CN118023791A - Welding method and system for precise shell - Google Patents
Welding method and system for precise shell Download PDFInfo
- Publication number
- CN118023791A CN118023791A CN202410433551.0A CN202410433551A CN118023791A CN 118023791 A CN118023791 A CN 118023791A CN 202410433551 A CN202410433551 A CN 202410433551A CN 118023791 A CN118023791 A CN 118023791A
- Authority
- CN
- China
- Prior art keywords
- welding
- parameter
- parameters
- path
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000003466 welding Methods 0.000 title claims abstract description 482
- 238000000034 method Methods 0.000 title claims abstract description 44
- 230000007547 defect Effects 0.000 claims abstract description 112
- 239000000463 material Substances 0.000 claims abstract description 106
- 238000012545 processing Methods 0.000 claims abstract description 10
- 238000013135 deep learning Methods 0.000 claims description 27
- 230000006978 adaptation Effects 0.000 claims description 22
- 230000002159 abnormal effect Effects 0.000 claims description 14
- 238000013527 convolutional neural network Methods 0.000 claims description 14
- 230000008569 process Effects 0.000 claims description 13
- 239000011159 matrix material Substances 0.000 claims description 11
- 238000011176 pooling Methods 0.000 claims description 10
- 238000012360 testing method Methods 0.000 claims description 10
- 238000004140 cleaning Methods 0.000 claims description 8
- 238000005219 brazing Methods 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 5
- 239000000203 mixture Substances 0.000 claims description 4
- 238000005516 engineering process Methods 0.000 abstract description 7
- 230000000704 physical effect Effects 0.000 abstract description 3
- 238000004519 manufacturing process Methods 0.000 description 7
- 238000006243 chemical reaction Methods 0.000 description 6
- 238000013136 deep learning model Methods 0.000 description 4
- 229910052751 metal Inorganic materials 0.000 description 4
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 229910000838 Al alloy Inorganic materials 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 239000000945 filler Substances 0.000 description 3
- 239000002184 metal Substances 0.000 description 3
- 238000007781 pre-processing Methods 0.000 description 3
- 239000004698 Polyethylene Substances 0.000 description 2
- 239000004743 Polypropylene Substances 0.000 description 2
- 239000000956 alloy Substances 0.000 description 2
- 239000002131 composite material Substances 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005304 joining Methods 0.000 description 2
- 239000004033 plastic Substances 0.000 description 2
- 229920003023 plastic Polymers 0.000 description 2
- -1 polyethylene Polymers 0.000 description 2
- 229920000573 polyethylene Polymers 0.000 description 2
- 229920001155 polypropylene Polymers 0.000 description 2
- 230000001105 regulatory effect Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 229910000967 As alloy Inorganic materials 0.000 description 1
- RYGMFSIKBFXOCR-UHFFFAOYSA-N Copper Chemical compound [Cu] RYGMFSIKBFXOCR-UHFFFAOYSA-N 0.000 description 1
- 206010063385 Intellectualisation Diseases 0.000 description 1
- FYYHWMGAXLPEAU-UHFFFAOYSA-N Magnesium Chemical compound [Mg] FYYHWMGAXLPEAU-UHFFFAOYSA-N 0.000 description 1
- 239000004793 Polystyrene Substances 0.000 description 1
- VYPSYNLAJGMNEJ-UHFFFAOYSA-N Silicium dioxide Chemical compound O=[Si]=O VYPSYNLAJGMNEJ-UHFFFAOYSA-N 0.000 description 1
- 229910001069 Ti alloy Inorganic materials 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 229910052802 copper Inorganic materials 0.000 description 1
- 239000010949 copper Substances 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 238000010894 electron beam technology Methods 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 238000009413 insulation Methods 0.000 description 1
- 229910052749 magnesium Inorganic materials 0.000 description 1
- 239000011777 magnesium Substances 0.000 description 1
- 239000000155 melt Substances 0.000 description 1
- 238000002844 melting Methods 0.000 description 1
- 230000008018 melting Effects 0.000 description 1
- 239000007769 metal material Substances 0.000 description 1
- 238000004377 microelectronic Methods 0.000 description 1
- 239000004417 polycarbonate Substances 0.000 description 1
- 229920000515 polycarbonate Polymers 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 239000000741 silica gel Substances 0.000 description 1
- 229910002027 silica gel Inorganic materials 0.000 description 1
- 238000005493 welding type Methods 0.000 description 1
Landscapes
- Investigating And Analyzing Materials By Characteristic Methods (AREA)
Abstract
The invention relates to the technical field of welding, in particular to a welding method and a system of a precise shell, wherein the method comprises the following steps: acquiring material information of the precise shell, and generating a welding path according to the structure of the precise shell; constructing a welding parameter submodel library, wherein the welding parameter submodel library provides a welding parameter model for welding the precise shell of each project; matching welding parameter models in a welding parameter model library according to material information and welding paths of the precise shell, and selecting execution welding parameters of the precise shell; and identifying the welding result of the precise shell, judging whether the welding result has welding defects, and if so, adjusting the welding parameters. According to the invention, the problem of high welding defect rate of the same processing technology due to the difference of materials and paths in precision welding is effectively solved, and welding technological parameters based on the matching of the physical properties of the materials and the welding paths are generated by referring to big data of the precision welding for different projects.
Description
Technical Field
The invention relates to the technical field of welding, in particular to a welding method and a welding system for a precise shell.
