Micropore defect detection and identification device and method

Abstract

The invention relates to the technical field of detection, in particular to a micropore defect detection and identification device and method. Performing feature analysis on the region feature model graph of the micropore region to obtain a defect concentration coefficient of the micropore region; judging and analyzing the micropore area according to the defect concentration coefficient, and judging whether the defect concentration of the micropore area meets the requirement or not; constructing a knowledge database, and if the micropore area is the area with the defect concentration not meeting the requirement, obtaining the defect name of the defect existing in the micropore area based on the knowledge database, so that the micropore area of the target device can be subjected to comprehensive defect detection and identification, and the quality of the micropore area is evaluated in a quantitative mode; therefore, the quality control level of the micropore area of the target device can be improved, and the defect problem of the micropore area can be found and processed in time.

Description

Technical Field

[0001] The present invention relates to the field of detection technology, and in particular to a device and method for detecting and identifying micro-pore defects. Background Art

[0002] With the growing demand for precision manufacturing in aerospace, automobile, energy and chemical industries, microporous structures play a vital role in many key components, and their quality and performance are directly related to the reliability and safety of equipment. However, in the manufacturing process of precision casting, laser processing, etc., the inner wall surface of the micropore is prone to tiny cracks, pores, pits and other defects. The existence of these defects will seriously affect the performance of the micropore and even lead to equipment failure or safety accidents. Acoustic wave detection technology has gradually become an important means of micropore defect detection due to its advantages of non-contact, high sensitivity and wide application range. By collecting the acoustic wave data fed back by the micropore area and combining it with signal processing and recognition technology, various defects on the inner wall surface of the micropore can be effectively identified. However, the actual collected acoustic wave data is often mixed with various noises, such as environmental noise, equipment noise, etc. These noises will interfere with the extraction of defect features and reduce the detection accuracy. In addition, different types of defects show different characteristics in the acoustic wave signal. How to effectively extract and identify these features and associate them with known defect types is also a challenge faced by the field of micropore defect detection and recognition. Summary of the invention

[0003] The present invention overcomes the problem of low detection accuracy in the prior art and provides a device and method for detecting and identifying micro-pore defects.

[0004] To achieve the above-mentioned purpose, the technical solution adopted by the present invention is:

[0005] The present invention discloses a method for detecting and identifying micropore defects, comprising the following steps:

[0006] Acquire acoustic wave data fed back from a micropore region in a target device, perform denoising on the acoustic wave data, acquire denoised acoustic wave data, and construct a regional feature model diagram of the micropore region based on the denoised acoustic wave data;

[0007] Performing characteristic analysis on the regional characteristic model diagram of the micropore region to obtain a defect concentration coefficient of the micropore region;

[0008] According to the defect concentration coefficient, the micropore region is judged and analyzed to determine whether the defect concentration of the micropore region meets the requirements;

[0009] A knowledge database is constructed. If the micropore region is a region where the defect concentration does not meet the requirements, the defect name of the defect in the micropore region is obtained based on the knowledge database; and the defect name of the defect in the micropore region is uploaded to a preset terminal.

[0010] Preferably, the acoustic wave data fed back from the micropore region in the target device is obtained, the acoustic wave data is denoised to obtain the denoised acoustic wave data, and a regional characteristic model diagram of the micropore region is constructed according to the denoised acoustic wave data, specifically:

[0011] Scanning and detecting the micro-hole area in the target device by using an ultrasonic detector, obtaining the acoustic wave data fed back by the micro-hole area, and dividing the acoustic wave data fed back by the micro-hole area into several frequency bands;

[0012] Obtain the frequency value of the acoustic wave data in each frequency band; and preset the preset frequency value range of the acoustic wave data, and traverse the acoustic wave data fed back from the micro-pore area in the order of the acquisition frequency bands;

[0013] Construct a multi-branch tree. If the frequency value of the sound wave data of a certain frequency band is not within the preset frequency value range, the sound wave data of the frequency band is divided into a child node of the multi-branch tree; if the frequency value of the sound wave data of a certain frequency band is within the preset frequency value range, the sound wave data of the frequency band is divided into another child node of the multi-branch tree; and a division criterion is formed in this way;

[0014] According to the division criterion, the child nodes are continuously divided, and after the division is completed, a multi-branch tree structure is formed; wherein each child node of the multi-branch tree structure represents a subset of the original sound wave data, and as the division proceeds, the sound wave data in the child nodes gradually tend to be similar in characteristics;

[0015] Acquire data features of normal sound wave data through a big data network, and acquire data features of sound wave data on each child node in a multi-branch tree structure; wherein the data features include frequency, amplitude, and phase;

[0016] Calculate the cosine similarity between the data features of the sound wave data on each child node and the data features of the normal sound wave data based on the cosine similarity algorithm;

[0017] The sub-nodes whose cosine similarity is not greater than the preset similarity threshold are marked as noise data sub-nodes, and the sub-nodes marked as noise data are pruned in the multi-branch tree structure;

[0018] Acquire the acoustic wave data of the remaining child nodes in the multi-branch tree structure to obtain denoised acoustic wave data; and construct a regional feature model diagram of the micropore area based on the denoised acoustic wave data.

