Sorting system and method for cemented carbide

Abstract

This disclosure provides a sorting system for cemented carbide, comprising a content detection device, a magnetic detection device, and a data processor. The content detection device detects and obtains cobalt content data of the alloy material; the magnetic detection device, connected to the content detection device, detects and obtains coercivity data of the alloy material; the data processor is coupled to both the content detection device and the magnetic detection device, and performs data analysis on the alloy material based on the cobalt content data and coercivity data. Furthermore, the data processor uses a clustering analysis model to determine the quality group to which the alloy material belongs. This data processing method assists personnel in sorting alloy materials, effectively improving the efficiency and quality of alloy material sorting and avoiding the impact of human judgment differences or lack of experience on the sorting results.

Description

Technical Field

[0001] This disclosure relates to a sorting system, and more particularly to a sorting system and method for cemented carbide. Prior Technology

[0002] Cemented carbide, also known as hard alloy, generally possesses properties such as high hardness, high melting point, wear resistance, corrosion resistance, and excellent thermal stability, making it widely used in wear-resistant parts such as cutting tools, molds, and drill bits. Tungsten carbide is the main component of cemented carbide, and tungsten is a scarce resource with high economic value. Therefore, the recycling and reuse of tungsten carbide materials has become a key objective for the cemented carbide industry.

[0003] To effectively utilize resources, industry players generally place great emphasis on the recycling, sorting, and remanufacturing processes of waste tungsten carbide parts. However, existing sorting technologies mostly rely on manual identification, which is not only inefficient but also prone to inconsistent sorting quality due to subjective judgment and experience differences, thus affecting the reliability of subsequent remanufacturing of cemented carbide. Summary of the Invention

[0004] The purpose of this disclosure is to provide a sorting system and method for cemented carbide that can efficiently determine the quality group to which the alloy material belongs, while also ensuring the accuracy of sorting the alloy material.

[0005] The foregoing objectives do not preclude the existence of other objectives. Objectives that can be derived by a person skilled in the art from the description, claims, or drawings are also included in the objectives disclosed herein.

[0006] To achieve the above objectives, this disclosure provides a sorting system for cemented carbide, comprising a content detection device, a magnetic detection device, and a data processor. The content detection device detects and obtains cobalt content data of an alloy material; the magnetic detection device is connected to the content detection device and detects and obtains coercivity data of the alloy material; the data processor is coupled to the content detection device and the magnetic detection device, and performs data analysis on the alloy material based on the cobalt content data and the coercivity data, and determines the quality group to which the alloy material belongs based on a clustering analysis model.

[0007] In one embodiment of this disclosure, a method for selecting cemented carbide is provided, comprising:

[0008] Step S1: Obtain cobalt content data for an alloy material.

[0009] Step S2: Obtain coercive magnetic force data of the alloy material.

[0010] Step S3: Analyze the alloy material based on the cobalt content data and coercivity data, and determine the quality group to which the alloy material belongs.

[0011] In summary, this disclosure assists personnel in selecting alloy materials through data processing, thereby effectively improving the efficiency and quality of alloy material classification and avoiding the impact of human judgment differences or lack of experience on classification results. Furthermore, it also helps control the stability of the quality of recycled alloy material powder, thus improving the reliability of alloy material recycling and reuse. Simple Explanation of the Diagram

[0012] Figure 1 is a block diagram of a picking system according to an embodiment of the present disclosure, especially a diagram showing the connection relationship of various objects. Figure 2 is a characteristic scattering diagram (a) of one embodiment of this disclosure. Figure 3 is a characteristic scattering diagram (II) of one embodiment of this disclosure. Figure 4 is a characteristic scattering diagram (III) of one embodiment of this disclosure. Figure 5 is a schematic flowchart of a sorting method according to one embodiment of this disclosure (I). Figure 6 is a schematic flowchart of a picking method according to one embodiment of this disclosure (II). Implementation

[0013] To facilitate the description of the central ideas expressed in the above-described technical content, specific embodiments are provided below. The various elements of these embodiments are depicted to a scale, size, deformation, or displacement suitable for illustration, and are not drawn to scale with actual elements; this will be stated prior to the description.

