LIBRARY ı6ıul 0) - Cranfield University

More documents

Recommendations

Info

According to the British Standards Institution [ref. 93], pose is defined as the position and attitude of one co-ordinate frame with relation to another. Position includes a 3-dimensional (3D) vector of translational parameters and attitude includes a 3D-vector of angular parameters. Command pose is defined as the pose specified through teach programming, manual data input or explicit programming. Attained pose is defined as the pose achieved by the robot under automatic mode in response to the command pose. The quantification of the deviations which occur between the command pose and the attained pose is defined as robot pose accuracy [ref. 93]. The quantification of the fluctuations in the attained pose which occur for a series of repeat visits to a command pose is defined as repeatability [ref. 93]. These are performance parameters that have evolved from on-line programming of industrial robots. When considering off-line programming, the absolute accuracy characteristics of the robot must be taken into account, since here externally generated co-ordinate data are used in much the same way as is done with CNC part programming [ref. 87]. The absolute accuracy of an industrial robot is the machine's ability to achieve a specified set of co- ordinates relative to its World Co-ordinate Frame [ref. 94]. Hence, the need for an accurate geometric model of the links and joints of the robot is essential. In order to improve the robot accuracy, it is necessary to perform a robot calibration. Bernhardt and Albright [ref. 94] have defined robot calibration as the term applied to the procedures used in determining actual values which describe the geometrical dimensions and mechanical characteristics of a robot structure. These are classified as kinematic and dynamic parameters. Kinematic parameters primarily describe a robot's arm lengths and relative joint-axis orientations while dynamic parameters primarily describe arm and joint masses and internal friction. There are two main types of robot calibration [ref. 94]: static and dynamic. Static calibration is an identification of the parameters which influence primarily the static positioning characteristics of a manipulator. It is used to identify the actual internal robot features such as joint-axis geometry, joint angle off-sets, actuator/link compliance, actuator transmission and coupling factors. All these factors may have an influence to a certain degree on the static positioning accuracy of the robot. On the other hand, dynamic calibration addresses the parameters which influence the motion characteristics. Once a robot's static parameters are identified, a dynamic calibration can take place. This type of calibration is used to determine the dynamic related characteristics of the robot structure (e. g. distribution of mass in the links, friction in actuators and joints, stiffness, etc. ). Internal characteristics such as friction tend to be difficult to identify accurately, due to their coupling with other dynamic parameters [ref. 94] Whatever the calibration type is (static or dynamic), a similar general procedure for the identification of the unknown parameters is used. This normally starts by modelling the robot system, followed by measurement of appropriate inputs/outputs or internal reactions (e. g. position, motion, force, torque, etc. ) and then ending with identification and verification of actual parameter values. The identification is performed by minimising an error function which is obtained by comparing the model computed tool positions with the measured positions. This is called forward calibration [ref. 91]. 23
Another type of robot static calibration reported in the literature is the inverse calibration [refs. 91,95,96], in which several points in a region of the workspace are measured precisely and compared with the points reported by the robot measuring system. An empirical (or table look up) relationship is then determined. This kind of calibration does not need system modelling, which can sometimes be very difficult. Furthermore, it may be possible to obtain a better match of the system. However, it usually requires more measurements and does not provide much insight to what caused the error. The implementation of the calibrated model can be performed in two different ways [ref. 91]: a) Real time inverse kinematics and b) Off-line inverse solution. The first method is based on the substitution of the incorrect model by the calibrated one in the robot controller. This results in a very accurate robot movement. The second method uses the calibrated model for adjusting the command poses in such a manner that the attained poses correspond to the required world co-ordinate points. The Real time inverse kinematics method cannot always be implemented, since the calibrated model is normally more complex than the nominal model. The off-line inverse solution is directly applicable in off-line programming systems. However, this technique does not provide the same levels of accuracy as the Real time inverse kinematics method [ref. 91]. One example of off-line inverse solution was described by Owens and Piatkowski [ref. 97] when applying Workspace® and Robotrak to calibrate a waterjet-cutting robot workcell. 2.4.3 Cell calibration The calibration issues discussed before addressed mainly the calibration of the robot arm. Another important factor in off-line programming is the calibration of the robot workcell. The relative positions of the several components of the cell must be known to the programmer for an accurate modelling of the robot environment. The measurement of the relative positions of the several cell components can be performed in several ways, the most simple being the one which uses the calibrated robot as a measuring device [refs. 87,88,98]. 2.4.4 Assessment of programming errors [ref. 87] According to the definition given in section 2.2.1, a robot needs to be programmed in order to perform a specific task. Depending on the method of programming, different types of programming errors may occur. Whatever the on-line programming method used (manual lead-through or teach pendant programming), the errors that can arise are normally due to variations in workpiece positioning and workpiece dimensional tolerances. In the case of lead through programming, additional errors due to mechanical flexing during programming may occur. The robot's relatively poor absolute accuracy is no longer an issue, since programmed points are set relative to the workpiece. When using off-line programming, further to the errors caused by misplaced workpiece and by workpiece tolerances, faults might also happen due to perfect kinematic models being used by both the simulation and the real robot controller to drive an imperfect robot arm. Other sources of error can be found, such as geometric 24
Page 1 and 2: CRANFIELD UNIVERSITY GC CARVALHO AN
Page 3 and 4: ABSTRACT The aim of this work was t
Page 5 and 6: I dedicate this work to the memory
Page 7 and 8: FIGURES TABLES APPENDICES NOTATION
Page 9 and 10: 3.3.2.1 Welding parameters generato
Page 11 and 12: References Further reading Appendic
Page 13 and 14: Figure 3.12 Offsets for weld start
Page 15 and 16: Figure 6.31 Voltage step input test
Page 17 and 18: Figure J. 2 Plot of stand-off varia
Page 19 and 20: Table A. 1 Welding extended entity
Page 22 and 23: NOTATION A, Electrode sectional are
Page 24 and 25: Pr(ign) Possibility measure of proc
Page 26 and 27: 1. Introduction Welding is the thir
Page 28: Chapter 4 describes the on-line pos
Page 31 and 32: In pulse transfer, the welding curr
Page 33 and 34: Pd = Pv - pzh-s (2.1) where Pd is t
Page 35 and 36: the bridge. Another factor that ind
Page 37 and 38: establish the stable conditions for
Page 39 and 40: Philpott [ref. 21] developed an on-
Page 41 and 42: current peak at the moment the tip
Page 43 and 44: obot on the line must be individual
Page 45 and 46: collision detection capabilities wh
Page 47: cause dynamic variation in the seam
Page 51 and 52: spatter the gas nozzle is particula
Page 53 and 54: Table 2.1: Joint positioning tolera
Page 55 and 56: In order to minimise the undesirabl
Page 57 and 58: c) Ultrasonic sensors; d) Through-t
Page 59 and 60: where K,, K2, K3 and K4 are constan
Page 61 and 62: Philpott [refs. 21,131], based on t
Page 63 and 64: centre. This is attributed to the m
Page 65 and 66: technique must be employed to preve
Page 67 and 68: " digital hardware, which is basica
Page 69 and 70: " Maximum value, W.: Wmý = max(W )
Page 71 and 72: process studied over a small range.
Page 73 and 74: In-process welding control is a muc
Page 75 and 76: volume caused by the presence of a
Page 77 and 78: SHIELDING GAS IN CONSUMABLE ELECTRO
Page 79 and 80: Arc voltage Anode Arc length I Anod
Page 81 and 82: computed ideal procedure optimisnti
Page 83 and 84: CONSTANT CURRENT POWER SOURCE `j' O
Page 85 and 86: Original Surface New Surface Electr
Page 87 and 88: otating welding wire direction weld
Page 89 and 90: 3.1.1.1 Robot errors A robot arm is
Page 91 and 92: 3.2.2 Programming error correction
Page 93 and 94: 3.3.2 Off-line programming module I
Page 95 and 96: Pen is the weld penetration (side p
Page 97 and 98: a. 2) Geometrical constraints from
Page 99 and 100:
Step 11 Step 12 Step 13 Step 14 Ste
Page 101 and 102:
Step 21 Step 22 If the wire feed sp
Page 103 and 104:
the Y'-axis results in a positive "
Page 105 and 106:
_ p1e0a - pORoba m31 Ilp! - poll (3
Page 107 and 108:
of a robot in its zero position12 w
Page 109 and 110:
frame points to the front of the ro
Page 111 and 112:
CAD Off-line (AutoCAD) programming
Page 113 and 114:
Z Surface 4 A Y 7 M2 Open Edge Join
Page 115 and 116:
WCF World Co-ordinates Frame 2a = J
Page 117 and 118:
x Torch Co-ordinates Frame o Indica
Page 119 and 120:
vo Ö 't7 Y7 't O m O 'S O O S O S
Page 121 and 122:
ýý aý 0 oA aW ö A O Zt A OO 't
Page 123 and 124:
Ti 9ý-! i4 *Tx t ap31 jx ý- ymin
Page 125 and 126:
errors', (i. e. joint positioning,
Page 127 and 128:
was considered to be outside the sc
Page 129 and 130:
Table 4.3 - Linguistic representati
Page 131 and 132:
The introduction of such a reduced
Page 133 and 134:
Table 4.8 - Final welding process c
Page 135 and 136:
the collection of welding data corr
Page 137 and 138:
Step 4- Calculate the confidence of
Page 139 and 140:
Z Table Co-ordinates Frame Y X Robo
Page 141 and 142:
5.3 Monitoring system The monitorin
Page 143 and 144:
the torch end. Only one degree of f
Page 145 and 146:
Current: 500 Position: 234.5 5.7.3
Page 147 and 148:
I" va JI ºr ö C 3 r" r" rl f_ £-
Page 149 and 150:
IgurC a. H -I UI qur vCiSUJ JpCCU C
Page 151 and 152:
126
Page 153 and 154:
Table 6.1 - Welding trials carried
Page 155 and 156:
Table 6.3 - Coefficients of Ogunbiy
Page 157 and 158:
Note that the calibration model sho
Page 159 and 160:
Table 6.6 - Welding parameters used
Page 161 and 162:
constant set-up welding parameters
Page 163 and 164:
stickout lengths and the temperatur
Page 165 and 166:
of the dip mode of metal transfer.
Page 167 and 168:
which would become active by settin
Page 169 and 170:
In section 4.2.5 it has been mentio
Page 171 and 172:
300 280 260 240 rd 35- 3D 25 20 15
Page 173 and 174:
Figure 6.5 - Measured "versus" Pred
Page 175 and 176:
i S 0 -0.1 -0.2 -0.3 -0.7- 68 10 12
Page 177 and 178:
32.0 E 31.5 y 31.0 30.5 30.0 j 29.5
Page 179 and 180:
35 - 30 p 15 10 0.00 2.60 5.42 8.23
Page 181 and 182:
40 35 30 25 20 15 10 A+iIII+F0.17 2
Page 183 and 184:
30 28 26 24 22 SO_act [mm] DipR [mo
Page 185 and 186:
22 20 18 SO_act [mm] DipR [ohm/100]
Page 187 and 188:
!: 30 25 20t 10 5-- 0 Dip Transfer
Page 189 and 190:
0.430- - 0.380- - :r 0.330 fPR L- 0
Page 191 and 192:
7.2 Tests with varying stand-off an
Page 193 and 194:
Table 7.3 - Bead geometry along the
Page 195 and 196:
ö a2 ., u" u,.., g .,. aucyuaac VI
Page 197 and 198:
0 24-- 22-- 18 210 1 90 0 v 170 mm
Page 199 and 200:
Q 34 32 3o mIm Viewing direction Fl
Page 201 and 202:
ýa 34 mm JO a .. ý nn ýr "u 00 M
Page 203 and 204:
22 20 18 Viewing direction Flange W
Page 205 and 206:
36 33 31 o 29 t 350 310 o 290 U due
Page 207 and 208:
I C, x -i 23 21 j 19 Q 240c WFS =10
Page 209 and 210:
en 34 32 - vsec 30- 26- 24 + .. ++
Page 211 and 212:
186
Page 213 and 214:
" to design and build the hardware
Page 215 and 216:
Although only linear joints were im
Page 217 and 218:
for predicting possibility of bad a
Page 219 and 220:
8.4 Process monitoring and control
Page 221 and 222:
satisfy the required quality criter
Page 223 and 224:
198
Page 225 and 226:
types of joints and multi-pass weld
Page 227 and 228:
[14] NEMCHINSKY, V. A. The effect o
Page 229 and 230:
[38] D1LTHEY, U. , REICHELL, T. , S
Page 231 and 232:
[64] KING, F-J. and HIRSCH, P. Seam
Page 233 and 234:
[86] CARGNELLI, M. and ROGOWSKI, A.
Page 235 and 236:
[109] KURKIN, N. S. and DRIKKER, V.
Page 237 and 238:
[130] USHIO, M. , LIU, W. , MAO, W.
Page 239 and 240:
[151] DAVIS, A. R. Orr-line gap det
Page 241 and 242:
[177] COOK, G. E. Feedback and adap
Page 243 and 244:
[201] WON, Y. J. and CHO, H. S. A f
Page 245 and 246:
220
Page 247 and 248:
222
Page 249 and 250:
wnum: local variable which defines
Page 251 and 252:
Table A. 1 - continuation List item
Page 253 and 254:
Imean = a2 + b2*WFS + c2*SO + d2*SO
Page 255 and 256:
TCP2 number are also defined. The o
Page 257 and 258:
f ROBSET. DAT: This file is created
Page 259 and 260:
Figure C. 3 - Main dialogue box of
Page 261 and 262:
Choose the set of welding parameter
Page 263 and 264:
Water Cooling optic fibre bundle 0
Page 265 and 266:
Figure E. 1 - Main graphical screen
Page 267 and 268:
242
Page 269 and 270:
Table G. 2 - Interface box external
Page 271 and 272:
Robot ii Controller +24W CO . +24Vd
Page 273 and 274:
248
Page 275 and 276:
Table 1.1 - continuation Run Vpk M
Page 277 and 278:
10- y =0.001ac 0 r. dSO [mm] $0 '10
Page 279 and 280:
5.00- 0.00 y=0.0019x 1000 1500 2000
Page 281 and 282:
Table J. 7 - Welding data collected
Page 283 and 284:
Table J. 10 - Welding data collecte
Page 285 and 286:
Page 287 and 288:
20.00- 15. M y=0.0047x - 1.7922 10.
Page 289 and 290:
Page 291 and 292:
266
Page 293 and 294:
Table K. 1- continuation Set-up par
Page 295 and 296:
Table K. 1 - continuation Set-up pa
Page 297 and 298:
Table K. 2 - continuation Set up pa
Page 299 and 300:
Table K. 2 - Continuation Set-up pa
Page 301 and 302:
Table K. 3 - Continuation Set-up pa
Page 303:
Table K. 3 - Continuation Setup par
show all

LIBRARY ı6ıul 0) - Cranfield University

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?