[0019] Preferably, a characteristic analysis is performed on the regional characteristic model diagram of the micropore region to obtain the defect concentration coefficient of the micropore region, specifically:

[0020] Obtaining standard structural drawing information of a target device, and obtaining a standard characteristic model diagram of a micro-hole region according to the standard structural drawing information;

[0021] Preset voxel size, voxelize the regional feature model map according to the preset voxel size; and voxelize the standard feature model map according to the preset voxel size;

[0022] Traverse all voxels in the two voxelized regional feature model maps and the standard feature model map;

[0023] If both the regional feature model map and the standard feature model map exist on a certain voxel, the voxel is marked as a coincident voxel;

[0024] If only a regional feature model map exists on a certain voxel, or only a standard feature model map exists, the voxel is marked as a non-coincident voxel;

[0025] If neither the regional feature model map nor the standard feature model map exists on a certain voxel, the voxel is marked as a blank voxel;

[0026] The total number of all non-overlapping voxels and the total number of overlapping voxels and non-overlapping voxels are calculated; the total number of non-overlapping voxels and the total number of overlapping voxels and non-overlapping voxels are ratio-processed to obtain the defect concentration coefficient of the micropore area.

[0027] Preferably, the micropore region is judged and analyzed according to the defect concentration coefficient to determine whether the defect concentration of the micropore region meets the requirements, specifically:

[0028] Presetting a defect concentration coefficient threshold, and comparing the defect concentration coefficient of the micropore area with the defect concentration coefficient threshold;

[0029] If the defect concentration coefficient of the micropore region is not greater than the defect concentration coefficient threshold, marking the micropore region as a region where the defect concentration meets the requirement;

[0030] If the defect concentration coefficient of the micropore region is greater than the defect concentration coefficient threshold, the micropore region is marked as a region where the defect concentration does not meet the requirements.

[0031] Preferably, the method for constructing the knowledge database is:

[0032] Retrieve the characteristic model diagrams corresponding to various micropore defects through the big data network and build a blank database;

[0033] Obtaining the number of defect types of micropore defects retrieved through the big data network, and dividing the blank database into a corresponding number of sub-databases according to the number of defect types;

[0034] and obtaining the defect type name of the micropore defect retrieved through the big data network, and converting the defect type name of each micropore defect into a unique binary digital code;

[0035] Each digital code is compressed and bundled with the corresponding feature model diagram to obtain a number of data packets, each of which is stored in a corresponding sub-database to form a knowledge database; and the knowledge database is updated regularly.

[0036] Preferably, if the micropore region is a region where the defect concentration does not meet the requirement, the defect name of the defect existing in the micropore region is obtained based on the knowledge database, specifically:

[0037] If the micropore region is a region where the defect concentration does not meet the requirements, then the model region corresponding to the voxels marked as non-overlapping between the regional feature model map and the standard feature model map is obtained to obtain a defect model map of the defects existing in the micropore region;

[0038] and extracting the characteristic model graph in the data packet from the knowledge database; introducing the cosine similarity algorithm, and calculating the cosine similarity between the defect model graph of the defect existing in the micropore area and the characteristic model graph in each data packet based on the cosine similarity algorithm;

[0039] Arrange multiple cosine similarities based on their numerical values, and after the arrangement is completed, obtain the maximum cosine similarity;

[0040] The data packet corresponding to the maximum cosine similarity is marked, and the digital code in the marked data packet is extracted. The extracted digital code is subjected to reverse identification processing to obtain the defect name of the defect existing in the micro-hole area.

[0041] Another aspect of the present invention discloses a micropore defect detection and identification device, which is applied to any of the micropore defect detection and identification methods described above, and the micropore defect detection and identification device comprises:

[0042] A test bench for fixing the target device;

[0043] A loading module, used to transfer the target device to the testing table;

[0044] The acoustic wave data acquisition module is responsible for collecting the acoustic wave data fed back by the micro-hole area in the target device. The acoustic wave data acquisition module includes an ultrasonic detector, a signal amplifier and a data acquisition card, which are used to obtain the acoustic information of the micro-hole area;

[0045] The sound wave data processing module performs denoising on the collected sound wave data to obtain the denoised sound wave data;

[0046] A feature model building module builds a regional feature model diagram of the micropore area based on the denoised acoustic wave data;

[0047] The feature analysis module performs feature analysis on the regional feature model diagram of the micropore area to obtain the defect concentration coefficient of the micropore area;

[0048] The defect judgment module judges and analyzes the micropore area according to the defect concentration coefficient to determine whether the defect concentration of the micropore area meets the requirements.

[0049] The present invention solves the technical defects existing in the background technology, and has the following beneficial effects: performing feature analysis on the regional feature model diagram of the micropore area to obtain the defect concentration coefficient of the micropore area; performing judgment analysis on the micropore area according to the defect concentration coefficient to determine whether the defect concentration of the micropore area meets the requirements; constructing a knowledge database, if the micropore area is an area where the defect concentration does not meet the requirements, then obtaining the defect name of the defect existing in the micropore area based on the knowledge database, the present invention can perform comprehensive defect detection and identification on the micropore area of ​​the target device, evaluate the quality of the micropore area in a quantitative manner, and accurately determine whether it is qualified, thereby helping to improve the quality control level of the micropore area of ​​the target device, and timely discover and deal with defect problems in the micropore area. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, drawings of other embodiments can be obtained based on these drawings without paying creative work.