[0014] The exemplary embodiments of this disclosure are described in detail below with reference to the accompanying drawings. It is not intended to limit the technical principles of this disclosure to the specific disclosed embodiments, but the scope of this disclosure is limited only by the scope of the claims and covers alternatives, modifications and equivalents.

[0015] Please refer to Figures 1 to 4. In this embodiment of the disclosure, a cemented carbide picking system 100 is provided.

[0016] Please refer to Figure 1. The sorting system 100 includes a content detection device 10, a magnetic detection device 20, and a data processor 30.

[0017] The content detection device 10 detects and obtains the cobalt content data of an alloy material. The cobalt content can be used to assess the toughness of the alloy material, and thus determine the manufacturing strength of the workpiece.

[0018] Magnetic detection device 20 is connected to content detection device 10. Magnetic detection device 20 detects and obtains coercive magnetic force data of the alloy material. The coercive magnetic force can be used to evaluate the quality of the alloy material, reflecting the changes in tungsten carbide particle size, distribution, and carbon content in the workpiece, thereby determining the workpiece's process quality.

[0019] The data processor 30 is coupled to the content detection device 10 and the magnetic detection device 20. The data processor 30 is configured in a terminal device. The data processor 30 performs data analysis on the alloy material based on the cobalt content data and coercivity data, and the data processor 30 determines the quality group to which the alloy material belongs based on a clustering analysis model, that is, selects alloy materials with different manufacturing strengths and process qualities.

[0020] Referring to Figure 1, the data processor 30 has a preprocessing unit 31. The preprocessing unit 31 preloads cobalt content data and coercivity data of a plurality of alloy materials as the basis for establishing a clustering analysis model. The preprocessing unit 31 converts the plurality of cobalt content data and coercivity data into a characteristic scatter plot; wherein, in the characteristic scatter plot, the X-axis represents the coercivity data, and the Y-axis represents the cobalt content data. The plurality of data will present a plurality of points distributed across the characteristic scatter plot. The preprocessing unit 31 identifies different quality groups in the characteristic scatter plot, and then establishes a clustering analysis model.

[0021] Please refer to Figure 2. In one embodiment of this disclosure, the preprocessing unit 31 randomly selects 10 points as the centers of each group in the characteristic scatter plot (the star points in Figure 2). The preprocessing unit 31 calculates the distance between each center and other points using the Euclidean distance algorithm. Based on the calculation results, the preprocessing unit 31 assigns each point to the nearest center, and the preprocessing unit 31 continues to assign points until different quality groups are classified, that is, each point finds its corresponding quality group. As shown in Figure 2, the alloy material is divided into six groups, and the standard deviation of the cobalt content needs to be within a certain range. The standard deviation of the coercivity needs to be between 0.5 and 0.5. Between 1.5 and 1.

[0022] Furthermore, the preprocessing unit 31 allocates the points using a centroid allocation equation. The centroid allocation equation is as follows: Let Ck represent the dataset of the k-th group, rik indicate that each point belongs to group k, xi represent each point, and μk represent the center of the k-th group. First, the preprocessing unit 31 assigns each point to the nearest center based on the calculation results; when rik = 1, k = argminj‖xi - μj‖, otherwise rik = 0. Next, the preprocessing unit 31 updates the center of each group, μk = Until the center of the community no longer changes.

[0023] Please refer to Figure 3. In one embodiment of this disclosure, the preprocessing unit 31 randomly selects 10 points as the centers of each group in the characteristic scatter plot (the cross points in Figure 3). The preprocessing unit 31 calculates the distance between each center and other points using the Euclidean distance algorithm, and obtains the membership value of each point for all groups based on the calculation results; wherein the membership value is between 0 and 1. When the sum of the membership values ​​of a point equals a preset value, the preprocessing unit 31 re-finds the centers of each group based on the membership values ​​of each point; wherein the preset value is 1. Furthermore, the preprocessing unit 31 recalculates the membership values ​​of each point until the preprocessing unit 31 classifies different quality groups, that is, the objective function converges. As shown in Figure 3, the characteristic scatter plot shows that the boundaries of each group are clear and there is no overlap. Therefore, this approach is suitable for processing groups with relatively ambiguous boundaries and allows each point to have multiple group characteristics simultaneously.

[0024] Furthermore, the preprocessing unit 31 calculates the membership value of each point using a centroid membership equation. The centroid membership equation is Jm= Jm is the objective function, uik represents the relationship of each point to the k-th group, xi represents each point, and vk represents the center of the k-th group. Next, the preprocessing unit 31 updates the centers of each group and the relationship of each point to the k-th group, vk = ,uik= The classification of quality groups is completed when the objective function converges.

[0025] Please refer to Figure 4. In one embodiment of this disclosure, the preprocessing unit 31 defines a neighborhood radius centered on each point in the characteristic scatter plot. As shown in Figure 4, when the number of points within the neighborhood radius is greater than a preset number of points, the preprocessing unit 31 in Figure 1 defines a point as a core point (the crosshair in Figure 4), and points falling within the neighborhood radius of the core point are classified into the same quality group. Points falling within the neighborhood radius of the core point but not core points are defined as boundary points. Points falling outside the neighborhood radius of the core point and unable to become a group are defined as noise points. Furthermore, if there are other core points within the neighborhood radius, the preprocessing unit 31 will further expand, incorporating a larger neighborhood centered on these new core points, forming a density-connected region expansion, until no more density points can be added.

[0026] It should be noted that the neighborhood radius can be adjusted according to the actual situation. Different neighborhood radii affect the size of the neighborhood, thus determining the density of the population. The preset number of points affects the density of the neighborhood; a higher preset number tends to find denser populations, while a lower preset number tends to find lower-density populations. In this way, this method can identify irregularly shaped populations and has good noise resistance.

[0027] Furthermore, the preprocessing unit 31 classifies each point into the same quality group using a centroid density equation. The centroid density equation is Nε(x)={y∣‖xy‖≦ε}, where Nε(x) is the core point and ε is the region enclosed by the neighborhood radius.

[0028] In one embodiment of this disclosure, the data processor 30 includes a marking unit 32 and an analysis unit 33. The marking unit 32 is coupled to the preprocessing unit 31 and the analysis unit 33. The marking unit 32 converts the detected cobalt content data and coercivity data into a target to be selected, and marks the target to be selected on a characteristic scatter plot, i.e., the target to be selected will appear in the characteristic scatter plot. The analysis unit 33 determines the quality group to which the alloy material belongs based on the target to be selected, the characteristic scatter plot, and a clustering analysis model. In this way, the quality group to which the alloy material belongs is selected efficiently.

[0029] In one embodiment of this disclosure, the data processor 30 has a suggestion unit 34. The suggestion unit 34 is coupled to the analysis unit 33. The suggestion unit 34 of this disclosure can provide picking suggestions, enabling operators to complete picking operations quickly and consistently even without professional experience.

[0030] Please refer to Figure 5. In one embodiment of this disclosure, a method 200 for selecting cemented carbide is provided, which includes:

[0031] Step S1: Obtain cobalt content data for an alloy material.

[0032] Step S2: Obtain coercive magnetic force data of the alloy material.

[0033] Step S3: Analyze the alloy material based on the cobalt content data and coercivity data, and determine the quality group to which the alloy material belongs.

[0034] Please refer to Figure 6. There is a step A1 before step S1 and a step A2 after step S3. In step A1, the alloy material is recovered and preliminarily sorted. In step A2, the sorted alloy material is remanufactured, for example, through a powdering process.

[0035] In summary, this disclosure has at least the following technical features:

[0036] I. This disclosure combines data processing technology to assist personnel in selecting and judging alloy materials. This not only effectively improves the efficiency and quality of alloy material classification but also avoids inconsistencies caused by subjective human judgment. Furthermore, it ensures the stability of alloy material reprocessing.

[0037] Second, this disclosure discloses that the previous processing unit 31 integrates multiple basic data and establishes a unified data standard, so that the picking system 100 has the same judgment criteria, thereby improving the reliability of the picking process.

[0038] Third, regardless of whether the centroid assignment algorithm, centroid membership algorithm, or centroid density algorithm is used in this disclosure, the central value, standard deviation, and maximum difference of cobalt content, as well as the central value, standard deviation, and maximum difference of coercivity, can be obtained for each group. Furthermore, each algorithm can optimize the data in the data scatter plot, thereby accurately classifying different quality groups.

[0039] The aforementioned technical features do not preclude the existence of other features. Features that can be derived by a person skilled in the art from the description, claims, or drawings disclosed herein are also included in the features disclosed herein.

[0040] In summary, the embodiments described herein are merely for illustrating the technology disclosed herein and are not intended to limit the scope of any patent application. All modifications or variations made without departing from the spirit of this disclosure are within the scope of protection intended by this disclosure.

[0041] 100: Picking System 10: Content detection device 20: Magnetic detection device 30: Data Processor 31: Preprocessing Unit 32: Marker unit 33: Analysis Unit 34: Recommendation Unit 200: Picking Method A1: Steps A2: Steps S1: Steps S2: Steps S3: Steps

Claims

1. A sorting system for cemented carbide, comprising: a content detection device for detecting and acquiring cobalt content data of an alloy material; a magnetic detection device connected to the content detection device, the magnetic detection device for detecting and acquiring coercivity data of the alloy material; and a data processor coupled to the content detection device and the magnetic detection device, the data processor performing data analysis on the alloy material based on the cobalt content data and the coercivity data, and the data processor determining the quality group to which the alloy material belongs based on a clustering analysis model; wherein, The data processor has a preprocessing unit that preloads a plurality of data on the cobalt content and coercivity of the alloy material. The preprocessing unit converts the cobalt content and coercivity data into a characteristic scatter plot. The preprocessing unit then identifies different quality groups in the characteristic scatter plot and establishes the cluster analysis model.

2. The cemented carbide sorting system as described in claim 1, wherein, The preprocessing unit randomly assigns 10 points as the center of each group in the characteristic scatter plot. The preprocessing unit calculates the distance between each center and other points. Based on the calculation results, the preprocessing unit assigns each point to the nearest center. The preprocessing unit continues to assign points until different quality groups are classified.

3. The carbide sorting system as described in claim 2, wherein, The preprocessing unit assigns each point using a centroid allocation equation, where Ck represents the data set of the k-th group, rik represents each point belonging to group k, xi represents each point, and μk represents the center of the k-th group.

4. The cemented carbide sorting system as described in claim 1, wherein, The preprocessing unit randomly selects 10 points as the center of each group in the characteristic scatter plot. The preprocessing unit calculates the distance between each center and other points, and obtains the membership value of each point for all groups based on the calculation results. When the sum of the membership values ​​of a point equals a preset value, the preprocessing unit finds the center of each group again based on the membership values ​​of each point, and recalculates the membership values ​​of each point until the preprocessing unit classifies different quality groups.

5. The carbide sorting system as described in claim 4, wherein, The preprocessing unit calculates the membership value of each point using a centroid membership equation, which is Jm = Jm, where Jm is the objective function, uik represents the relationship of each point to the k-th group, xi represents each point, and vk represents the center of the k-th group.

6. The cemented carbide sorting system as described in claim 1, wherein, The preprocessing unit defines a neighborhood radius for each point in the characteristic scatter plot. When the number of points in the area enclosed by the neighborhood radius is greater than a preset number of points, the preprocessing unit defines the point as a core point, and the points in the area enclosed by the neighborhood radius of the core point are classified into the same quality group.

7. The carbide sorting system as described in claim 6, wherein, The preprocessing unit classifies each point into the same quality group using a centroid density equation, which is Nε (x)=‖x-y‖≦ε, where Nε (x) is the core point and ε is the region enclosed by the neighborhood radius.

8. A cemented carbide sorting system as described in claim 1, wherein, The data processor has a marking unit and an analysis unit. The marking unit is coupled to the preprocessing unit and the analysis unit. The marking unit converts the detected cobalt content data and the coercivity data into a target to be selected. The marking unit marks the target to be selected on the characteristic scatter plot. The analysis unit determines the quality group to which the alloy material belongs based on the target to be selected, the characteristic scatter plot and the cluster analysis model.

9. A method for selecting cemented carbide, comprising: step S1: detecting and obtaining cobalt content data of an alloy material; step S2: detecting and obtaining coercivity data of the alloy material; and step S3: pre-loading a plurality of cobalt content data and coercivity data of the alloy material, converting the cobalt content data and coercivity data into a characteristic scatter plot, identifying different quality groups in the characteristic scatter plot, establishing a clustering analysis model, performing data analysis on the alloy material based on the detected cobalt content data and the detected coercivity data, and determining the quality group to which the alloy material belongs based on the clustering analysis model.

Images