CN103264216B - Based on controlled plasma arc welding system and the technique of back side aperture vision-based detection - Google Patents
Based on controlled plasma arc welding system and the technique of back side aperture vision-based detection Download PDFInfo
- Publication number
- CN103264216B CN103264216B CN201310179677.1A CN201310179677A CN103264216B CN 103264216 B CN103264216 B CN 103264216B CN 201310179677 A CN201310179677 A CN 201310179677A CN 103264216 B CN103264216 B CN 103264216B
- Authority
- CN
- China
- Prior art keywords
- welding
- back side
- value
- current
- image
- 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.)
- Expired - Fee Related
Links
Abstract
The invention discloses a kind of controlled plasma arc welding system based on back side aperture vision-based detection, comprise welder system, electrical quantity acquisition system, visual sensor system, four parts of regulation and control system. The invention also discloses a kind of controlled plasma arc welding technique based on back side aperture vision-based detection, step is as follows: (1) starts: choose suitable current waveform parameter according to material, the thickness of slab for the treatment of perforated material; (2) choose and set after parameter, start welding, Real-time Collection current waveform and back side aperture image, then by communication to computer, calculate orifice size, and calculate and the deviation of setting value, to welding end. The present invention by the bond pattern in " arteries and veins one hole " stainless sheet steel below 4mm welding expanded to 8mm thick plates welding; Can ensure workpiece perforation, can make again orifice size maintain near predetermined value, effectively prevent the generation that molten bath subsides and burns.
Description
Technical field
The present invention relates to a kind of controlled plasma arc welding system and technique based on back side aperture vision-based detection, belong to PLASMA ARC WELDING Quality Control Technology field.
Background technology
Compare with electron beam welding with Laser Welding, keyhole plasma welding equipment cost is low, simple to operate, seam fit-up gap strong adaptability, thereby have broad prospect of application in the processing and manufacturing of various alloy material structures. Plasma arc welding is a kind of high-energy-density welding technique, but due to the physical characteristic of plasma arc uniqueness, there are some problems in conventional plasma arc welding, as the poor stability of molten bath aperture, variation to welding condition is more responsive, and narrow, the adjustable nargin of reasonable standard parameter area that obtains good joint quality is little. For addressing these problems, have researcher to develop the pulse plasma arc welding technology in " arteries and veins one hole ", but its common square wave pulsed current waveform adopting have certain limitation, is generally only applicable to plate sheet welding.
For by plasma arc welding process application in the processing and manufacturing of middle thickness metalwork, this laboratory independent research goes out the controlled pulse plasma arc welding system of a set of low cost, high efficiency welding. This system is made further improvement to common square-wave pulse, designs the pulse welding current waveform of two slow descending grade degree of pulse back edge band, by regulating the size of these two descending slopes, can control more neatly heat input and plasma jet power. For realizing the real-time adjustment of paired pulses current waveform, place a metal in back of work and measure plate for detection of efflux voltage signal, whether its size can characterize the perforation state of workpiece and (bore a hole; The degree of perforation). But adopt such efflux voltage detection method can only obtain indirectly aperture open and-shut mode, the aperture information that can provide is very limited, cannot provide the concrete data of hole shape and size.
Vision sensor can provide more direct information. Compare with other sensor, vision sensor because of with welding circuit without the electric image that connects and can directly absorb back side aperture, have do not contact, non-inertia, information content are abundant, the interference performance of anti-highfield and high-intensity magnetic field is strong, reliability strong, sensitivity and the advantage such as certainty of measurement is high, dynamic response characteristic is good, and have avoided the adverse effect of efflux voltage sensor. If can develop suitable vision sensor, measure the direct information of hole shape and size, realize the closed-loop control of plasma arc welding process, there is undoubtedly important engineering using value.
Summary of the invention
For above-mentioned prior art, for plasma arc welding technique, the present invention researches and develops the vision sensor of filtering plasma arc wake flame, takes the image of aperture from back of work; Process and demarcate through image, obtain the data of hole shape and size; According to hole shape and the size of actual measurement, adjust online numerical value and the descending slope of pulse current, control on one's own initiative the effect of heat-power and the perforation state of plasma arc; Realize the automatic control of plasma arc welding quality, expand its applicable welding condition interval. Here it is---based on controlled plasma arc welding system and the technique of back side aperture vision-based detection.
The present invention is achieved by the following technical solutions:
A kind of controlled plasma arc welding system based on back side aperture vision-based detection, comprise welder system, electrical quantity acquisition system, visual sensor system, four parts of regulation and control system, wherein, welder system comprises plasma generator and the source of welding current, plasma generator is connected with the source of welding current, is connected with water cooling box on the source of welding current; Electrical quantity acquisition system comprises image pick-up card and data collecting card, and data collecting card is connected with workpiece with the source of welding current by Hall element, is also connected with the source of welding current by ROB interface module simultaneously; Visual sensor system comprises ccd video camera, and ccd video camera is connected with image pick-up card, is equipped with composite filter device on ccd video camera, composite filter device is made up of optical filter and neutral light damping plate, optical filter centre wavelength 638nm, bandwidth 36nm, neutral light damping plate transmitance 65%; Regulation and control system is computer, and image pick-up card is connected with computer respectively with data collecting card, messaging software is housed in computer, controls software etc.