[0051] Figure 1 This is the overall method flow chart of the micro-pore defect detection and identification method;

[0052] Figure 2 This is a partial method flow chart of the micropore defect detection and identification method;

[0053] Figure 3 This is a simplified structural diagram of the micro-pore defect detection and identification device. DETAILED DESCRIPTION

[0054] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0055] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0056] The present invention discloses a method for detecting and identifying micropore defects. Figure 1 As shown, the following steps are included:

[0057] S102: Acquire acoustic wave data fed back from a micropore region in a target device, perform denoising on the acoustic wave data to acquire denoised acoustic wave data, and construct a regional feature model diagram of the micropore region according to the denoised acoustic wave data;

[0058] S104: performing feature analysis on the regional feature model diagram of the micropore region to obtain a defect concentration coefficient of the micropore region;

[0059] S106: analyzing the micropore region according to the defect concentration coefficient to determine whether the defect concentration of the micropore region meets the requirement;

[0060] S108: construct a knowledge database. If the micropore region is a region where the defect concentration does not meet the requirements, obtain the defect name of the defect in the micropore region based on the knowledge database; and upload the defect name of the defect in the micropore region to a preset terminal.

[0061] It should be noted that, first, the acoustic wave data fed back from the micropore area in the target device is obtained. The acoustic wave data is the original information source for detecting micropore defects. Then, the acquired acoustic wave data is denoised. Denoising is to improve data quality and remove noise components that may interfere with subsequent analysis, thereby obtaining denoised acoustic wave data. Then, the denoised acoustic wave data is used to construct a regional characteristic model diagram of the micropore area. This model diagram can present the characteristics of the micropore area in an intuitive way, laying the foundation for subsequent analysis. The constructed regional characteristic model diagram of the micropore area is subjected to feature analysis, and the defect concentration coefficient of the micropore area is obtained through specific algorithms and analysis methods. This coefficient is a quantitative indicator for measuring the degree of defects in the micropore area, which can reflect information such as the relative concentration of defects in the area. According to the obtained defect concentration coefficient, the micropore area is judged and analyzed, and the coefficient is compared with the pre-set requirements to determine whether the defect concentration of the micropore area meets the requirements. This step can determine the overall quality status of the micropore area, whether it is qualified or unqualified. A knowledge database is constructed, which contains relevant knowledge about various defects that may exist in micropores. When the micropore area is determined to be an area where the defect concentration does not meet the requirements, the defect name of the defect in the micropore area can be obtained based on this knowledge database. Finally, the defect name of the defect in the micropore area is uploaded to a preset terminal, which can be a monitoring device, a management system, etc., so that relevant personnel can obtain the defect information of the micropore area in a timely manner and take corresponding measures.

[0062] In summary, this micropore defect detection and identification method obtains acoustic wave data, denoises and constructs a model diagram, and analyzes the characteristics to obtain the defect concentration coefficient, and then determines whether the defect concentration meets the requirements. When the defect concentration does not meet the requirements, the defect name is identified using the knowledge database and uploaded to the preset terminal. It can perform comprehensive defect detection and identification on the micropore area of ​​the target device, evaluate the quality of the micropore area in a quantitative manner (defect concentration coefficient), accurately determine whether it is qualified, and when it is unqualified, it can clarify the defect name and promptly feedback the information to the designated terminal, thereby helping to improve the quality control level of the micropore area of ​​the target device and promptly discover and deal with defects in the micropore area.

[0063] Preferably, the acoustic wave data fed back from the micropore region in the target device is obtained, the acoustic wave data is denoised to obtain the denoised acoustic wave data, and a regional characteristic model diagram of the micropore region is constructed according to the denoised acoustic wave data, specifically:

[0064] Scanning and detecting the micro-hole area in the target device by using an ultrasonic detector, obtaining the acoustic wave data fed back by the micro-hole area, and dividing the acoustic wave data fed back by the micro-hole area into several frequency bands;

[0065] Obtain the frequency value of the acoustic wave data in each frequency band; and preset the preset frequency value range of the acoustic wave data, and traverse the acoustic wave data fed back from the micro-pore area in the order of the acquisition frequency bands;

[0066] Construct a multi-branch tree. If the frequency value of the sound wave data of a certain frequency band is not within the preset frequency value range, the sound wave data of the frequency band is divided into a child node of the multi-branch tree; if the frequency value of the sound wave data of a certain frequency band is within the preset frequency value range, the sound wave data of the frequency band is divided into another child node of the multi-branch tree; and a division criterion is formed in this way;

[0067] According to the division criterion, the child nodes are continuously divided, and after the division is completed, a multi-branch tree structure is formed; wherein each child node of the multi-branch tree structure represents a subset of the original sound wave data, and as the division proceeds, the sound wave data in the child nodes gradually tend to be similar in characteristics;

