CN109283378A - A kind of rotating arc welding is seamed into shape parameter detection method, system, device and medium - Google Patents
A kind of rotating arc welding is seamed into shape parameter detection method, system, device and medium Download PDFInfo
- Publication number
- CN109283378A CN109283378A CN201810999989.XA CN201810999989A CN109283378A CN 109283378 A CN109283378 A CN 109283378A CN 201810999989 A CN201810999989 A CN 201810999989A CN 109283378 A CN109283378 A CN 109283378A
- Authority
- CN
- China
- Prior art keywords
- parameter
- forming
- model
- detector
- forming parameter
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R19/00—Arrangements for measuring currents or voltages or for indicating presence or sign thereof
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B21/00—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant
- G01B21/02—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant for measuring length, width, or thickness
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B21/00—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant
- G01B21/10—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant for measuring diameters
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B21/00—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant
- G01B21/22—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant for measuring angles or tapers; for testing the alignment of axes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- Length Measuring Devices With Unspecified Measuring Means (AREA)
Abstract
The invention discloses a kind of rotating arc weldings to be seamed into shape parameter detection method, system, device and medium, the method includes obtaining signal related with actual measurement technological parameter in real time during using rotating the arc weld seam, signal related with actual measurement technological parameter is input to the forming parameter model of fit pre-established after optimization processing, uses output forming parameter after the progress regression analysis of forming parameter model of fit；The system comprises process parameter measurement module, technological parameter processing module and forming parameter computing modules；Described device includes welding torch detector, current detector, voltage detector, torch height detector, welding torch tilt angle detector, groove data detector, data collecting card and computer.The present invention overcomes the insecure disadvantages of forming parameter that the prior art is predicted, and the real-time detection of forming parameter may be implemented.The present invention is widely used in welding technology field.
Description
Technical field
The present invention relates to welding technology field, especially a kind of rotating arc welding be seamed into shape parameter detection method, system and
Device.
Background technique
Rotating the arc welding is a kind of welding method for forming high-speed rotating electric arc using special welding gun and carrying out,
It is widely used in industrial welding with superiority outstanding.
For the follow-up of quality and quality evaluation of rotating the arc welding, a main aspect is concern rotating the arc welding institute
The forming quality of the weld seam of formation.The forming quality of weld seam can indicate that most important one is at parameter by forming parameter
Number is weld pool width and weld penetration.
The complicated, flow chart of data processing that measures during weld pool width and weld penetration that there are complex steps by instrument
The defects of complicated, therefore more troublesome to the actual measurement of weld pool width and weld penetration, it is difficult in rotating the arc welding process into
Real-time detection of the row to weld pool width and weld penetration.
Studies have shown that rotating the arc welding is formed by weld pool width and weld penetration and rotating the arc welding process
Welding torch rotation speed, welding torch radius of turn, welding current, weldingvoltage, torch height, welding torch inclination angle, bevel angle and groove
The technological parameters such as size are related, that is, theoretically can predict or calculate forming parameter by technological parameter.It is entitled " a kind of
The patent document of the molding detection device of A-TIG welding line and method " (Publication No. CN103551709A) discloses a kind of logical
The molten wide and fusion penetration under measurement different technical parameters are crossed, appearance of weld parameter is predicted using neural network mathematical model
Method.Paper " rotating the arc non gas shielded welding appearance of weld and the research automatically tracked " discloses a kind of by choosing rotating the arc
Welding current, speed, welding gun height and welding gun radius of turn carry out orthogonal test, utilize regression analysis to establish rotation
Turn electric arc forming math equation, to predict the method that rotating arc welding is seamed into shape parameter.
It is before welding by mathematical model or number that above-mentioned published rotating arc welding, which is seamed into shape parameter detection method,
Equation is learned to predict appearance of weld parameter.But since technological parameter is continually changing, root in the whole welding process
The forming parameter predicted according to the technological parameter before welding is unreliable, while existing method also cannot achieve the reality of forming parameter
When detect.
Summary of the invention
In order to solve the above-mentioned technical problem, the purpose of the present invention is to provide a kind of rotating arc weldings to be seamed into shape parameter detection
Method, system, device and medium.
First technical solution adopted by the present invention is:
A kind of rotating arc welding is seamed into shape parameter detection method, and the forming parameter includes weld pool width and weld penetration,
The following steps are included:
Obtain signal related with actual measurement technological parameter in real time during using rotating the arc weld seam；The technological parameter
Including welding torch rotation speed, welding torch radius of turn, welding current, weldingvoltage, torch height, welding torch inclination angle, bevel angle and
Groove size；
Signal related with actual measurement technological parameter is input to the forming parameter pre-established to be fitted after optimization processing
Model；The forming parameter model of fit has recorded the corresponding relationship of technological parameter and forming parameter using random forests algorithm；
Forming parameter is exported after carrying out regression analysis using forming parameter model of fit.