Background
The precision welding technology is generated under the background that requirements on welding quality, precision and reliability are increasingly improved, and the technology takes geometric dimensional precision as a key quality characteristic value, achieves the aim of accurate forming and manufacturing by adopting a high-energy density welding method such as laser welding, electron beam welding and the like, can reduce subsequent processing procedures, reduce cost and improve quality, and is a key technology for manufacturing and defeating microstructure products especially in important fields such as microelectronics, aerospace and the like.
Because the physical characteristics of the materials are different, and each project has a unique welding path, in precision welding processing, the optimization of a welding process is an important means for regulating and controlling the formation of a welding seam and eliminating welding defects, and along with the change of welding materials, welding structures and lasers, the number of process parameters which can be regulated and controlled by laser micro-welding is numerous, and how to select proper welding parameters for different materials and different welding paths becomes a key for improving the precision welding quality.
The information disclosed in this background section is only for enhancement of understanding of the general background of the disclosure and is not to be taken as an admission or any form of suggestion that this information forms the prior art that is well known to a person skilled in the art.
Disclosure of Invention
The invention provides a welding method and a welding system for a precise shell, which can effectively solve the problems in the background technology.
In order to achieve the above purpose, the technical scheme adopted by the invention is as follows:
A method of welding a precision housing, the method comprising:
Acquiring material information of a precise shell, and generating a welding path according to the structure of the precise shell;
constructing a welding parameter sub-model library, wherein the welding parameter sub-model library provides a welding parameter model for the precise shell for welding each item;
matching the welding parameter model in the welding parameter sub-model library according to the material information of the precise shell and the welding path, and selecting the executing welding parameters of the precise shell;
And identifying the welding result of the precise shell, judging whether the welding result has a welding defect, if so, adjusting the welding executing parameters, and if not, carrying out batch welding according to the welding executing parameters.
Further, matching the welding parameter model in the welding parameter submodel library according to the material information of the precise shell and the welding path, and selecting the execution welding parameters of the precise shell, including:
the welding parameter submodel library comprises a material matching submodel and a welding path matching submodel;
Matching the material information of the precise shell through the material matching sub-model, and determining a material bias welding parameter matching result according to the matching similarity;
Matching the welding path of the precise shell through the welding path matching sub-model, and determining a path bias welding parameter matching result according to the matching similarity;
And recombining a plurality of welding parameters according to the material bias welding parameter matching result and the path bias welding parameter matching result to obtain the execution welding parameters of the precise shell.
Further, constructing the material matching sub-model includes:
Collecting historical material information of the welding of the precise shell, and cleaning data of the historical material information;
clustering the cleaned historical material information according to material composition, generating a plurality of material clustering subsets, and performing deep learning on the material clustering subsets to obtain the bias influence of material properties on welding parameters;
The deep learning comprises transverse learning and longitudinal learning, wherein the transverse learning is used for comparing a plurality of material cluster subsets and deeply learning the influence bias of different materials on welding parameters, and the longitudinal learning is used for comparing a plurality of welding parameter items in a single material cluster subset and deeply learning the influence bias degree of the same materials on the welding parameters.
Further, constructing a welding path matching sub-model, comprising:
collecting historical welding path information of the precise shell structure, and cleaning data of the historical welding path information;
clustering the cleaned historical welding path information according to a precise shell structure to generate a plurality of path clustering subsets, and performing deep learning on the path clustering subsets to obtain the bias influence of welding paths on welding parameters;
The deep learning comprises transverse learning and longitudinal learning, wherein the transverse learning is used for comparing the path clustering subsets and deeply learning the influence bias of different welding paths on welding parameters, and the longitudinal learning is used for comparing the welding parameter items in a single path clustering subset and deeply learning the influence bias degree of the same welding path on the welding parameters.
Further, identifying the welding result of the precision housing includes:
Image acquisition is carried out on multiple angles of the precise shell, and the acquired images are preprocessed;
converting the preprocessed image into a matrix which can be identified by a convolutional neural network, generating a convolutional layer according to the defect type, and performing convolutional operation on the converted matrix;
And setting a regional interest pooling layer for the convolutional neural network according to the welding path, wherein the regional interest pooling layer is used for outputting the welding path region corresponding to the convolutional operation result to obtain a welding identification result, and the welding identification result is obtained by adjusting the welding path region to be of a fixed size, wherein the fixed size is the maximum pooling.
Further, a welding defect database is constructed for the convolutional neural network, historical reject and return factory data information is collected by the welding defect database, corresponding defect types and corresponding welding abnormal parameter items are obtained, and the determination of the welding recognition result is based on the welding defect database.
Further, if the welding defect exists, adjusting the welding parameter, including:
judging defect types according to the welding defects, and acquiring corresponding welding abnormal parameters based on the welding defect database;
Deep learning is carried out on the welding defect database, a defect adaptation adjustment model is constructed, and the defect adaptation adjustment model is used for converting a welding recognition result with defects into a parameter adjustment strategy based on a corresponding sub model in the welding parameter sub model library as a reference for adjustment;
setting a parameter adjustment strategy according to the defect type and the sub-model reference, and adjusting the welding parameters according to the parameter adjustment strategy;
Wherein the reference of the sub-model and the corresponding welding abnormal parameters are used as an input layer of the defect adaptation adjustment model, and the deep learning and the parameter adjustment strategy are processing layers of the defect adaptation adjustment model.
