• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    The present invention discloses a robot welding seam tracking method for all—position welding of a large—curvature pipe fitting, which relates to the field of welding processing. The present invention is to solve the problem that the existing welding seam tracking technology cannot realize all—position welding and welding seam tracking when a pipe cannot be selected accurately or a positioner cannot be applied. The method of the present invention includes: obtaining a transformation matrix from a laser sensor coordinate system to a robot tool coordinate system; obtaining a teaching transformation matrix from an arc coordinate system to a robot base coordinate system; obtaining a welding seam point deviation in the arc coordinate system according to the obtained matrix; respectively smoothing a deviation lul in a normal direction ijf and a deviation ARV in a vertical direction af by using an SG smoothing algorithm; inputting the smoothed deviation Au of a radius and the deviation Ahv in the direction if into a PID controller to obtain a smooth and stable welding seam tracking deviation; and obtaining a welding seam point after a welding seam is corrected in the robot base coordinate system according to the obtained welding seam tracking deviation. The present invention is used for welding seam tracking of a robot.
    公开号:NL2029506A
    申请号:NL2029506
    申请日:2021-10-25
    公开日:2021-12-21
    发明作者:Liu Zhiheng；Li Ruifeng；Wang Ke；Zhao Lijun；Ge Lianzheng
    申请人:Harbin Inst Technology；
    IPC主号:
    专利说明:

    ROBOT WELDING SEAM TRACKING METHOD FOR ALL-POSITION WELDING OF LARGE-CURVATURE PIPE FITTING
    TECHNICAL FIELD The present invention belongs to the field of welding pro- cessing, and in particular, to a robot welding seam tracking meth- od for all-position welding of a large-curvature pipe fitting.
    BACKGROUND ART With the rapid development of artificial intelligence, a welding robot has been introduced into industrial operations. A welding robot is an industrial robot engaged in welding. The in- dustrial robot is a multi-purpose and re-programnable automatic control manipulator. Therefore, welding process parameters and welding paths of the welding robot must be manually set in advance when the welding robot operates, which has a requirement on the consistency of welding conditions. Most of the existing welding seam tracking methods use wall-climbing robots, which can be suit- able for welding seam tracking of pipe fittings, basically for large-diameter pipe fittings, but are not suitable for convention- al industrial robots, and are not suitable for welding small- radius pipe fittings. Therefore, a method for welding small-radius pipe fittings becomes a research focus in the art.
    At present, the welding and welding seam tracking of the small-radius pipe fittings are performed by a method combining a positioner. The robot is fixed at a determined position, welded pipelines rotate at a set rotating speed by an axially applied ro- tation force, and then a welding gun welds continuously at the same position, so as to achieve a complete welding seam forming technology. However, when pipes cannot rotate or the positioner cannot be applied, the welding method combining the positioner cannot be applied, so all-position welding and welding seam track- ing cannot be realized.
    SUMMARY An objective of the present invention is to provide a robot welding seam tracking method for all-position welding of a large- curvature pipe fitting to solve the problem that the existing welding seam tracking technology cannot realize all-position weld- ing and welding seam tracking when a pipe cannot be selected accu- rately or a positioner cannot be applied.