Further, the forming parameter model of fit is pre-established by following steps:
Set multiple groups test technology parameter；
The test of rotating the arc appearance of weld is carried out, to obtain test corresponding to each group of test technology parameter into parameter
Number；
Random Forest model is constructed using random forests algorithm；
Random Forest model is trained using the multiple groups test technology parameter and corresponding test forming parameter, it will
The Random Forest model obtained after training is as forming parameter model of fit.
Further, the Random Forest model has determining maximum decision tree number, by following steps to random
Forest model is trained:
Generate decision tree；
The multiple groups test technology parameter is input in Random Forest model, so that Random Forest model output and every group
Test technology parameter predicts forming parameter accordingly；
The deviation between prediction forming parameter and corresponding test forming parameter is calculated, if the deviation reaches default precision
In range, then increase the decision tree number of random forest, otherwise, is back to the step of generating decision tree and re-executes；
The decision tree number of current random forest is compared with maximum decision tree number, if decision tree number be greater than or
Equal to maximum decision tree number, then the training to Random Forest model is completed, otherwise, is back to the step of generating decision tree again
It executes.
Further, the step of generation decision tree specifically includes:
From the multiple groups test technology parameter selected section test technology parameter as Random Forest model node into
Line splitting, until cannot divide, to generate decision tree.
Further, the deviation is mean square deviation.
Further, the optimization processing comprises at least one of the following processing step:
Signal denoising, signal enhancing, signal division, signal normalization and signal regularization.
Second technical solution adopted by the present invention is:
A kind of rotating arc welding is seamed into shape parameter detection system, comprising:
Process parameter measurement module is had for being obtained in real time during using rotating the arc weld seam with actual measurement technological parameter
The signal of pass；The technological parameter includes welding torch rotation speed, welding torch radius of turn, welding current, weldingvoltage, welding torch height
Degree, welding torch inclination angle, bevel angle and groove size；
Technological parameter processing module, it is pre- for that will be input to after optimization processing with the related signal of actual measurement technological parameter
The forming parameter model of fit first established；The forming parameter model of fit using random forests algorithm have recorded technological parameter with
The corresponding relationship of forming parameter；
Forming parameter computing module, for exporting forming parameter after using forming parameter model of fit to carry out regression analysis.
Further, it further includes forming parameter model of fit that a kind of rotating arc welding of the present invention, which is seamed into shape parameter detection system,
Module is established, it includes following submodule that the forming parameter model of fit, which establishes module:
Test technology parameter setting module, for setting multiple groups test technology parameter；
Forming parameter measurement module is tested, for carrying out rotating the arc appearance of weld test, to obtain every battery of tests
Test forming parameter corresponding to technological parameter；
Random Forest model constructs module, for constructing Random Forest model using random forests algorithm；
Random Forest model training module, for utilizing the multiple groups test technology parameter and corresponding test forming parameter
Random Forest model is trained, using the Random Forest model obtained after training as forming parameter model of fit.
Third technical solution adopted by the present invention is:
A kind of rotating arc welding is seamed into shape parameter detection device, comprising:
Welding torch detector, for detecting welding torch rotation speed and welding torch radius of turn；
Current detector, for detecting welding current；
Voltage detector, for detecting weldingvoltage；
Torch height detector, for detecting torch height；
Welding torch tilt angle detector, for detecting welding torch inclination angle；
Groove data detector, for detecting bevel angle and groove size；
Data collecting card, for acquiring welding torch detector, current detector, voltage detector, torch height detector, weldering
Data that torch tilt angle detector and groove data detector detect simultaneously upload to computer；
Computer, for running program to execute if any one of claim 1-7 the method is to data collecting card
Data handled.
4th technical solution adopted by the present invention is:
A kind of medium, the medium have store function and are stored with the executable instruction of processor, and the processor can
The instruction of execution is used to execute when executed by the processor such as first technical solution the method.
The beneficial effects of the present invention are: forming parameter model of fit used in the present invention is to utilize rotating the arc technology
Just set up before actual production, only needs when carrying out actual production using rotating the arc technology to the technique measured
Parameter carries out regression analysis, therefore can survey process to avoid cumbersome weld pool width and weld penetration.Forming parameter is fitted mould
It can recorde the relationship of multiple groups technological parameter and corresponding forming parameter in type, so that it covers all operating conditions, in this way, logical
Actual condition can be reflected by crossing the forming parameter that forming parameter model of fit is calculated, and overcome the prior art and predicted
The insecure disadvantage of forming parameter out.Since the used time that forming parameter model of fit carries out regression analysis is seldom, and it is input to
Running parameter in forming parameter model of fit measures in real time, therefore the forming parameter of forming parameter model of fit output
It is in real time, to realize the real-time detection of forming parameter.
Detailed description of the invention
Fig. 1 is the flow chart of 1 method of embodiment；
Fig. 2 is the collected signal waveform related with weldingvoltage of instrument；
Fig. 3 is that resulting signal waveform after signal denoising and signal enhancing is carried out to signal waveform shown in Fig. 2；
Fig. 4 is that resulting signal waveform after signal division is carried out to signal waveform shown in Fig. 3；
Fig. 5 is that resulting signal waveform after signal normalization and signal regularization is carried out to signal waveform shown in Fig. 4；
Fig. 6 is the structural block diagram of 2 system of embodiment；
Fig. 7 is the structural block diagram of 3 device of embodiment.
Specific embodiment
Embodiment 1
A kind of rotating arc welding is seamed into shape parameter detection method, and the forming parameter includes weld pool width and weld penetration,
Referring to Fig.1, comprising the following steps:
Obtain signal related with actual measurement technological parameter in real time during using rotating the arc weld seam；The technological parameter
Including welding torch rotation speed, welding torch radius of turn, welding current, weldingvoltage, torch height, welding torch inclination angle, bevel angle and
Groove size；
Signal related with actual measurement technological parameter is input to the forming parameter pre-established to be fitted after optimization processing
Model；The forming parameter model of fit has recorded the corresponding relationship of technological parameter and forming parameter using random forests algorithm；
Forming parameter is exported after carrying out regression analysis using forming parameter model of fit.
Forming parameter model of fit is the mathematical forecasting model established based on random forests algorithm, wherein having recorded technique ginseng
Several corresponding relationships with forming parameter pass through therefore it may only be necessary to which the technological parameter measured is input in forming parameter model of fit
Corresponding forming parameter can be exported after crossing regression analysis.Technological parameter includes welding torch rotation speed, welding torch radius of turn, weldering
Electric current, weldingvoltage, torch height, welding torch inclination angle, bevel angle and groove size etc. are connect, rotating the arc welding can be used
Existing instrument detected in technology.Technological parameter be in the form of pertinent instruments measuring signal generated existing for,
Technological parameter signal, which is input to before forming parameter model of fit, can be optimized processing.
Forming parameter model of fit used in the present invention be using rotating the arc technology carry out actual production before just
It sets up, is carrying out only needing to carry out regression analysis to the technological parameter measured when actual production using rotating the arc technology,
Therefore process can be surveyed to avoid cumbersome weld pool width and weld penetration.It can recorde multiple groups work in forming parameter model of fit
The relationship of skill parameter and corresponding forming parameter, so that it covers all operating conditions, in this way, passing through forming parameter model of fit
The forming parameter calculated can reflect actual condition, and it is unreliable to overcome the forming parameter that the prior art is predicted
The shortcomings that.Since the used time that forming parameter model of fit carries out regression analysis is seldom, and it is input in forming parameter model of fit
Running parameter measure in real time, therefore forming parameter model of fit output forming parameter be also in real time, to realize
The real-time detection of forming parameter.
It is further used as preferred embodiment, pre-establishes the forming parameter model of fit by following steps:
Set multiple groups test technology parameter；
The test of rotating the arc appearance of weld is carried out, to obtain test corresponding to each group of test technology parameter into parameter
Number；
Random Forest model is constructed using random forests algorithm；
Random Forest model is trained using the multiple groups test technology parameter and corresponding test forming parameter, it will
The Random Forest model obtained after training is as forming parameter model of fit.
The process for pre-establishing forming parameter model of fit mainly includes establishing model and training pattern two parts.Make first
Random Forest model is established with random forests algorithm, Random Forest model indiscipline at this time is equivalent to and does not record technique
The correct corresponding relationship of parameter and forming parameter.Then using multiple groups test technology parameter and corresponding test forming parameter to
Machine forest model is trained, and is equivalent to so that the correct corresponding pass that Random Forest model records technological parameter with forming parameter
System.
Wherein, test technology parameter is setting, and multiple groups test technology parameter can be set so that its covering is as more as possible
Operating condition.By carrying out the test of rotating the arc appearance of weld, it can determine that each group is tested in practical rotating the arc welding condition
Corresponding test forming parameter under technological parameter.
It is further used as preferred embodiment, the Random Forest model has determining maximum decision tree number, leads to
Following steps are crossed to be trained Random Forest model:
Generate decision tree；
The multiple groups test technology parameter is input in Random Forest model, so that Random Forest model output and every group
Test technology parameter predicts forming parameter accordingly；
The deviation between prediction forming parameter and corresponding test forming parameter is calculated, if the deviation reaches default precision
In range, then increase the decision tree number of random forest, otherwise, is back to the step of generating decision tree and re-executes；
The decision tree number of current random forest is compared with maximum decision tree number, if decision tree number be greater than or
Equal to maximum decision tree number, then the training to Random Forest model is completed, otherwise, is back to the step of generating decision tree again
It executes.
Random Forest model just has determining maximum decision tree number after initially setting up.To Random Forest model
In training process, every group of test technology parameter, which is input to after Random Forest model all, can export a corresponding prediction into parameter
Number, can by predict forming parameter and test forming parameter between comparative quantity be used as feedback quantity thus to Random Forest model into
The corresponding adjustment of row, what the adjustment was carried out particular by the generation of decision tree.
In the method for the present invention, battery of tests running parameter corresponds to battery of tests forming parameter, while battery of tests work ginseng
Number also corresponds to one group of prediction forming parameter, therefore tests also to have to correspond between forming parameter and prediction forming parameter and close
System.
It is further used as preferred embodiment, the step of generation decision tree specifically includes:
From the multiple groups test technology parameter selected section test technology parameter as Random Forest model node into
Line splitting, until cannot divide, to generate decision tree.
It is further used as preferred embodiment, it is inclined between the prediction forming parameter and corresponding test forming parameter
Difference is mean square deviation.
In the case where calculating mean square deviation, need to multiple prediction forming parameters and multiple corresponding test forming parameters into
Row statistics.Plurality of prediction forming parameter can be the first prediction weld pool width, the first prediction weld penetration, the second prediction weldering
It stitches molten wide, the second prediction weld penetration, third prediction weld pool width, third and predicts the multiple test forming parameters of weld penetration ...
It can be the first pilot seam molten wide, the first pilot seam fusion penetration, the second pilot seam molten wide, the second pilot seam fusion penetration,
Three pilot seam molten wides, third pilot seam fusion penetration ... thus can according to first prediction weld pool width, the first pilot seam
The meters such as molten wide, the second prediction weld pool width, the second pilot seam molten wide, third prediction weld pool width, third pilot seam molten wide
Molten wide mean square deviation is calculated, according to the first prediction weld penetration, the first pilot seam fusion penetration, the second prediction weld penetration, the second examination
It tests weld penetration, third prediction weld penetration, third pilot seam fusion penetration etc. and calculates fusion penetration mean square deviation, then comprehensively consider molten
Whether wide mean square deviation and fusion penetration mean square deviation have respectively reached required precision.
Predict the deviation between forming parameter and corresponding test forming parameter, the optimization processing includes following at least one
Kind processing step:
Signal denoising, signal enhancing, signal division, signal normalization and signal regularization.
Before it will be input to forming parameter model of fit with the related signal of actual measurement technological parameter, at least one can be passed through
Kind optimization processing.It preferably, can be according to signal denoising, signal enhancing, signal division, signal normalization and signal regularization
Sequence optimizes processing to signal.
The method of the present invention is illustrated with more specifical embodiment below.
Pre-establish the process of forming parameter model of fit:
1. setting multiple groups test technology parameter, and test forming parameter corresponding to each group of test technology parameter is obtained,
Each group of test technology parameter constitutes a data sample with corresponding test forming parameter, shares N number of data sample；
2. have M test technology parameter in each data sample, random selection wherein m test technology parameter (meet m < <
M), lasting division is then carried out as node from selected one test technology parameter of m test technology parameter selection, directly
Until it cannot divide, to establish decision tree for Random Forest model；
3. multiple groups test technology parameter is input in Random Forest model, to export corresponding multiple predictions into parameter
Number；
4. the mean square deviation between multiple prediction forming parameters and corresponding test forming parameter is calculated, if mean square deviation is pre-
If in accuracy rating, then the decision tree number of random forest is added 1, otherwise, it is back to step and 2. re-executes；
5. the decision tree number of current random forest is compared with maximum decision tree number, if decision tree number is equal to
Maximum decision tree number then completes the training to Random Forest model and is otherwise back to step and 2. re-executes.
Carry out using the forming parameter model of fit pre-established the process of appearance of weld parameter real-time detection:
1. as shown in Figure 2 by the collected signal waveform related with weldingvoltage of instrument；
2. carrying out signal denoising to signal shown in Fig. 2 and signal enhancing, gained signal waveform being as shown in Figure 3；
3. dividing to the signal of signal further progress shown in Fig. 3, gained signal waveform is as shown in Figure 4；
4. carrying out signal normalization to signal shown in Fig. 4 and signal regularization, gained signal waveform being as shown in Figure 5；
5. extracting weldingvoltage from signal shown in Fig. 5, the specific of other technological parameters is extracted using similar method
Numerical value, as shown in table 1；
Table 1
| Technological parameter | Survey sampled value |
| Welding torch rotation speed | 30Hz |
| Welding current | 360A |
| Weldingvoltage | 40V |
| Torch height | 20mm |
| Welding gun inclination angle | 15° |
| Bevel angle | 45° |
| Groove size (width) | 25mm |
| Groove size (depth) | 10mm |
6. data shown in table 1 are input in forming parameter model of fit and carry out regression analysis, thus output accordingly at
Shape parameter.
By taking the detection of weld penetration as an example, it is assumed that the result of each tree calculating weld penetration is in forming parameter model of fit
D1=P1(C), wherein D1For one tree calculating weld penetration as a result, P1For the regression function of one tree, C is tree life
At when randomly selected m test technology parameter.So, the overall result of the weld penetration of forming parameter model of fit output isWherein D is weld penetration as a result, PnThe regression function set for n-th, N are the sum of tree, and C is each tree
The randomly selected m test technology parameter of institute.
Embodiment 2
A kind of rotating arc welding of the present invention is seamed into shape parameter detection system, referring to Fig. 6, comprising:
Process parameter measurement module is had for being obtained in real time during using rotating the arc weld seam with actual measurement technological parameter
The signal of pass；The technological parameter includes welding torch rotation speed, welding torch radius of turn, welding current, weldingvoltage, welding torch height
Degree, welding torch inclination angle, bevel angle and groove size；
Technological parameter processing module, it is pre- for that will be input to after optimization processing with the related signal of actual measurement technological parameter
The forming parameter model of fit first established；The forming parameter model of fit using random forests algorithm have recorded technological parameter with
The corresponding relationship of forming parameter；
Forming parameter computing module, for exporting forming parameter after using forming parameter model of fit to carry out regression analysis.
It is further used as preferred embodiment, a kind of rotating arc welding of the present invention is seamed into shape parameter detection system and further includes
Forming parameter model of fit establishes module, and it includes following submodule that the forming parameter model of fit, which establishes module:
Test technology parameter setting module, for setting multiple groups test technology parameter；
Forming parameter measurement module is tested, for carrying out rotating the arc appearance of weld test, to obtain every battery of tests
Test forming parameter corresponding to technological parameter；
Random Forest model constructs module, for constructing Random Forest model using random forests algorithm；
Random Forest model training module, for utilizing the multiple groups test technology parameter and corresponding test forming parameter
Random Forest model is trained, using the Random Forest model obtained after training as forming parameter model of fit.
Each functional module can be the software module for realizing corresponding function, be also possible to have corresponding function
Hardware module.
Embodiment 3
A kind of rotating arc welding of the present invention is seamed into shape parameter detection device, referring to Fig. 7, comprising:
Welding torch detector, for detecting welding torch rotation speed and welding torch radius of turn；
Current detector, for detecting welding current；
Voltage detector, for detecting weldingvoltage；
Torch height detector, for detecting torch height；
Welding torch tilt angle detector, for detecting welding torch inclination angle；
Groove data detector, for detecting bevel angle and groove size；
Data collecting card, for acquiring welding torch detector, current detector, voltage detector, torch height detector, weldering
Data that torch tilt angle detector and groove data detector detect simultaneously upload to computer；
Computer executes method as described in Example 1 for running program to which the data to data collecting card carry out
Processing.
Preferably, computer can select PC machine；Data collecting card can select the PCIE-7822R number of NI company to adopt
Truck, welding torch detector can select the velocity sensor 3040A of Honeywell, and current detector can select Honeywell
Induced current senses CSDA1BA, and voltage detector can select VTV-500DB voltage sensor, and torch height detector can be with
Laser sensor GLL170 is selected, welding gun tilt angle detector can select the obliquity sensor BWS5500 of northern micro sensing, groove number
5 scanning laser sensor of Leica T-Scan of Hai Kesikang can be selected according to detector.
Embodiment 4
A kind of medium of the present invention, the medium have store function and are stored with the executable instruction of processor, the place
The executable instruction of reason device is used to execute when executed by the processor 1 the method for embodiment.Invention medium in the present embodiment
It can be used for 3 described device of embodiment.
It is to be illustrated to preferable implementation of the invention, but the implementation is not limited to the invention above
Example, those skilled in the art can also make various equivalent variations on the premise of without prejudice to spirit of the invention or replace
It changes, these equivalent deformations or replacement are all included in the scope defined by the claims of the present application.
Claims (10)
1. a kind of rotating arc welding is seamed into shape parameter detection method, the forming parameter includes weld pool width and weld penetration,
It is characterized in that, comprising the following steps:
Obtain signal related with actual measurement technological parameter in real time during using rotating the arc weld seam；The technological parameter includes
Welding torch rotation speed, welding torch radius of turn, welding current, weldingvoltage, torch height, welding torch inclination angle, bevel angle and groove
Size；
Signal related with actual measurement technological parameter is input to the forming parameter model of fit pre-established after optimization processing；
The forming parameter model of fit has recorded the corresponding relationship of technological parameter and forming parameter using random forests algorithm；
Forming parameter is exported after carrying out regression analysis using forming parameter model of fit.
2. a kind of rotating arc welding according to claim 1 is seamed into shape parameter detection method, which is characterized in that by following
Step pre-establishes the forming parameter model of fit:
Set multiple groups test technology parameter；
The test of rotating the arc appearance of weld is carried out, to obtain test forming parameter corresponding to each group of test technology parameter；
Random Forest model is constructed using random forests algorithm；
Random Forest model is trained using the multiple groups test technology parameter and corresponding test forming parameter, will be trained
The Random Forest model obtained afterwards is as forming parameter model of fit.
3. a kind of rotating arc welding according to claim 2 is seamed into shape parameter detection method, which is characterized in that described random
Forest model has determining maximum decision tree number, is trained by following steps to Random Forest model:
Generate decision tree；
The multiple groups test technology parameter is input in Random Forest model, so that Random Forest model output and every group of test
Technological parameter predicts forming parameter accordingly；
The deviation between prediction forming parameter and corresponding test forming parameter is calculated, if the deviation reaches default accuracy rating
It is interior, then increase the decision tree number of random forest, otherwise, is back to the step of generating decision tree and re-executes；
The decision tree number of current random forest is compared with maximum decision tree number, if decision tree number is greater than or equal to
Maximum decision tree number then completes the training to Random Forest model, otherwise, is back to the step of generating decision tree and holds again
Row.
4. a kind of rotating arc welding according to claim 3 is seamed into shape parameter detection method, which is characterized in that the generation
The step of decision tree, specifically includes:
Selected section test technology parameter is divided as the node of Random Forest model from the multiple groups test technology parameter
It splits, until cannot divide, to generate decision tree.
5. a kind of rotating arc welding according to claim 3 is seamed into shape parameter detection method, which is characterized in that the deviation
For mean square deviation.
6. a kind of rotating arc welding according to claim 1 is seamed into shape parameter detection method, which is characterized in that the optimization
Processing comprises at least one of the following processing step:
Signal denoising, signal enhancing, signal division, signal normalization and signal regularization.
7. a kind of rotating arc welding is seamed into shape parameter detection system characterized by comprising
Process parameter measurement module, it is related with actual measurement technological parameter for being obtained in real time during using rotating the arc weld seam
Signal；The technological parameter includes welding torch rotation speed, welding torch radius of turn, welding current, weldingvoltage, torch height, weldering
Torch inclination angle, bevel angle and groove size；
Technological parameter processing module is built in advance for signal related with actual measurement technological parameter to be input to after optimization processing
Vertical forming parameter model of fit；The forming parameter model of fit has recorded technological parameter and forming using random forests algorithm
The corresponding relationship of parameter；
Forming parameter computing module, for exporting forming parameter after using forming parameter model of fit to carry out regression analysis.
8. a kind of rotating arc welding according to claim 7 is seamed into shape parameter detection system, which is characterized in that further include into
Shape parameter model of fit establishes module, and it includes following submodule that the forming parameter model of fit, which establishes module:
Test technology parameter setting module, for setting multiple groups test technology parameter；
Forming parameter measurement module is tested, for carrying out rotating the arc appearance of weld test, to obtain each group of test technology
Test forming parameter corresponding to parameter；
Random Forest model constructs module, for constructing Random Forest model using random forests algorithm；
Random Forest model training module, for using the multiple groups test technology parameter and corresponding test forming parameter to
Machine forest model is trained, using the Random Forest model obtained after training as forming parameter model of fit.
9. a kind of rotating arc welding is seamed into shape parameter detection device characterized by comprising
Welding torch detector, for detecting welding torch rotation speed and welding torch radius of turn；
Current detector, for detecting welding current；
Voltage detector, for detecting weldingvoltage；
Torch height detector, for detecting torch height；
Welding torch tilt angle detector, for detecting welding torch inclination angle；
Groove data detector, for detecting bevel angle and groove size；
Data collecting card inclines for acquiring welding torch detector, current detector, voltage detector, torch height detector, welding torch
Data that angle detector and groove data detector detect simultaneously upload to computer；
Computer, for running program to execute if any one of claim 1-7 the method is to the number to data collecting card
According to being handled.
10. a kind of medium, the medium has store function and is stored with the executable instruction of processor, which is characterized in that institute
The executable instruction of processor is stated to be used to execute such as any one of claim 1-7 the method when executed by the processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810999989.XA CN109283378A (en) | 2018-08-30 | 2018-08-30 | A kind of rotating arc welding is seamed into shape parameter detection method, system, device and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810999989.XA CN109283378A (en) | 2018-08-30 | 2018-08-30 | A kind of rotating arc welding is seamed into shape parameter detection method, system, device and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109283378A true CN109283378A (en) | 2019-01-29 |
Family
ID=65184226
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810999989.XA Pending CN109283378A (en) | 2018-08-30 | 2018-08-30 | A kind of rotating arc welding is seamed into shape parameter detection method, system, device and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109283378A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110405388A (en) * | 2019-08-05 | 2019-11-05 | 蕴硕物联技术(上海)有限公司 | Predict the method, apparatus and electronic equipment of welding quality |
| CN111177976A (en) * | 2019-12-25 | 2020-05-19 | 广东省焊接技术研究所（广东省中乌研究院） | Arc welding seam forming accurate prediction method based on deep learning |
| CN111451608A (en) * | 2020-04-08 | 2020-07-28 | 北京博清科技有限公司 | Welding method, welding device, storage medium and processor |
| CN112183555A (en) * | 2020-09-23 | 2021-01-05 | 唐山松下产业机器有限公司 | Method and system for detecting welding quality, electronic device and storage medium |
| CN115609180A (en) * | 2022-12-01 | 2023-01-17 | 江苏新恒基特种装备股份有限公司 | Strengthening parameter generation method and system based on welding process prediction and storage medium |
| CN116561710A (en) * | 2023-05-12 | 2023-08-08 | 西咸新区大熊星座智能科技有限公司 | Welding parameter transfer learning prediction method based on data space conversion |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4366362A (en) * | 1980-03-18 | 1982-12-28 | Ishikawajima-Harima Jukogyo Kabushiki Kaisha | All position TIG welding process |
| CN103551709A (en) * | 2013-11-16 | 2014-02-05 | 沈阳工业大学 | Welding joint forming detection device and method for A-TIG welding |
| CN104636814A (en) * | 2013-11-14 | 2015-05-20 | 中国科学院深圳先进技术研究院 | Method and system for optimizing random forest models |
| CN107179503A (en) * | 2017-04-21 | 2017-09-19 | 美林数据技术股份有限公司 | The method of Wind turbines intelligent fault diagnosis early warning based on random forest |
| CN107563425A (en) * | 2017-08-24 | 2018-01-09 | 长安大学 | A kind of method for building up of the tunnel operation state sensor model based on random forest |
| CN108027911A (en) * | 2015-07-29 | 2018-05-11 | 伊利诺斯工具制品有限公司 | Promote the system and method that welding is service software |
| CN108031954A (en) * | 2017-12-12 | 2018-05-15 | 湘潭市电力工程机械厂 | A kind of welding quality inspection system based on arc-welding machine |
| CN108062720A (en) * | 2017-12-05 | 2018-05-22 | 贵州电网有限责任公司电力科学研究院 | A kind of load forecasting method based on similar day selection and random forests algorithm |
-
2018
- 2018-08-30 CN CN201810999989.XA patent/CN109283378A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4366362A (en) * | 1980-03-18 | 1982-12-28 | Ishikawajima-Harima Jukogyo Kabushiki Kaisha | All position TIG welding process |
| CN104636814A (en) * | 2013-11-14 | 2015-05-20 | 中国科学院深圳先进技术研究院 | Method and system for optimizing random forest models |
| CN103551709A (en) * | 2013-11-16 | 2014-02-05 | 沈阳工业大学 | Welding joint forming detection device and method for A-TIG welding |
| CN108027911A (en) * | 2015-07-29 | 2018-05-11 | 伊利诺斯工具制品有限公司 | Promote the system and method that welding is service software |
| CN107179503A (en) * | 2017-04-21 | 2017-09-19 | 美林数据技术股份有限公司 | The method of Wind turbines intelligent fault diagnosis early warning based on random forest |
| CN107563425A (en) * | 2017-08-24 | 2018-01-09 | 长安大学 | A kind of method for building up of the tunnel operation state sensor model based on random forest |
| CN108062720A (en) * | 2017-12-05 | 2018-05-22 | 贵州电网有限责任公司电力科学研究院 | A kind of load forecasting method based on similar day selection and random forests algorithm |
| CN108031954A (en) * | 2017-12-12 | 2018-05-15 | 湘潭市电力工程机械厂 | A kind of welding quality inspection system based on arc-welding machine |
Non-Patent Citations (2)
| Title |
|---|
| 刘翠霞等: "《旋转电弧焊接工艺参数与焊缝成形的关系》", 《焊接技术》 * |
| 李方方等: "《面向对象随机森林方法在湿地植被分类的应用》", 《遥感信息》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110405388A (en) * | 2019-08-05 | 2019-11-05 | 蕴硕物联技术(上海)有限公司 | Predict the method, apparatus and electronic equipment of welding quality |
| CN110405388B (en) * | 2019-08-05 | 2022-02-11 | 蕴硕物联技术(上海)有限公司 | Method and device for predicting welding quality and electronic equipment |
| CN111177976A (en) * | 2019-12-25 | 2020-05-19 | 广东省焊接技术研究所（广东省中乌研究院） | Arc welding seam forming accurate prediction method based on deep learning |
| CN111177976B (en) * | 2019-12-25 | 2024-04-26 | 广东省科学院中乌焊接研究所 | Arc welding seam forming accurate prediction method based on deep learning |
| CN111451608A (en) * | 2020-04-08 | 2020-07-28 | 北京博清科技有限公司 | Welding method, welding device, storage medium and processor |
| CN112183555A (en) * | 2020-09-23 | 2021-01-05 | 唐山松下产业机器有限公司 | Method and system for detecting welding quality, electronic device and storage medium |
| CN112183555B (en) * | 2020-09-23 | 2022-12-02 | 唐山松下产业机器有限公司 | Method and system for detecting welding quality, electronic device and storage medium |
| CN115609180A (en) * | 2022-12-01 | 2023-01-17 | 江苏新恒基特种装备股份有限公司 | Strengthening parameter generation method and system based on welding process prediction and storage medium |
| CN116561710A (en) * | 2023-05-12 | 2023-08-08 | 西咸新区大熊星座智能科技有限公司 | Welding parameter transfer learning prediction method based on data space conversion |
| CN116561710B (en) * | 2023-05-12 | 2024-02-02 | 西咸新区大熊星座智能科技有限公司 | Welding parameter transfer learning prediction method based on data space conversion |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109283378A (en) | A kind of rotating arc welding is seamed into shape parameter detection method, system, device and medium | |
| CN112154040B (en) | Method for automatically determining optimal welding parameters for performing a weld on a workpiece | |
| CN107088892A (en) | A kind of industrial robot motion accuracy checking method based on binocular vision | |
| CN112632724A (en) | Test design and structured data acquisition method for metal additive manufacturing process system | |
| CN102168945B (en) | System and method for image measurement | |
| CN105400915A (en) | Method and system for quantitatively evaluating distribution of gas flows on blast furnace top | |
| CN107796718A (en) | Brineling system and method | |
| CN112950771B (en) | Detection method and system for surrounding rock structure and related components | |
| CN105224941B (en) | Process identification and localization method | |
| US11982998B2 (en) | Manufacturing monitoring assistance device, manufacturing monitoring assistance method, and manufacturing monitoring assistance program | |
| CN102095388A (en) | System and method for automatically measuring best matching angle of crystals | |
| KR20220085827A (en) | Prediction systems, information processing devices and recording media | |
| CN105651202A (en) | Three-dimensional scanning method and device used for measuring volume of mine | |
| JP5381941B2 (en) | Particle size distribution measuring device | |
| Park et al. | Productivity analysis for the 3D digitization of earthwork sites based on scanning conditions | |
| CN104169960A (en) | Operation support system, operation support device, operation support method and program | |
| CN115907008A (en) | Hydraulic support position and pose adjustment virtual planning system | |
| CN116026487A (en) | Liquid level temperature measuring method, liquid level temperature measuring device, computer equipment and storage medium | |
| CN106558086A (en) | A kind of velocity error correction method of forest stress wave fault imaging | |
| CN114726962A (en) | Method and device for determining working state of pathological section scanner | |
| Cox et al. | Concise process improvement definition with case studies | |
| CN100383778C (en) | System and method for measuring figure file geometric feature | |
| US20220382247A1 (en) | Estimation method and estimation system | |
| JP7286049B1 (en) | Learning device, numerical control device, machining result prediction method and machining system | |
| Fattoruso et al. | A new composite indicator for Manufacturing efficiency |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190129 |