CN109014544A - Miniature resistance spot welding quality on-line monitoring method - Google Patents
Miniature resistance spot welding quality on-line monitoring method Download PDFInfo
- Publication number
- CN109014544A CN109014544A CN201810938371.2A CN201810938371A CN109014544A CN 109014544 A CN109014544 A CN 109014544A CN 201810938371 A CN201810938371 A CN 201810938371A CN 109014544 A CN109014544 A CN 109014544A
- Authority
- CN
- China
- Prior art keywords
- spot welding
- miniature
- resistance spot
- quality
- monitoring
- 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
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
- B23K—SOLDERING OR UNSOLDERING; WELDING; CLADDING OR PLATING BY SOLDERING OR WELDING; CUTTING BY APPLYING HEAT LOCALLY, e.g. FLAME CUTTING; WORKING BY LASER BEAM
- B23K11/00—Resistance welding; Severing by resistance heating
- B23K11/24—Electric supply or control circuits therefor
- B23K11/25—Monitoring devices
- B23K11/252—Monitoring devices using digital means
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
- B23K—SOLDERING OR UNSOLDERING; WELDING; CLADDING OR PLATING BY SOLDERING OR WELDING; CUTTING BY APPLYING HEAT LOCALLY, e.g. FLAME CUTTING; WORKING BY LASER BEAM
- B23K11/00—Resistance welding; Severing by resistance heating
- B23K11/10—Spot welding; Stitch welding
- B23K11/11—Spot welding
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
- B23K—SOLDERING OR UNSOLDERING; WELDING; CLADDING OR PLATING BY SOLDERING OR WELDING; CUTTING BY APPLYING HEAT LOCALLY, e.g. FLAME CUTTING; WORKING BY LASER BEAM
- B23K31/00—Processes relevant to this subclass, specially adapted for particular articles or purposes, but not covered by only one of the preceding main groups
- B23K31/12—Processes relevant to this subclass, specially adapted for particular articles or purposes, but not covered by only one of the preceding main groups relating to investigating the properties, e.g. the weldability, of materials
- B23K31/125—Weld quality monitoring
Landscapes
- Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- Quality & Reliability (AREA)
- Investigating Or Analyzing Materials Using Thermal Means (AREA)
- Resistance Welding (AREA)
Abstract
The present invention relates to a kind of miniature resistance spot welding quality on-line monitoring methods, belong to miniature resistance quality monitoring of resistance spot welding technical field.First with single-photon detector in miniature resistance spot welding production process electrode and workpiece engagement edge number of photons change with time situation carry out real-time monitoring, draw out photon change curve.The photon characteristic quantity of the miniature joint for resistance spot welding quality of different quality grade is extracted using data mining means, generate miniature resistance spot welding quality disaggregated model, the database of the miniature resistance quality monitoring of resistance spot welding based on photon signal is established, realizes miniature joint for resistance spot welding quality automatic on-line monitoring under different technology conditions.On-line monitoring for miniature resistance spot welding quality provides new way, and the single-photon detector used is high to the detectivity of nugget change of temperature field, improves the high reliablity of the online quality-monitoring of miniature resistance spot welding.
Description
Technical field
The present invention relates to miniature resistance quality monitoring of resistance spot welding technical field, in particular to a kind of miniature resistance spot welding quality exists
Line monitoring method.By monitoring the photon variable signal in nearly nugget region in miniature resistance spot welding process, by data mining side
Method extracts the nearly nugget region photon variation characteristic of different miniature joint for resistance spot welding credit ratings, generates miniature resistance spot welding matter
Monitoring model online is measured, the database of the miniature joint for resistance spot welding quality-monitoring based on photon signal is established.
Background technique
Miniature resistance spot welding is widely used in fields such as medical instrument, electronic equipment, battery packages, especially current new energy
The power battery of source automobile is packed, common lithium battery it is series-parallel in miniature resistance spot welding technique be used widely.
The comprehensive function of the multiple physical fields such as the thermopower as involved in spot welding production process causes tack-weld quality unstable
Fixed, especially miniature resistance spot welding, weld interval generally concentrates on 2-3ms, and extremely short weld interval is to miniature resistance spot welding matter
Amount on-line monitoring is put forward higher requirements.
The workpiece of miniature resistance spot welding is relatively thin, and electrode displacement, the variation of electrode force are unobvious in nugget forming process, and
Electrode displacement sensor and electrode dynamic pressure transducer anti-interference ability are poor, are that electrode displacement monitoring method and electrode force monitor
Method brings certain difficulty.
The temperature field in nugget region and tack-weld quality are closely related, existing infrared temperature field on-line monitoring side
Method, the infrared temperature monitor response time used and photon sensitivity are poor, are difficult to capture miniature nugget formation in resistance spot welding and are formed
All processes temperature information.
Summary of the invention
The purpose of the present invention is to provide a kind of miniature resistance spot welding quality on-line monitoring methods, solve present micro electric
Hinder point quality on-line monitoring technique there are the problem of, miniature resistance spot welding matter is established based on nearly nugget region photon signal feature
It measures monitoring model online and establishes miniature resistance quality monitoring of resistance spot welding database in conjunction with a variety of miniature resistance spot welding parameters.
Above-mentioned purpose of the invention is achieved through the following technical solutions:
Miniature resistance spot welding quality on-line monitoring method produced miniature resistance spot welding first with single-photon detector
The number of photons of electrode in the journey and workpiece engagement edge situation that changes with time carries out real-time monitoring, and it is bent to draw out photon variation
Line；By metallographic test determine spot welding different technology conditions under miniature joint for resistance spot welding credit rating, different quality it is miniature
Photon change curve is different in resistance spot welding process, and the miniature resistance spot welding of different quality grade is extracted using data mining means
The photon characteristic quantity of quality generates miniature resistance spot welding quality disaggregated model, establishes the miniature resistance spot welding based on photon signal
The database of quality-monitoring realizes miniature resistance spot welding quality automatic on-line monitoring under different technology conditions；It specifically includes as follows
Step:
Selection of the step (1) to the characteristic quantity of photon signal；
Step (2) analyzes the characteristic quantity of acquisition, and the threshold value for obtaining each characteristic quantity is initial judgment basis；
Step (3), the characteristic quantity according to obtained in step (1) about photon variation, establish miniature resistance spot welding quality
Model, to guarantee miniature resistance quality monitoring of resistance spot welding accuracy rate, using the method for dual model cross validation to miniature resistance spot welding
Quality is monitored.During miniature resistance quality monitoring of resistance spot welding model foundation, at random by the miniature resistance spot welding of initial input
Sample number N points are two parts, respectively training sample set N1 and test sample set N2；
The monitoring model of step (4), the Decision boundaries that step (2) are obtained and step (3) foundation corresponds to technological parameter
It is stored in miniature resistance spot welding monitoring system database.
It is to the selection of the characteristic quantity of monitoring signals described in step (1):
(1.1) welding current is started to the number of photons in welding current end time section to add up, when obtaining welding
Interior total number of photons P；
(1.2) data mining is carried out with the change curve of weld interval t to number of photons c, extracts and reflects miniature resistance spot welding
The photon characteristic quantity of joint quality: the number of photons c of welding current time started is respectively obtainedstart, generally with minimum number of photons one
It causes, maximum photon number cmax, transfer in photon curve number m, and each time corresponding number of photons c that transferstr1、ctr2、
ctr3.。。。。。ctrm, transfer corresponding time t twicetr1、ttr2、ttr3……ttrmAnd balancing photon number institute duration tb；
(1.3) in current duration t, every 0.2ms time interval calculates separately the total number of light photons c in every time1、
c2….cn-1、cnAs characteristic quantity, wherein n=t/0.2；
(1.4) according to formula of varianceCalculate the Variance feature amount σ of each miniature resistance spot welding number of photons2,
Wherein c is to change number of photons with weld interval t, and μ is that welding current started into the welding current end time, all photons
Average；
(1.5) test the photon signal in simultaneously monitoring process to common quality problem in miniature resistance spot welding process,
The miniature joint for resistance spot welding of different quality is obtained, by metallographic test, determines the joint quality etc. of miniature resistance spot welding sample
Grade is respectively labeled as excellent 2, good 3, qualified 4, unqualified 0；
(1.6) the quality classification grade in step (1.5) is only used for during miniature resistance spot welding quality model foundation
Classification, quality model in actual application will excellent 2, good 3, qualified 4 categories combinations be grade 1, that is, lead to and belong to qualified spot welding
Connector 1, unqualified tack-weld are 0；
(1.7) by the characteristic quantity in step (1.1), (1.2), (1.3), (1.4) and four quality in step (1.5) etc.
Grade does correlation analysis, rejects the relevant characteristic quantity of partial linear, obtains and miniature joint for resistance spot welding credit rating phase relation
The biggish characteristic quantity of number.
Step analyzes the characteristic quantity of acquisition described in (2), and the threshold value for obtaining each characteristic quantity is initial judgment basis,
Specifically:
(2.1) probability according to each characteristic quantity of the different miniature joint for resistance spot welding credit ratings of Gaussian Profile statistics is close
Degree；
(2.2) according to the different application of miniature joint for resistance spot welding, different miniature joint for resistance spot welding quality are rejected
The probability density overlapping interval of each characteristic quantity of grade, between determining the suitable high density area of each characteristic quantity, as initial decision threshold
Value；
Above-mentioned credit rating is: excellent 2, good 3, qualification 4, unqualified 0.
Miniature resistance quality monitoring of resistance spot welding model is established described in step (3), specifically:
(3.1) it is based on multiple regression analysis and least square method error criterion, establishes miniature joint for resistance spot welding quality prison
Survey model A；
(3.1.1) divides the photon signal in miniature resistance spot welding process using N1 miniature resistance spot welding samples
Section fitting, and according to the matched curve of least square method criterion searching photon signal；
(3.1.2) is based on N2 miniature resistance spot welding samples and verifies to the matched curve in step (3.1.1), works as survey
When trying accuracy higher than setting value, the photon signal curve of qualified spot welding sample is confirmed；
(3.2) miniature joint for resistance spot welding quality-monitoring Model B is established based on combination nearest neighbour classification device KNN algorithm；
(3.2.1) is used different using N1 miniature resistance spot welding samples in the case where number of classifying determines respectively
The measurement criterion of KNN algorithm, binding characteristic amount establish miniature resistance spot welding monitoring model, use N2 miniature resistance spot welding samples
Different measurement criterion models are verified, record each model to the monitoring result of N2 miniature joint for resistance spot welding；
(3.2.2) gives each the monitoring accuracy of different measurement criterions according to the verification result of step (3.2.1)
The different weight of disaggregated model, to optimize the accuracy rate of miniature resistance spot welding monitoring model.
When new miniature resistance spot welding characteristic quality of sample enters system, first determine whether the sample all characteristic quantities whether
In initial decision threshold section all in step (2), it can directly judge that the tack-weld belongs to acceptable splice 1 or unqualified
Connector 0；
If all characteristic quantities are all in the initial decision threshold interval that step (2) obtains, miniature joint for resistance spot welding
Mass monitoring system output identification 1 illustrates that the miniature joint for resistance spot welding belongs to qualified sample；
If there are Partial Feature amounts not in the initial decision threshold interval of step (2) for the characteristic quantity of sample, need
To the miniature resistance quality monitoring of resistance spot welding model that miniature resistance spot welding sample further utilizes step (3) to establish, micro electric is judged
Hinder the credit rating of spot welding sample；When the monitoring result of miniature joint for resistance spot welding quality-monitoring model A, B are consistent, if micro-
The monitoring result of type joint for resistance spot welding quality-monitoring model A, B belong to acceptable splice 1, then monitor system and export 1 automatically, such as
The monitoring result of miniature joint for resistance spot welding quality-monitoring model A, the B of fruit belongs to unqualified connector 0, then it is automatically defeated to monitor system
Out 0；
When the monitoring result of miniature joint for resistance spot welding quality-monitoring model A, B are inconsistent, then need for the micro electric
Resistance tack-weld gives abnormality processing, reminds operator using other detection methods or manually checks the miniature resistance spot welding
Whether connector meets requirement.