Further, performing welding test again on the adjusted welding parameters, judging whether a welding result has defects, if so, repeating the welding parameter adjustment process or performing test and evaluation on the parameter adjustment strategy of the defect adaptation adjustment model, and if not, performing batch welding according to the adjusted welding parameters.
Further, the welding path performs contact surface welding on a part of the structure of the precision housing by using brazing.
A welding system for precision casings, the system comprising:
the welding information acquisition module acquires material information of the precise shell and generates a welding path according to the structure of the precise shell;
the sub-model library generation module is used for constructing a welding parameter sub-model library, and the welding parameter sub-model library is used for providing a welding parameter model for the precise shell for welding each item;
The welding parameter matching module is used for matching the welding parameter model in the welding parameter sub-model library according to the material information of the precise shell and the welding path, and selecting the execution welding parameters of the precise shell;
And the welding parameter optimization module is used for identifying the welding result of the precise shell, judging whether the welding result has a welding defect or not, adjusting the welding execution parameters if the welding defect exists, and performing batch welding according to the welding execution parameters if the welding defect does not exist.
By the technical scheme of the invention, the following technical effects can be realized:
The problem that the welding defect rate of the same processing technology is high due to the difference of materials and paths in precision welding is effectively solved, welding technological parameters based on the matching of the physical properties of the materials and the welding paths are generated according to big data of the precision welding for different projects, the precision welding quality is improved, and the rejection rate is reduced.
The foregoing description is only an overview of the present application, and is intended to be implemented in accordance with the teachings of the present application in order that the same may be more clearly understood and to make the same and other objects, features and advantages of the present application more readily apparent.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments described in the present invention, and other drawings may be obtained according to the drawings without inventive effort to those skilled in the art.
FIG. 1 is a flow chart of a method of welding a precision shell;
FIG. 2 is a flow chart of selecting parameters for performing welding of a precision shell;
FIG. 3 is a schematic diagram of a welding parameter submodel library;
FIG. 4 is a schematic diagram of a convolutional neural network;
FIG. 5 is a flow chart for identifying and adjusting welding parameters.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and/or" as used herein includes any and all combinations of one or more of the associated listed items.
Example 1
As shown in fig. 1, the present application provides a method for welding a precision housing, the method comprising:
S10: acquiring material information of the precise shell, and generating a welding path according to the structure of the precise shell;
s20: constructing a welding parameter submodel library, wherein the welding parameter submodel library provides a welding parameter model for welding the precise shell of each project;
S30: matching welding parameter models in a welding parameter model library according to material information and welding paths of the precise shell, and selecting execution welding parameters of the precise shell;
S40: and identifying the welding result of the precise shell, judging whether the welding result has welding defects, if so, adjusting the welding execution parameters, and if not, carrying out batch welding according to the welding execution parameters.
Specifically, for a material of a precision casing, for example, the precision casing is a mobile phone casing, then a material which may become a mobile phone casing, for example, a plastic material includes Polycarbonate (PC), polyethylene (PE), polypropylene (PP), polystyrene (PS) and the like, and a metal material such as aluminum alloy, titanium alloy and the like, the composite material of which can also specifically obtain composition information of the composite material if necessary with accuracy, for example, some casings consider that metal elements such as copper, silica gel or magnesium and the like may be added as alloy materials, for example, heat resistance, heat insulation or strength and the like, then parameter adjustment for welding of different materials is also slightly changed, and a welding route is generated according to a structure in which the casings are required to be welded, a precision welding casing to be processed can be established by three-dimensional modeling scanning, and a welding position is then generated, and a coordinate system can be generated, and the welding route is represented by coordinates, and a welding parameter sub-model library collects and processes a large amount of relevant precision casing welding data, and provides a certain degree of reference for the current welding parameter from two aspects of the material and the welding route, a method such as image processing technique, a point cloud data analysis and the like can be used, and a welding precision casing completed can be detected and welded to identify whether or not a welding parameter has a defect exists, and a welding parameter can be adjusted according to a welding parameter and a welding defect condition can be adjusted according to a welding defect condition, and a welding parameter can be adjusted if a welding defect condition is different in terms such as a welding parameter is adjusted; when no welding defect exists or the defect is confirmed to be eliminated after adjustment, the system can perform batch welding according to the welding parameters, so that the welding quality and consistency of the whole batch can be ensured.
According to the technical scheme, the problem that the welding defect rate of the same processing technology is high due to the difference of materials and paths in precision welding is effectively solved, welding technological parameters based on matching of the physical properties of the materials and the welding paths are generated according to big data of precision welding of different projects, precision welding quality is improved, and rejection rate is reduced.
Further, as shown in fig. 2 to3, according to the material information of the precise housing and the welding path, the welding parameter model in the welding parameter submodel library is matched, and the performing welding parameters of the precise housing are selected, including:
S31: the welding parameter sub-model library comprises a material matching sub-model and a welding path matching sub-model;
S32: matching the material information of the precise shell through a material matching sub-model, and determining a material weight bias welding parameter matching result according to the matching similarity;
s33: matching the welding path of the precise shell through a welding path matching sub-model, and determining a path overweight welding parameter matching result according to the matching similarity;
s34: and recombining the welding parameters according to the material bias welding parameter matching result and the path bias welding parameter matching result to obtain the execution welding parameters of the precise shell.
As a preference of the above embodiment, a welding parameter submodel library is built in the system, wherein the welding parameter submodel library comprises a material matching submodel and a welding path matching submodel, the material matching submodel comprises welding parameter models of various materials, such as welding parameters of aluminum alloy, plastic and the like, the welding path matching submodel comprises welding parameter models of different welding paths or structures, such as welding parameters of edge welding, angle welding, surface welding and the like, the system calculates matching similarity according to material information of a precise shell, each material has a corresponding welding parameter model, and determines which material welding parameter is more suitable for the current precise shell; the system is matched with a welding path matching sub-model according to a welding path or structure of a precise shell, each welding path is provided with a corresponding welding parameter model, the system calculates matching similarity, determines which welding parameters of the welding path are more suitable for the current precise shell, the system reorganizes a plurality of welding parameter models according to the material matching result and the welding path matching result, and each sub-model is provided with bias adjustment of the welding parameters, for example, when an aluminum alloy material is welded, which welding parameters are required to be adjusted, when the similarity welding path is needed to be adjusted, and certain default established welding parameters are added to form complete welding parameters; according to the specific materials and structures of the precise shell, the system can provide personalized welding parameters according to the matching result, is better suitable for precise shells of different materials and structures, and recombines the welding parameters according to the matching result, so that the welding quality can be optimized and improved, the incidence of welding defects is reduced, and along with continuous use and matching, the system can accumulate more welding parameter data, continuously optimize and perfect a welding parameter submodel library, and the intellectualization and adaptability of the system are improved.
Further, constructing the material matching sub-model includes:
collecting historical material information of welding of the precise shell, and cleaning data of the historical material information;
Clustering the cleaned historical material information according to the material composition, generating a plurality of material clustering subsets, and performing deep learning on the material clustering subsets to obtain the bias influence of the material properties on welding parameters;
the deep learning comprises transverse learning and longitudinal learning, wherein the transverse learning is used for comparing a plurality of material clustering subsets and deeply learning the influence of different materials on welding parameters, and the longitudinal learning is used for comparing a plurality of welding parameter items in a single material clustering subset and deeply learning the influence degree of the same material on the welding parameters.
Based on the above embodiment, collecting relevant material information from the historical data of the precise shell welding can include components, characteristics, performances and the like of different materials, and performing data cleaning and preprocessing on the collected historical material information, including removing error data, filling missing values and the like, so as to ensure the accuracy and the integrity of the data, using a clustering algorithm to divide the materials into a plurality of clustering subsets according to similarity, each subset including material samples with similar material components, facilitating the learning of systematicness and purposefulness by deep learning, and performing horizontal learning to learn the bias of influence of different materials on welding parameters for a deep learning model, such as learning which materials are more suitable for which welding parameters, and performing vertical learning to learn the bias of influence of the same materials on different welding parameters for a deep learning model, such as learning which welding parameters are more important or greatly influenced in a certain material, and capturing the influence of different material properties on welding parameters in the deep learning process, wherein the material properties include but are not limited to components, density, strength, thermal conductivity and the like, and the properties can influence finally selected welding parameters, and combining with the horizontal learning and the longitudinal learning to implement the learning to realize the learning of the optimum performance of the precise shell welding parameters, and the optimal performance of the welding parameters, and the performance of the intelligent shell is improved, and the system is capable of realizing the matching of the quality is improved by performing the intelligent system, and the system is capable of implementing the matching by the deep learning and the learning to realize the optimum quality and the performance of the quality-dependent learning.
Further, constructing a welding path matching sub-model, comprising:
Collecting historical welding path information of the precise shell structure, and cleaning data of the historical welding path information;
Clustering the cleaned historical welding path information according to a precise shell structure to generate a plurality of path clustering subsets, and performing deep learning on the path clustering subsets to obtain the bias influence of welding paths on welding parameters;
The deep learning comprises transverse learning and longitudinal learning, wherein the transverse learning is used for comparing a plurality of path clustering subsets and deeply learning the influence of different welding paths on welding parameters, and the longitudinal learning is used for comparing a plurality of welding parameter items in a single path clustering subset and deeply learning the influence degree of the same welding path on the welding parameters.
Based on the above embodiment, relevant welding path information is collected from historical data of a precise shell structure, including edge welding, fillet welding, surface welding and the like of different structures, the collected historical welding path information is subjected to data cleaning and preprocessing, the cleaned historical welding path information is clustered according to the precise shell structure, a clustering algorithm is used for dividing the welding paths into a plurality of path clustering subsets according to similarity, each subset comprises welding path samples with similar structures, the welding path samples among the different path clustering subsets are transversely learned, comparison is carried out and deep learning is carried out, and a deep learning model learns that different welding paths have a bias on the influence of welding parameters, such as learning which paths are more suitable for which welding parameter settings; longitudinal learning is to compare a plurality of welding parameter items and perform deep learning on welding path samples in a single path clustering subset, a deep learning model learns the influence bias degree of the same welding path on different welding parameters, for example, learning which welding parameters in a certain welding path are more important or have larger influence, in the deep learning process, the model learns and captures the influence degree of different welding path structures on the welding parameters, wherein the welding path structures comprise but are not limited to edge welding, fillet welding, surface welding and the like, the structures influence the finally selected welding parameters, and the system finally obtains a welding path matching sub-model by combining the deep learning results of transverse learning and longitudinal learning.
Further, as shown in fig. 4, identifying the welding result of the precision housing includes:
Image acquisition is carried out on multiple angles of the precise shell, and the acquired images are preprocessed;
converting the preprocessed image into a matrix which can be identified by a convolutional neural network, generating a convolutional layer according to the defect type, and performing convolutional operation on the converted matrix;
And setting a regional interest pooling layer for the convolutional neural network according to the welding path, and adjusting the welding path region corresponding to the result of the convolutional operation to be a fixed size output by the regional interest pooling layer to obtain a welding identification result, wherein the fixed size is the maximum pooling.
As a preference of the above embodiment, the method includes performing multi-angle image acquisition on welding results of a precise housing at different angles, performing preprocessing on the acquired images, including denoising, graying, edge detection, and the like, to improve image quality and reduce interference, converting the preprocessed images into matrix form capable of being identified by a convolutional neural network, typically converting the images into a matrix of pixel values, determining the size and channel number of an input matrix, for example, converting the gray images into a two-dimensional matrix, designing appropriate convolutional kernels according to defect types, for capturing different types of welding defect characteristics, setting a plurality of convolutional layers according to requirements in the design of the convolutional neural network, each layer being used for extracting information of different levels and characteristics, performing convolutional operations on the converted matrix, sliding calculation on the images by using the convolutional kernels, extracting characteristics in the images, the convolutional operations generating a series of characteristic graphs including characteristic information on different positions and dimensions of the images, setting a region interest pool layer according to welding path information, adjusting a corresponding to a convolutional operation region to a fixed size, outputting a maximum conversion region, namely, setting a maximum conversion region to a maximum conversion region, and a maximum conversion region, connecting the maximum conversion region to a maximum region of the region interest pool, and a maximum conversion region, thereby preserving the maximum interest value as a maximum region, connecting the region interest pool, and the region can be used for preserving the characteristics, and the maximum region interest can be processed by connecting the critical region, and the region is the region-saving, and the region-saving the region-interest is the important region-saving, and the region-saving the region-interest is better by the method, and finally, determining a welding identification result, namely which type of defect or normal the welding result belongs to, according to the highest score or probability.
Further, a welding defect database is constructed for the convolutional neural network, historical reject and return factory data information is collected by the welding defect database, corresponding defect types and corresponding welding abnormal parameter items are obtained, and the determination of the welding recognition result is based on the welding defect database.
As a preference of the above embodiment, the historical reject and return factory data may include images of reject products, related welding parameters and process information, and related information of return factory products, the data collection should cover products of different batches, different models and different production time periods, so as to ensure that the database has representativeness, analyze the collected reject and return factory data, determine types of various welding defects, such as welding cracks, welding defects, welding deformation, and the like, determine related welding abnormal parameter items according to the types of defects, determine architecture, layer number, convolution kernel size, pooling mode, and the like of the convolutional neural network according to the constructed welding defect database, and train the designed convolutional neural network model by using the collected reject and return factory data as training sets, automatically correlate corresponding welding abnormal parameter items according to the types of defects, realize quick recognition of welding defects, perform corresponding processing and feedback, improve production efficiency, and facilitate management and management of the constructed defect database, provide reference to the production history data, optimize the process, and provide basis for the analysis of production history data.
Further, adjusting the performing welding parameters if the welding defect exists comprises:
judging defect types according to the welding defects, and acquiring corresponding welding abnormal parameters based on the welding defect database;
Deep learning is carried out on the welding defect database, a defect adaptation adjustment model is constructed, and the defect adaptation adjustment model is used for converting a welding recognition result with defects into a parameter adjustment strategy based on a corresponding sub model in the welding parameter sub model library as a reference for adjustment;
setting a parameter adjustment strategy according to the defect type and the sub-model reference, and adjusting the welding parameters according to the parameter adjustment strategy;
Wherein the reference of the sub-model and the corresponding welding abnormal parameters are used as an input layer of the defect adaptation adjustment model, and the deep learning and the parameter adjustment strategy are processing layers of the defect adaptation adjustment model.
In this implementation, the defects are classified according to the detected welding defects, specific defect types are determined, corresponding welding abnormal parameters are obtained according to the defect types, the abnormal parameters describe possible reasons for causing the corresponding defects, the defect adaptation adjustment model maps the welding defects and the corresponding abnormal parameters into parameter adjustment strategies, the corresponding abnormal parameters and the mapped parameter adjustment strategies are based on a welding defect database by a deep learning method, historical defect data are learned and the degree that the adjusted parameters comprise adjustment parameters is set, the corresponding parameter adjustment strategies are possibly different according to factors such as the severity degree, the type and the position of the defects, the welding parameters are adjusted according to the formulated parameter adjustment strategies, the welding parameters can be individually adjusted according to the welding defects of different types through deep learning and establishment of the defect adaptation adjustment model, the adaptation and repair capability of the different defects are improved, the established defect adaptation adjustment model can automatically adjust the welding parameters, the burden of manual intervention is lightened, and the automation level and the intelligent degree of the production line are improved.
Further, as shown in fig. 5, the welding test is performed again on the adjusted welding parameters, and it is determined whether the welding result has a defect, if so, the above welding parameter adjustment process is repeated or the parameter adjustment strategy of the defect adaptive adjustment model is tested and evaluated, and if not, batch welding is performed according to the adjusted welding parameters.
As a preference of the above embodiment, according to the adjusted welding parameters, performing welding test on the precision shell, detecting and judging a welding result, and if a welding defect exists, entering a next adjustment process; if no defects exist, batch welding can be continued, if welding defects are found in a re-welding test, welding parameters need to be readjusted, or parameter adjustment strategies of a defect adaptation adjustment model need to be tested and evaluated, training data of the model can be updated or adjusted to improve accuracy of the model in identifying different types of defects and corresponding adjustment strategies, the defect adaptation adjustment model can be verified and optimized through testing and evaluation, the defect adaptation adjustment model can be ensured to be effectively identified and parameter adjusted, the welding parameters and the parameter adjustment strategies of the defect adaptation adjustment model are continuously tested and adjusted in a cyclic iteration mode, after each test and evaluation, the welding test and defect judgment are carried out again according to the result adjustment parameters until an expected welding quality standard is achieved, and iterative optimization of the welding parameters and the adjustment model can be realized through continuous welding test, defect judgment and parameter adjustment, so that welding quality and stability are continuously improved.
Further, the welding path performs contact surface welding of a part of the structure of the precision housing using brazing.
In this embodiment, brazing is a joining process using filler metal, and in general, since the melting point of the filler metal is lower than that of the material to be joined, during the brazing process, the filler metal melts after heating and flows into the joint, so that the thermal influence on the material in the welding of the precision shell is small, and the joint gap can be filled, and the method is suitable for joining different thicknesses and different materials, wherein part of the structure of the precision shell refers to a relatively narrow portion, the narrow portion structure can be defined by the limit value of the welding width of the laser spot welding, the firmness of the welding of the narrow portion of the precision shell is improved by using means of brazing and surface welding in this specific scene, and the contact surface welding refers to that the continuity of the welding is kept continuous.
Example two
Based on the same inventive concept as the welding method of a precision housing in the foregoing embodiment, the present invention also provides a welding system of a precision housing, including:
The welding information acquisition module acquires material information of the precise shell and generates a welding path according to the structure of the precise shell;
The sub-model library generation module is used for constructing a welding parameter sub-model library which provides a welding parameter model for welding the precise shell of each project;
the welding parameter matching module is used for matching welding parameter models in the welding parameter sub-model library according to the material information and the welding path of the precise shell, and selecting the execution welding parameters of the precise shell;
And the welding parameter optimization module is used for identifying the welding result of the precise shell, judging whether the welding result has welding defects or not, adjusting the execution welding parameters if the welding defects exist, and performing batch welding according to the execution welding parameters if the welding defects do not exist.
The adjusting system of the present invention can effectively realize the welding method of the precise housing, and can achieve the technical effects described in the above embodiments, and will not be repeated here.
Although the application has been described in connection with specific features and embodiments thereof, it will be apparent that various modifications and combinations can be made without departing from the spirit and scope of the application. Accordingly, the specification and figures are merely exemplary illustrations of the present application as defined in the appended claims and are considered to cover any and all modifications, variations, combinations, or equivalents that fall within the scope of the application. It will be apparent to those skilled in the art that various modifications and variations can be made to the present application without departing from the scope of the application. Thus, the present application is intended to include such modifications and alterations insofar as they come within the scope of the application or the equivalents thereof.
Claims (10)
1. A method of welding a precision housing, the method comprising:
Acquiring material information of a precise shell, and generating a welding path according to the structure of the precise shell;
constructing a welding parameter sub-model library, wherein the welding parameter sub-model library provides a welding parameter model for the precise shell for welding each item;
matching the welding parameter model in the welding parameter sub-model library according to the material information of the precise shell and the welding path, and selecting the executing welding parameters of the precise shell;
And identifying the welding result of the precise shell, judging whether the welding result has a welding defect, if so, adjusting the welding executing parameters, and if not, carrying out batch welding according to the welding executing parameters.
2. The method of welding a precision shell according to claim 1, wherein matching the welding parameter model in the welding parameter submodel library according to the material information of the precision shell and the welding path, selecting the execution welding parameters of the precision shell comprises:
the welding parameter submodel library comprises a material matching submodel and a welding path matching submodel;
Matching the material information of the precise shell through the material matching sub-model, and determining a material bias welding parameter matching result according to the matching similarity;
Matching the welding path of the precise shell through the welding path matching sub-model, and determining a path bias welding parameter matching result according to the matching similarity;
And recombining a plurality of welding parameters according to the material bias welding parameter matching result and the path bias welding parameter matching result to obtain the execution welding parameters of the precise shell.
3. The method of welding precision shells of claim 2, wherein constructing a material matching sub-model comprises:
Collecting historical material information of the welding of the precise shell, and cleaning data of the historical material information;
clustering the cleaned historical material information according to material composition, generating a plurality of material clustering subsets, and performing deep learning on the material clustering subsets to obtain the bias influence of material properties on welding parameters;
The deep learning comprises transverse learning and longitudinal learning, wherein the transverse learning is used for comparing a plurality of material cluster subsets and deeply learning the influence bias of different materials on welding parameters, and the longitudinal learning is used for comparing a plurality of welding parameter items in a single material cluster subset and deeply learning the influence bias degree of the same materials on the welding parameters.
4. The method of welding precision shells according to claim 2, wherein constructing a welding path matching sub-model comprises:
collecting historical welding path information of the precise shell structure, and cleaning data of the historical welding path information;
clustering the cleaned historical welding path information according to a precise shell structure to generate a plurality of path clustering subsets, and performing deep learning on the path clustering subsets to obtain the bias influence of welding paths on welding parameters;
The deep learning comprises transverse learning and longitudinal learning, wherein the transverse learning is used for comparing the path clustering subsets and deeply learning the influence bias of different welding paths on welding parameters, and the longitudinal learning is used for comparing the welding parameter items in a single path clustering subset and deeply learning the influence bias degree of the same welding path on the welding parameters.
5. The method of welding a precision housing according to claim 1, wherein identifying a welding result of the precision housing comprises:
Image acquisition is carried out on multiple angles of the precise shell, and the acquired images are preprocessed;
converting the preprocessed image into a matrix which can be identified by a convolutional neural network, generating a convolutional layer according to the defect type, and performing convolutional operation on the converted matrix;
And setting a regional interest pooling layer for the convolutional neural network according to the welding path, wherein the regional interest pooling layer is used for outputting the welding path region corresponding to the convolutional operation result to obtain a welding identification result, and the welding identification result is obtained by adjusting the welding path region to be of a fixed size, wherein the fixed size is the maximum pooling.
6. The method of claim 5, wherein a welding defect database is constructed for the convolutional neural network, wherein the welding defect database collects historical reject and return data information and obtains corresponding defect categories and corresponding welding anomaly parameter terms, and wherein the determination of the welding recognition result is based on the welding defect database.
7. The method of claim 6, wherein adjusting the welding parameters if the welding defect exists comprises:
judging defect types according to the welding defects, and acquiring corresponding welding abnormal parameters based on the welding defect database;
Deep learning is carried out on the welding defect database, a defect adaptation adjustment model is constructed, and the defect adaptation adjustment model is used for converting a welding recognition result with defects into a parameter adjustment strategy based on a corresponding sub model in the welding parameter sub model library as a reference for adjustment;
setting a parameter adjustment strategy according to the defect type and the sub-model reference, and adjusting the welding parameters according to the parameter adjustment strategy;
Wherein the reference of the sub-model and the corresponding welding abnormal parameters are used as an input layer of the defect adaptation adjustment model, and the deep learning and the parameter adjustment strategy are processing layers of the defect adaptation adjustment model.
8. The method according to claim 7, wherein the welding test is performed again on the adjusted welding parameters, and whether the welding result has a defect is determined, if so, the above welding parameter adjustment process is repeated or the parameter adjustment strategy of the defect adaptation adjustment model is tested and evaluated, and if not, batch welding is performed according to the adjusted welding parameters.
9. The method of welding a precision housing according to any one of claims 1 to 8, wherein the welding path welds contact surfaces of portions of the structure of the precision housing using brazing.
10. A welding system for precision casings, the system comprising:
the welding information acquisition module acquires material information of the precise shell and generates a welding path according to the structure of the precise shell;
the sub-model library generation module is used for constructing a welding parameter sub-model library, and the welding parameter sub-model library is used for providing a welding parameter model for the precise shell for welding each item;
The welding parameter matching module is used for matching the welding parameter model in the welding parameter sub-model library according to the material information of the precise shell and the welding path, and selecting the execution welding parameters of the precise shell;
And the welding parameter optimization module is used for identifying the welding result of the precise shell, judging whether the welding result has a welding defect or not, adjusting the welding execution parameters if the welding defect exists, and performing batch welding according to the welding execution parameters if the welding defect does not exist.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410433551.0A CN118023791B (en) | 2024-04-11 | 2024-04-11 | Welding method and system for precise shell |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410433551.0A CN118023791B (en) | 2024-04-11 | 2024-04-11 | Welding method and system for precise shell |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118023791A true CN118023791A (en) | 2024-05-14 |
| CN118023791B CN118023791B (en) | 2024-06-18 |
Family
ID=90989819
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410433551.0A Active CN118023791B (en) | 2024-04-11 | 2024-04-11 | Welding method and system for precise shell |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118023791B (en) |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105912747A (en) * | 2016-03-30 | 2016-08-31 | 内蒙古第机械集团有限公司 | Method of establishing general simulation material parameter library |
| WO2017084186A1 (en) * | 2015-11-18 | 2017-05-26 | 华南理工大学 | System and method for automatic monitoring and intelligent analysis of flexible circuit board manufacturing process |
| CN109175608A (en) * | 2018-09-30 | 2019-01-11 | 华南理工大学 | Weld bead feature points position On-line Measuring Method and seam track automatic measurement system |
| CN109472358A (en) * | 2018-10-17 | 2019-03-15 | 深圳市微埃智能科技有限公司 | Welding condition recommended method, device and robot neural network based |
| KR102414534B1 (en) * | 2021-02-17 | 2022-06-28 | 조윤기 | A method for providing search and transaction service for unused materials |
| CN114789307A (en) * | 2021-12-24 | 2022-07-26 | 郑州轻工业大学 | Plate welding quality real-time monitoring method based on digital twinning |
| CN116197070A (en) * | 2023-05-05 | 2023-06-02 | 常州市闳晖科技发展股份有限公司 | Automatic reciprocating spraying equipment for plate-type workpiece |
| CN116275742A (en) * | 2023-04-11 | 2023-06-23 | 扬州工业职业技术学院 | Welding method based on digital modeling and flexible weld joint recognition |
| CN116363089A (en) * | 2023-03-23 | 2023-06-30 | 天地上海采掘装备科技有限公司 | Automatic detection method for welding defects of roller tooth holder of coal mining machine |
| CN116644667A (en) * | 2023-06-05 | 2023-08-25 | 江苏迅兴重工设备有限公司 | Steel structure welding process optimization method and system based on big data processing |
| WO2024018736A1 (en) * | 2022-07-19 | 2024-01-25 | 株式会社神戸製鋼所 | Robot control method, and robot control device |
| CN117808835A (en) * | 2024-02-06 | 2024-04-02 | 苏州大学 | Image processing method based on deep learning and local pre-fitting |
| CN117817111A (en) * | 2023-12-13 | 2024-04-05 | 山东奥太电气有限公司 | Method and system for intelligently identifying and matching process parameters in laser welding |
-
2024
- 2024-04-11 CN CN202410433551.0A patent/CN118023791B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017084186A1 (en) * | 2015-11-18 | 2017-05-26 | 华南理工大学 | System and method for automatic monitoring and intelligent analysis of flexible circuit board manufacturing process |
| CN105912747A (en) * | 2016-03-30 | 2016-08-31 | 内蒙古第机械集团有限公司 | Method of establishing general simulation material parameter library |
| CN109175608A (en) * | 2018-09-30 | 2019-01-11 | 华南理工大学 | Weld bead feature points position On-line Measuring Method and seam track automatic measurement system |
| CN109472358A (en) * | 2018-10-17 | 2019-03-15 | 深圳市微埃智能科技有限公司 | Welding condition recommended method, device and robot neural network based |
| KR102414534B1 (en) * | 2021-02-17 | 2022-06-28 | 조윤기 | A method for providing search and transaction service for unused materials |
| CN114789307A (en) * | 2021-12-24 | 2022-07-26 | 郑州轻工业大学 | Plate welding quality real-time monitoring method based on digital twinning |
| WO2024018736A1 (en) * | 2022-07-19 | 2024-01-25 | 株式会社神戸製鋼所 | Robot control method, and robot control device |
| CN116363089A (en) * | 2023-03-23 | 2023-06-30 | 天地上海采掘装备科技有限公司 | Automatic detection method for welding defects of roller tooth holder of coal mining machine |
| CN116275742A (en) * | 2023-04-11 | 2023-06-23 | 扬州工业职业技术学院 | Welding method based on digital modeling and flexible weld joint recognition |
| CN116197070A (en) * | 2023-05-05 | 2023-06-02 | 常州市闳晖科技发展股份有限公司 | Automatic reciprocating spraying equipment for plate-type workpiece |
| CN116644667A (en) * | 2023-06-05 | 2023-08-25 | 江苏迅兴重工设备有限公司 | Steel structure welding process optimization method and system based on big data processing |
| CN117817111A (en) * | 2023-12-13 | 2024-04-05 | 山东奥太电气有限公司 | Method and system for intelligently identifying and matching process parameters in laser welding |
| CN117808835A (en) * | 2024-02-06 | 2024-04-02 | 苏州大学 | Image processing method based on deep learning and local pre-fitting |
Non-Patent Citations (2)
| Title |
|---|
| 张胜跃;卿黎;张宇栋;冯秋霞;牛: "基于DMADOV的12Cr1MoV异质接头焊接缺陷分析", 焊管, no. 12, 28 December 2015 (2015-12-28) * |
| 杨理践;曹辉;: "基于深度学习的管道焊缝法兰组件识别方法", 仪器仪表学报, no. 02, 15 February 2018 (2018-02-15) * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118023791B (en) | 2024-06-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105118044B (en) | A kind of wheel shape cast article defect automatic testing method | |
| CN114240891B (en) | Welding spot quality identification method integrating knowledge graph and graph convolution neural network | |
| CN114354639B (en) | Weld defect real-time detection method and system based on 3D point cloud | |
| CN112329860B (en) | Mixed deep learning visual detection method, device, equipment and storage medium | |
| CN103927534A (en) | Sprayed character online visual detection method based on convolutional neural network | |
| CN112557445B (en) | Defect online detection method, device and system based on additive manufacturing | |
| CN116618764B (en) | Production optimization method and system of mirror surface electric discharge machine | |
| Li et al. | Towards intelligent monitoring system in wire arc additive manufacturing: a surface anomaly detector on a small dataset | |
| CN108008006B (en) | Welding seam defect detection method, device, equipment and system | |
| CN116740044B (en) | Copper pipe milling surface processing method and system based on visual detection and control | |
| Kao et al. | Laser cladding quality monitoring using coaxial image based on machine learning | |
| CN113906451A (en) | AI-based pre-training model decision system and AI-based visual inspection management system for product production line using the same | |
| CN115439694A (en) | High-precision point cloud completion method and device based on deep learning | |
| CN115775249A (en) | Additive manufacturing part forming quality monitoring method and system and storage medium | |
| CN116618878A (en) | Pre-welding process parameter determination method, welding quality online prediction method, device and storage medium | |
| CN113298190A (en) | Weld image recognition and classification algorithm based on large-size unbalanced samples | |
| CN110414388B (en) | Hump and fusion penetration online early warning method based on depth prediction network | |
| CN115965960A (en) | Weld joint identification method based on deep learning and 3D point cloud | |
| CN118023791B (en) | Welding method and system for precise shell | |
| García-Moreno | A fast method for monitoring molten pool in infrared image streams using gravitational superpixels. | |
| CN116909208B (en) | Shell processing path optimization method and system based on artificial intelligence | |
| Hallmann et al. | A method for analyzing the influence of process and design parameters on the build time of additively manufactured components | |
| Vater et al. | Development of a Cloud-and Edge-Architecture for adaptive model weight optimization of a CNN exemplified by optical detection of hairpin welding | |
| CN112967267B (en) | Laser directional energy deposition sputtering counting method of full convolution neural network | |
| CN115294061A (en) | Method and system for detecting surface defects of paper-plastic composite bag based on machine vision |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |