premixed slurry to be screened under each initial test wet vibrating sieving scheme, the screening operation status information under each initial test wet vibrating sieving scheme is determined; based on the screening operation status information under each initial test wet vibrating sieving scheme, multiple first wet vibrating sieving schemes are generated; the operation video of the wet vibrating sieving device for the powder premixed slurry to be screened under each first wet vibrating sieving scheme is obtained; based on the operation video of the wet vibrating sieving device for the powder premixed slurry to be screened under each first wet vibrating sieving scheme, a preferred wet vibrating sieving scheme is determined; based on the preferred wet vibrating sieving scheme, the wet vibrating sieving device is controlled to perform full wet vibrating sieving control on the powder premixed slurry to be screened.

[0014] According to a fourth aspect, this embodiment provides a computer-readable storage medium storing a computer program thereon. When executed by a processor, the program implements the aforementioned control method for a powder wet vibrating sieving device based on image analysis. The method includes: acquiring a scanned image of the original dried powder to be sieved, characteristic data of the premixed slurry of the powder to be sieved, and structural configuration information of the powder wet vibrating sieving device; determining multiple initial test wet excitation screening schemes based on the scanned image of the original dried powder to be sieved, the characteristic data of the premixed slurry of the powder to be sieved, and the structural configuration information of the powder wet vibrating sieving device, wherein the wet excitation screening scheme includes a vibration frequency and a spray flow rate; and acquiring the test results of the powder wet vibrating sieving device under each initial test wet excitation screening scheme for the premixed slurry of the powder to be sieved. The process involves: 1. Analyzing the operation video of the wet vibrating sieving device for the powder premixed slurry under each initial wet vibrating sieving scheme; 2. Determining the sieving operation status information under each initial wet vibrating sieving scheme; 3. Generating multiple first wet vibrating sieving schemes based on the sieving operation status information under each initial wet vibrating sieving scheme; 4. Obtaining the operation video of the wet vibrating sieving device for the powder premixed slurry under each first wet vibrating sieving scheme; 5. Determining the preferred wet vibrating sieving scheme based on the operation video of the wet vibrating sieving device for the powder premixed slurry under each first wet vibrating sieving scheme; 6. Controlling the wet vibrating sieving device to perform full wet vibrating sieving control on the powder premixed slurry based on the preferred wet vibrating sieving scheme.

[0015] This invention provides a control method and system for a wet vibrating sieving device for powder based on image analysis. The method includes acquiring a scanned image of the original dried powder to be sieved, characteristic data of the premixed slurry of the powder to be sieved, and structural configuration information of the wet vibrating sieving device; determining multiple initial test wet vibration screening schemes based on the scanned image of the original dried powder to be sieved, the characteristic data of the premixed slurry of the powder to be sieved, and the structural configuration information of the wet vibrating sieving device, wherein the wet vibration screening scheme includes vibration frequency and spray flow rate; acquiring trial operation videos of the wet vibrating sieving device under each initial test wet vibration screening scheme for the premixed slurry of the powder to be sieved; and determining the wet vibration screening scheme based on the premixed slurry of the powder to be sieved under each initial test wet vibration screening scheme. The method involves using trial operation videos of a vibrating sieve device to determine the screening operation status information under each initial test wet vibration screening scheme; generating multiple first wet vibration screening schemes based on the screening operation status information under each initial test wet vibration screening scheme; acquiring operation videos of the powder premixed slurry to be screened under each first wet vibration screening scheme; determining the preferred wet vibration screening scheme based on the operation videos of the powder premixed slurry to be screened under each first wet vibration screening scheme; and controlling the powder wet vibration sieve device to perform full-volume wet vibration screening of the powder premixed slurry to be screened based on the preferred wet vibration screening scheme. This method can efficiently and accurately determine the preferred vibration screening scheme of the powder wet vibration sieve device. Attached Figure Description

[0016] Figure 1 A schematic flowchart illustrating a control method for a powder wet vibrating sieving device based on image analysis, provided in an embodiment of the present invention;

[0017] Figure 2 This is a schematic diagram of a wet vibrating sieve device for powder provided in an embodiment of the present invention;

[0018] Figure 3 This is a schematic diagram of a process for generating multiple first wet excitation screening schemes provided in an embodiment of the present invention;

[0019] Figure 4 A flowchart illustrating a preferred wet excitation screening scheme provided in an embodiment of the present invention;

[0020] Figure 5 This is a schematic diagram of a control system for a powder wet vibrating sieve device based on image analysis, provided as an embodiment of the present invention. Detailed Implementation

[0021] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings. Similar elements in different embodiments are referred to by associated similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of the invention. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other elements, materials, or methods. In some cases, certain operations related to the present invention are not shown or described in the specification. This is to avoid obscuring the core parts of the invention with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.

[0022] In this embodiment of the invention, the following are provided: Figure 1 The diagram illustrates a control method for a wet powder vibrating sieve based on image analysis. The control method includes steps S1 to S8:

[0023] Step S1: Obtain the original dry scan image of the powder to be screened, the premixed slurry characteristic data of the powder to be screened, and the structural configuration information of the wet vibrating sieve device for powder.

[0024] The raw, dried scan image of the powder to be sieved is obtained by scanning the powder in a dry state using industrial scanning imaging equipment. The raw, dried scan image of the powder to be sieved can show the outline, particle size, shape, surface texture, particle spacing, and agglomeration area of ​​the powder particles.

[0025] The premixed slurry characteristic data of the powder to be screened is obtained by measuring the premixed slurry of the powder to be screened using a particle size analyzer, a viscosity analyzer, and a density analyzer. The premixed slurry characteristic data of the powder to be screened includes solid mass concentration, slurry density, slurry viscosity, average particle size, maximum particle size, and percentage of agglomerated particles.

[0026] The structural configuration information of the wet powder vibrating screen is obtained by reading the equipment design drawings, technical parameter documents, and control system configuration files. This information describes the structural composition and controllable operating range of the wet powder vibrating screen. The structural configuration information includes screen aperture, screen surface area, screen surface inclination angle, allowable vibration frequency range, allowable spray flow rate range, number of spray nozzles, and the coordinates of the spray nozzles above the screen surface. Figure 2 This is a schematic diagram of a wet vibrating sieve device for powder provided in an embodiment of the present invention.

[0027] Step S2: Based on the original dry scan image of the powder to be screened, the premixed slurry characteristic data of the powder to be screened, and the structural configuration information of the wet vibration screening device, determine multiple initial test wet excitation screening schemes, including vibration frequency and spray flow rate.

[0028] In some embodiments, an initial test scheme determination model can be used to determine multiple initial test wet vibration screening schemes. The initial test scheme determination model is a generative adversarial network (GAN). The inputs to the initial test scheme determination model are the original dry scan image of the powder to be screened, the premixed slurry characteristic data of the powder to be screened, and the structural configuration information of the wet vibration screening device for the powder. The output of the initial test scheme determination model is multiple initial test wet vibration screening schemes.

[0029] Generative Adversarial Networks (GANs) are generative models consisting of a generator and a discriminator. The generator produces candidate samples based on the input, while the discriminator judges the differences between the candidate samples and the distribution of real samples. The generator and discriminator continuously update the model parameters through adversarial training, enabling the generator to gradually generate new samples that conform to the training data distribution and input constraints.

[0030] Multiple initial test wet vibration screening schemes are several alternative control schemes generated by the initial test scheme determination model for the premixed slurry of powder to be screened, used to control the wet vibration screening device for small-scale trial operation. Each initial test wet vibration screening scheme includes vibration frequency and spray flow rate.

[0031] Vibration frequency refers to the parameter indicating the speed of vibration of the screen surface in a wet powder vibrating screen.

[0032] Spray flow rate refers to the amount of liquid sprayed onto the screen surface per unit time in a wet vibrating powder screening device.

[0033] The original dry scan image of the powder to be screened reflects the morphology, particle size differences, and agglomeration distribution of the powder particles, enabling the model to determine the basic tendencies of powder passing through the sieve openings, spreading on the sieve surface, and forming agglomerates. The premixed slurry characteristic data of the powder to be screened reflects the ease of flow, particle size, and degree of agglomeration after the powder forms a slurry, allowing the vibration frequency and spray flow rate settings to adapt to the actual state of the premixed slurry. The structural configuration information of the wet vibrating powder screening device defines the sieve surface, sieve openings, and spray structure upon which the screening operation relies, and provides the controllable range of vibration frequency and spray flow rate, ensuring that the generated parameter combinations fall within the permissible range of the wet vibrating powder screening device.

[0034] After reading the original dry scan image of the powder to be screened, the generator of the generative adversarial network (GAN) extracts visual features such as particle outline, particle size, particle spacing, and agglomeration regions from the image. It then establishes a correspondence between these visual features and the flowability, particle size, and agglomeration state reflected in the premixed slurry characteristics of the powder to be screened, forming a conditional representation of the diffusion, penetration, and accumulation tendency of the premixed slurry on the screen surface. Next, the generator reads the structural configuration information of the wet vibrating sieving device, incorporating the operating conditions formed by the screen aperture, screen surface structure, and spray nozzle arrangement into the parameter generation process, and restricting the value boundaries of candidate parameters with the allowable ranges of vibration frequency and spray flow rate. Based on this, the generator can generate multiple combinations of vibration frequency and spray flow rate. The discriminator, based on the distribution of wet vibrating screening schemes learned during the GAN training process, judges the degree of matching between each candidate parameter combination and the current powder state, slurry state, and device structural conditions, and feeds the judgment result back to the generator. The generator adjusts the combination of vibration frequency and spray flow rate based on feedback, so that the generated results meet the allowable range of the device while corresponding to different testable operating conditions, and finally form multiple initial test wet excitation screening schemes.

[0035] Step S3: Obtain trial operation videos of the wet vibrating sieving device for the powder premixed slurry under each initial test wet vibrating sieving scheme.

[0036] The trial operation video of the wet vibrating sieving device for the powder premixed slurry under each initial test wet vibrating sieving scheme was obtained by continuously filming the sieving process of the wet vibrating sieving device during operation according to each initial test wet vibrating sieving scheme using industrial camera equipment set above the wet vibrating sieving device and at the discharge end.

[0037] The trial operation video of the wet vibrating sieving device for the powder premixed slurry under each initial test wet vibrating sieving scheme recorded the dynamic changes of the slurry flow diffusion trajectory, screen penetration process, and powder accumulation phenomenon on the screen surface over time during the operation of the wet vibrating sieving device under the corresponding initial test wet vibrating sieving scheme.

[0038] Step S4: Based on the trial operation video of the wet vibrating sieving device for the powder premixed slurry to be screened under each initial test wet vibrating sieving scheme, determine the screening operation status information under each initial test wet vibrating sieving scheme.

[0039] In some embodiments, a screening operation analysis model can be used to determine the screening operation status information under each initial test wet vibration screening scheme. The screening operation analysis model is a recurrent neural network. The input to the screening operation analysis model is the trial operation video of the wet vibration screening device for the powder premixed slurry to be screened under each initial test wet vibration screening scheme, and the output of the screening operation analysis model is the screening operation status information under each initial test wet vibration screening scheme.

[0040] Recurrent Neural Networks (RNNs) are neural network models used to process sequential data. RNNs pass the hidden state from the previous time step to the next time step through recurrent connections, allowing the computation result at the current time step to combine with the input content from previous time steps, thereby learning the dependencies between different time positions in the sequence. RNNs are suitable for processing video data with consecutive frame sequences and identifying changes in the state of a target in the video over time.

[0041] The screening operation status information under each initial test wet vibration screening scheme is determined by the screening operation analysis model and is used to characterize the actual screening operation effect of the powder wet vibrating screening device under the corresponding initial test wet vibration screening scheme.

[0042] The screening operation status information for each initial test wet vibration screening scheme includes the vibration frequency, spray flow rate, percentage of screen-permeable area per unit time, percentage of screen hole blockage area, percentage of slurry accumulation area on the screen surface, width of slurry spreading on the screen surface, percentage of non-screen-permeable powder area at the discharge end, and percentage of material run-out area at the discharge end.

[0043] A recurrent neural network (RNN) sequentially reads frame-by-frame video recordings of the wet vibratory screening device for each initial wet vibratory screening scheme, processing the premixed powder to be screened. When processing the current frame, the model receives and integrates the flow and diffusion trajectory of the slurry on the screen surface and the characteristics of powder accumulation from the previous frame through hidden states, thus accurately capturing the dynamic changes in the screen penetration process. The RNN identifies the edge boundaries of the slurry covering the screen openings and the movement positions of the powder that has not penetrated the screen in consecutive frames, and calculates the percentage of the area of ​​the screen-penetrating region, the percentage of the screen opening blockage area, and the percentage of the slurry accumulation area on the screen surface per unit time. The model further extracts the pixel width of the slurry's lateral spread to map to the slurry spread width on the screen surface, while simultaneously monitoring the powder pixel aggregation at the discharge end to calculate the percentage of the area of ​​the powder that has not penetrated the screen and the percentage of the area of ​​the material running off the screen at the discharge end. Finally, the model integrates these quantitative indicators extracted based on time-series dynamic changes and outputs the screening operation status information for each initial wet vibratory screening scheme.

[0044] Step S5: Generate multiple first wet vibration screening schemes based on the screening operation status information under each initial test wet vibration screening scheme.

[0045] In some embodiments, Figure 3 This is a schematic flowchart illustrating the generation of multiple first wet vibration screening schemes according to an embodiment of the present invention. The generation of multiple first wet vibration screening schemes includes steps S51 to S53:

[0046] Step S51: Based on the screening operation status information under each initial test wet excitation screening scheme, cluster K mesh-through condition characterization clusters are obtained.

[0047] K-means clustering is an unsupervised clustering algorithm that divides multiple samples into K clusters based on the numerical distance between them. By repeatedly performing sample allocation and cluster center updates, K-means clustering gradually reduces the differences between samples within the same cluster and distinguishes the cluster centers of different clusters. K-means clustering is suitable for classifying the status of screening operations with multiple continuous values.

[0048] In some embodiments, the value of K can be obtained by pre-setting it manually.

[0049] The K mesh-through condition characterization clusters are formed by dividing the screening operation status information under each initial test wet-excitation screening scheme according to the similarity using the K-means clustering algorithm.

[0050] The screening operation status information within each screen-penetrating condition characterization cluster exhibits high consistency in terms of screen-penetrating efficiency and operational stability, while significant differences exist in the screening operation status information between different screen-penetrating condition characterization clusters.

[0051] The screening operation status information for each initial wet vibratory screening scheme is simultaneously recorded, along with the corresponding parameter combinations and the actual screen penetration, clogging, accumulation, spreading, and discharge end statuses. The K-means clustering algorithm can classify the operating conditions based on the similarity of screening performance under different parameter combinations. Initial wet vibratory screening schemes with similar operating states are grouped into the same screen penetration condition cluster, while initial wet vibratory screening schemes with significant differences in screen penetration efficiency or operational stability are grouped into different screen penetration condition clusters.

[0052] The process of clustering the screening operation status information under each initial wet-vibration screening scheme using the K-means clustering algorithm is as follows: First, K initial cluster center coordinates are randomly projected into the multidimensional numerical space containing the screening operation status information under each initial wet-vibration screening scheme. Then, the algorithm measures the Euclidean distance from the screening operation status information of each scheme to each cluster center, and assigns each scheme to the nearest cluster center to form a preliminary classification. Next, the algorithm traverses each newly formed classification group and calculates the arithmetic mean of various indicators such as the percentage of screen penetration area per unit time and the percentage of screen hole blockage area for all schemes within the group. These arithmetic means are used to recalibrate the spatial position of the cluster centers, thus the new cluster centers more accurately represent the overall operation performance trend within the group. When a cluster center shifts, the algorithm prompts all schemes to reassess their affiliation based on the new cluster center positions. After several alternating executions of distance calculation and center position calibration, the affiliation of all schemes will gradually stabilize. At this point, the algorithm will converge and finally output K clusters representing the working conditions of the through-mesh operation.

[0053] K clusters of mesh-penetrating condition characterization can categorize a large number of initial wet-vibration screening schemes with varying operational performance into sets of conditions with similar conditions within each cluster and differences between clusters. This allows for a centralized presentation of changes in mesh penetration, clogging, accumulation, and material runoff corresponding to different vibration frequencies and spray flow rates, categorized by condition type. Subsequent processing can analyze the changes in parameters based on the state differences between each mesh-penetrating condition characterization cluster, transforming insufficient parameters into those capable of suppressing clogging and suitable parameters into those prone to causing material runoff. This reduces the complexity of comparing all initial wet-vibration screening schemes one by one.

[0054] Step S52: Based on the K mesh-permeable condition characterization clusters, determine the critical minimum vibration frequency for preventing sticking, the critical minimum spray flow rate for preventing sticking, the critical maximum vibration frequency for preventing material leakage, and the critical maximum spray flow rate for preventing material leakage.

[0055] In some embodiments, a critical analysis model can be used to determine the minimum critical vibration frequency for preventing clogging, the minimum critical spray flow rate for preventing clogging, the maximum critical vibration frequency for preventing material leakage, and the maximum critical spray flow rate for preventing material leakage. The critical analysis model is a deep neural network. The input to the critical analysis model is the K mesh-permeable condition characterization clusters, and the output of the critical analysis model is the minimum critical vibration frequency for preventing clogging, the minimum critical spray flow rate for preventing clogging, the maximum critical vibration frequency for preventing material leakage, and the maximum critical spray flow rate for preventing material leakage.

[0056] A deep neural network is a multi-layered neural network model consisting of an input layer, multiple hidden layers, and an output layer stacked sequentially. Through layer-by-layer combinations of linear transformations and non-linear activation functions, deep neural networks can progressively abstract and represent the features of input data. Deep neural networks exhibit strong fitting capabilities in processing high-dimensional data and uncovering complex non-linear mapping relationships.

[0057] The critical minimum vibration frequency for preventing sticking and clogging is the lower limit of the vibration frequency required by a quantitative wet vibrating sieve device to suppress the adhesion and clogging of the premixed slurry of the powder to be sieved.

[0058] The critical minimum spray flow rate for preventing sticking is the lower limit of the spray flow rate required for a quantitative powder wet vibrating screen to flush away the slurry adhering to the screen surface and maintain the permeability of the screen holes.

[0059] The critical maximum vibration frequency for preventing material spillage is the upper limit of the allowable vibration frequency for a quantitative powder wet vibrating screen to prevent the premixed slurry of the powder to be screened from moving too quickly toward the discharge end.

[0060] The critical maximum spray flow rate for preventing material leakage is the upper limit of the allowable spray flow rate of a quantitative powder wet vibrating screen to prevent excessive spray liquid from pushing unpenetrated powder to the discharge end.

[0061] When the vibration frequency is lower than the critical minimum vibration frequency for anti-clogging or the spray flow rate is lower than the critical minimum spray flow rate for anti-clogging, the proportion of screen hole clogging area and the proportion of slurry accumulation area on the screen surface increase.

[0062] When the vibration frequency is higher than the critical maximum vibration frequency for preventing material leakage or the spray flow rate is higher than the critical maximum spray flow rate for preventing material leakage, the migration speed of the premixed powder slurry to be screened on the screen surface increases and the amount of material leakage at the discharge end increases.

[0063] K permeable screen condition characterization clusters record the distribution of screening states corresponding to different combinations of vibration frequencies and spray flow rates, enabling the model to identify the locations where screening states change within the overall evolution of the working conditions. As the vibration frequency and spray flow rate gradually change, the different permeable screen condition characterization clusters exhibit a continuous process of change, from gradual relief of screen clogging, gradual reduction of slurry accumulation, gradual expansion of the permeable screen area, to a sustained increase in slurry migration speed, ultimately leading to the risk of material leakage. By comparing the state differences among the various permeable screen condition characterization clusters, the deep neural network can identify the parameter ranges corresponding to the transitions from unstable to stable and from stable to unstable processes during screening. This allows it to determine which parameter ranges can effectively prevent screen clogging and which parameter ranges are approaching the risk zone of rapid slurry migration.

[0064] Deep neural networks, through their multi-layered nonlinear mapping structure, deeply deconstruct the parameter evolution thresholds hidden within K clusters representing screen-through operating conditions. The first layer of neurons in the model absorbs the vibration frequency and spray flow rate values ​​for each cluster and correlates them with the corresponding percentage of screen clogging area and the percentage of material leakage area at the discharge end. During information transmission in the hidden layers, the model can focus its analysis on screen-through operating condition clusters exhibiting high levels of clogging, fitting a function curve between the decrease in vibration frequency and spray flow rate and the surge in clogging area using a nonlinear activation function. When the model detects an extreme point on the curve, below which a slight decrease would cause an exponential deterioration in the percentage of slurry accumulation on the screen surface, the model locks this coordinate point as the critical minimum vibration frequency and critical minimum spray flow rate for preventing clogging. Higher-order sensing units within the deep neural network examine screen-through operating condition clusters representing high material leakage risk, analyzing the phenomenon of a sudden increase in the percentage of non-screen-through powder area due to spray overload or vibration frequency. Deep neural networks will draw a gradient map of material runaway risk in high-dimensional space, and find the upper limit parameter band that causes the powder migration speed to completely get out of control through gradient descent optimization. Then, they will accurately extract and output the critical maximum vibration frequency and critical maximum spray flow rate for preventing material runaway from the boundary band.

[0065] Step S53: Based on the screening operation status information under each initial test wet vibration screening scheme, the critical minimum vibration frequency for preventing sticking, the critical minimum spray flow rate for preventing sticking, the critical maximum vibration frequency for preventing material leakage, and the critical maximum spray flow rate for preventing material leakage, multiple first wet vibration screening schemes are generated.

[0066] In some embodiments, a first scheme generation model can be used to generate multiple first wet vibration screening schemes. The first scheme generation model is a generative adversarial network. The inputs of the first scheme generation model are the screening operation status information under each initial test wet vibration screening scheme, the critical minimum vibration frequency for preventing sticking, the critical minimum spray flow rate for preventing sticking, the critical maximum vibration frequency for preventing material leakage, and the critical maximum spray flow rate for preventing material leakage. The output of the first scheme generation model is multiple first wet vibration screening schemes.

[0067] Multiple first wet excitation screening schemes are sets of wet excitation screening parameter schemes generated within the critical parameter constraints for further testing and verification. Each first wet excitation screening scheme includes vibration frequency and spray flow rate.

[0068] In each of the first wet vibration screening schemes, the vibration frequency shall not be lower than the critical minimum vibration frequency for anti-clogging and not higher than the critical maximum vibration frequency for anti-material leakage, and the spray flow rate shall not be lower than the critical minimum spray flow rate for anti-clogging and not higher than the critical maximum spray flow rate for anti-material leakage.

[0069] The screening operation status information of each initial test wet vibration screening scheme recorded the screening performance generated by different combinations of vibration frequencies and spray flow rates during actual trial operation. This reflected the actual influence of various parameter ratios within the initial parameter space on screen penetration efficiency, clogging degree, and material leakage tendency, and enabled the model to grasp the correspondence between parameter changes and screening performance. The critical minimum vibration frequency, critical minimum spray flow rate, critical maximum vibration frequency, and critical maximum spray flow rate for preventing material leakage together define the upper and lower boundaries of the safe parameter range that simultaneously suppresses clogging and material leakage, narrowing the range of possible schemes to within the practically feasible screening operation control space.

[0070] The generator of the generative adversarial network (GAN) is based on the correspondence between parameters and screening performance contained in the screening operation status information of each initial wet-excitation screening scheme. Simultaneously, it embeds the feasible vibration frequency range (comprising the critical minimum vibration frequency for preventing clogging and the critical maximum vibration frequency for preventing material leakage) and the feasible spray flow range (comprising the critical minimum spray flow rate for preventing clogging and the critical maximum spray flow rate for preventing material leakage) as hard constraints into the generation process. Referring to the inherent relationship between parameters and screen penetration efficiency verified in the initial testing phase, the generator generates candidate schemes with different parameter emphases within the feasible range. These include schemes biased towards the lower limit of vibration frequency to reduce the risk of material leakage, schemes biased towards the upper limit of spray flow rate to enhance the screen scouring effect, and schemes seeking a balance between screen penetration efficiency and operational stability in the middle of the range. The discriminator then filters the generated candidate schemes, determining whether the vibration frequency is not lower than the critical minimum vibration frequency for anti-clogging and not higher than the critical maximum vibration frequency for anti-material leakage, and whether the spray flow rate is not lower than the critical minimum spray flow rate for anti-clogging and not higher than the critical maximum spray flow rate for anti-material leakage. Simultaneously, it determines whether this parameter combination matches the actual distribution of the high-permeability mesh performance in the initial test. Through continuous iteration of adversarial training, the generator gradually generates multiple first-stage wet-excitation screening schemes with diverse distributions that conform to the actual screening operation rules within the critical parameter constraints.

[0071] Step S6: Obtain the operation video of the wet vibrating sieving device for the powder premixed slurry under each first wet vibrating sieving scheme.

[0072] The operation video of the wet vibrating sieving device for the powder premixed slurry under each first wet excitation screening scheme is obtained by continuously filming the screening process of the wet vibrating sieving device according to each first wet excitation screening scheme through industrial camera equipment set above the wet vibrating sieving device and at the discharge end.

[0073] The video recording of the operation of the wet vibrating sieving device for the powder premixed slurry under each first wet vibrating screening scheme shows the dynamic changes over time in the flow and diffusion trajectory of the slurry on the screen surface, the screen penetration process, and the powder accumulation phenomenon of the wet vibrating sieving device under the corresponding first wet vibrating screening scheme.

[0074] Step S7: Determine the preferred wet vibration screening scheme based on the operation video of the wet vibration screening device of the powder premixed slurry to be screened under each first wet vibration screening scheme.

[0075] In some embodiments, Figure 4 This is a flowchart illustrating a preferred wet vibration screening scheme according to an embodiment of the present invention. The process of determining the preferred wet vibration screening scheme includes steps S71 to S74:

[0076] Step S71: Construct a wet vibration screening map. The wet vibration screening map includes multiple nodes of the first wet vibration screening scheme and multiple edges between the nodes. The node features of each first wet vibration screening scheme node include the operation video of the powder premix slurry to be screened under the corresponding first wet vibration screening scheme of the powder wet vibration screening device. The edges are the vibration frequency difference and spray flow rate difference between the nodes of the first wet vibration screening scheme.

[0077] The wet vibration screening map is a graph data composed of multiple nodes of the first wet vibration screening scheme and multiple edges between the nodes of the multiple first wet vibration screening scheme. The wet vibration screening map is used to represent the actual operation process of multiple first wet vibration screening schemes and the differences in control parameters between different first wet vibration screening schemes.

[0078] Multiple first wet excitation screening scheme nodes are nodes in the wet excitation screening map used to represent multiple first wet excitation screening schemes respectively. Each first wet excitation screening scheme node corresponds to one first wet excitation screening scheme.

[0079] The node characteristics of each first wet vibration screening scheme node include the operation video of the wet vibration screening device of the powder premix slurry to be screened under each first wet vibration screening scheme.

[0080] The edges are used to connect nodes of different first wet vibration screening schemes, and represent the differences in control parameters between different first wet vibration screening schemes. Each edge includes the difference in vibration frequency and the difference in spray flow rate between the two first wet vibration screening scheme nodes it connects.

[0081] The vibration frequency difference is the numerical difference between the vibration frequencies corresponding to the two connected nodes of the first wet-type vibration screening scheme. The spray flow rate difference is the numerical difference between the spray flow rates corresponding to the two connected nodes of the first wet-type vibration screening scheme.

[0082] Step S72: Process the wet excitation screening spectrum based on graph neural network to determine multiple second wet excitation screening schemes.

[0083] Graph Neural Networks (GNNs) are neural network models used to process graph-based structural data. GNNs can perform information transfer and feature aggregation based on node features, features of adjacent nodes, and edges between nodes, and update the feature representations of each node through multi-layer propagation, thereby analyzing the nodes and the structural relationships between them. The input to the GNN is the wet-excitation screening graph, and the output is multiple second wet-excitation screening schemes.

[0084] Multiple second wet-vibration screening schemes were determined by analyzing the actual operation process and control parameter differences of multiple first wet-vibration screening schemes using graph neural networks. These schemes serve as candidate control schemes for further screening operation simulation. Each second wet-vibration screening scheme includes vibration frequency and spray flow rate.

[0085] The wet-excited screening diagram organizes multiple first-stage wet-excited screening schemes into nodes and edges, enabling the graph neural network to simultaneously process the operational performance of each scheme and the parameter differences between schemes. The node features of each first-stage wet-excited screening scheme node are derived from the corresponding operational video, reflecting the actual movement of the slurry on the screen surface under that scheme. The edges between nodes are derived from the difference in vibration frequency and the difference in spray flow rate, reflecting the parameter adjustment range between the two schemes. Therefore, the graph neural network can determine the direction of the influence of adjacent parameter changes on the screening effect; for example, whether a slight increase in vibration frequency improves screen penetration, whether increasing spray flow rate alleviates clogging, or whether excessive parameter changes cause increased material leakage at the discharge end.

[0086] When processing the wet vibration screening graph, the graph neural network embeds the node features of each first wet vibration screening scheme node—that is, the operating video of the wet vibration screening device of the powder premix slurry to be screened under the corresponding first wet vibration screening scheme—as the initial features of the node. During the message passing phase, the model reads the vibration frequency difference and spray flow rate difference between the two ends of each edge to determine the influence weight when aggregating the features of adjacent nodes. Adjacent schemes with smaller vibration frequency and spray flow rate differences indicate highly similar parameters, and the screening differences reflected in their operating videos are more likely to stem from subtle parameter adjustments, thus having a greater mutual influence weight. Adjacent schemes with larger vibration frequency or spray flow rate differences have smaller weights. After multiple rounds of message passing, the features of each node incorporate the screening performance of other first wet vibration screening schemes with similar parameters under similar operating conditions. When multiple nodes within a certain parameter region consistently exhibit a high proportion of transparent area and a low risk of screen clogging and material leakage in the aggregated features, the graph neural network determines that the parameters in this region are worth further refined exploration. Then, based on the parameter distribution of the existing first wet vibration screening scheme in this region, it determines new candidate parameter combinations and finally outputs multiple second wet vibration screening schemes for further screening operation simulation.

[0087] Step S73: Based on the trial operation video of the powder wet vibrating sieve device under each initial test wet vibrating sieve scheme, the operation video of the powder wet vibrating sieve device under each first wet vibrating sieve scheme, and the multiple second wet vibrating sieve schemes, generate an operation simulation video of the powder wet vibrating sieve device under each second wet vibrating sieve scheme.

[0088] In some embodiments, a vibration screening simulation model can be used to generate operational simulation videos of the powder wet vibrating sieve under each second wet vibration screening scheme. The vibration screening simulation model is a variational autoencoder. The inputs to the vibration screening simulation model are the trial operation videos of the powder wet vibrating sieve under each initial test wet vibration screening scheme, the operational videos of the powder wet vibrating sieve under each first wet vibration screening scheme, and the plurality of second wet vibration screening schemes. The output of the vibration screening simulation model is the operational simulation video of the powder wet vibrating sieve under each second wet vibration screening scheme.

[0089] A variational autoencoder (VAE) is a generative model consisting of an encoder, a latent probability space, and a decoder. The encoder maps input data to probability distribution parameters in the latent probability space, and the decoder samples from the latent probability space and reconstructs the data. A VAE can learn the continuous distribution patterns of known samples and generate new images, videos, and continuous numerical data that conform to the known sample distribution based on given conditions.

[0090] The operation simulation video of the powder wet vibrating screen under each second wet excitation screening scheme is generated by the excitation screening simulation model. It is used to simulate the screening process of the powder wet vibrating screen under each second wet excitation screening scheme. Each operation simulation video of the powder wet vibrating screen under each second wet excitation screening scheme corresponds to one second wet excitation screening scheme.

[0091] The simulation video of the operation of the wet vibrating screen device for powder under each second wet excitation screening scheme can simulate the dynamic changes of the flow and diffusion trajectory of the slurry on the screen surface, the screen penetration process, and the powder accumulation phenomenon over time under the corresponding vibration frequency and spray flow rate.

[0092] The trial operation videos of the powder wet vibrating sieving device under each initial test wet vibrating sieving scheme and the operation videos of the powder wet vibrating sieving device under each first wet vibrating sieving scheme together provide a real sieving process of the powder premixed slurry to be sieved under verified parameters. These videos cover the slurry's flow and diffusion trajectory on the sieve surface, the screen penetration process, and powder accumulation phenomena within a range of parameters from a broad initial range to narrowed constraints. These real videos form the physical reference basis for the simulation generation, enabling the model to grasp the actual impact of different vibration frequencies and spray flow rates on the slurry behavior. Multiple second wet vibrating sieving schemes provide target parameter combinations that have not yet been actually run, serving as trajectory constraints during simulation generation and determining the vibration frequency and spray flow rate conditions that the simulation videos need to reproduce.

[0093] The variational autoencoder encodes the screening process in the trial operation video of the powder wet vibrating screening device under each initial test wet vibrating screening scheme and the operation video of the powder wet vibrating screening device under each first wet vibrating screening scheme. It compresses and maps the flow diffusion pattern, screen penetration rate, and accumulation change pattern of the slurry under different vibration frequencies and spray flow rates into a latent probability space, forming a continuous probability distribution representation describing the dynamic law of slurry screening. Adjacent distribution positions in the latent space correspond to screening processes with similar parameters. In the generation stage, the decoder uses the vibration frequency and spray flow rate of a certain second wet vibrating screening scheme as conditions to locate and sample the distribution area corresponding to that parameter combination in the latent probability space, decoding the sampling results to restore a video frame sequence. During decoding, the model assigns the flow diffusion trajectory of the slurry on the screen surface in each frame based on the learned correspondence between vibration frequency and slurry spreading speed, and determines the expansion rhythm of the screen penetration area based on the correlation between spray flow rate and screen penetration rate. Simultaneously, it generates images of accumulation and material flow at the discharge end by referring to the evolution pattern of accumulation phenomena with parameter changes. By applying their respective parameter conditions to multiple second wet vibration screening schemes and repeating the above sampling and decoding process, the model finally outputs a simulation video of the operation of the powder wet vibration screening device under each second wet vibration screening scheme.

[0094] Step S74: Determine the preferred wet vibration screening scheme based on the operation simulation video of the powder wet vibration screening device under each of the second wet vibration screening schemes.

[0095] In some embodiments, a preferred scheme analysis model can be used to determine the preferred wet vibration screening scheme. The preferred scheme analysis model is a recurrent neural network. The input to the preferred scheme analysis model is the operation simulation video of the powder wet vibration screening device under each second wet vibration screening scheme, and the output of the preferred scheme analysis model is the preferred wet vibration screening scheme.

[0096] The preferred wet vibration screening scheme was determined from multiple second wet vibration screening schemes through an optimal scheme analysis model. It is a control scheme used to control the wet vibration screening device for the powder premix slurry to be screened, performing full-volume wet vibration screening. The preferred wet vibration screening scheme includes the vibration frequency and spray flow rate.

[0097] The simulation video of the operation of the wet vibrating sieving device for powder under each second wet excitation screening scheme simulates the dynamic changes of the slurry flow and diffusion trajectory on the screen surface, the screen penetration process, and the powder accumulation phenomenon under the corresponding vibration frequency and spray flow rate conditions in the form of a continuous frame sequence. The simulation videos of different second wet excitation screening schemes show comparable differences in screen penetration efficiency and operational stability, enabling the model to determine which parameter ratio has the best comprehensive screening performance from multiple candidate schemes by analyzing the long-term screening process footage.

[0098] A recurrent neural network (RNN) processes the simulation video of the powder wet vibrating screen under each second wet vibrating screening scheme in chronological order, converting the simulation video into a continuous sequence of video frames and passing hidden states between adjacent time steps. The hidden states continuously retain the states of slurry spreading on the screen surface, water permeability and blockage of the screen pores, slurry accumulation on the screen surface, and powder outward movement at the discharge end. This allows the RNN to judge the performance of each second wet vibrating screening scheme within a complete operating cycle, following the sequence of the screening process. In the simulation video corresponding to a certain second wet vibrating screening scheme, if the slurry spreading speed on the screen surface is fast but the water permeability and blockage index of the screen pores continues to increase, the RNN will retain the temporal information of the gradual blockage of the screen pores in subsequent time steps and reduce the anti-clogging evaluation of the corresponding second wet vibrating screening scheme. In another simulation video corresponding to the second wet vibratory screening scheme, the slurry spreading speed on the screen surface was not the highest, but the water permeability of the screen mesh remained stable, the slurry accumulation thickness on the screen surface did not continuously increase, and the powder outward movement at the discharge end was also at a low level. The recurrent neural network will improve the comprehensive screening evaluation of the corresponding second wet vibratory screening scheme in continuous time-series judgment. After completing the simulation video analysis of the powder wet vibratory screening device under each second wet vibratory screening scheme, the recurrent neural network compares the comprehensive performance of each second wet vibratory screening scheme in terms of mesh continuity, anti-clogging ability, anti-material leakage ability, and continuous operation stability, and determines the optimal wet vibratory screening scheme.

[0099] In some embodiments, determining the preferred wet vibration screening scheme based on the operation simulation video of the powder wet vibration screening device under each of the second wet vibration screening schemes includes steps S741~S743:

[0100] Step S741: Based on the operation simulation video of the powder wet vibrating sieve device under each second wet vibrating sieve scheme, determine the screen pore permeability blockage index curve and the cumulative sequence of slurry thickness on the screen surface corresponding to each second wet vibrating sieve scheme.

[0101] In some embodiments, a recurrent neural network can be used to determine the screen pore permeability blockage index curve and the cumulative sequence of slurry thickness on the screen surface corresponding to each second wet excitation screening scheme.

[0102] The screen pore permeability blockage index curve is used to represent the degree of obstruction of the screen pore permeability over time during the simulation operation of each second wet-excitation screening scheme.

[0103] The screen pore water permeability blockage index curve reflects the continuous process of screen pores changing from unobstructed to blocked. A higher screen pore water permeability blockage index indicates a greater degree of obstruction by powder agglomerates and slurry deposits on the screen pores, while a lower screen pore water permeability blockage index indicates that the screen pores still maintain good liquid permeability.

[0104] The cumulative slurry thickness sequence on the screen surface is a sequence used to represent the continuous cumulative results of the slurry accumulation thickness on the screen surface at different time points during the simulation operation of each second wet-excitation screening scheme.

[0105] The cumulative sequence of slurry thickness on the screen surface reflects the time process of slurry residence, spreading, piling up, and dissipation on the screen surface. A continuous increase in the value of the cumulative sequence of slurry thickness on the screen surface indicates insufficient slurry discharge capacity under the corresponding second wet vibrating screening scheme. A stable value in the cumulative sequence of slurry thickness on the screen surface indicates that the slurry can maintain a relatively stable flow state on the screen surface under the corresponding second wet vibrating screening scheme.

[0106] When the recurrent neural network processes the simulation video of the powder wet vibrating screen under each second wet excitation screening scheme, the initial screen surface spreading image is converted into a hidden state in the early time steps. The image of the screen pores gradually being covered by slurry will continue to affect the update of the hidden state in subsequent time steps. After slurry stagnation occurs locally on the screen surface, the recurrent neural network will not judge the slurry accumulation on the screen surface based on a single frame, but will correct the screen pore permeability blockage index curve based on whether the slurry thickness repeatedly increases in consecutive time steps, whether the screen pore water permeability image continuously darkens, and whether the effective screen permeability area gradually shrinks. When a brief liquid flow scour occurs in the simulation video of the powder wet vibrating screen, the recurrent neural network will continue to pass the degree of screen surface recovery after the scour to subsequent time steps to avoid misjudging the temporarily thinned screen surface slurry as a cumulative decrease in screen surface slurry thickness. After the time-series recursion of the complete simulation cycle, the recurrent neural network can calculate the screen pore permeability blockage index curve and the cumulative sequence of screen surface slurry thickness for each second wet excitation screening scheme.

[0107] Step S742: Based on the screen pore permeability blockage index curve and screen surface slurry thickness accumulation sequence corresponding to each second wet vibration screening scheme, determine the expected screen penetration completion rate, screen anti-blocking compliance score, and discharge end anti-material runoff compliance level corresponding to each second wet vibration screening scheme.

[0108] In some embodiments, a recurrent neural network can be used to determine the expected screen penetration rate, screen anti-clogging compliance score, and discharge end anti-material runoff compliance level for each second wet vibratory screening scheme.

[0109] The expected screen penetration rate is used to represent the proportion of powder that is expected to pass through the screen during the full-volume wet vibration screening process for each second wet vibration screening scheme.

[0110] The higher the expected screen penetration rate, the more target powder can pass through the screen openings and enter the under-screen area. The lower the expected screen penetration rate, the more likely the powder will remain on the screen surface or drift towards the discharge end with the slurry.

[0111] The screen anti-clogging compliance score is a score used to indicate the degree of compliance of each second wet vibratory screening scheme in preventing the screen pores from being covered by slurry and powder agglomerates.

[0112] The discharge end anti-leakage standard level is used to indicate the level of each second wet excitation screening scheme in terms of preventing unpenetrated powder from being carried to the discharge end by the liquid flow and vibration. The higher the discharge end anti-leakage standard level, the lower the possibility of unpenetrated powder being concentrated and moved outward under the corresponding second wet excitation screening scheme.

[0113] When determining the expected screen penetration rate for each second wet vibration screening scheme, the recurrent neural network (RNN) monitors the continuous level of the screen pore permeability blockage index curve during the simulation cycle and combines this with the cumulative sequence of slurry thickness on the screen surface to determine whether the slurry has been occupying the effective screening area for an extended period. If the screen pore permeability blockage index curve continuously increases in the later stages, and the cumulative sequence of slurry thickness also increases synchronously, the RNN will determine that the corresponding second wet vibration screening scheme is prone to insufficient screen penetration in full-scale wet vibration screening and will reduce the expected screen penetration rate. If the cumulative sequence of slurry thickness on the screen surface briefly increases and then falls back, while the screen pore permeability blockage index curve remains stable, the RNN will identify the corresponding second wet vibration screening scheme as having good screen recovery ability and improve the screen anti-clogging compliance score. The determination of the discharge end's anti-material leakage compliance level does not solely rely on the highest value of the cumulative slurry thickness sequence on the screen surface. The recurrent neural network also considers whether the cumulative slurry thickness sequence on the screen surface decreases rapidly in the later stages and whether the screen pore permeability blockage index curve rises synchronously to determine whether the slurry carries unpenetrated powder to the discharge end. After completing the temporal correlation judgment between screen pore blockage, slurry accumulation on the screen surface, and the risk of powder migration at the discharge end, the recurrent neural network outputs the expected screen penetration rate, screen anti-clogging compliance score, and discharge end anti-material leakage compliance level for each second wet vibration screening scheme.

[0114] Step S743: Based on the expected screen penetration rate, screen anti-clogging score, and discharge end anti-material run-out level of each second wet vibration screening scheme, determine the preferred wet vibration screening scheme.

[0115] In some embodiments, deep neural networks can be used to determine the preferred wet excitation screening scheme.

[0116] When processing the expected screen penetration rate, screen anti-clogging score, and discharge end anti-material leakage level for each second wet vibratory screening scheme, the deep neural network establishes a judgment boundary for the difference in screening effect between different second wet vibratory screening schemes. Second wet vibratory screening schemes with higher expected screen penetration rates will receive a stronger judgment of screen penetration advantage, but the deep neural network will further examine whether the screen anti-clogging score can support the long-term operation of the powder wet vibratory screening device. When the screen anti-clogging score is at a high level, the deep neural network will further determine whether the discharge end anti-material leakage level is sufficient to prevent powder from being carried away from the effective screening area by the spray liquid and vibration. If a second wet vibratory screening scheme simultaneously possesses a high expected screen penetration rate, a high screen anti-clogging score, and a high discharge end anti-material leakage level, the deep neural network will identify the corresponding second wet vibratory screening scheme as the more suitable scheme for full-volume wet vibratory screening. When there is a significant imbalance between the expected screen penetration rate, the screen anti-clogging compliance score, and the discharge end anti-material leakage compliance level, the model will reduce the likelihood of selecting the corresponding second wet vibration screening scheme. After comparing the compliance levels of all second wet vibration screening schemes, the model can finally determine the optimal wet vibration screening scheme that takes into account the requirements of screen penetration, anti-clogging, and anti-material leakage.

[0117] Step S8: Based on the preferred wet vibration screening scheme, control the wet vibration screening device to perform full wet vibration screening of the powder premix slurry to be screened.

[0118] Once the preferred wet vibration screening scheme is determined, the screen surface and spray pipeline of the wet vibration screening device are controlled according to the vibration frequency and spray flow rate in the preferred wet vibration screening scheme to carry out the overall wet vibration screening operation of the powder premixed slurry to be screened.

[0119] Based on the same inventive concept Figure 5 This invention provides a schematic diagram of a control system for a powder wet vibrating sieving device based on image analysis. The control system of the powder wet vibrating sieving device based on image analysis includes:

[0120] The acquisition module 91 is used to acquire the original dry scan image of the powder to be screened, the premixed slurry characteristic data of the powder to be screened, and the structural configuration information of the wet vibrating sieve device for powder.

[0121] The initial scheme determination module 92 is used to determine multiple initial test wet excitation screening schemes based on the original dry scan image of the powder to be screened, the premixed slurry characteristic data of the powder to be screened, and the structural configuration information of the powder wet vibrating screening device. The wet excitation screening schemes include vibration frequency and spray flow rate.

[0122] The excitation screening trial operation module 93 is used to acquire trial operation videos of the wet vibrating screening device for the powder premixed slurry to be screened under each initial test wet excitation screening scheme;

[0123] The status information determination module 94 is used to determine the screening operation status information under each initial test wet vibration screening scheme based on the trial operation video of the powder wet vibration screening device under each initial test wet vibration screening scheme of the powder premixed slurry to be screened.

[0124] The first scheme generation module 95 is used to generate multiple first wet vibration screening schemes based on the screening operation status information under each initial test wet vibration screening scheme.

[0125] The video acquisition module 96 is used to acquire the operation video of the wet vibrating sieving device for the powder premixed slurry to be screened under each first wet vibrating screening scheme.

[0126] The preferred scheme determination module 97 is used to determine the preferred wet vibration screening scheme based on the operation video of the wet vibration screening device of the powder premixed slurry to be screened under each first wet vibration screening scheme.

[0127] The screening control module 98 is used to control the wet vibration screening device to perform full-volume wet vibration screening of the powder premixed slurry to be screened based on the preferred wet vibration screening scheme.

[0128] It should be noted that, in order to simplify the descriptions disclosed herein and thus aid in the understanding of one or more embodiments of the invention, the foregoing description of embodiments of this specification may sometimes combine multiple features into a single embodiment, drawing, or description thereof. However, this method of disclosure does not imply that the subject matter of this specification requires more features than those mentioned in the claims. In fact, the embodiments contain fewer features than all the features of a single embodiment disclosed above.

[0129] Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments described herein. Other variations may also fall within the scope of this specification. Therefore, alternative configurations of the embodiments described herein are intended to be illustrative rather than limiting, and should be considered consistent with the teachings of this specification. Accordingly, the embodiments described herein are not limited to those explicitly introduced and described herein.

Claims

1. A control method for a wet vibrating sieve device for powder based on image analysis, characterized in that, include: Acquire the original dry scan image of the powder to be screened, the premixed slurry characteristic data of the powder to be screened, and the structural configuration information of the wet vibrating sieve device for powder. Based on the original dry scan image of the powder to be screened, the premixed slurry characteristic data of the powder to be screened, and the structural configuration information of the wet vibrating sieving device for the powder, multiple initial test wet excitation screening schemes are determined. The wet excitation screening schemes include vibration frequency and spray flow rate. Obtain trial operation videos of the wet vibrating sieving device for the powder premixed slurry under each initial test wet vibrating sieving scheme; Based on the trial operation video of the wet vibrating sieving device for the powder premixed slurry under each initial test wet vibrating sieving scheme, the screening operation status information under each initial test wet vibrating sieving scheme is determined. Multiple first wet vibration screening schemes are generated based on the screening operation status information under each initial test wet vibration screening scheme. Obtain operation videos of the wet vibrating sieving device for the powder premixed slurry to be screened under each first wet vibrating screening scheme; Based on the operation video of the wet vibrating sieving device for the powder premixed slurry under each first wet vibrating sieving scheme, the preferred wet vibrating sieving scheme is determined. Based on the preferred wet vibration screening scheme, the powder wet vibration screening device is controlled to perform full wet vibration screening of the premixed slurry of the powder to be screened.

2. The control method for the image analysis-based wet vibrating sieve device for powder as described in claim 1, characterized in that, The generation of multiple first wet vibration screening schemes based on the screening operation status information under each initial test wet vibration screening scheme includes: Based on the screening operation status information under each initial test wet excitation screening scheme, K mesh-through condition characterization clusters are obtained. Based on the K mesh-through condition characterization clusters, the critical minimum vibration frequency for preventing sticking, the critical minimum spray flow rate for preventing sticking, the critical maximum vibration frequency for preventing material leakage, and the critical maximum spray flow rate for preventing material leakage are determined. Based on the screening operation status information under each initial test wet vibration screening scheme, the critical minimum vibration frequency for preventing sticking, the critical minimum spray flow rate for preventing sticking, the critical maximum vibration frequency for preventing material leakage, and the critical maximum spray flow rate for preventing material leakage, multiple first wet vibration screening schemes are generated.

3. The control method for the powder wet vibrating sieve device based on image analysis as described in claim 1, characterized in that, The determination of the preferred wet vibration screening scheme based on the operating video of the powder wet vibrating screening device under each first wet vibration screening scheme for the premixed slurry of the powder to be screened includes: A wet vibration screening map is constructed, which includes multiple nodes of the first wet vibration screening scheme and multiple edges between the nodes. The node features of each first wet vibration screening scheme node include the operation video of the wet vibration screening device of the powder premixed slurry to be screened under each first wet vibration screening scheme, and the edges are the vibration frequency difference and spray flow rate difference between the nodes of the first wet vibration screening scheme. The wet excitation screening spectrum is processed based on a graph neural network to determine multiple second wet excitation screening schemes. Based on the trial operation video of the powder wet vibrating sieve device under each initial test wet excitation screening scheme, the operation video of the powder wet vibrating sieve device under each first wet excitation screening scheme, and the multiple second wet excitation screening schemes, an operation simulation video of the powder wet vibrating sieve device under each second wet excitation screening scheme is generated. The optimal wet vibration screening scheme is determined based on the operation simulation video of the powder wet vibration screening device under each of the second wet vibration screening schemes.

4. The control method for the image analysis-based wet vibrating sieving device for powder as described in claim 2, characterized in that, The K mesh-through condition characterization clusters obtained by clustering the screening operation status information based on each initial test wet-excitation screening scheme include: Based on the screening operation status information under each initial test wet excitation screening scheme, K-means clustering algorithm is used to obtain K mesh-through condition characterization clusters.

5. A control system for a wet vibrating sieve device for powder based on image analysis, characterized in that, include: The acquisition module is used to acquire the original dry scan image of the powder to be screened, the premixed slurry characteristic data of the powder to be screened, and the structural configuration information of the wet vibrating sieve device for powder. The initial scheme determination module is used to determine multiple initial test wet excitation screening schemes based on the original dry scan image of the powder to be screened, the premixed slurry characteristic data of the powder to be screened, and the structural configuration information of the wet vibration screening device for the powder. The wet excitation screening schemes include vibration frequency and spray flow rate. The vibration screening trial operation module is used to acquire trial operation videos of the wet vibration screening device for the powder premixed slurry to be screened under each initial test wet vibration screening scheme; The status information determination module is used to determine the screening operation status information under each initial test wet vibration screening scheme based on the trial operation video of the powder wet vibration screening device under each initial test wet vibration screening scheme of the powder premixed slurry to be screened. The first scheme generation module is used to generate multiple first wet vibration screening schemes based on the screening operation status information under each initial test wet vibration screening scheme. The video acquisition module is used to acquire the operation video of the wet vibrating sieving device for the powder premixed slurry to be screened under each first wet excitation screening scheme. The preferred scheme determination module is used to determine the preferred wet vibration screening scheme based on the operation video of the wet vibration screening device of the powder premixed slurry to be screened under each first wet vibration screening scheme. The screening control module is used to control the wet vibration screening device to perform full-volume wet vibration screening of the powder premixed slurry to be screened based on the preferred wet vibration screening scheme.

6. The control system of the image analysis-based wet vibrating sieving device for powder as described in claim 5, characterized in that, The first scheme generation module is also used for: Based on the screening operation status information under each initial test wet excitation screening scheme, K mesh-through condition characterization clusters are obtained. Based on the K mesh-through condition characterization clusters, the critical minimum vibration frequency for preventing sticking, the critical minimum spray flow rate for preventing sticking, the critical maximum vibration frequency for preventing material leakage, and the critical maximum spray flow rate for preventing material leakage are determined. Based on the screening operation status information under each initial test wet vibration screening scheme, the critical minimum vibration frequency for preventing sticking, the critical minimum spray flow rate for preventing sticking, the critical maximum vibration frequency for preventing material leakage, and the critical maximum spray flow rate for preventing material leakage, multiple first wet vibration screening schemes are generated.

7. The control system of the image analysis-based wet vibrating sieving device for powder as described in claim 5, characterized in that, The preferred solution determination module is also used for: A wet vibration screening map is constructed, which includes multiple nodes of the first wet vibration screening scheme and multiple edges between the nodes. The node features of each first wet vibration screening scheme node include the operation video of the wet vibration screening device of the powder premixed slurry to be screened under each first wet vibration screening scheme, and the edges are the vibration frequency difference and spray flow rate difference between the nodes of the first wet vibration screening scheme. The wet excitation screening spectrum is processed based on a graph neural network to determine multiple second wet excitation screening schemes. Based on the trial operation video of the powder wet vibrating sieve device under each initial test wet excitation screening scheme, the operation video of the powder wet vibrating sieve device under each first wet excitation screening scheme, and the multiple second wet excitation screening schemes, an operation simulation video of the powder wet vibrating sieve device under each second wet excitation screening scheme is generated. The optimal wet vibration screening scheme is determined based on the operation simulation video of the powder wet vibration screening device under each of the second wet vibration screening schemes.

8. The control system of the image analysis-based wet vibrating sieve for powder as described in claim 5, characterized in that, The K mesh-through condition characterization clusters obtained by clustering the screening operation status information based on each initial test wet-excitation screening scheme include: Based on the screening operation status information under each initial test wet excitation screening scheme, K-means clustering algorithm is used to obtain K mesh-through condition characterization clusters.

9. An electronic device, characterized in that, include: processor; Memory; And a computer program; wherein the computer program is stored in the memory and configured to be executed by the processor to implement the control method of the image analysis-based wet vibrating sieve for powder as described in any one of claims 1 to 4.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the control method for the image analysis-based wet vibrating sieve device for powder as described in any one of claims 1 to 4.

Images