    A robot welding seam tracking method for all-position welding of a large-curvature pipe fitting includes a specific process as follows: step 1, obtaining a transformation matrix from a laser sensor coordinate system to a robot tool coordinate system; step 2, obtaining a teaching transformation matrix from an arc coordinate system to a robot base coordinate system, which in- cludes a specific process is as follows: step 21, obtaining an initial path of robot full-position la- ser welding through online teaching; step 22, inputting the initial path of laser welding into a controller to obtain a spatial arc coordinate system H; , aT step 23, obtaining a transformation matrix H# from the arc spatial coordinate system H to a robot base coordinate system B; step 3, obtaining a welding seam point deviation in an arc coordinate system according to the matrices obtained in step 1 and step 2, which includes a specific process as follows: step 31, transforming a welding seam point detected in a la- ser sensor coordinate system into the arc coordinate system; step 32, obtaining a deviation Au in a normal direction U and a deviation 4W in a vertical direction W of an arc between a welding seam point detected by a laser sensor and a welding seam point on a teaching trajectory; step 4, respectively smoothing the deviation Au in the nor- mal direction Ù and the deviation 4W in the vertical direction Ww’ by using a Savizkg-Golag (SG) smoothing algorithm; step 5, inputting the smoothed deviation A of a radius and the deviation 4W in direction W into a Proportion Integration Differentiation (PID) controller to obtain a smooth and stable welding seam tracking deviation; and step 6, obtaining a welding seam point after a welding seam is corrected in the robot base coordinate system according to the welding seam tracking deviation obtained in step 5. The present invention has the following beneficial effects: According to the present invention, the robot is taught to transform and smooth position data information collected by the laser sensor and input the transformed and smoothed position data information into the PID controller along a movement path from a lowest point to a highest point of a pipe fitting, and then the position data information is transformed into the robot base coor- dinate system through coordinate transformation to control the end of the robot to correct the welding seam deviation online. There- fore, the all-position welding and welding seam tracking can also be realized even if the pipe cannot be selected accurately or the positioner cannot be applied.
    BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a robot tool coordinate system and a laser sensor coordinate system; FIG. 2(a) is a position of a teaching starting point PO of a pipe fitting welding path; FIG. 2(b) is a position of a teaching end point P2 of the pipe fitting welding path; FIG. 3 is a spatial arc coordinate system; FIG. 4 is a corrected welding seam trajectory deviation Au in a normal direction W; and FIG. 5 is a corrected welding seam trajectory deviation Aw in a vertical direction W'.
    DETAILED DESCRIPTION OF THE EMBODIMENTS Specific implementation mode 1: a robot welding seam tracking method for all-position welding of a large-curvature pipe fitting of specific implementation mode 1 includes a specific process as fellows: step 1, a transformation matrix from a laser sensor coordi- nate system to a robot tool coordinate system is obtained, which includes the specific process as follows: step 11, coordinates Pr (pr op 1) of the welding seam point in the laser sensor coordinate system and coordinates Pr= (pf pe pi 1) of the welding seam point tested in the robot tool coordinate system are obtained.
    Step 12, the transformation matrix from the laser sensor co- ordinate system to the robot tool coordinate system is set as: 1, S, oa.
    Pp, ir "= p_ _ : 5, 4, p | p ns, a Pp, 0 0 0 1 (1) re : : Where, 5 1s the set transformation matrix from the laser sensor coordinate system to the robot tool coordinate system, S is the laser sensor coordinate system, x is the x-directional posi- tion information of a welding seam in the laser sensor coordinate system S, and z is the z-directional position information of a welding seam in the laser sensor coordinate system S; the S is obtained by directly detecting and extracting weld- ing seam information through a laser sensor; and the x-directional position information and the z-directional position information in the laser sensor coordinate system S are directly obtained through the laser sensor.
    Step 13, the coordinates of the spatial welding seam point in the laser sensor coordinate system and the coordinates of the welding seam point in the robot tool coordinate system are input into the set transformation matrix from the laser sensor coordi- nate system to the robot tool coordinate system: nop +a pr +p =p n,- Jou +a, pi+ p= pr nop ta PAP.
    Formula (2) is further transformed to obtain: sx sz ix ) nl P Pr a, Pi sx sz 1 a, 7 Ix pi P l DP. bi props pr . - - A, . sr sz 1 a, = ly ’ ’ ’ ’ (3) sx 5 t Pi Pi lk Pr . . . 1, . sx sz 1 a; = Iz bobo | a=Cn_n no)a=Ca a a), p= ppp) .
    5 Step 14, parameters VVT 0p dP =P Py P: in the transformation matrix are obtained by solving formula (3) by a least square method.
    S= SS) . Step 15, “ # is obtained according to the step 14: S=axn (4) Step 16, the obtained parameters n=(n nn), =(a 9,0), D =C ‚p) y=(s_,5,5) ee dpd hp Do Py Ps and PT are input into Ly the set transformation matrix I from the laser sensor coordinate to the robot base coordinate system to obtain the transformation oh matrix 5 from the laser sensor coordinate system to the robot tool coordinate system.
    Step 2, a teaching transformation matrix from an arc coordi- nate system to a robot base coordinate system is obtained, which includes a specific process as follows: step 21, an initial path of robot full-position laser welding is obtained through online teaching.
    The initial path of the laser welding is from a lowest posi- tion point to a highest position point of a pipe fitting, covers half of the whole butt pipe fitting welding seam, and includes processes of overhead position welding, vertical position welding,
    and flat position welding of all-position welding (as shown in FIG. 3). Step 22, the initial path of the laser welding is input into a controller to obtain an arc spatial coordinate system H. oT Step 23, a transformation matrix # from the arc spatial co- ordinate system H to a robot base coordinate system B is obtained: u, Ve Ww. Pp x U v, W, Pp, Bp | 7 ¥ y + wl = U Vv. Ww. Pp, 0 0 OO 1 (5) Step 3, a welding seam deviation in an arc coordinate system is obtained according to the matrices obtained in step 1 and step 2, which includes a specific process as follows: step 31, a welding seam point detected by the laser sensor is transformed into the arc coordinate system: H Bp yt Bp Lp ps PO) TT (6)
    BT Where, L is the transformation matrix from the robot tool coordinate system to the robot base coordinate system, and is ob-
    LT tained from an end pose of a robot; S is a transformation matrix from the laser sensor coordinate system to the robot tool coordi- Ho nate system, Pp =(u,v‚w) is a welding seam point in an arc coor-
    S dinate system, and PP isa welding seam point detected by the la- Ser sensor.
    Step 32, a deviation Au in a normal direction U and a devi- ation 4W in a vertical direction W of an arc between a welding seam point detected by a laser sensor and a welding seam point on a teaching trajectory are obtained: Au=u u Aw =w —w (7) Where, A is a deviation of the arc in the normal direction u, AW is a deviation of the arc in the vertical direction w, u is a normal coordinate value of the welding seam point detected by the laser sensor, # is a normal coordinate value of the welding seam point on a teaching trajectory, W is a vertical coordinate value of the welding seam point detected by the laser sensor, and W is a vertical coordinate value of the welding seam point on the teaching trajectory.
    Step 4, the deviation Au in the normal direction U and the deviation AW in the vertical direction W are respectively smoothed by using an SG smoothing algorithm: out(i)=C=2*[in(i—3)+in(i+3)]+3*[in(i-2)+in(i+2)] +6*fin(i—1)+in(i+1)]+7*in(i)] 21 (8) Where, i is a sampling sequence number, in(i) is the ith in- put, and out(i) is the ith output.
    Step 5, the smoothed deviation AM of a radius r and the de- viation 4W in a direction W are input into a PID controller to obtain a smooth and stable welding seam tracking deviation: the PID controller is: k U =k, e, +k De vk, (Ce) J=0 ee (9) Where, k is a sampling sequence number, Uc is the output of . Le . . . e. | the kth sampling, © is the input of the kth sampling, + is the input of the jth sampling, “rt is the input of the (k-1)th sam- k, . : : Lo. k : : pling, + is a proportionality coefficient, '/ is an integration coefficient, and ky is a differential coefficient.
    Step 6, a welding seam point after a welding seam is correct- ed in the robot base coordinate system is obtained according to the welding seam tracking deviation obtained in step 5, which in- cludes a specific process as follows: step 61, the deviations in the normal direction W and the vertical direction W obtained in step 5 are substituted into a point P1 on a corresponding arc to obtain a corrected welding seam point P1’ in an arc coordinate system:
    Pl = (r+ Mt)cosO (r+ Mu)sing, Aw) (10) Where, f is a rotation angle of the point Pl, r is a radius of the arc, the position deviation information A js the devia- tion of the radius r, and AW is the deviation in the direction W 5 .
    Step 62, the corrected welding seam point in the arc coordi- nate system is transformed into the robot base coordinate system oT according to the transformation matrix #4 : | pi =u, (r+ Aw cosf+v, (r+ Ausin@+w_- Aw +p, pi =u, (r+ Au)cos@ + var (r+A4u)sin@ + w‚ Aw+p, lp! =u_-(r+Au)cos@+v,-(r+Au)sin@+w_- Aw+p, (11) where, PP=Cp pls Ps) is a corrected welding seam point in the robot base coordinate system.
    During welding, the deviations of the welding seam and the teaching trajectory are caused by the changes of the position and the size of the welding seam due to factors, such as deformation, a variable clearance, and an assembly error, so the robot needs to track and correct quickly and stably in real time. The welding seam tracking during the welding is different from other automatic tracking technologies, stability is very important in welding. If a welding gun is not stable enough during the welding and the cor- rection is not quick and stable, it not only affects the attrac- tiveness of welding seam formation, but also is not conducive to subsequent welding. Therefore, the design of a welding seam track- ing controller is to make welding seam tracking control smooth as much as possible and reduce oscillation on the premise of ensuring a correction force.
    Embodiment: An experiment is performed according to a method recorded in the specific implementation mode: the robot is taught to transform and smooth the position data information collected by the laser sensor through a smoothing algorithm and input the transformed and smoothed position data information into the PID controller along a movement path from a lowest point to a highest point of a pipe fitting, and then the position data information is transformed in- to the robot base coordinate system through coordinate transfor- mation to control the end of the robot to correct the welding seam deviation online along an arc trajectory. The robot is stable without obvious jitter during the welding.
    A corrected welding seam trajectory is recorded. The robot moves according to the recorded trajectory. The laser sensor col- lects and records deviation information Aland aw for example, a 4W value recorded by a curve as shown in FIG. 4 and a Au value recorded by a curve as shown in FIG. 5. AW fluctuates between -
    0.6-0.6, and a data center value is near 0; and Au fluctuates be- tween -0.5-0.7, and a data center value is near 0.
    权利要求:
    Claims (1)
    [1]
    CONCLUSIONS
    A robotic weld tracking method for welding a high curvature pipe fitting in all positions, characterized by comprising a specific process as follows: step 1, obtaining a transformation matrix from a laser sensor coordinate system to a coordinate system of the robot tool; step 2, obtaining a learning transformation matrix from an arc coordinate system to a base coordinate system of the robot, comprising a specific process as follows: step 21, obtaining a first path for laser welding in all positions with a robot by online teaching; step 22, inputting the initial path of laser welding into a controller to obtain an arc spatial coordinate system H; e OT step 23, obtaining a transformation matrix H from the arc spatial coordinate system H to a base coordinate system of the robot B; step 3, obtaining a weld point deviation in an arc coordinate system in accordance with the matrices obtained in step 1 and step 2, comprising a specific process as follows: step 31, transforming a weld point detected in a coordinate system of the laser sensor into the arc coordinate system; step 32, obtaining a deviation Au in a normal direction U and a deviation AW in a vertical direction W' of an arc between a weld spot detected by a laser sensor and a weld spot on a learning path; step 4, smoothing the deviation Au in the normal direction U and the deviation AW in the vertical direction W, respectively, using a Savizkg-Golag (SG) smoothing algorithm; step 5, inputting the smoothed deviation Au of a radius and the deviation AW in a direction W into a Proportion Integration Differentiation (PID) controller to obtain a smoothed and stable weld tracking deviation; and step 6, obtaining a weld spot after a weld seam is corrected in the robot's base coordinate system in accordance with the weld tracking deviation obtained in step
    5.
    A robotic weld tracking method for all position welding of a high curvature pipe fitting according to claim 1, characterized in that obtaining a transformation matrix from a coordinate system of the laser sensor to a coordinate system of the robot tool in step 1 involves a specific process as follows: 8 s x 53 / step 11, obtaining coordinates Pp = p) 0 Pp, 1) of the weld point in the coordinate system of the laser sensor and coordinates P-(pf Pp pi 1) of the weld point tested in the coordinate system of the robot tool; step 12, setting the transformation matrix from the laser sensor coordinate system to the robot tool coordinate system as: ns. 4a. Pp pr p _ s, 4, * pi IL SS, ap, 0 0 0 1 (1) LT Lo where 9 represents the set transformation matrix is from the coordinate system of the laser sensor to the coordinate system of the robot tool, S represents the coordinate system of the laser sensor, x represents the x-directional position information of a weld seam in the coordinate system of the laser sensor S, and z represents the z-directional position information of a weld seam in the coordinate system of the laser sensor S ; wherein the coordinate system of the laser sensor S is obtained by directly detecting and extracting weld seam information through the laser sensor; wherein the x-directional position information and the z-directional position information in the coordinate system of the laser sensor S are obtained directly through the laser sensor;
    step 13, entering the coordinates of a spatial weld point of the laser sensor coordinate system and the coordinates of the weld point of the coordinate system of the robot tool into the transformation matrix from the laser sensor coordinate system to the coordinate system of the robot tool
    tool: | no pita pp =p" sx SZ he 1, pj) ta, pp, =p; no pr tap" +P.=P > (2) further transforming formula (2) to obtain: sx SZ Ix D pl ) Pi Pi a, Pi : : Na l=] sx SZ x Ix Lond) sx SZ Iy pi Pi | Pi : coh : | u 1 a, = IN le] ’ * ’ ’ ee (3) sx sz Iz Pi Pi 1 n Di | . a, = Pp PP Hp | : : : : step 14, solving formula (3) by means of a least squares method to find parameters i=(n,n,n),ä=(a, a, a), p=p Pp, P:) . ee 0p dP PoP» Pe in the transformation matrix; - S=ls ss) . step 15, obtaining " » 7 j in accordance with step 14; and S=axn (4) step 16, entering the obtained n=n,n,n)d=(a,a,a) ,p="pp,p.) and S=(s,8,8 ,) in the set Tr transformation matrix $ from the laser sensor coordinate to the
    IT bone tool coordinate system to obtain the transformation matrix 5» from the coordinate system from the laser sensor to the robotic tool coordinate system.
    A robotic weld tracking method for welding a high-curvature pipe fitting in all positions in accordance with claim 2, characterized in that transforming a weld spot detected in a coordinate system of the laser sensor into the arc coordinate system in step 31 involves a specific process as follows:
    POTT P PT at where, represents the transformation matrix from the coordinate system of the robot tool to the base coordinate system of the robot, and is obtained from an end position of a ro-or bot; ¥ represents a transformation matrix from the laser sensor coordinate system to the coordinate system of the robot tool, PY =u, vw) is a weld point in the arc coordinate system, P* is a weld point detected by the laser sensor, and PT ee is PD a transformation matrix from the arc-spatial coordinate system H to the base coordinate system of the robot B; and Bp and £ are obtained from the robot's end position.
    A robotic welding seam tracking method for welding a high-curvature pipe fitting in all positions according to claim 3, characterized in that obtaining a deviation Au in a normal direction U and a deviation 4W in a vertical direction W of an arc between a weld point detected by a laser sensor and a weld point on a learning path in step 32 involves a specific process as follows:
    Au=u —u Aw =w —w (7) where, Au is a deviation in the normal direction U of an arc, Aw is a deviation in the vertical direction W' of the arc, ¥ is a normal coordinate value of the weld point detected by the laser sensor, ¥ is a normal coordinate value of the weld point on a learning path, W is a vertical coordinate value of the weld point detected by the laser sensor, and W is a vertical coordinate value of the weld point on the learning path.
    A robotic welding seam tracking method for all-position welding of a high curvature pipe fitting according to claim 4, characterized in that smoothing the deviation Au in the normal direction U and the deviation Aw in the vertical direction, respectively W' by using a SG smoothing algorithm in step 4 a specific process includes as follows: out(i)=C=2%[in(i 3)+ii +3) +3* [ini 2 )+i{i+2)] +6*fin(i—1)+in(i+1)]+7*in(i)] 21 (8) where, 1 is a sample sequence number, in(i) the is ith input and out(i) is the ith output.
    A robotic weld tracking method for all position welding of a high curvature pipe fitting according to claim 5, characterized in that it comprises inputting the smoothed deviation Au of a radius and the deviation AW in a direction W in a PID controller to obtain a smoothed and stable weld trace error in step 5: where the PID controller is: ku, =k, e +k De, +k, (tej) J=0 me (9)
    where, k is a sample sequence number, “ is the output of the . 2, . . €; . k-th sample, Ck is the input of the k-th sample, + is the input of the jth sample, Cr! is the input of the (k-1)th sample, + is a proportionality coefficient, 3 is an integration coefficient and kq is a differential coefficient.
    A robotic weld tracking method for all position welding of a high curvature pipe fitting according to claim 6, characterized in that obtaining a weld spot after a weld has been corrected in the base coordinate system of the robot in accordance with the weld tracking deviation obtained in step 5, in step 6 includes a specific process as follows: step 61, replacing the deviations in the normal direction U and the vertical direction Ww' obtained in step 5 at a point P1 at a corresponding arc about a corrected weld point Pl! in the arc coordinate system; PI =((r + Au)cos6,(r + Au)sin 8, Aw) (10) where, 0 is an angle of rotation of the point Pl, r is a radius of the arc, the position deviation information Au is the deviation of the radius r is, and Aw is ge deviation in the direction W; step 62, transforming the corrected weld point of the arc coordinate system into the base coordinate system of the
    ST robot according to the transformation matrix f# ; pi =14, (r+ AucosO+v, (r+ Asin +w_-Aw+ p. pi =u, (r+ dudcos@+v, (r+ Auisin@+w Aw+p, pl =u (r+ A)cos@+ v, (r+ A)sin@+w_- Aw +p, * : * : * (11) PP =p”, pl ps) where, Chr Tyr Prd is a corrected weld point in the robot's base coordinate system.
    类似技术:
    公开号 | 公开日 | 专利标题
    CN104384765B|2016-04-27|Based on the automatic soldering method of threedimensional model and machine vision
    CN105772905A|2016-07-20|Skew offset pipe trajectory planning method based on arc welding robot system
    CN109927047B|2021-12-03|Axial tracking system and method for straight butt weld of arc welding robot
    US20110106308A1|2011-05-05|Method and a device for optimizing a programmed movement path for an industrial robot
    Fang et al.2017|Robot path planning optimization for welding complex joints
    CN206263418U|2017-06-20|A kind of real-time seam tracking system of six degree of freedom welding robot line laser
    CN103418950A|2013-12-04|Automatic posture adjusting method for industrial welding robot in seam tracking process
    Chen et al.2014|A robust visual servo control system for narrow seam double head welding robot
    Pan et al.2011|Automated offline programming for robotic welding system with high degree of freedoms
    Efimov et al.2019|Improving the efficiency of automated precision robotics-enabled positioning and welding
    NL2029506A|2021-12-21|Robot welding seam tracking method for all-position welding of large-curvature pipe fitting
    CN108311835A|2018-07-24|A kind of dynamic electroplating equipment wielding machine arm system that view-based access control model measures
    CN111941421A|2020-11-17|Self-adaptive fuzzy force tracking control method based on multi-robot cooperative operation
    CN106002014A|2016-10-12|Welding system suitable for right-angled welding line laser tracking and welding method thereof
    CN113334018A|2021-09-03|Visual guide robot automatic welding pipeline intersecting line device and method
    Lei et al.2021|The development of tube-to-tubesheet welding from automation to digitization
    Chen et al.2016|A path-planning algorithm of the automatic welding robot system for three-dimensional arc welding using image processing
    Gao et al.2015|Study on arc welding robot weld seam touch sensing location method for structural parts of hull
    Wang et al.2021|Fuzzy-PI double-layer stability control of an online vision-based tracking system
    Davila-Rios et al.2016|A fuzzy approach for on-line error compensation during robotic welding
    Shanmugasundar et al.2016|Method of Trajectory Generation of a Generic Robot using Bresenham’s Circle Algorithm
    Yin et al.2019|Monitoring-based visual servoing of wheeled mobile robots
    CN112692470A|2021-04-23|Robot off-line adaptive programming processing system and method based on visual sensing
    CN215787900U|2022-02-11|Visual guide robot automatic welding pipeline intersecting line device
    CN113240753A|2021-08-10|Sphere fitting method for calibrating base coordinate system of robot and double-shaft deflection mechanism
    同族专利:
    公开号 | 公开日
    CN112809167A|2021-05-18|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    法律状态:
    优先权:
    申请号 | 申请日 | 专利标题
    CN202011641612.0A|CN112809167A|2020-12-31|2020-12-31|Robot weld joint tracking method for all-position welding of large-curvature pipe fitting|
    [返回顶部]