Controlled plasma arc welding system based on back side aperture vision-based detection of the present invention, can realize in gathering welding current and back side aperture image, and on-line control pulse welding current waveform parameter, forms closed-loop control. Fig. 1 is the general structure schematic diagram of this system. Visual sensor system is taken the image of back of work aperture in real time, and machine image is processed and demarcated as calculated, provides the measurement data of hole shape and size. Regulation and control system is according to the measured value of hole shape and size, and paired pulses current waveform and numerical value are adjusted, and realizes the pattern of " aperture of a pulse ", has both ensured the complete penetration of weldment, has avoided again molten bath to subside and generation is burnt.
The difficult point that adopts image vision sensor to take aperture image is how to eliminate the interference of the plasma arc wake flame arc light appearing in back of work after perforation. Independent research of the present invention has gone out composite filter device (light damping plate and optical filter by special parameter form), the effectively interference of filtering plasma arc wake flame arc light. Composite filter is installed on to common CCD camera and is assemblied in one, obtained back side aperture image clearly.
Based on a controlled plasma arc welding technique for back side aperture vision-based detection, step is as follows:
(1) start: choose suitable current waveform parameter according to material, the thickness of slab for the treatment of perforated material, comprise background current IBAnd action time TB, peak point current IPAnd action time TP, the descending slope K of pulse back edge1、K2And transit time TK1、TK2；
(2) choose and set after parameter, start welding, ccd video camera, composite filter device and Hall element Real-time Collection current waveform and back side aperture image, then by communication to computer, calculate orifice size, and calculate and the deviation setting value: if 1. deviation is less than set-point (set-point is the desired value of back side orifice size, set in advance according to concrete mother metal and thickness of slab), judge and whether reach predetermined value weld interval, if reach predetermined value weld interval, welding finishes, if do not reach predetermined value weld interval, get back to " Real-time Collection current waveform and back side aperture image ", carry out next weld cycle, if 2. deviation is greater than set-point, adopt Fuzzy neural control device to make a policy, adjust the descending slope K of peak point current IP, pulse back edge1、K2And IK1、IK2, after adjustment, get back to " starting welding ", carry out next weld cycle; Ik1The numerical value after peak value declines at first slow depression of order section welding current of pulse back edge; Ik2Be second slow depression of order section welding current of pulse back edge fromIk1Numerical value after declining.
Preferably, as shown in Figure 2, the current waveform parameter relating to has the controlled pulse current waveform that the present invention adopts: background current IBAnd action time TB, peak point current IPAnd action time TP, the descending slope K of pulse back edge1、K2And transit time TK1、TK2. Between these pulse parameters, reasonably coupling is to ensure that controlled pulse perforation control strategy is successfully crucial.
As shown in Figure 2, in the complete pulse period changing at welding current waveform, when welding current is after peak value continues for some time, at toIn the moment, workpiece forms perforation just, in this moment by welding current with certain slope K1Slowly reduce. Due to K1The impact of the less and thermal inertia of absolute value, orifice size still can continue slow increase. When orifice size meets when setting value, that strengthens welding current underspeeds that (descending slope is K2); Along with welding current continue reduce, also corresponding reducing of heat input, aperture slowly diminishes, at moment tCClosed. Through background current action time (TB) after, be again the beginning of next cycle. Repeat said process and cycle. Adopt such welding current waveform, both can ensure molten bath perforation, workpiece penetration, realize " single-sided welding, two-sides forming ", also can ensure that the molten bath that adjacent 2 pulses form overlaps mutually, realizes good welding forming simultaneously.
When unlike material, the welding of different thickness of slab, pulse current waveform parameter is different. This research for 6-8mm thickness corrosion resistant plate, TPAnd IPSelected under the prerequisite that ensures through welding, perforation; First system chooses sufficiently high I according to thickness of slabP, to ensure that penetration and perforation choose TP; Then the condition that the penetration molten bath producing according to adjacent pulse can suitably overlap, determines TBAnd IB; And IK1、IK2Value depend on peak point current IP, be shown below:
IK1＝IB+a%×(IP-IB)
IK2＝IB+(1-a)％×(IP-IB)
A is for determining Ik1And Ik2Value intermediate parameters, its size between 0 to 1, i.e. 0 < a < 1.
Welding current descending slope K1And K2Being in welding process, adjusting in real time according to control algolithm, is dynamic change, and each pulse period is also different. But, in each pulse period, once know welding current descending slope K1And K2, according to Fig. 2, above formula, K1And K2, calculated by following formula:
TK1＝K1×[-(IP-IK1)]
TK2＝K2×[-(IK1-IK2)]
Unit explanation: time: second S, electric current: ampere A, orifice size: area (square millimeter) mm2。
The present invention adopts controlled pulse perforation control strategy, designs special pulse current waveform; Hole shape based on vision-based detection and size are carried out to real-time regulating impulse welding current waveform parameter as feedback signal, form closed-loop control, ensure the Proper Match of controlled pulse current waveform parameter, thereby controlling aperture periodically forms and closure, both ensure the complete penetration of weldment, and avoided again molten bath to subside and generation is burnt. Improve stability and the robustness of plasma arc welding process, ensure welding quality.
The input quantity (controlled quentity controlled variable) of described regulation and control system is the area of aperture; Output quantity (controlled variable) is along descending slope after pulse current peak and action time thereof and two.
Independent research of the present invention goes out fuzzy Neuro Controller, the controller parameter being optimized by a large amount of soldering tests. Adjust accordingly according to the deviate between little hole area and its setting value of reality test, ensure that orifice size is in expected range, thereby ensure penetration and quality of weld seam molding. Fig. 3 is control system theory diagram, and Fig. 4 is system controlling software flow chart.
Fuzzy neural control device of the present invention:
Select pulse welding current peak IPAs controlled quentity controlled variable, back side orifice size (area) SKAs controlled variable. Controller adopts dual input, single export structure, as shown in Figure 5. The input variable of controller is error and the error rate of the little hole area in the back side; The output variable of controller is the variation of pulse welding current peak.
The exact value of error represents with e, and the exact value of error rate represents with ce. ?
e＝SK(t)-SG
ce＝SK(t)-SK(t-1)
S in formulaK(t) back side aperture area value of measuring for the t moment;
SK(t-1) back side aperture area value of measuring for the t-1 moment;
SGFor the set-point of the little hole area in the back side.
The fuzzy set of error represents with E, and the fuzzy set of error rate represents with CE. In this controller, all adopt 7 linguistic variable values to describe error and the error rate of the little hole area in the back side, negative large (NB), negative in (NM), negative little (NS), zero (O), just little (PS), center (PM), honest (PB). If the domain of error linguistic variable, error rate linguistic variable is [6 ,+6]. Its membership function is all taken as the form of isosceles triangle, sees Fig. 6.
Prerequisite variable represents with fuzzy variable form, and conclusion variable represents with constant. All control laws are as shown in table 1.
Fuzzy rule represents with following form:
Fuzzy rule: IFEisNBandCEisNBthenuisu1
U in fuzzy control rule1、u2、…u49Be exact value, this exact value can obtain by the off-line of monolayer neural networks and online self study.
The control law of table 1 neural network fuzzy control device
If the exact value of the back side aperture area error now obtaining through over-sampling and error rate is converted to e through domaini、cei, the pulse welding current change quantity u now obtaining is tried to achieve by following method:
The activity of k rule:
E in formulak1,CEk2Be respectively the corresponding error fuzzy set of this rule and error rate fuzzy set.
The average activity of k rule is:
In formulaIt is the activity sum of all 49 rules.
Can push away to obtain conclusion:
The present invention develops applicable aperture image processing algorithm and scaling method:
Vision detection system can photograph typical aperture image as shown in Figure 7, is from left to right aperture antetheca, aperture duct, aperture rear wall, wake flame successively, the clutter of having powerful connections around aperture. An optional scan line, draws the grey value profile along this line, as shown in Figure 8, can see, background clutter frequency is higher, but peak value is lower than 50; Orifice region left hand edge and right hand edge have respectively gray scale to skyrocket and suddenly fall, and aperture inside has gray scale Sudden change region but amplitude of variation does not have fringe region so violent, therefore can distinguish marginal point by the variation severe degree (step) that compares gray value.
Concrete processing method is first background clutter to be removed, and strengthen contrast; Reflect the variation of gray scale by difference, the mathematic(al) representation of one dimension difference in x direction:
▽f(xi,y)＝f(xi+1,y)-f(xi,y)
In formula, f (xi,y)，f(xi+1, y) represent respectively point (xi,y)，(xi+1, the gray value of y) locating, ▽ f (xi, y) represent (xi, the gray scale difference score value of y) locating.
Grey scale curve in Fig. 8 is done to difference and obtains grey scale change curve, as shown in Figure 9, can see left hand edge put corresponding grey scale change peak value be maximum on the occasion of, it is minimum negative value that right hand edge is put corresponding grey scale change peak value. In order to eliminate presumable pixel catastrophe point, can do a smothing filtering to shade of gray curve catastrophe point is removed, filtering algorithm as shown in the formula:
▽f(xi,y)=▽f(xi2,y)/3+▽f(xi1,y)/3+▽f(xi,y)/3
In formula, the implication of each parameter is the same, represents the gray scale difference score value at corresponding coordinate points place.
Smooth effect as shown in figure 10. According to this curve, the aperture image boundary marginal point that can extract. On collection image, line by line scan, obtain left and right two marginal points; The coordinate pair of a series of marginal points, describes out the border of aperture image.
Scaling method: as shown in Figure 11 (a), place the object (concentric rectangles, spacing 1.0mm between adjacent segments) of a known form and size before video camera, be called demarcation thing. Obtained the image of this object by video camera, as shown in Figure 11 (b). Then, according to the Pixel Dimensions on concentric rectangles actual size and image, determine on image along X the actual size of a pixel representative separately in Y-direction.
Accordingly, can calculate the area of back side aperture.
The present invention, according to the aperture image sequence of taking, has extracted not orifice size parameter and location parameter in the same time exactly.
The present invention, by researching and developing and design " based on controlled plasma arc welding technique and the system of back side aperture vision-based detection ", carries out welding procedure test, and analytical test result has obtained following some conclusion:
(1) controlled plasma arc welding technique and the system based on back side aperture vision-based detection, the welding of stainless sheet steel by the bond pattern in " arteries and veins one hole " below 4mm expanded to 8mm thick plates welding; Can ensure workpiece perforation, can make again orifice size maintain near predetermined value, effectively prevent the generation that molten bath subsides and burns.
(2) native system can be controlled heat input and plasma jet power more neatly, sensitivity and certainty of measurement are high, dynamic response characteristic good, can more effectively make response to perforation state, " open-close-opening " state conversion of aperture is more steady, welding process stability improves, appearance of weld is good, and wide uniformity is melted in positive and negative.
(3) when the variation of thickness of workpiece generation 50%, while causing radiating condition generation great variety, system adjustment process with effective, has ensured steadily carrying out of welding process rapidly, and quality of weld seam molding is good.
Brief description of the drawings
Fig. 1: the structural representation of the controlled plasma arc welding system based on back side aperture vision-based detection of the present invention.
Fig. 2: controlled pulse current waveform schematic diagram of the present invention.
Fig. 3: control system theory diagram.
Fig. 4: system controlling software flow chart.
Fig. 5: neural network fuzzy control device block diagram.
Fig. 6: the fuzzy set definition of input variable.
Fig. 7: the original aperture image of shooting.
Fig. 8: typical gray value curve (Y=237).
Fig. 9: grey scale difference curve (Y=237).
Figure 10: by the curve after grey scale difference curve mean filter and the marginal point (Y=237) that searches.
Figure 11: the calibration of vision sensor, wherein, (a): calibrated reference; (b): the demarcation of camera.
Figure 12: the welding effect of 8mm thickness corrosion resistant plate, wherein, (a): the appearance of weld at the back side; (b): the welding effect of 8mm thickness corrosion resistant plate.
Figure 13: the structure diagram of the controlled plasma arc welding system based on back side aperture vision-based detection of the present invention, wherein, 1, plasma generator; 2, the source of welding current; 3, water cooling box; 4, image pick-up card; 5, data collecting card; 6, Hall element; 7, ROB interface module; 8, ccd video camera; 9, computer.
Detailed description of the invention
Below in conjunction with drawings and Examples, the present invention is further illustrated.
A kind of controlled plasma arc welding system based on back side aperture vision-based detection, comprise welder system, electrical quantity acquisition system, visual sensor system, four parts of regulation and control system, as shown in Fig. 1, Figure 13, wherein, welder system comprises plasma generator 1 and the source of welding current 2, plasma generator 1 source of welding current 2 connects, and is connected with water cooling box 3 on the source of welding current 2; Electrical quantity acquisition system comprises image pick-up card 4 and data collecting card 5, and data collecting card 5 is connected with workpiece with the source of welding current 2 by Hall element 6, is also connected with the source of welding current 2 by ROB interface module 7 simultaneously; Visual sensor system comprises ccd video camera 8, and ccd video camera 8 is connected with image pick-up card 4, is equipped with composite filter device on ccd video camera 8, composite filter device is made up of optical filter and neutral light damping plate, optical filter centre wavelength 638nm, bandwidth 36nm, neutral light damping plate transmitance 65%; Regulation and control system is computer 9, and image pick-up card 4 is connected with computer 9 respectively with data collecting card 5, messaging software is housed in computer 9, controls software etc.
The above-mentioned controlled plasma arc welding system based on back side aperture vision-based detection, can realize in gathering welding current and back side aperture image, and on-line control pulse welding current waveform parameter, forms closed-loop control. Fig. 1 is the general structure schematic diagram of this system. Visual sensor system is taken the image of back of work aperture in real time, and machine image is processed and demarcated as calculated, provides the measurement data of hole shape and size. Regulation and control system is according to the measured value of hole shape and size, and paired pulses current waveform and numerical value are adjusted, and realizes the pattern of " aperture of a pulse ", has both ensured the complete penetration of weldment, has avoided again molten bath to subside and generation is burnt.
A kind of controlled plasma arc welding technique based on back side aperture vision-based detection of embodiment
Step is as follows:
(1) start: choose suitable current waveform parameter according to material, the thickness of slab for the treatment of perforated material, comprise background current IBAnd action time TB, peak point current IPAnd action time TP, the descending slope K of pulse back edge1、K2And transit time TK1、TK2；
(2) choose and set after parameter, start welding, ccd video camera, composite filter device and Hall element Real-time Collection current waveform and back side aperture image, then by communication to computer, calculate orifice size, and calculate and the deviation setting value: if 1. deviation is less than set-point (set-point is the desired value of back side orifice size, set in advance according to concrete mother metal and thickness of slab), judge and whether reach predetermined value weld interval, if reach predetermined value weld interval, welding finishes, if do not reach predetermined value weld interval, get back to " Real-time Collection current waveform and back side aperture image ", carry out next weld cycle, if 2. deviation is greater than set-point, adopt Fuzzy neural control device to make a policy, adjust peak point current IP, pulse back edge descending slope K1、K2And IK1、IK2, after adjustment, get back to " starting welding ", carry out next weld cycle.
As shown in Figure 2, the current waveform parameter relating to has the controlled pulse current waveform that the present invention adopts: background current IBAnd action time TB, peak point current IPAnd action time TP, the descending slope K of pulse back edge1、K2And transit time TK1、TK2. Between these pulse parameters, reasonably coupling is to ensure that controlled pulse perforation control strategy is successfully crucial.
As shown in Figure 2, in the complete pulse period changing at welding current waveform, when welding current is after peak value continues for some time, at toIn the moment, workpiece forms perforation just, in this moment by welding current with certain slope K1Slowly reduce. Due to K1The impact of the less and thermal inertia of absolute value, orifice size still can continue slow increase. When orifice size meets when setting value, that strengthens welding current underspeeds that (descending slope is K2); Along with welding current continue reduce, also corresponding reducing of heat input, aperture slowly diminishes, at moment tCClosed. Through background current action time (TB) after, be again the beginning of next cycle. Repeat said process and cycle. Adopt such welding current waveform, both can ensure molten bath perforation, workpiece penetration, realize " single-sided welding, two-sides forming ", also can ensure that the molten bath that adjacent 2 pulses form overlaps mutually, realizes good welding forming simultaneously.
When unlike material, the welding of different thickness of slab, pulse current waveform parameter is different. This research for 6-8mm thickness corrosion resistant plate, TPAnd IPSelected under the prerequisite that ensures through welding, perforation; First system chooses sufficiently high I according to thickness of slabP, to ensure that penetration and perforation choose TP; Then the condition that the penetration molten bath producing according to adjacent pulse can suitably overlap, determines TBAnd IB; And IK1、IK2Value depend on peak point current IP, be shown below:
IK1＝IB+a%×(IP-IB)
IK2＝IB+(1-a)%×(IP-IB)
Welding current descending slope K1And K2Being in welding process, adjusting in real time according to control algolithm, is dynamic change, and each pulse period is also different. But, in each pulse period, once know welding current descending slope K1And K2, according to Fig. 2, above formula, K1And K2, calculated by following formula:
TK1＝K1×[-(IP-IK1)]
TK2＝K2×[-(IK1-IK2)]
Unit explanation: time: second S, electric current: ampere A, orifice size: area (square millimeter) mm2。
The input quantity (controlled quentity controlled variable) of described regulation and control system is the area of aperture; Output quantity (controlled variable) is along descending slope after pulse current peak and action time thereof and two.
Fuzzy neural control device of the present invention: Fig. 3 is control system theory diagram, and Fig. 4 is system controlling software flow chart.
Select pulse welding current peak IPAs controlled quentity controlled variable, back side orifice size (area) SKAs controlled variable. Controller adopts dual input, single export structure, as shown in Figure 5. The input variable of controller is error and the error rate of the little hole area in the back side; The output variable of controller is the variation of pulse welding current peak.
The exact value of error represents with e, and the exact value of error rate represents with ce. ?
e＝SK(t)-SG
ce＝SK(t)-SK(t-1)
S in formulaK(t) back side aperture area value of measuring for the t moment;
SK(t-1) back side aperture area value of measuring for the t-1 moment;
SGFor the set-point of the little hole area in the back side.
The fuzzy set of error represents with E, and the fuzzy set of error rate represents with CE. In this controller, all adopt 7 linguistic variable values to describe error and the error rate of the little hole area in the back side, negative large (NB), negative in (NM), negative little (NS), zero (O), just little (PS), center (PM), honest (PB). If the domain of error linguistic variable, error rate linguistic variable is [6 ,+6]. Its membership function is all taken as the form of isosceles triangle, sees Fig. 6.
Prerequisite variable represents with fuzzy variable form, and conclusion variable represents with constant. All control laws are as shown in table 1.
Fuzzy rule represents with following form:
Fuzzy rule: IFEisNBandCEisNBthenuisu1
U in fuzzy control rule1、u2、…u49Be exact value, this exact value can obtain by the off-line of monolayer neural networks and online self study.
If the exact value of the back side aperture area error now obtaining through over-sampling and error rate is converted to e through domaini、cei, the pulse welding current change quantity u now obtaining is tried to achieve by following method:
The activity of k rule:
E in formulak1,CEk2Be respectively the corresponding error fuzzy set of this rule and error rate fuzzy set.
The average activity of k rule is:
In formulaIt is the activity sum of all 49 rules.
Can push away to obtain conclusion:
Applicable aperture image processing algorithm and the scaling method of described calculating orifice size is as follows:
Vision detection system can photograph typical aperture image as shown in Figure 7, is from left to right aperture antetheca, aperture duct, aperture rear wall, wake flame successively, the clutter of having powerful connections around aperture.
Concrete processing method is first background clutter to be removed, and strengthen contrast; Reflect the variation of gray scale by difference, the mathematic(al) representation of one dimension difference in x direction:
▽f(xi,y)＝f(xi+1,y)-f(xi,y)
In formula, f (xi,y)，f(xi+1, y) represent respectively point (xi,y)，(xi+1, the gray value of y) locating, ▽ f (xi, y) represent (xi, the gray scale difference score value of y) locating.
Grey scale curve in Fig. 8 is done to difference and obtains grey scale change curve, as shown in Figure 9, can see left hand edge put corresponding grey scale change peak value be maximum on the occasion of, it is minimum negative value that right hand edge is put corresponding grey scale change peak value. In order to eliminate presumable pixel catastrophe point, can do a smothing filtering to shade of gray curve catastrophe point is removed, filtering algorithm as shown in the formula:
▽f(xi,y)=▽f(xi2,y)/3+▽f(xi1,y)/3+▽f(xi,y)/3
In formula, the implication of each parameter is the same, represents the gray scale difference score value at corresponding coordinate points place.
Smooth effect as shown in figure 10. According to this curve, the aperture image boundary marginal point that can extract. On collection image, line by line scan, obtain left and right two marginal points; The coordinate pair of a series of marginal points, describes out the border of aperture image.
Scaling method: as shown in Figure 11 (a), place the object (concentric rectangles, spacing 1.0mm between adjacent segments) of a known form and size before video camera, be called demarcation thing. Obtained the image of this object by video camera, as shown in Figure 11 (b). Then, according to the Pixel Dimensions on concentric rectangles actual size and image, determine on image along X the actual size of a pixel representative separately in Y-direction.
Accordingly, can calculate the area of back side aperture.
Experiment adopts 304 corrosion resistant plates of 8mm thickness to carry out the control experiment of plasma arc welding technique
Plasma (welding) torch nozzle diameter and orifice throat length be 2.8mm and 3mm respectively, tungsten electrode neck-in 2mm, nozzle to workpiece apart from 5mm. Ion-gas and protective gas are straight argon, and flow is respectively 3.2L/min and 20L/min. Speed of welding is 120mm/min. Ip=200A,IB=130A，TPAnd TBIn experimentation, determined by program. The back side aperture area set-point of implementing controlled pulse control strategy is 1.5mm2. Adopt the present invention, the appearance of weld of plasma arc welding is good, perforation state evenly, penetration, the back side are molten wide consistent completely, typical weld seam is as shown in figure 12. Illustrate that technical scheme of the present invention can realize, can be applicable to the middle thick plates of 8mm.
Claims (4)
1. the controlled plasma arc welding system based on back side aperture vision-based detection, is characterized in that:Comprise welder system, electrical quantity acquisition system, visual sensor system, four parts of regulation and control system,Wherein, welder system comprises plasma generator and the source of welding current, and plasma generator is connected with the source of welding current,On the source of welding current, be connected with water cooling box; Electrical quantity acquisition system comprises image pick-up card and data collecting card, dataCapture card is connected with the source of welding current by Hall element, also connects by ROB interface module and the source of welding current simultaneouslyConnect; Visual sensor system comprises ccd video camera, and ccd video camera is connected with image pick-up card, CCDOn video camera, be equipped with composite filter device, composite filter device is made up of optical filter and neutral light damping plate, filtersSheet centre wavelength 638nm, bandwidth 36nm, neutral light damping plate transmitance 65%; Regulation and control system isComputer, image pick-up card is connected with computer respectively with data collecting card.
2. the controlled plasma arc welding technique based on back side aperture vision-based detection, is characterized in that:Step is as follows:
(1) start: choose suitable current waveform parameter according to material, the thickness of slab for the treatment of perforated material, compriseBackground current IBAnd action time TB, peak point current IPAnd action time TP, the lower drop angle of pulse back edgeRate K1、K2And transit time TK1、TK2；
(2) choose and set after parameter, start welding, ccd video camera, composite filter device and HallElement Real-time Collection current waveform and back side aperture image, then by communication to computer, calculate aperture chiVery little, and the deviation of calculating and setting value: if 1. deviation is less than set-point, judges and whether reach pre-weld intervalDefinite value, if reach predetermined value weld interval, welding finishes, if do not reach predetermined value weld interval, gets back to" Real-time Collection current waveform and back side aperture image ", carries out next weld cycle; If 2. deviation be greater than toDefinite value, adopts Fuzzy neural control device to make a policy, and adjusts peak point current IP, pulse back edge declineSlope K1、K2And IK1、IK2, after adjustment, get back to " starting welding ", carry out next weld cycle;
Ik1The numerical value after peak value declines at first slow depression of order section welding current of pulse back edge; Ik2BeSecond slow depression of order section welding current of pulse back edge is from Ik1Numerical value after declining.