[0068] Acquire data features of normal sound wave data through a big data network, and acquire data features of sound wave data on each child node in a multi-branch tree structure; wherein the data features include frequency, amplitude, and phase;

[0069] Calculate the cosine similarity between the data features of the sound wave data on each child node and the data features of the normal sound wave data based on the cosine similarity algorithm;

[0070] The sub-nodes whose cosine similarity is not greater than the preset similarity threshold are marked as noise data sub-nodes, and the sub-nodes marked as noise data are pruned in the multi-branch tree structure;

[0071] Acquire the acoustic wave data of the remaining child nodes in the multi-branch tree structure to obtain denoised acoustic wave data; and construct a regional feature model diagram of the micropore area based on the denoised acoustic wave data.

[0072] It should be noted that an ultrasonic detector is used to scan and detect the micro-hole area in the target device to obtain the acoustic wave data fed back from the micro-hole area. This is the starting data source for the entire detection process. The acquired acoustic wave data is then divided into several frequency bands. This frequency band division provides a more detailed basis for subsequent analysis. Then the frequency values ​​of the acoustic wave data of each frequency band are obtained, and the preset frequency value range of the acoustic wave data is preset. Traversing the acoustic wave data fed back from the micro-hole area in the order of the acquisition frequency bands is helpful for systematically analyzing the acoustic wave data, and the preset frequency value range provides a basis for judging whether the data is normal. Construct a multi-branch tree structure, and divide the frequency band acoustic wave data into different child nodes of the multi-branch tree according to the relationship between the frequency value of the frequency band acoustic wave data and the preset frequency value range. If the frequency value of the acoustic wave data of a certain frequency band is not within the preset frequency value range, it is divided into a child node; if it is within the range, it is divided into another child node, thereby forming a division criterion. The child nodes are continuously divided according to the division criterion, and as the division proceeds, the multi-branch tree structure is gradually formed. In this process, each child node represents a subset of the original sound wave data, and the sound wave data in the child nodes gradually become similar in features. This feature convergence helps to classify and analyze the data more accurately. The data features (including frequency, amplitude and phase) of normal sound wave data are obtained through the big data network, and the data features of the sound wave data on each child node in the multi-branch tree structure are obtained at the same time. Then, the cosine similarity between the data features of the sound wave data on each child node and the data features of the normal sound wave data is calculated based on the cosine similarity algorithm. Cosine similarity is an indicator of the similarity between two vectors (here, data feature vectors). The child nodes whose cosine similarity is not greater than the preset similarity threshold are marked as noise data child nodes, which means that the sound wave data in these child nodes are judged to be noise. Then, the child nodes marked as noise data are cut in the multi-branch tree structure to remove these noise data, thereby obtaining purer sound wave data. The sound wave data of the remaining child nodes in the multi-branch tree structure are obtained, and these data are the denoised sound wave data. Finally, the regional characteristic model diagram of the micropore area is constructed based on the denoised acoustic wave data. This model diagram will provide a basis for subsequent micropore area defect analysis and other operations.

[0073] In summary, this step obtains the acoustic wave data of the micropore area through ultrasonic detection, and after a series of complex operations such as frequency band segmentation, multi-branch tree construction and division, data feature acquisition, and cosine similarity calculation, accurately identifies and removes noise data, obtains denoised acoustic wave data, and then constructs a regional feature model diagram of the micropore area. It can effectively remove noise interference in acoustic wave data, accurately extract effective information related to the characteristics of the micropore area, provide a high-quality data foundation for subsequent micropore defect detection and other work, and improve the accuracy and reliability of micropore area analysis.

[0074] Preferably, a characteristic analysis is performed on the regional characteristic model diagram of the micropore region to obtain the defect concentration coefficient of the micropore region, specifically:

[0075] Obtaining standard structural drawing information of a target device, and obtaining a standard characteristic model diagram of a micro-hole region according to the standard structural drawing information;

[0076] Preset voxel size, voxelize the regional feature model map according to the preset voxel size; and voxelize the standard feature model map according to the preset voxel size;

[0077] Traverse all voxels in the two voxelized regional feature model maps and the standard feature model map;

[0078] If both the regional feature model map and the standard feature model map exist on a certain voxel, the voxel is marked as a coincident voxel;

[0079] If only a regional feature model map exists on a certain voxel, or only a standard feature model map exists, the voxel is marked as a non-coincident voxel;

[0080] If neither the regional feature model map nor the standard feature model map exists on a certain voxel, the voxel is marked as a blank voxel;

[0081] The total number of all non-overlapping voxels and the total number of overlapping voxels and non-overlapping voxels are calculated; the total number of non-overlapping voxels and the total number of overlapping voxels and non-overlapping voxels are ratio-processed to obtain the defect concentration coefficient of the micropore area.

[0082] It should be noted that in order to perform feature analysis on the regional feature model map of the micropore area, it is necessary to obtain the standard structural drawing information of the target device. Based on this standard structural drawing information, a standard feature model map of the micropore area can be obtained. This standard feature model map is a benchmark for comparative analysis with the actual regional feature model map. The voxel size is preset, and then the regional feature model map and the standard feature model map are voxelized according to the preset voxel size. Voxelization is the discretization of the model map into a collection of voxels (volume elements), so that the continuous model map can be converted into a discrete form that is easy for computer processing. Each voxel can be regarded as a basic unit of the model map at the microscopic level. Traverse all voxels in the voxelized regional feature model map and the standard feature model map. Mark the voxels according to their presence in the two model maps. If a voxel exists in both the regional feature model map and the standard feature model map, it is marked as an overlapping voxel; if a voxel only exists in the regional feature model map or only exists in the standard feature model map, it is marked as a non-overlapping voxel; if a voxel does not exist in both model maps, it is marked as a blank voxel. This marking method can clearly distinguish different types of voxels, thus providing a basis for the subsequent calculation of the defect concentration coefficient. Calculate the total number of all non-overlapping voxels, and at the same time calculate the total number of overlapping voxels and non-overlapping voxels. Finally, the total number of non-overlapping voxels is ratioed with the total number of overlapping voxels and non-overlapping voxels to obtain the defect concentration coefficient of the micropore area. This coefficient reflects the degree of difference between the regional feature model map and the standard feature model map. The more non-overlapping voxels, the greater the defect concentration coefficient, which means that the defects in the micropore area may be more serious.

[0083] In summary, this step constructs a standard feature model diagram by acquiring the standard structural drawing information of the target device, voxelizes the actual regional feature model diagram and the standard feature model diagram and marks different types of voxels, and finally calculates the ratio of the number of non-overlapping voxels to the total number of related voxels to obtain the defect concentration coefficient. The degree of defect in the micropore area relative to the standard structure can be measured in a quantitative manner. Through voxelization and voxel comparison analysis, the coefficient reflecting the defect situation in the micropore area can be accurately extracted, which provides a key quantitative indicator for the subsequent judgment of whether the micropore area meets the requirements and further defect analysis.

[0084] Preferably, the micropore region is analyzed and judged according to the defect concentration coefficient to determine whether the defect concentration of the micropore region meets the requirements, such as Figure 2 As shown, specifically:

[0085] S202: Preset a defect concentration coefficient threshold, and compare the defect concentration coefficient of the micropore area with the defect concentration coefficient threshold;

[0086] S204: if the defect concentration coefficient of the micropore region is not greater than the defect concentration coefficient threshold, marking the micropore region as a region where the defect concentration meets the requirement;

[0087] S206: If the defect concentration coefficient of the micropore region is greater than the defect concentration coefficient threshold, the micropore region is marked as a region where the defect concentration does not meet the requirement.

[0088] Preferably, the method for constructing the knowledge database is:

[0089] Retrieve the characteristic model diagrams corresponding to various micropore defects through the big data network and build a blank database;

[0090] Obtaining the number of defect types of micropore defects retrieved through the big data network, and dividing the blank database into a corresponding number of sub-databases according to the number of defect types;

[0091] and obtaining the defect type name of the micropore defect retrieved through the big data network, and converting the defect type name of each micropore defect into a unique binary digital code;

[0092] Each digital code is compressed and bundled with the corresponding feature model diagram to obtain a number of data packets, each of which is stored in a corresponding sub-database to form a knowledge database; and the knowledge database is updated regularly.

[0093] It should be noted that the feature model diagrams corresponding to various micropore defects are retrieved with the help of the big data network, which is the basic data source for building the knowledge database. Then a blank database is constructed, which is like a framework to prepare for subsequent data storage and classification. The number of defect types of micropore defects retrieved through the big data network is obtained, and the corresponding number of sub-databases are cut out in the blank database according to the number of defect types. The purpose of this is to classify and store different types of micropore defects, and each sub-database can be used specifically to store information related to a certain type of micropore defects. Then the defect type name of the micropore defect is obtained, and the defect type name of each micropore defect is converted into a unique binary digital code. This encoding method helps to uniquely identify different defect types and facilitate computer recognition and processing. Then each digital code is compressed and bundled with the corresponding feature model diagram to form several data packets. Through this bundling process, the defect type name (in the form of digital code) and its corresponding feature model diagram are closely linked, so that each data packet contains the key information of a specific micropore defect. Each data packet is stored in the corresponding sub-database to form a knowledge database. The knowledge database constructed in this way has a clear structure and is easy to query and use. In addition, in order to ensure the validity and accuracy of the knowledge database, it should be updated regularly to adapt to the emergence of new micropore defect types or new understandings of existing defect types. Through this step, a knowledge database with a clear structure, complete information and continuous updating is constructed, which can effectively store and manage information related to micropore defects, facilitate the subsequent rapid and accurate identification of defect types according to the defect feature model diagram in micropore area defect detection, and improve the efficiency and accuracy of micropore defect detection and identification.

[0094] Preferably, if the micropore region is a region where the defect concentration does not meet the requirement, the defect name of the defect existing in the micropore region is obtained based on the knowledge database, specifically:

[0095] If the micropore region is a region where the defect concentration does not meet the requirements, then the model region corresponding to the voxels marked as non-overlapping between the regional feature model map and the standard feature model map is obtained to obtain a defect model map of the defects existing in the micropore region;

[0096] and extracting the characteristic model graph in the data packet from the knowledge database; introducing the cosine similarity algorithm, and calculating the cosine similarity between the defect model graph of the defect existing in the micropore area and the characteristic model graph in each data packet based on the cosine similarity algorithm;

[0097] Arrange multiple cosine similarities based on their numerical values, and after the arrangement is completed, obtain the maximum cosine similarity;

[0098] The data packet corresponding to the maximum cosine similarity is marked, and the digital code in the marked data packet is extracted. The extracted digital code is subjected to reverse identification processing to obtain the defect name of the defect existing in the micro-hole area.

[0099] It should be noted that when the micropore area is determined to be an area where the defect concentration does not meet the requirements, the model area corresponding to the non-overlapping voxels between the regional feature model map and the standard feature model map is first obtained. Since the non-overlapping voxels represent the difference from the standard model, the model area corresponding to these difference parts can reflect the defects in the micropore area, thereby obtaining the defect model map of the defects in the micropore area. The feature model map in the data packet is extracted from the knowledge database, and then the cosine similarity algorithm is introduced. This algorithm is used to calculate the cosine similarity between the defect model map of the defects in the micropore area and the feature model map in each data packet. Cosine similarity is an indicator of the similarity between two vectors (here it can be regarded as the feature vector of the defect model map and the feature model map in the data packet), and its value range is between -1 and 1. The closer the value is to 1, the more similar it is. The multiple cosine similarities calculated are arranged according to the numerical value. After the arrangement is completed, the maximum cosine similarity is obtained. The data packet corresponding to the maximum cosine similarity is likely to contain the feature model map that is most similar to the defects in the micropore area. Mark the data packet corresponding to the maximum cosine similarity, and then extract the digital code in the data packet. Since the defect type name was previously converted into a unique binary digital code, it is necessary to perform reverse identification processing on the extracted digital code here to restore the digital code to the corresponding defect name, so as to obtain the defect name of the defect existing in the micropore area. In summary, when the defect concentration in the micropore area is judged to be not meeting the requirements, the defect model diagram is obtained, the cosine similarity is calculated and sorted with the feature model diagram in the knowledge database, the most similar data packet is found, and then the defect name is obtained by reversely identifying the digital code. The defect name existing in the micropore area can be accurately identified. Through matching based on cosine similarity and reverse identification of digital codes, with the support of the knowledge database, specific defect type information is provided for further analysis and processing of defects in the micropore area, which helps to take targeted measures to solve the defect problems in the micropore area.

[0100] In addition, the method further comprises the following steps:

[0101] If the defect concentration coefficient of the micropore region is not greater than the defect concentration coefficient threshold, obtaining the defect name of the defect existing in the micropore region based on the knowledge database;

[0102] According to the defect name of the defect existing in the micropore area and in combination with the finite element analysis method, a finite element analysis is performed on the regional characteristic model diagram of the micropore area to obtain a finite element analysis result;

[0103] Obtaining predicted expansion stress values ​​of defects existing in the micro-hole region according to the finite element analysis results; and obtaining the processing steps required for subsequent processing of the micro-hole region, and obtaining preset processing parameters of the processing equipment corresponding to the processing steps required;

[0104] Determine a preset load value for subsequent processing of the micro-hole region by the corresponding processing equipment according to preset processing parameters of the processing equipment corresponding to the processing steps required to be experienced;

[0105] Determine whether a preset load value when the micro-hole region is subsequently processed by a corresponding processing device is greater than the predicted expansion stress value;

[0106] If it is greater than, a genetic algorithm is introduced, and a genetic algebra is set according to the genetic algorithm, and the preset processing parameters of the corresponding processing equipment are adjusted by the genetic algebra in combination with the preset load force value, until the preset load force value when the corresponding processing equipment performs subsequent processing on the micro-hole area is not greater than the predicted expansion stress value, and the adjusted processing parameters are output;

[0107] When the target device flows into the corresponding processing equipment, the corresponding processing equipment is controlled to process the micro-hole area of ​​the target device according to the adjusted processing parameters.

[0108] It should be noted that when the defect concentration coefficient of the micropore area is not greater than the defect concentration coefficient threshold, the defect name of the defect in the area is obtained based on the knowledge database. Then, the regional characteristic model diagram of the micropore area is analyzed in combination with the finite element analysis method. Finite element analysis is a numerical analysis method that discretizes a continuous structure into a finite number of units to solve an approximate solution to a physical problem. Here, the defect name is used to assist the finite element analysis to obtain the finite element analysis results. According to the finite element analysis results, the predicted expansion stress value of the defect in the micropore area is obtained. At the same time, the processing steps required for subsequent processing of the micropore area and the preset processing parameters of the processing equipment corresponding to each processing step are obtained. These processing parameters are parameters set by the processing equipment under normal circumstances, including parameters such as cutting speed and feed rate. According to the preset processing parameters of the processing equipment corresponding to the processing step, the preset load force value when the micropore area is subsequently processed by the corresponding processing equipment is determined. The load force value reflects the magnitude of the force applied to the micropore area during the processing process. Determine whether the preset load force value is greater than the predicted expansion stress value. If it is greater than, it means that under the current processing parameters, the force applied during the processing may cause the defect to expand further, thereby affecting the quality of the micro-pore area. At this time, the genetic algorithm is introduced. The genetic algorithm is an optimization algorithm based on the principle of biological evolution. Set the genetic algebra, and use the genetic algebra combined with the preset load force value to adjust the preset processing parameters of the corresponding processing equipment. This adjustment process is an iterative optimization process until the preset load force value is not greater than the predicted expansion stress value, and finally output the adjusted processing parameters. When the target device flows into the corresponding processing equipment, the corresponding processing equipment is controlled to use the adjusted processing parameters to process the micro-pore area of ​​the target device. In this way, it can be ensured that the force applied during the processing will not cause the defect to expand further, thereby improving the processing quality of the micro-pore area. When the defect concentration coefficient in the micro-pore area meets the conditions, this method first obtains the defect name and performs finite element analysis to obtain information such as stress value, and then determines the load force value and makes a judgment according to the processing steps and equipment parameters, adjusts the processing parameters through the genetic algorithm until the requirements are met, and finally processes with the adjusted parameters. By comprehensively considering multiple factors such as the defects in the micro-hole area, processing steps, and processing equipment parameters, finite element analysis and genetic algorithms are used to optimize the processing parameters to ensure that defects will not expand due to excessive load during the processing process, thereby improving the processing quality of the micro-hole area of ​​the target device, achieving precise control of the processing process in the micro-hole area, and effectively reducing the scrap rate.

[0109] Another aspect of the present invention discloses a micro-pore defect detection and identification device, which is applied to any of the micro-pore defect detection and identification methods described above, such as Figure 3 As shown, the micro-pore defect detection and identification device comprises:

[0110] A testing table 1011, used for fixing a target device;

[0111] A loading module 1102 is used to transfer the target device to the testing table;

[0112] The acoustic wave data acquisition module 1103 is responsible for collecting acoustic wave data fed back from the micro-hole region in the target device. The acoustic wave data acquisition module includes an ultrasonic detector, a signal amplifier and a data acquisition card, which are used to obtain acoustic information of the micro-hole region;

[0113] The sound wave data processing module 1104 performs denoising on the collected sound wave data to obtain denoised sound wave data;

[0114] A feature model building module 1105 builds a regional feature model diagram of the micropore region based on the denoised acoustic wave data;

[0115] A feature analysis module 1106 performs feature analysis on a regional feature model diagram of a micropore region to obtain a defect concentration coefficient of the micropore region;

[0116] The defect judgment module 1107 judges and analyzes the micropore area according to the defect concentration coefficient to determine whether the defect concentration of the micropore area meets the requirements.

[0117] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0118] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0119] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0120] Those skilled in the art can understand that: all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiments; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), disks or optical disks, and other media that can store program codes.

[0121] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods of each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0122] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for detecting and identifying micropore defects, characterized in that: The following steps are involved: Acquire acoustic wave data fed back from a micropore region in a target device, perform denoising on the acoustic wave data, acquire denoised acoustic wave data, and construct a regional feature model diagram of the micropore region based on the denoised acoustic wave data; Performing characteristic analysis on the regional characteristic model diagram of the micropore region to obtain a defect concentration coefficient of the micropore region; According to the defect concentration coefficient, the micropore region is judged and analyzed to determine whether the defect concentration of the micropore region meets the requirements; Building a knowledge database, if the micropore region is a region where the defect concentration does not meet the requirement, obtaining the defect name of the defect existing in the micropore region based on the knowledge database; And the defect name of the defect existing in the micro-hole area is uploaded to the preset terminal.

2. A micro-pore defect detection and identification method according to claim 1, characterized in that: Acquire the acoustic wave data fed back by the micro-hole region in the target device, perform denoising on the acoustic wave data, obtain the denoised acoustic wave data, and construct a regional feature model diagram of the micro-hole region according to the denoised acoustic wave data, specifically: Scanning and detecting the micro-hole area in the target device by using an ultrasonic detector, obtaining the acoustic wave data fed back by the micro-hole area, and dividing the acoustic wave data fed back by the micro-hole area into several frequency bands; Get the frequency value of sound wave data in each frequency band; And preset a preset frequency value range of the acoustic wave data, and traverse the acoustic wave data fed back from the micro-pore area in the order of the acquisition frequency band; Construct a multi-branch tree. If the frequency value of the sound wave data of a certain frequency band is not within the preset frequency value range, the sound wave data of the frequency band is divided into a child node of the multi-branch tree; if the frequency value of the sound wave data of a certain frequency band is within the preset frequency value range, the sound wave data of the frequency band is divided into another child node of the multi-branch tree; and a division criterion is formed in this way; According to the division criterion, the child nodes are continuously divided, and after the division is completed, a multi-branch tree structure is formed; wherein each child node of the multi-branch tree structure represents a subset of the original sound wave data, and as the division proceeds, the sound wave data in the child nodes gradually tend to be similar in characteristics; Acquire data features of normal sound wave data through a big data network, and acquire data features of sound wave data on each child node in a multi-branch tree structure; wherein the data features include frequency, amplitude, and phase; Calculate the cosine similarity between the data features of the sound wave data on each child node and the data features of the normal sound wave data based on the cosine similarity algorithm; The sub-nodes whose cosine similarity is not greater than the preset similarity threshold are marked as noise data sub-nodes, and the sub-nodes marked as noise data are pruned in the multi-branch tree structure; Acquire the acoustic wave data of the remaining child nodes in the multi-branch tree structure to obtain denoised acoustic wave data; and construct a regional feature model diagram of the micropore area based on the denoised acoustic wave data.

3. A micro-pore defect detection and identification method according to claim 1, characterized in that: The characteristic analysis is performed on the regional characteristic model diagram of the micropore region to obtain the defect concentration coefficient of the micropore region, specifically: Obtaining standard structural drawing information of a target device, and obtaining a standard characteristic model diagram of a micro-hole region according to the standard structural drawing information; Preset voxel size, voxelize the regional feature model map according to the preset voxel size; and voxelize the standard feature model map according to the preset voxel size; Traverse all voxels in the two voxelized regional feature model maps and the standard feature model map; If both the regional feature model map and the standard feature model map exist on a certain voxel, the voxel is marked as a coincident voxel; If only a regional feature model map exists on a certain voxel, or only a standard feature model map exists, the voxel is marked as a non-coincident voxel; If neither the regional feature model map nor the standard feature model map exists on a certain voxel, the voxel is marked as a blank voxel; The total number of all non-overlapping voxels and the total number of overlapping voxels and non-overlapping voxels are calculated; the total number of non-overlapping voxels and the total number of overlapping voxels and non-overlapping voxels are ratio-processed to obtain the defect concentration coefficient of the micropore area.

4. A micro-pore defect detection and identification method according to claim 1, characterized in that: The micropore region is judged and analyzed according to the defect concentration coefficient to determine whether the defect concentration of the micropore region meets the requirements, specifically: Presetting a defect concentration coefficient threshold, and comparing the defect concentration coefficient of the micropore area with the defect concentration coefficient threshold; If the defect concentration coefficient of the micropore region is not greater than the defect concentration coefficient threshold, marking the micropore region as a region where the defect concentration meets the requirement; If the defect concentration coefficient of the micropore region is greater than the defect concentration coefficient threshold, the micropore region is marked as a region where the defect concentration does not meet the requirements.

5. The method for detecting and identifying micropore defects according to claim 1, characterized in that: The method for constructing the knowledge database is as follows: Retrieve the characteristic model diagrams corresponding to various micropore defects through the big data network and build a blank database; Obtaining the number of defect types of micropore defects retrieved through the big data network, and dividing the blank database into a corresponding number of sub-databases according to the number of defect types; and obtaining the defect type name of the micropore defect retrieved through the big data network, and converting the defect type name of each micropore defect into a unique binary digital code; Each digital code is compressed and bundled with the corresponding feature model diagram to obtain a number of data packets, each of which is stored in a corresponding sub-database to form a knowledge database; and the knowledge database is updated regularly.

6. A method for detecting and identifying micropore defects according to claim 3, characterized in that: If the micropore region is a region where the defect concentration does not meet the requirement, the defect name of the defect existing in the micropore region is obtained based on the knowledge database, specifically: If the micropore region is a region where the defect concentration does not meet the requirements, then the model region corresponding to the voxels marked as non-overlapping between the regional feature model map and the standard feature model map is obtained to obtain a defect model map of the defects existing in the micropore region; and extracting the characteristic model graph in the data packet from the knowledge database; introducing the cosine similarity algorithm, and calculating the cosine similarity between the defect model graph of the defect existing in the micropore area and the characteristic model graph in each data packet based on the cosine similarity algorithm; Arrange multiple cosine similarities based on their numerical values, and after the arrangement is completed, obtain the maximum cosine similarity; The data packet corresponding to the maximum cosine similarity is marked, and the digital code in the marked data packet is extracted. The extracted digital code is subjected to reverse identification processing to obtain the defect name of the defect existing in the micro-hole area.

7. A micropore defect detection and identification device, applied to the micropore defect detection and identification method according to any one of claims 1 to 6, characterized in that: The micropore defect detection and identification device comprises: A test bench for fixing the target device; A loading module, used to transfer the target device to the testing table; The acoustic wave data acquisition module is responsible for collecting the acoustic wave data fed back by the micro-hole area in the target device. The acoustic wave data acquisition module includes an ultrasonic detector, a signal amplifier and a data acquisition card, which are used to obtain the acoustic information of the micro-hole area; The sound wave data processing module performs denoising on the collected sound wave data to obtain the denoised sound wave data; A feature model building module builds a regional feature model diagram of the micropore area based on the denoised acoustic wave data; The feature analysis module performs feature analysis on the regional feature model diagram of the micropore area to obtain the defect concentration coefficient of the micropore area; The defect judgment module judges and analyzes the micropore area according to the defect concentration coefficient to determine whether the defect concentration of the micropore area meets the requirements.

Images