The beneficial effects of the present invention are: using the response time up to the highly sensitive single-photon detector monitoring of nanosecond
Weld interval extremely short miniature resistance spot welding process, according to the photon in nearly nugget region with weld interval changing rule to micro electric
Resistance tack-weld quality is monitored.There are correlation, tack-weld forming process and temperature fields for the variation of number of photons and temperature
Closely related, the correlation of the two is that the present invention provides Important Theoretic Foundations.Present invention firstly provides be based on photon signal
The characteristic quantity of point quality proposes to establish with Gaussian Profile aiming at the problem that initial decision boundary can not be set based on experience value
The method combined is rejected with overlay space in the probability density space of the miniature joint for resistance spot welding of different brackets, determines initial judgement
Boundary.The present invention proposes to guarantee effective detection of unqualified solder joint using the method for dual model cross validation, and establishes micro electric
Hinder quality monitoring of resistance spot welding database.
Detailed description of the invention
The drawings described herein are used to provide a further understanding of the present invention, constitutes part of this application, this hair
Bright illustrative example and its explanation is used to explain the present invention, and is not constituted improper limitations of the present invention.
Fig. 1 is system module flow chart of the invention；
Fig. 2 is that miniature resistance spot welding of the invention monitors system initialization interface；
Fig. 3 is that miniature resistance spot welding of the invention monitors system working interface.
Specific embodiment
Detailed content and its specific embodiment of the invention are further illustrated with reference to the accompanying drawing.
Referring to shown in Fig. 1 to Fig. 3, miniature resistance spot welding quality on-line monitoring method of the invention, when first with response
Between up to nanosecond single-photon detector in miniature resistance spot welding production process electrode and workpiece engagement edge number of photons
The situation that changes with time carries out real-time monitoring, draws out photon change curve；It is tested by metallographic and determines spot welding different process
Under the conditions of miniature joint for resistance spot welding credit rating, since photon variation reflects the variation in temperature field to a certain extent, because
Photon change curve is different in the miniature resistance spot welding process of this different quality, extracts different quality etc. using data mining means
The photon characteristic quantity of the miniature joint for resistance spot welding quality of grade, generates miniature joint for resistance spot welding quality classification model, establishes base
In the database of the miniature joint for resistance spot welding quality-monitoring of photon signal, miniature resistance spot welding under different technology conditions is realized
Head quality automatic on-line monitoring.The present invention provides new way for the on-line monitoring of miniature resistance spot welding quality, the monochromatic light used
Sub- detector is high to the detectivity of nugget change of temperature field, and improve the online quality-monitoring of miniature resistance spot welding can
By property height.
Referring to shown in Fig. 1 to Fig. 3, miniature resistance spot welding quality on-line monitoring method of the invention includes the following steps:
Step (1) establishes human-machine interactive information module: for searching for whether miniature resistance quality monitoring of resistance spot welding database is deposited
In the miniature resistance spot welding parameter that will be used；
Step (2), for having existed identical spot welding parameter in database, only need to be selected；
Step (3), if there is no identical technological parameter, need to input new miniature resistance spot welding process ginseng
Number: workpiece material 1, workpiece material 2,1 plate thickness d of workpiece1, 2 plate thickness d of workpiece2, selection welding current mode (Pc pulse current, Ac
Alternating current), electric current continue conduction time t, electrode pressure F, environment temperature T, initial model establish miniature resistance spot welding sample
This number N
Step (4), in miniature resistance spot welding process, the real-time acquisition of photon signal: the list being exceedingly fast using the response time
Photon detector detects the photon signal in nearly nugget region, and wherein the spectral response range of single-photon detector selects infrared waves
Section；
Step (5), in photon signal monitoring process, avoid the interference of extraneous infrared waves, influence miniature resistance spot welding
Quality-monitoring accuracy reduces interference of the clutter to miniature resistance quality monitoring of resistance spot welding result by three kinds of method comprehensive functions:
It is filtered in monitoring device using the infrared fileter to match with single-photon detector spectral band, photon detector is obtained
The photon signal taken filters out the radio-frequency component unrelated with miniature joint for resistance spot welding quality, at data by filter circuit
During reason, threshold filter is carried out to the singular point signal of photonic data；
The number of photons that step (6), single-photon detector detect, which changes with time, can reflect out the variation of temperature, draw
Photon signal processed is with the change curve of miniature resistance spot welding time, and wherein time interval is that welding current starts to welding current knot
Beam；
A coefficient ε is introduced for the property and surface state of different materials, according to Planck law absolute black body
Known to relational expression between radianting capacity and wavelength and temperature:
M (λ, T)=ε * C1λ-5/[exp(C2/λT)-1]
Wherein, λ-wavelength；T-absolute temperature；C1、C2- radiation constant.According to infrared measurement of temperature principle it is found that single-photon detecting
The number of photons that survey device detects, which changes with time, can reflect out the variation of temperature, draw photon versus time curve,
Wherein time interval is that welding current starts to welding current to terminate.
The selection of step (7), characteristic quantity:
(7.1) temperature of the nugget size of miniature joint for resistance spot welding, nugget interior tissue and resistance heat is closely related, will
Welding current starts to the number of photons in welding current end time section to add up, and obtains total number of photons P in weld interval；
(7.2) data mining is carried out with the change curve of welding duration t to photon c, extracts and reflects miniature point of resistance
The photon characteristic quantity of plumb joint quality: the number of photons c of welding current time started is respectively obtainedstart, generally with minimum number of photons
Unanimously, maximum photon number cmax, transfer in photon curve number m, and each time corresponding number of photons c that transferstr1、ctr2、
ctr3.。。。。。ctrm, transfer corresponding time t twicetr1、ttr2、ttr3……ttrmAnd balancing photon number institute duration tb；
(7.3) in current duration t, every 0.2ms time interval calculates separately the total number of light photons p in every time1、
p2….pn-1、pnAs characteristic quantity, wherein n=t/0.2；
(7.4) according to formula of varianceCalculate the Variance feature amount σ of each miniature resistance spot welding number of photons2,
Wherein c is to change number of photons with the welding duration, and μ is that welding current started into the welding current end time, all photons
Average；
(7.5) test the photon signal in simultaneously monitoring process to common quality problem in miniature resistance spot welding process,
The miniature joint for resistance spot welding of different quality is obtained, by metallographic test, determines the joint quality etc. of miniature resistance spot welding sample
Grade is respectively labeled as excellent 2, good 3, qualified 4, unqualified 0；
(7.6) the quality classification grade in step 7.5) is only used for point during miniature resistance spot welding quality model foundation
Class, quality model in actual application by excellent 2, good 3, qualified 4 categories combinations be grade 1, i.e., it is logical to belong to qualified spot welding
First 1, unqualified tack-weld is 0；
(7.7) by the characteristic quantity in step (7.1), (7.2), (7.3), (7.4) and four quality in step (7.5) etc.
Grade does correlation analysis, rejects the relevant characteristic quantity of partial linear, obtains and miniature joint for resistance spot welding credit rating phase relation
The biggish characteristic quantity of number.
Step (8) analyzes the characteristic quantity of acquisition, and the threshold value for obtaining each characteristic quantity is initial judgment basis:
(8.1) according to the different miniature joint for resistance spot welding credit rating of Gaussian Profile statistics (excellent 2, it is good 3, qualification 4, do not conform to
Lattice 0) each characteristic quantity probability density；
(8.2) according to the different application of miniature joint for resistance spot welding, different miniature joint for resistance spot welding quality are rejected
Grade (excellent 2, good 3, qualification 4, unqualified 0) each characteristic quantity probability density overlapping interval, determine that each characteristic quantity is suitably high
Density section, as initial decision threshold value.
Step (9), the characteristic quantity according to obtained in step (7) about photon variation, establish miniature resistance spot welding quality
Model, to guarantee miniature resistance quality monitoring of resistance spot welding accuracy rate, using the method for dual model cross validation to miniature resistance spot welding
Connector is verified, during miniature joint for resistance spot welding quality-monitoring model foundation, at random by the miniature resistance of initial input
Spot welding sample number N points are two parts, respectively training sample set N1 and test sample set N2:
(9.1) it is based on multiple regression analysis and least square method error criterion, establishes miniature joint for resistance spot welding quality prison
Survey model A；
(9.1.1) divides the photon signal in miniature resistance spot welding process using N1 miniature resistance spot welding samples
Section fitting, and according to the matched curve of least square method criterion searching photon signal；
(9.1.2) is based on N2 miniature resistance spot welding samples and verifies to the matched curve in step (9.1.1), works as survey
When trying accuracy higher than setting value, the photon signal curve of qualified spot welding sample is confirmed；
(9.2) miniature joint for resistance spot welding quality-monitoring Model B is established based on combination nearest neighbour classification device KNN algorithm；
(9.2.1) is used different using N1 miniature resistance spot welding samples in the case where number of classifying determines respectively
The measurement criterion of KNN algorithm:
Euclidean distance formula:
Manhatton distance:
And vector space cosine similarity:
Miniature resistance spot welding monitoring model is established in conjunction with features described above amount, using N2 miniature resistance spot welding samples to difference
The verifying of measurement criterion model records each model to the monitoring result of N2 miniature joint for resistance spot welding；
(9.2.2) gives each the monitoring accuracy of different measurement criterions according to the verification result of step (9.2.1)
The different weight of disaggregated model, to optimize the accuracy rate of miniature resistance spot welding monitoring model.
The monitoring model of step (10), the Decision boundaries that step (8) are obtained and step (9) foundation corresponds to technological parameter
It is stored in miniature resistance spot welding monitoring system database.
When new miniature resistance spot welding characteristic quality of sample enters system, first determine whether the sample all characteristic quantities whether
In initial decision threshold section all in step (8), it can directly judge that the tack-weld belongs to acceptable splice 1 or unqualified
Connector 0；
If all characteristic quantities are all in the initial decision threshold interval that step (8) obtains, spot welding system output identification
1, that is, illustrate that the miniature joint for resistance spot welding belongs to qualified sample；
If there are Partial Feature amounts not in the initial decision threshold interval of step (8) for the characteristic quantity of sample, need
To the miniature resistance quality monitoring of resistance spot welding model that miniature resistance spot welding sample further utilizes step (9) to establish, micro electric is judged
Hinder the credit rating of spot welding sample；When the monitoring result of miniature joint for resistance spot welding quality-monitoring model A, B are consistent, if micro-
The monitoring result of type joint for resistance spot welding quality-monitoring model A, B belong to acceptable splice 1, then monitor system and export 1 automatically, such as
The monitoring result of miniature resistance quality monitoring of resistance spot welding model A, the B of fruit belongs to unqualified connector 0, then monitors system and export 0 automatically；
When the monitoring result of miniature resistance quality monitoring of resistance spot welding model A, B are inconsistent, then need for the miniature point of resistance
Plumb joint gives abnormality processing, reminds operator using other detection methods or manually checks the miniature joint for resistance spot welding
Whether requirement is met.
Embodiment 1:
With the single photon monitoring in the miniature resistance spot welding process of lithium battery series and parallel in new-energy automobile power battery group
For, wherein the positive and negative extremely stainless steel case of 0.3mm of battery core, busbar connector are the nickel sheet of 0.1mm, welding current 1.5kA, weldering
Connecing the time is 2ms, electrode diameter 1.5mm.Operating wavelength range is used to detect for the near-infrared single photon of 900~1700nm
Device, fastest response speed are up to 10ns, and ultrared single-photon detector overall dimensions are the cylinders that radius is a height of 100mm of 50mm
Body, the frequency acquisition of data acquisition module are 500Mhz.
It is shown in Figure 1, it is data acquisition and procession module flow diagram of the present invention, is entering miniature resistance spot welding single photon
After monitoring system, system interface is shown in Figure 2, and interface center shows current welding material details and miniature resistance spot welding work
Skill condition.
If the miniature resistance spot welding parameter that will be used does not change, clicks directly on middle button and enter prison
Examining system.If using new miniature resistance spot welding parameter, need to click lower left search technological parameter by
Button scans for model existing in database.
When searching the monitoring model under corresponding process conditions in the database, can directly select into monitoring state.
When the monitoring model of corresponding process parameters is not present in database, technological parameter required for inputting in systems, to lithium electricity
The miniature resistance spot welding process of pond series and parallel carries out single photon signal monitoring.
1) characteristic quantity of monitoring signals is chosen:
The temperature of the nugget sizes of 1.1 miniature joint for resistance spot welding, nugget interior tissue and resistance heat is closely related, from weldering
It connects electric current to start to the number of photons in welding current end time section to add up, obtains total number of photons P in weld interval.
The change curve of 1.2 couples of photon c t at any time carries out data mining, extracts and reflects miniature joint for resistance spot welding quality
Photon characteristic quantity: respectively obtain the number of photons c of welding current time startedstart(generally consistent with minimum number of photons), it is maximum
Number of photons cmax, transfer in photon curve number m, and to each time corresponding number of photons c that transferstr1、ctr2、ctr3.。。。。。ctrm,
Transfer corresponding time t twicetr1、ttr2、ttr3……ttrmAnd balancing photon number institute duration tb。
1.3 in weld interval t, and every 0.2ms time interval calculates separately the total number of light photons p in every time1、p2…
.pn-1、pnAs characteristic quantity, wherein n=t/0.2.
1.4 according to formula of varianceCalculate the Variance feature amount σ of each miniature resistance spot welding number of photons2,
Middle c is to change number of photons with the welding duration, and μ is that welding current started into the welding current end time, all photons
Average.
Common quality problem test the photon signal in simultaneously monitoring process in 1.5 pairs of miniature resistance spot welding processes, obtains
The joint quality grade of miniature resistance spot welding sample is determined by metallographic test to the miniature joint for resistance spot welding of different quality,
It is respectively labeled as excellent 2, good 3, qualified 4, unqualified 0.
1.6 it is above-mentioned 1.5) in quality classification grade be only used for point during miniature resistance spot welding quality model foundation
Class, quality model in actual application by excellent 2, good 3, qualified 4 categories combinations be grade 1, i.e., it is logical to belong to qualified spot welding
First 1, unqualified tack-weld is 0.
1.7 by it is above-mentioned 1.1) 1.2) 1.3) 1.4) in each characteristic quantity to it is above-mentioned 7.5) in four credit ratings do it is related
Property analysis, reject the relevant characteristic quantity of partial linear, obtain biggish with miniature joint for resistance spot welding credit rating related coefficient
Characteristic quantity.
2) above-mentioned 1) the middle characteristic quantity obtained is analyzed, the threshold value for obtaining each characteristic quantity is initial judgment basis:
2.1 according to the different miniature joint for resistance spot welding credit rating of Gaussian Profile statistics (excellent 2, it is good 3, qualification 4, unqualified
0) each characteristic quantity is in probability density.
2.2, according to the different application of miniature joint for resistance spot welding, reject different miniature joint for resistance spot welding quality etc.
Grade (excellent 2, good 3, qualification 4, unqualified 0) each characteristic quantity probability density overlapping interval, determine that each characteristic quantity is suitably highly dense
Section is spent, as initial decision threshold value.
3) the various characteristic quantities about photon variation according to obtained in 1), establish miniature resistance spot welding quality model, are
Guarantee that miniature resistance quality monitoring of resistance spot welding accuracy rate, the present invention are spot welded miniature resistance using the method for dual model cross validation
Head, during miniature joint for resistance spot welding quality-monitoring model foundation, at random by the miniature resistance spot welding sample number N of initial input
It is divided into two parts, respectively training sample set N1 and test sample set N2:
3.1 are based on multiple regression analysis and least square method error criterion, establish miniature joint for resistance spot welding quality-monitoring
Model A.
3.1.1 using N1 miniature resistance spot welding samples, the photon signal in miniature resistance spot welding process is segmented
Fitting, and according to the matched curve of the photon signal under least square method criterion searching specific process conditions.
3.1.2 the matched curve in 3.1.1 is verified based on N2 miniature resistance spot welding samples, works as test accuracy
When higher than setting value, the photon signal curve of qualified spot welding sample is confirmed.
3.2 establish miniature joint for resistance spot welding quality-monitoring Model B based on combination KNN (nearest neighbour classification device) algorithm.
3.2.1 using N1 miniature resistance spot welding samples, in the case where number of classifying determines, respectively using different
The measurement criterion of KNN algorithm establishes miniature resistance spot welding monitoring model in conjunction with features described above amount, uses N2 miniature resistance spot weldings
Sample verifies different measurement criterion models, records each model to the monitoring result of N2 miniature joint for resistance spot welding.
3.2.2 this point is given for the monitoring accuracy of different each measurement criterions according to the verification result of above-mentioned 3.2.1
The different weight of class model, to optimize the accuracy rate of miniature resistance spot welding monitoring model.
4) by above-mentioned steps 2) obtain Decision boundaries and above-mentioned steps 3) establish monitoring model correspond to technological parameter
It is stored in miniature resistance spot welding monitoring system database.
5) when new miniature each characteristic quantity of resistance spot welding sample enters system, all characteristic quantities of the sample are first determined whether
Whether all in above-mentioned steps 8) in initial decision threshold section in, can directly judge that the tack-weld belongs to acceptable splice
1 or unqualified connector 0.
If 6) all characteristic quantities are all in above-mentioned steps 2) in obtained initial decision threshold interval, spot welding system output
Mark 1, that is, illustrate that the miniature joint for resistance spot welding belongs to qualified sample.
If 7) there are Partial Feature amounts not in above-mentioned steps 2 for the characteristic quantity of the sample) in initial decision threshold interval
Interior, then need further to utilize above-mentioned steps 3 to the miniature resistance spot welding sample) in the miniature resistance quality monitoring of resistance spot welding established
Model judges the credit rating of the miniature resistance spot welding sample.When the monitoring result of model A, B are consistent, if model A, B
Monitoring result monitoring belongs to acceptable splice 1, then monitors system and export 1 automatically, if the monitoring result of model A, B belong to not
Acceptable splice 0 then monitors system and exports 0 automatically.
It when the monitoring result of model A, B are inconsistent, then needs to give abnormality processing to the miniature resistance spot welding, reminds behaviour
Make personnel using other detection methods or manually checks whether the tack-weld meets requirement.
The foregoing is merely preferred embodiments of the invention, are not intended to restrict the invention, for the technology of this field
For personnel, the invention may be variously modified and varied.All any modification, equivalent substitution, improvement and etc. made for the present invention,
It should all be included in the protection scope of the present invention.
Claims (5)
1. a kind of miniature resistance spot welding quality on-line monitoring method, it is characterised in that: first with single-photon detector to miniature
The number of photons of electrode in the resistance spot welding production process and workpiece engagement edge situation that changes with time carries out real-time monitoring, draws
Produce photon change curve；It is tested by metallographic and determines miniature joint for resistance spot welding credit rating under spot welding different technology conditions,
Photon change curve is different in the miniature resistance spot welding process of different quality, extracts different quality grade using data mining means
Miniature resistance spot welding quality photon characteristic quantity, generate miniature resistance spot welding quality disaggregated model, establish and be based on photon signal
Miniature resistance quality monitoring of resistance spot welding database, realize miniature resistance spot welding quality automatic on-line prison under different technology conditions
It surveys；Specifically comprise the following steps:
Selection of the step (1) to the characteristic quantity of photon signal；
Step (2) analyzes the characteristic quantity of acquisition, and the threshold value for obtaining each characteristic quantity is initial judgment basis；
Step (3), the characteristic quantity according to obtained in step (1) about photon variation, establish miniature resistance spot welding quality model,
To guarantee miniature resistance quality monitoring of resistance spot welding accuracy rate, using dual model cross validation method to miniature resistance spot welding quality into
Row monitoring；During miniature resistance quality monitoring of resistance spot welding model foundation, at random by the miniature resistance spot welding sample number N of initial input
It is divided into two parts, respectively training sample set N1 and test sample set N2；
The monitoring model of step (4), the Decision boundaries that step (2) are obtained and step (3) foundation is stored in corresponding to technological parameter
Miniature resistance spot welding monitors system database.
2. miniature resistance spot welding quality on-line monitoring method according to claim 1, it is characterised in that: step (1) is described
The selection of the characteristic quantity to monitoring signals be:
(1.1) welding current is started to the number of photons in welding current end time section to add up, is obtained in weld interval
Total number of photons P；
(1.2) data mining is carried out with the change curve of weld interval t to number of photons c, extracts and reflects miniature joint for resistance spot welding
The photon characteristic quantity of quality: the number of photons c of welding current time started is respectively obtainedstart, it is generally consistent with minimum number of photons, most
Big number of photons cmax, transfer in photon curve number m, and each time corresponding number of photons c that transferstr1、ctr2、ctr3.。。。。。ctrm,
Transfer corresponding time t twicetr1、ttr2、ttr3……ttrmAnd balancing photon number institute duration tb；
(1.3) in current duration t, every 0.2ms time interval calculates separately the total number of light photons c in every time1、c2…
.cn-1、cnAs characteristic quantity, wherein n=t/0.2；
(1.4) according to formula of varianceCalculate the Variance feature amount σ of each miniature resistance spot welding number of photons2, wherein
C is to change number of photons with weld interval t, and μ is that welding current started into the welding current end time, and all photons are averaged
Number；
(1.5) test the photon signal in simultaneously monitoring process to common quality problem in miniature resistance spot welding process, obtain
The miniature joint for resistance spot welding of different quality determines the joint quality grade of miniature resistance spot welding sample by metallographic test, point
It is not labeled as excellent 2, good 3, qualified 4, unqualified 0；
(1.6) the quality classification grade in step (1.5) is only used for the classification during miniature resistance spot welding quality model foundation,
Quality model in actual application by excellent 2, good 3, qualified 4 categories combinations be grade 1, i.e., it is logical to belong to qualified tack-weld 1,
Unqualified tack-weld is 0；
(1.7) characteristic quantity in step (1.1), (1.2), (1.3), (1.4) is done with four credit ratings in step (1.5)
Correlation analysis rejects the relevant characteristic quantity of partial linear, obtain with miniature joint for resistance spot welding credit rating related coefficient compared with
Big characteristic quantity.
3. miniature resistance spot welding quality on-line monitoring method according to claim 1, it is characterised in that: step (2) is described
The characteristic quantity of acquisition is analyzed, obtain each characteristic quantity threshold value be initial judgment basis, specifically:
(2.1) according to the probability density of each characteristic quantity of the different miniature joint for resistance spot welding credit ratings of Gaussian Profile statistics；
(2.2) according to the different application of miniature joint for resistance spot welding, different miniature joint for resistance spot welding credit ratings are rejected
Each characteristic quantity probability density overlapping interval, between determining the suitable high density area of each characteristic quantity, as initial decision threshold value；
Above-mentioned credit rating is: excellent 2, good 3, qualification 4, unqualified 0.
4. miniature resistance spot welding quality on-line monitoring method according to claim 1, it is characterised in that: step (3) is described
Establish miniature resistance quality monitoring of resistance spot welding model, specifically:
(3.1) it is based on multiple regression analysis and least square method error criterion, establishes miniature joint for resistance spot welding quality-monitoring mould
Type A；
(3.1.1) carries out segmentation to the photon signal in miniature resistance spot welding process and intends using N1 miniature resistance spot welding samples
It closes, and finds the matched curve of photon signal according to least square method criterion；
(3.1.2) is based on N2 miniature resistance spot welding samples and verifies to the matched curve in step (3.1.1), when test is quasi-
When exactness is higher than setting value, the photon signal curve of qualified spot welding sample is confirmed；
(3.2) miniature joint for resistance spot welding quality-monitoring Model B is established based on combination nearest neighbour classification device KNN algorithm；
(3.2.1) uses different KNN in the case where number of classifying determines using N1 miniature resistance spot welding samples respectively
The measurement criterion of algorithm, binding characteristic amount establish miniature resistance spot welding monitoring model, use N2 miniature resistance spot welding samples pair
Different measurement criterion model verifyings, record each model to the monitoring result of N2 miniature joint for resistance spot welding；
(3.2.2) gives each classification for the monitoring accuracy of different measurement criterions according to the verification result of step (3.2.1)
The different weight of model, to optimize the accuracy rate of miniature resistance spot welding monitoring model.
5. miniature resistance spot welding quality on-line monitoring method according to any one of claim 1 to 4, it is characterised in that:
When new miniature resistance spot welding characteristic quality of sample enters system, first determine whether all characteristic quantities of the sample whether all in step
(2) in the initial decision threshold section in, it can directly judge that the tack-weld belongs to acceptable splice 1 or unqualified connector 0；
If all characteristic quantities are all in the initial decision threshold interval that step (2) obtains, miniature joint for resistance spot welding quality
Monitoring system output identification 1 illustrates that the miniature joint for resistance spot welding belongs to qualified sample；
If there are Partial Feature amounts not in the initial decision threshold interval of step (2) for the characteristic quantity of sample, need to micro-
The miniature resistance quality monitoring of resistance spot welding model that type resistance spot welding sample further utilizes step (3) to establish, judges miniature point of resistance
Weld the credit rating of sample；When the monitoring result of miniature joint for resistance spot welding quality-monitoring model A, B are consistent, if micro electric
The monitoring result of resistance tack-weld quality-monitoring model A, B belong to acceptable splice 1, then monitor system and export 1 automatically, if micro-
The monitoring result of type joint for resistance spot welding quality-monitoring model A, B belong to unqualified connector 0, then monitor system and export 0 automatically；
When the monitoring result of miniature joint for resistance spot welding quality-monitoring model A, B are inconsistent, then need for the miniature point of resistance
Plumb joint gives abnormality processing, reminds operator using other detection methods or manually checks the miniature joint for resistance spot welding
Whether requirement is met.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810938371.2A CN109014544B (en) | 2018-08-17 | 2018-08-17 | Micro resistance spot welding quality on-line monitoring method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810938371.2A CN109014544B (en) | 2018-08-17 | 2018-08-17 | Micro resistance spot welding quality on-line monitoring method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109014544A true CN109014544A (en) | 2018-12-18 |
| CN109014544B CN109014544B (en) | 2020-07-28 |
Family
ID=64630790
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810938371.2A Active CN109014544B (en) | 2018-08-17 | 2018-08-17 | Micro resistance spot welding quality on-line monitoring method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109014544B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110227878A (en) * | 2019-06-27 | 2019-09-13 | 天津七所高科技有限公司 | A kind of method that electric resistance welding equipment detection nut misplaces or loses |
| CN111390367A (en) * | 2020-05-13 | 2020-07-10 | 天津七所高科技有限公司 | Method for identifying welding spot splashing in resistance spot welding |
| CN111681235A (en) * | 2020-06-11 | 2020-09-18 | 广东工业大学 | IC welding spot defect detection method based on learning mechanism |
| CN112453750A (en) * | 2019-09-09 | 2021-03-09 | 英业达科技有限公司 | System and method for establishing detection model according to standard value to confirm welding state |
| CN112557416A (en) * | 2019-09-09 | 2021-03-26 | 英业达科技有限公司 | System and method for detecting whether welding spots are bridged or not by using deep learning model |
| CN112756759A (en) * | 2021-01-11 | 2021-05-07 | 上海智能制造功能平台有限公司 | Spot welding robot workstation fault judgment method |
| CN114147382A (en) * | 2021-12-14 | 2022-03-08 | 吉林大学 | Electric and ultrasonic signal composite resistance spot welding quality on-line monitoring method and device |
| CN117436769A (en) * | 2023-12-20 | 2024-01-23 | 山东方垠智能制造有限公司 | Structural part welding quality monitoring method, system, storage medium and equipment |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4079225A (en) * | 1976-08-04 | 1978-03-14 | Warner Allan S | Fiber optic/photon detector for brazing machine |
| CN103862136A (en) * | 2014-03-14 | 2014-06-18 | 吉林大学 | Monitoring device and monitoring method of fusion-welding process |
| CN105750754A (en) * | 2016-05-06 | 2016-07-13 | 广州市精源电子设备有限公司 | Method and system for identifying resistance spot welding quality affecting factors |
| US20170151628A1 (en) * | 2014-03-24 | 2017-06-01 | James Eldon Craig | Additive manufacturing temperature controller/sensor apparatus and method of use thereof |
| CN107160048A (en) * | 2017-05-22 | 2017-09-15 | 哈尔滨工业大学（威海） | A kind of inline diagnosis method of the laser welding defect based on spectral information |
| KR20170136111A (en) * | 2016-05-31 | 2017-12-11 | 피에프케이 (주) | Smart welding machine monitoring system |
| CN108188560A (en) * | 2017-12-12 | 2018-06-22 | 广州亨龙智能装备股份有限公司 | Hand-held electric resistance welding Quality Monitoring Control System based on linux system |
-
2018
- 2018-08-17 CN CN201810938371.2A patent/CN109014544B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4079225A (en) * | 1976-08-04 | 1978-03-14 | Warner Allan S | Fiber optic/photon detector for brazing machine |
| CN103862136A (en) * | 2014-03-14 | 2014-06-18 | 吉林大学 | Monitoring device and monitoring method of fusion-welding process |
| US20170151628A1 (en) * | 2014-03-24 | 2017-06-01 | James Eldon Craig | Additive manufacturing temperature controller/sensor apparatus and method of use thereof |
| CN105750754A (en) * | 2016-05-06 | 2016-07-13 | 广州市精源电子设备有限公司 | Method and system for identifying resistance spot welding quality affecting factors |
| KR20170136111A (en) * | 2016-05-31 | 2017-12-11 | 피에프케이 (주) | Smart welding machine monitoring system |
| CN107160048A (en) * | 2017-05-22 | 2017-09-15 | 哈尔滨工业大学（威海） | A kind of inline diagnosis method of the laser welding defect based on spectral information |
| CN108188560A (en) * | 2017-12-12 | 2018-06-22 | 广州亨龙智能装备股份有限公司 | Hand-held electric resistance welding Quality Monitoring Control System based on linux system |
Non-Patent Citations (2)
| Title |
|---|
| 万晓东: "微电阻点焊条件下的质量控制问题研究", 《中国博士学位论文全文数据库 工程科技Ⅰ辑》 * |
| 文静: "电阻点焊质量监测智能化及系统研究", 《中国博士学位论文全文数据库 工程科技Ⅰ辑》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110227878A (en) * | 2019-06-27 | 2019-09-13 | 天津七所高科技有限公司 | A kind of method that electric resistance welding equipment detection nut misplaces or loses |
| CN112453750A (en) * | 2019-09-09 | 2021-03-09 | 英业达科技有限公司 | System and method for establishing detection model according to standard value to confirm welding state |
| CN112557416A (en) * | 2019-09-09 | 2021-03-26 | 英业达科技有限公司 | System and method for detecting whether welding spots are bridged or not by using deep learning model |
| CN111390367A (en) * | 2020-05-13 | 2020-07-10 | 天津七所高科技有限公司 | Method for identifying welding spot splashing in resistance spot welding |
| CN111390367B (en) * | 2020-05-13 | 2021-09-03 | 天津七所高科技有限公司 | Method for identifying welding spot splashing in resistance spot welding |
| CN111681235A (en) * | 2020-06-11 | 2020-09-18 | 广东工业大学 | IC welding spot defect detection method based on learning mechanism |
| CN111681235B (en) * | 2020-06-11 | 2023-05-09 | 广东工业大学 | IC welding spot defect detection method based on learning mechanism |
| CN112756759A (en) * | 2021-01-11 | 2021-05-07 | 上海智能制造功能平台有限公司 | Spot welding robot workstation fault judgment method |
| CN112756759B (en) * | 2021-01-11 | 2022-04-08 | 上海智能制造功能平台有限公司 | Spot welding robot workstation fault judgment method |
| CN114147382A (en) * | 2021-12-14 | 2022-03-08 | 吉林大学 | Electric and ultrasonic signal composite resistance spot welding quality on-line monitoring method and device |
| CN117436769A (en) * | 2023-12-20 | 2024-01-23 | 山东方垠智能制造有限公司 | Structural part welding quality monitoring method, system, storage medium and equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109014544B (en) | 2020-07-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109014544A (en) | Miniature resistance spot welding quality on-line monitoring method | |
| CN104457841B (en) | Resistance spot welding quality on-line monitoring method | |
| Zhang et al. | Real-time seam penetration identification in arc welding based on fusion of sound, voltage and spectrum signals | |
| US9120186B2 (en) | Method and system for online quality monitoring and control of a vibration welding process | |
| Liu et al. | Defect intelligent identification in resistance spot welding ultrasonic detection based on wavelet packet and neural network | |
| Zhang et al. | A new method for nondestructive quality evaluation of the resistance spot welding based on the radar chart method and the decision tree classifier | |
| CA2337294C (en) | Method and apparatus for evaluation of eddy current testing signal | |
| CN102809611A (en) | System and method for detecting damage of metal component nondestructively | |
| CN106442525A (en) | Online detecting method for wizened defects of insides of walnuts | |
| CN106557637B (en) | Safety performance evaluation method of energy storage product | |
| KR102230886B1 (en) | Projection welding defect detection system based on image recognition | |
| CN104007182B (en) | A kind of method utilizing acoustic emission signal frequency spectrum quantitatively to detect nugget formation in resistance spot welding forming core quality | |
| CN107335908A (en) | The automation random device found for feature and its use in repetitive process | |
| CN105866690A (en) | Method and system for detecting battery state based on capacitive transducer | |
| Yu et al. | Monitoring of butt weld penetration based on infrared sensing and improved histograms of oriented gradients | |
| CN105928979B (en) | A kind of agitating friction welding hole hole measurement method and equipment | |
| CN106018405A (en) | Online detection method for laser welding penetration | |
| Boos et al. | Melt pool monitoring using fuzzy based anomaly detection in laser beam melting | |
| CN108857027B (en) | Quality monitoring device and monitoring method in micro spot welding process | |
| CN112649101A (en) | Battery module early warning method and system and fire detection device | |
| CN103675103A (en) | Method for quantitative detection of nugget crack by using energy equivalent of acoustic emission signal in resistance spot-welding process | |
| CN114663425A (en) | Ultrasonic welding quality on-line monitoring method based on generation of countermeasure network | |
| CN103994821A (en) | Method for evaluating resistance spot welding spattering quantitatively in real time through acoustic emission signal frequency spectrum | |
| CN114240896A (en) | Method for detecting welding quality of stud welding based on convolutional neural network | |
| CN109444265B (en) | Laser ultrasonic vibration detection device and method |
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 | ||
| GR01 | Patent grant |