3. the controlled plasma arc welding based on back side aperture vision-based detection according to claim 2Technique, is characterized in that: described current waveform parameter meets the following conditions:
First choose sufficiently high I according to thickness of slabP, to ensure that penetration and perforation choose TP; Then foundationThe condition that the penetration molten bath that adjacent pulse produces can suitably overlap, determines TBAnd IB; And IK1、IK2'sValue depends on peak point current IP, be shown below:
IK1＝IB+a×(IP-IB)
IK2＝IB+(1-a)×(IP-IB)
In each pulse period, once know welding current descending slope K1And K2, by following formula meterCalculate:
TK1＝K1×[-(IP-IK1)]
TK2＝K2×[-(IK1-IK2)]
Unit explanation: time: second S, electric current: ampere A;
Wherein, a is for determining Ik1And Ik2Value intermediate parameters, its size between 0 to 1, i.e. 0 < a < 1.
4. the controlled plasma arc welding work based on back side aperture vision-based detection according to claim 2Skill, is characterized in that: applicable aperture image processing algorithm and the scaling method of described calculating orifice size is as follows:
Vision detection system photographs aperture image, the clutter of having powerful connections around aperture;
First background clutter is removed, and strengthen contrast; Reflect the variation of gray scale by difference, in x directionThe mathematic(al) representation of upper one dimension difference:
In formula, f (xi,y)，f(xi+1, y) represent respectively point (xi,y)，(xi+1, the gray value of y) locating,Represent(xi, the difference value of y) locating;
Grey scale curve is done to difference and obtain grey scale change curve, grey scale change curve is done to a smothing filtering and will dash forwardHeight is removed, filtering algorithm as shown in the formula:
In formula, the implication of each parameter is the same, represents the gray scale difference score value at corresponding coordinate points place;
On collection image, line by line scan, obtain left and right two marginal points; The coordinate pair of a series of marginal points,The border of aperture image is described out;
Scaling method: place the object of a known form and size before video camera, be called demarcation thing, by taking the photographCamera obtains the image of this object, then, and according to the Pixel Dimensions on concentric rectangles actual size and image, reallyMake on image along X, the actual size of a pixel representative separately in Y-direction, accordingly, calculates the back sideThe area of aperture.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310179677.1A CN103264216B (en) | 2013-05-15 | 2013-05-15 | Based on controlled plasma arc welding system and the technique of back side aperture vision-based detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310179677.1A CN103264216B (en) | 2013-05-15 | 2013-05-15 | Based on controlled plasma arc welding system and the technique of back side aperture vision-based detection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103264216A CN103264216A (en) | 2013-08-28 |
| CN103264216B true CN103264216B (en) | 2016-05-04 |
Family
ID=49007902
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310179677.1A Expired - Fee Related CN103264216B (en) | 2013-05-15 | 2013-05-15 | Based on controlled plasma arc welding system and the technique of back side aperture vision-based detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103264216B (en) |
Families Citing this family (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103831516B (en) * | 2013-12-16 | 2016-05-25 | 上海工程技术大学 | The sequential welding method of the discontinuous solder joint of stainless steel cloth based on visual sensing technology |
| CN104227191B (en) | 2014-09-10 | 2016-04-20 | 北京工业大学 | A kind of low back pressure penetrating arc welder and method |
| CN104858557B (en) * | 2015-06-01 | 2016-08-17 | 山东大学 | A kind of combined welding method of austenitic stainless steel cut deal |
| CN106530269A (en) * | 2015-09-15 | 2017-03-22 | 苏州中启维盛机器人科技有限公司 | Weld detection method |
| CN105234539B (en) * | 2015-11-18 | 2017-06-30 | 山东大学 | The plasma arc welding (PAW) connection device and technique of a kind of ultrasonic wave added perforation |
| CN105478976B (en) * | 2016-01-26 | 2017-10-24 | 清华大学 | Termination MICROBEAM PLASMA WELDING shaping control method based on Dynamic System Identification |
| CN106624266B (en) * | 2016-12-31 | 2018-08-07 | 东莞职业技术学院 | A kind of weld seam deviation and penetration signal monitoring method for Automobile Welding |
| CN107552932A (en) * | 2017-10-20 | 2018-01-09 | 北京工业大学 | It is a kind of based on the variable-polarity plasma welding method of quality control controlled keyhole profile and device |
| CN108544068B (en) * | 2018-04-19 | 2021-02-26 | 北京工业大学 | Welding quality control system and method for arc temperature difference before and after plasma arc perforation |
| CN108873700B (en) * | 2018-07-13 | 2020-06-30 | 中国科学院力学研究所 | Laser cladding closed-loop control method based on stereoscopic vision |
| CN109175747B (en) * | 2018-11-12 | 2019-09-10 | 湖南大学 | A kind of comprehensive method directly observed of metal material penetration fustion welding keyhole profile |
| CN110238487B (en) * | 2019-05-22 | 2020-09-22 | 北京航空航天大学 | Medium plate aluminum alloy deep fusion welding system based on visual sensing and control method |
| CN113199184B (en) * | 2021-07-05 | 2021-10-08 | 北京航空航天大学 | Weld joint shape prediction method based on improved self-adaptive fuzzy neural network |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1262158A (en) * | 2000-03-17 | 2000-08-09 | 清华大学 | Arc voltage method and system for sensing perforation state of plasma welding's molten bath |
| CN1264822A (en) * | 2000-03-17 | 2000-08-30 | 清华大学 | Arc method for measuring orifice size of molten bath for plasma welding and signal acquisition system |
| CN103071902A (en) * | 2013-01-11 | 2013-05-01 | 北京工业大学 | Method and device for controlling welding process based on stability of perforated molten bath |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5869259B2 (en) * | 2011-08-24 | 2016-02-24 | 株式会社ディスコ | Drilling method and laser processing apparatus |
-
2013
- 2013-05-15 CN CN201310179677.1A patent/CN103264216B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1262158A (en) * | 2000-03-17 | 2000-08-09 | 清华大学 | Arc voltage method and system for sensing perforation state of plasma welding's molten bath |
| CN1264822A (en) * | 2000-03-17 | 2000-08-30 | 清华大学 | Arc method for measuring orifice size of molten bath for plasma welding and signal acquisition system |
| CN103071902A (en) * | 2013-01-11 | 2013-05-01 | 北京工业大学 | Method and device for controlling welding process based on stability of perforated molten bath |
Non-Patent Citations (2)
| Title |
|---|
| PAW焊接熔池—小孔流场与热场动态行为的数值分析;张涛;《中国博士学位论文全文数据库 工程科技Ⅰ辑》;20120615(第6期);B022-49 * |
| PAW熔池正面形状参数的视觉检测;张云威;《中国优秀硕士学位论文全文数据库 信息科技辑》;20130215(第2期);I138-1472 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103264216A (en) | 2013-08-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103264216B (en) | Based on controlled plasma arc welding system and the technique of back side aperture vision-based detection | |
| CN106001926B (en) | The laser of view-based access control model sensing-real-time automaton of electric arc combined welding and its welding method | |
| Chang et al. | Penetration quality prediction of asymmetrical fillet root welding based on optimized BP neural network | |
| Chen et al. | Intelligent methodology for sensing, modeling and control of pulsed GTAW: Part 1--Bead-on-plate welding | |
| Zhang et al. | Single vision system for simultaneous observation of keyhole and weld pool in plasma arc welding | |
| Van Anh et al. | Experimental investigation on the weld pool formation process in plasma keyhole arc welding | |
| Liu et al. | Stable keyhole welding process with K-TIG | |
| Cui et al. | Keyhole process in K-TIG welding on 4 mm thick 304 stainless steel | |
| CN105033408A (en) | Double passive vision sensing detecting device in GMA material increase manufacturing and detecting method thereof | |
| Wu et al. | Understanding of the weld pool convection in twin-wire GMAW process | |
| CN107552932A (en) | It is a kind of based on the variable-polarity plasma welding method of quality control controlled keyhole profile and device | |
| Huang et al. | Correlation between gas-dynamic behaviour of a vapour plume and oscillation of keyhole size during laser welding of 5083 Al-alloy | |
| Yu et al. | Application of arc plasma spectral information in the monitor of Al–Mg alloy pulsed GTAW penetration status based on fuzzy logic system | |
| Gao et al. | Multi-sensor information fusion for monitoring disk laser welding | |
| CN104772554B (en) | The pulsed tungsten argon arc weld(ing) arc voltage sensing penetration control method of stepping electric arc welding type | |
| CN109290664B (en) | K-TIG welding penetration monitoring system and method based on sound sensing and current control | |
| CN108672893A (en) | A kind of control method of asymmetry angle joint penetration form and fusion penetration | |
| CN108544068B (en) | Welding quality control system and method for arc temperature difference before and after plasma arc perforation | |
| Liu et al. | Control of keyhole exit position in plasma arc welding process | |
| CN108213651B (en) | Penetration control method for initial stage of pulse arc welding | |
| CN110202264A (en) | A kind of accuracy controlling method of titanium alloy covering-stringer siding double laser beam two-side synchronous welding microstructure | |
| DE19927803A1 (en) | Device for checking the focus position during laser beam welding | |
| Xiong et al. | Arc voltage measurements for height control in pulsed arc additive manufacturing | |
| Sun et al. | Applications of molten pool visual sensing and machine learning in welding quality monitoring | |
| Vilarinho et al. | An alternative algorithm for synergic pulsed GMAW of aluminium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20160504 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |