GB2027936A - Automatic arc welding system - Google Patents

Abstract

Apparatus for rapid and automatic welding of joints in a workpiece or in interconnected workpieces, such as oil or gas transmission pipe. The system includes at least one torch transport assembly which simultaneously moves a plural number of welding torches along a path parallel to the pipe joint being welded. The torch transport assembly and selected operating parameters of each welding torch are adjusted to accurately position each torch with respect to the pipe joint, and in response to an indication of the location of the torch along the path to provide a desired value of a welding parameter previously programmed for that location resulting in a uniform pipe joint weld. The present system is disclosed in the context of hot-wire gas-tungsten arc welding torches, and in the operating environment of out-of-position joint welding. <IMAGE>

Description

SPECIFICATION Automatic welding system This invention relates in general to welding and in particular to off-axis welding of pipeline and other workpieces in which a welding torch is traversed with respect to a relatively stationary workpiece.

Transmission pipelines are frequently used to transport fluid products for substantial distances, with oil and gas transmission pipelines being but two well-known examples. Such pipelines are generally constructed of individual pipe sections that are joined together by welding, and it is important that each welded pipe joint meet the criteria necessary for safe operation of the pipeline. Not only is the detection and repair of defective joints an extremely expensive undertaking, particularly where the pipeline is buried underground or submerged beneath the sea, but the existence of a defective weld joint creates a hazard of catastrophe in the case of pipelines which carry flammable products.

The separate lengths of pipe which make up an oil or a gas transmission pipeline are typically interconnected by electric-arc welding, and many types of welding techniques are used or proposed in the art for that purpose. Perhaps the simplest welding technique is shielded metal arc welding, commonly known as "stick welding", in which one or more persons manually weld each joint using hand-held welding torches of conventional design. Since each welded joint of a typical pipeline actually consists of several separate weld segments or layers, known as "passes", manual welding techniques are time-consuming and costly in view of the large number of welded pipe joints to be welded in a pipeline extending for many miles.

The relative slowness of manual and other conventional welding techniques is particularly troublesome in pipe laying operations where wages and equipment expenses provide a costly overhead which must be rationalized by maximizing the number of acceptable welds that can be produced per unit of time.

Welding devices have been proposed which traverse a welding torch about the circumference of a pipe joint while welding the joint. Since it is obviously impossible to rotate the joint end of a pipeline which may extend for many miles from a laydown barge (or any other joint welding site), automatic pipe welding equipment generally traverse a weldng torch circumferentially about the pipe joint. Those skilled in the art will recognize that 'off-axis" welding, that is, welding applications where the welding torch departs from an upright vertical position, becomes increasingly difficult as the welding torch departs the 12 o'clock or upright position in its passage around the pipe joint.The weld puddle is subjected to natural forces including gravity, surface tension and capillary attraction within the grooved joint being welded, and the net force acting on the puddle constantly changes as the torch traverses about the circumference of a pipe joint which is in a nonhorizontal plane. When the torch is welding at the 6 o'clock position, maximun care must be exercised to prevent the weld puddle from falling out of the weld by gravity.

It has been proposed to overcome gravitational pullout of the weld puddle during off-axis welding by applying pulsed welding current to the torch, so that the weld puddle will slightly congeal during each "off" portion of the pulsed weld current. The operating speed of welding torches receiving pulsed welding current must be correspondingly reduced, however, and it is possible that the welds which are produced by the pulse-induced intermittent partial cooling may have undesirable metallurgical properties.

The speed at which a pipe joint can be welded is determined by the maximum rate at which the welding torch can deposit weld metal while traversing a pipe joint which is in a nonhorizontal plane. While gas metallic arc (MIG) welding torches generally have a relatively high rate of metal deposition, such torches generally produce welds that are nonuniform and difficult to repeatably obtain. Gas tungsten arc (TIG) welding torches are known to produce a pipe joint weld of superior and more repeatable quality, although the metal deposition rate for TIG welding torches is relatively slow. A development known as the hot-wire TIG torch, in which electric current is passed through the filler wire to preheat the filler wire which is melted in the weld puddle, is known to produce a substantially increased rate of metal deposition, relative to conventional TIG welding.The lack of sufficiently precise and repeatable weld parameter control of prior-art hot-wire TIG welding torches, however, along with the aforementioned problem of off-axis weld puddle control and related problems, have heretofore kept the metal-deposition rates of hot-wire TIG torches from being fully realized in pipeline welding applications. A description of hot-wire TIG welding is set forth in U.S.

Patent No, 3,122,629.

Accordingly, it is a object of the present invention to provide a improved apparatus for welding along a predetermined path of workpieces such as pipe joints or the like.

It is another object of the present invention to provide apparatus for automatically welding pipe joints and the like, at an improved rate of metal deposition and with an improved quality of weld.

Still another object of the present invention is to provide a programmable welding system in which selected welding parameters are automatically varied in relation to welding torch position or other factors.

Stated in general terms, the present invention comprises a welding system in which one or more welding parameters of single or plural welding torches are independently varied as a function of the location of each torch on a path, as the torch traverses the path. The system includes track sections which are positionable parallel to a joint to be welded, and carriages which are movable along each track. Each carriage supports one or more welding torches in angular offset relation, relative to the circumference of the pipe joint to be welded. Each carriage, as well as each welding torch carried by the plural carriages, is independently operable to provide preselected optimal welding parameters for the particular circumferential position of each welding torch relative to the circumferential pipe joint being welded.

The foregoing and other objects and advantages of the present invention will become more readily apparent from the disclosed preferred embodiment as described below with respect to the drawings, in which: Figure l is a pictorial view of a two-station pipeline welding system according to the disclosed embodiment of the present invention; Figure 2 is a fragmentary pictorial view showing the spatial relation of the clamp assemblies and associated welding torch heads which comprise the welding assembly portion of the welding station in Figure 2; Figure 3 is a block diagram showing the overall control system of the disclosed embodiment; Figure 4 is a section view, taken transverse to the weld pass direction, showing one example of a multiple-pass welded joint made with the present apparatus:: Figure 5 is a partial circumferential section view, shown somewhat enlarged, of the weld of Figure 4; Figure 6shows an example of an operating sequence which the disclosed apparatus undergoes during a typical weld pass; Figure 7is a block diagram of the control circuits for one of the fourwelding head assemblies Figure 8 is a schematic diagram of the AC filler wire heating power control for one of the welding head control circuits of the present invention; and Figure 9 is a schematic diagram of the filler wire feed control for one of the welding head control circuits of the present invention.

Turning to Figure 1, there is shown generally a two-section pipeline welding system comprising a first welding station A and a second welding station B. Each of the welding stations A and B is shown as being suspended from a rail R which is supported above and generally parallel to the several depicted sections of pipe P. The pipe section P' may arbitrarily be considered to be the head-end section of an already-completed pipeline of indefinite length while the other sections of pipe P depicted in Figure 1 represent pipe sections which are in the process of being welded onto the pipe section P' to constitute an extension of the pipelie.

Support member S may be provided for supporting the several pipe sections P, and it will be understood that suitable support means (not shown) is provided to maintain the rail R in the depicted position.

Those skilled in the art of pipeline welding will understand that a weld interconnecting two sections of pipe is typically composed of several separately-applied welding passes. The first such pass, typically called the "root pass", mechanically joins together the adjacent ends of two pipe sections. The root pass may be followed by a "hot pass", and then by one or more "fill passes" which primarily serve to fill the space between abutting ends of the pipe sections with filler metal, and the final or cap pass provides the exterior surface of the welded joint. It is frequently desirable to provide the root pass and one or more fill passes at a first welding station, A, for example, while the remaining fill passes and the cap pass may be provided at a second welding station B.It will be understood, however, that more than the two disclosed welding stations may be employed in a typical production pipeline welding situation. It should also be understood that each of the welding stations A and B disclosed herein can be identical in construction; alternatively, only a single such station may be provided for all welding passes of each pipe joint such as joint J.

Each welding station has a clamp assembly which can grip the pipe ends adjacent a joint J, and each clamp assembly carries an inner assembly IS. The inner assembly is movably positionable relative to the clamp assembly to permit positioning the inner assembly and its weld heads in position relative to the joint J. The general configuration and arrangement of the inner assembly IS is best seen in Figure 2, in which the two clamp assemblies CA1 and CA2 each include the corresopnding half-clamps CH1, Chi' and CH2, CH2'.

Each of the half-clamps CH2 and CH2' is supported by a corresponding pair of inner clamp arms 80 and 81 which are suspended below the clamp assembly 82. The half-clamps CH1 and CH1' of the clamp assembly CA1 are likewise suspended from the pair of inner clamp arms 86 and 87, and these two inner clamp arms are suspended from the clamp assembly. Each of the half-clamps CH1' and CH2' supports clamp feet F, F' which extend radially inwardly from the half-clamp for engaging the respective pipe joint ends.

Each of the half-clamps CH1 and CH2 is constructed to provide a track which is parallel to a circumferential portion of the pipe joint J. Each track supports a movable carriage TC for traverse in an arc about the pipe joint J. The carriage TC1 is mounted on the track associated with the half-clamp CH1, and carries a pair of welding head assemblies H1 and H2. The carriage TC2 mounted on the half-clamp CH2 carries another pair of welding head assemblies H3 and H4. The two welding head assemblies on each carriage are positioned at a right angle to each other and are longitudinally offset from the respective half-clamp a distance which places all of the four welding head assemblies in alignment with a common plane which is alignable, by appropriate positioning of inner assembly IS, with the joint J.

Each welding head assembly carries a hot-wire welding torch, and each such torch is supplied with filler wire from a separate spool by way of a wire drive and the flexible hollow guide tube which conveys the feed wire to a point adjacent the lower end of the torch assembly. Those skilled in the art will realize that a flow of electrical current, preferably AC, for 12R heating purposes is established between the filler wire and the weld puddle at the pipe joint J by passing the wire through an electrical contact tube located adjacent the torch.

The wire drive includes a motor which is operated at a selectably variable rate, in a manner described below, to feed wire from the spool to the torch assembly T. The motor preferably includes a velocity-feedback servo or is otherwise equipped to convey filler wire at a steady rate which is determined by operating signals supplied to the motor.

The movement of each carriage TC about its respective clamp half occurs independently of the other carriage, although the relative positions of the carriages TC1 and TC2 may be controlled so as to be interrelated for certain purposes. The operating parameters of each welding head assembly H1 through H4 are controlled independently of any other welding head assembly, with the values of the operating parameters being independently determined by factors such as the angular position of the carriage which supports the head assembly, sensed parameters, and preselected inputs for various parameters.Control of the overall system disclosed herein is provided through the control system shown generally in Figure 3, in which a central processor 214 interfaces through the input-output means 215 with the control circuits 21 6a--216d i ndependently associated with each of the welding head assem blies H1--H4. The central processor 214 also supplies signals through the interface 215 to the carriage control circuits 218--1 and 218--2, which individually drive the separate carriage positioning motors and which receive carriage position signals from the separate digital carriage position encoders 91--1 and 97--2. Each carraige drive motor is a servomotor which employs rate feedback to drive the carriage at the velocity commanded by the postion signals.

A transverse section of a typical welded pipeline joint of the type attainable with the present apparatus is shown in Figure 4, wherein the abutting ends 77a and 77b of pipe sections P and P' have been prepared by grinding or otherwise removing metal to form the beveled suface 452 extending from the outer surface of the pipe inwardly to the land 453. The lands 453 extend circumferentially around unremoved portions 454 of pipe metal which are abutted to provide the root of the joint J to be welded.

The initial welding pass, or root pass, mechanically joins together the two abutting portions 454, and may penetrate the pipe sufficiently to provide the slightly convex bead 455 about the joint J within the pipe. The root pass need not extend completely between the confronting beveled surfaces 452. The next welding pass, or "hot pass", is applied over the root pass to reshape and anneal the root pass. In typical operation of the present welding apparatus, the welding torch assemblies do not oscillate while welding the root pass and the hot pass.

Following the hot pass, one or more filler passes are applied to substantially fill the joint J with weld metal.

It will be understood that the weld parameters of each welding torch assembly should be selected to maximize the rate of filler metal deposition during the hot passes. The joint J is completed by covering the final filler pass with a pair of cap passes which overlap one another as shown as 457.

A section view of a typical joint J welded with the present apparatus is shown in Figure 4. The root pass commences at 460 as the welding arc is established, and slopes outwardly to attain maximum thickness at point 461 as the welding head assembly commences to traverse a sector of the pipe, as indicated by the arrow 462. The apparent terminal overlapping portion 463 of the root pass is actually the upsloping initial portion of the root pass produced by welding head assembly H4 which was moved to the 12 o'clock position to commence welding at point 465 after H1 is moved out of position, as discussed above. The complete root pass and each subsequent pass includes four such overlapped joints disposed about the circumference of the joint J.

The hot pass commences at point 464, and each subsequent welding pass similarly starts and finishes on slopes to provide overlapping joints as shown in Figure 5, so that the several weld passes smoothly overlap one another without abrupt step-shaped discontinuities. The starting and finishing slopes are provided by apparatus described below.

A typical example of a weld pass sequence with the present apparatus is shown in Figure 6. A number of the weld parameters for any welding sequence, including the sequence depicted in Figure 6, are predetermined for the particular weld pass and the type of joint, and are programmed to take place at selected points during the particular weld sequence. It should be understood that the weld parameters are "programmed" in the sense of being operator-predetermined, possibly by trial-and-error experimentation for a particular type and diameter of pipe, for example, as well as for each kind of welding pass to be used, and input into the system either by appropriate control settings or by information stored in a suitable memory and utilized throughout the selected weld sequence according to an appropriate operational program of the central processor 214.The level of computer programming ability necessary to accomplish such sequential operational programming of predetermined stored weld parameters is well within ordinary skill of the art.

The depicted weld sequence starts at a time indicated at 468 on Figure 6, and may be initiated by any suitable manual control. The supply of coolant and shielding gas to each torch assembly commences at time 468 and continues throughout the welding sequence. At the same time, each torch moves vertically inwardly as shown as 469, through operation of the autmoatic voltage control, until the torch contacts the joint J at time 470. After contacting the joint, each torch is independently retracted as indicated at 471 for a predetermined time and at a known rate which causes each torch to have a predetermined spacing from the joint, even though the clamp assemblies CA1 and CA2 may be essentially clamped onto the pipe.

Weld current is applied to each torch assembly at time 473, and an arc is initiated by any suitable technique such as a high4requency arc starting circuit (not shown) or the like. The weld current starts at a relatively low value indicated at 474 and increases over a predetermined upslope time to reach a programmed value at time 472. The filler wire feed and heating power, as well as the torch automatic voltage control, remain off at this time, and carriage movement has not yet commenced.

The carriage commences movement at a programmed time 475 soon after maximum programmed weld current is attained, and the automatic voltage control is initiated at time 476 to maintain the arc voltage of each torch at a programmed level. The torch oscillation control, if required by the particular weld pass, is also initiated at time 476. The automatic voltage control allows each torch assembly to be alternately retracted and extended as the torch oscillates across the nonuniform depth of the weld joint J, as seen in Figure 4, so that the arc spacing is automatically adjusted to maintain the programmed arc voltage throughout each oscillation of the torch assembly.

Filler wire feed and AC filler wire heating power commence at time 477, soon after carriage movement commences. It is apparent from Figure 6 that both the speed at which the filler wire is fed into the weld puddle, and the AC power applied to heat the filler wire, are increased during an upslope time to the maximum programmed values of speed and voltage.

The depicted weld sequence is now fully underway, and the programmed value of carriage speed, hot-wire heating current and/or feed rate, or any other weld parameter may be varied during the weld pass in response to the circumferential positions of the carriage assemblies, in accordance with "programmed" weld parameters that have previously been determined to be appropriate for the particular torch at various circumferential positions.The programmed weld parameters typically are separately predetermined, programmed, and controlled for each of the torch assemblies so that, for example, the size of an inverted weld puddle on the underside of the joint can be reduced to a point where surface tension and other natural forces acting on the molten metal exceeds the gravitational pull on the metal and prevents the puddle from dropping out of the joint, without similarly restricting the operating parameters of torch assemblies positioned elsewhere about the circumference of the joint.

At time 478 approaching the end of the depicted weld sequence, the automatic voltage control and torch oscillation are turned off and the weld current commences to decline from a programmed level to reach the final current level at time 479, at which point the carriage has reached the end of its programmed arc of travel. The filler wire feed and heating current also commence at time 478 to decrease over a downslope time to terminal levels at time 480, whereupon the feed wire is momentarily reversed to retract the wire from the still-molten weld puddle. It will be understood that the gradual downslope of depicted weld parameters, along with the programmed upslopes at the start of the weld sequence, produces the sloped and overlapped ends of the weld passes shown in Figure 5.The weld sequence is completed at time 479, although coolant and shielding gas continue to flow through each torch assemblyforthe cool-down period 481 to allow the weld puddle to solidify. Each torch assembly may be retracted during this time, as shown at 482, and it will be understood that each carriage can now be returned to its programmed starting position to await commencement of the next weld sequence.

Each of the control circuits 216a--216d may be basically identical, and only a single such circuit 216a is described herein with reference to Figure 7. The control circuit indicated generally art 216a includes a DC current control 222 which supplies a welding current control signal along the line 223 to the DC power supply for the torch assembly associated with welding head assembly Hi. The hot-wire power control 225a provides an output signal along the line 226 controlling AC power to heat the filler wire being supplied to the torch of the welding head assembly H1, and the wire feed control 225b provides a feed rate signal along the line 227 to the wire feed servomotor. Both the power control 225a and the feed control 225b for the hot wire are described in greater detail below.

The automatic voltage control 229 provides signals along line 230 to a motor which controls the vertical movement of welding head assembly H1. The automatic voltage control 229 receives an input signal on line 344 corresponding to the actual welding arc voltage of the torch associated with welding head assembly H1, and provides an output signal along line 230 to adjust the vertical position of the torch as necessary to maintain a programmed arc voltage determined by an input signal supplied on line 340 to the automatic voltage control.

The oscillation control 233 supplies operating signals along the line 234 to a motor which oscillates the welding head assembly H1 on a transverse path relative to the joint J.

Each of the DC current control 222, the hot-wire power supply control 225a, the wire feed control 225b, the automatic voltage control 229, and the oscillation control 233 receives command signals from the central processor 214 and input/output means 215 in response to factors such as the angular position of the carriage TC1, corresponding to desired welding parameters that have been manually or programmatically placed into the central processor. Each desired parameter change value is stored in digital format in the digital memory along with the desired carriage position (also in digital format) at which that parameter change is to occur.

For example, the following values might be stored for arc voltage: POSITION ARC VOLTAGE 000 10.7 080 10.8 170 10.9 450 10.8 570 10.7 800 10.6 When welding, the actual carriage position (as determined by the digital encoder position encoder 97-1 or 97-2) is continuously compared with the stored positions by means of an electronic digital comparator.

When the actual carriage position becomes equal to a stored position (for example, 080) that corresponding parameter value (Arc Voltage = 10.8) is output to the arc voltage control 229. Between the stored positions (for example, 080 and 170) the circuit may be programmed to linearly interpolate between the stored setpoint values (10.8 and 10.9) or it may be programmed to remain at 10.8 from position 080 to position 170 and then change abruptly to 10.9. The setpoint outputs may remain in digital format, or they may be converted to analog format for analog servos.

An Allen-Bradley programmable logic controller containing a 4096-word read/write memory provides the central processing function in an actual embodiment of the present invention, although any suitable programmable general-purpose computer can be used. The inputs to the control circuit 216a are received from the input-output means in digital form, and so each of the aforementioned controls is preceded by a digital-to-analog converter labeled "D/A" in Figure 7. The hot-wire power control 225a receives an input for wire-heating current (which provides a coarse adjustment of filler wire deposition), and the filler wire feed control 225b receives an input for the filler wire feed rate )which provides a fine or trim adjustment of filler wire deposition).The oscillation control 233 similarly receives separate inputs for the rate and width of torch oscillation, and the center of oscillation can be electronically adjusted through the oscillation control. Those skilled in the art will realize that the welding torch may not be oscillated during the initial passes of the pipe joint J, and that subsequent weld passes of the joint may require differing width and/or rate of oscillation of the torch.

The DC current control 222 includes an amplifier which receives a weld current command signal on the line 241 from the corresponding digital/analog converter. A feed back signal corresponding to the actual DC welding current is obtained from a shunt (not shown) in series with the welding current supplied to the torch, and the signal at 245 is also supplied to the weld current control.

The hot-wire current control 225a is shown in Figure 8, and includes the amplifier 282 receiving an input signal supplied by line 283 and corresponding to the actual AC voltage supplied to heat the filler wire at the torch associated with the particular welding head assembly. Output from the amplifier 282 goes to the RMS circuit 284, which supplies along line 285 an output signal proportional to the RMS value of the AC heating current signal on line 283. The RMS signal on line 285 is supplied to the summing point 286 of the summing amplifier 287.

An AC heating voltage command signal supplied from the central processor 214, through the appropriate digital/analog converter, is supplied on the line 291 to the amplifier 292, the output of which is supplied to the multiplier circuit 293. The output from the multiplier circuit 293 is amplified at 294 to provide a filler wire heating command signal which is supplied to the summing point 286.

The sudden application of full commanded wire heating current to the filler wire heating circuit can produce excessive heating at the start of a welding sequence, and the sudden removal of wire heating at the end of the welding sequence can also produce adverse effects, and it is desirable that the wire heating current be increased and subsequently decreased with the relatively gradual upslope and downslope depicted in Figure 6. The upslope/downslope control is provided by the circuit 296 in Figure 8, which includes an amplifier 262 for supplying a control signal to the multiplier 242. The amplifier 262 receives either a positive up-slope input signal through resistance 263 and a normally-open contactor (not shown), or a negative down-slope signal through resistance 265 and a normally-open contactor (not shown).The amplifier 262 functions as a integrator circuit, and the output on line 295 is clamped by the zener diode 268 across the amplifier 262 to be not less than zero volts.

The contactor for resistance 263 is closed when a typical programmed welding sequence is initiated, as described above in greater detail, and the voltage on line 295 is integrated from zero up to a level and at a rate determined by the voltage across resistance 263. The voltage on the line 295 is applied to the multiplier 392, so that the output from the multiplier to the summing point 286 is a product of the hot-wire current command signal and the upwardly-increasing signal on line 295. A predetermined minimum starting value for the hot-wire heating current may be provided, if desired.

Approaching the end of a programmed welding sequence, the contactorfor resistance 263 is opened and the contactor for resistance 265 is closed to cause the amplifier 262 to integrate the voltage on line 295 downwardly toward a final value and at a rate determined by the voltage across resistance 272.

The summing amplifier 287 is bypassed by the series resistance-capacitance circuit 299 to provide integral plus proportional compensated output to a conventional firing circuit 300 which sets the firing angle of SCRs 301 connected in the primary circuit of the transformer 302. The secondary winding of the transformer 302 is connected in series with the inductor 303 to supply heating current to the filler wire heating circuit. The inductor 303 provides the desired voltage-current slope of power supplied to heat the filler wire, and such inductors are known to those skilled in the art. The voltage across the secondary winding of the transformer 302 is fed back along the line 283 to provide the input signal to the amplifier 282.

The AC power control circuit 225a functions to provide an output signal from the summing amplifier 287 to the firing circuit 300 which will maintain the RMS output voltage from the transformer 302 at a level which is determined by the AC power command signal supplied on the line 291. The AC power supply to heat the filler wire is applied on an upslope at the commencement of welding and is terminated on a downslope near the end of welding, by the circuit 296, as described above. The described AC power control circuit provides AC line compensation while supplying a constant RMS AC voltage to the filler wire for 12R heating. Feedback control of the AC power is accomplished through relatively low voltage signals that are compatible with the command signals provided by the input/output means 215 associated with the central processor.

The wire feed control 225b, shown in Figure 9, includes an amplifier 309 connected to receive an input command signal on the line 310 from the appropriate digital/analog converter. The filler wire feed rate command signal on the line 310 is interconnected through the line 311 with the filler wire heating power command signal supplied on line 291, Figure 8, so that the control of the filler wire heating power and feed rate are operationally interrelated. It will be understood that a commanded increase in the feed rate of the filler wire will require a corresponding increase in the AC power necessary to maintain the desired heating rate of the filler wire; a commanded change in the AC heating power will necessitate a corresponding adjustment in the feed rate of the filler wire.

The potentiometer 312 provides a scale adjustment input to the amplifier 309, and the output of the amplifier goes to the multiplier circuit 313. The upslope/downslope circuit shown generally at 314 also provides an input to the multiplier circuit 313 to modify the wire feed command signal in a manner similar to the slope adjustment circuit described above, so that the output of the multiplier circuit on line 315 provides a reference signal corresponding to a commanded or desired wire feed rate.

The signal on the reference line 315 is amplified at 318 and 319, and supplied through diode 320 to a servoamplifier which drives the wire feed servomotor corresponding to the particular weld head assembly.

The wire feed servoamplifier may include an operational input which receives the feed command signal as aforementioned, and also a velocity feedback signal generated by a tachometer integral with the wire feed servomotor.

When the filler wire feed is terminated at or near the end of a welding operation, the filler wire conventionally remains in the molten weld puddle and adheres thereto, or forms a ball of molten metal at the end of the filler wire, either of which requires manual intervention of an operator. That problem is overcome with the automatic filler wire retrac circuit as shown generally at 321, in which the wire feed signal from amplifier 319 is supplied along line 322 to an input of the amplifier 323. An adjustable bias from the potentiometer 324 is maintaned on the amplifier 323, and the output of that amplifier is supplied through the RC differentiator circuit 325 to the timing circuit 326. Potentiometer 327 adjusts the period of the timer 326.

During normal commanded wire feed, the signal on the line 322 overcomes the bias on the amplifier 323 and no operating output is supplied through the differentiator 325 to the timer 326. As programmed downslope of the commanded wire feed rate occurs while approaching the end of a weld cycle, however, the voltage on the line 322 decreases to a point where the amplifier 323 commences providing an output signal which is differentiated at 325 and activates the timing circuit 326 to provide an output signal on the line 328 for a time determined by the setting of potentiometer 327. The signal on the line 328, as amplified at 329 and passed through diode 330 to the filler wire servoamplifier, has opposite polarity to the wire heating signal supplied through the diode 320.The filler wire servomotor responds to the signal from the servoamplifier by reversing direction, for a time determined by potentiometer 327, to retract the filler wire from the weld puddle. In that manner, the filler wire is withdrawn from the weld puddle before forming a ball or becoming entrapped in the solidified puddle.

The automatic voltage control 229 in relation to the motor 117 which drives a typical weld head assembly H1 for movement along a vertical path which is perpendicular to the joint J, so as to adjust the arc distance between the torch and the joint. The arc voltage command signal is received on line 340 from the appropriate digital/analog converter. Measured arc voltage may also be applied to the automatic voltage control 229 along the line 344 and compared with the commanded arc voltage to change the vertical spacing between welding torch and pipe joint J in response to a difference between the arc voltage command signal on line 340 and the measured arc voltage signal on line 344.The vertical position of welding head assembly H1 is thus automatically adjusted as necessary to maintain a commanded arc voltage as that welding head assembly is traversed around the pipe joint J, so that predetermined arc voltage for the particular welding head assembly is maintained notwithstanding slight eccentricity between the pipe axis and the particular clamp assembly CA. Of course, the commanded arc voltage input on line 340 can be changed for various circumferential positions of each welding head assembly, and/or for particular welding passes, independently of any other welding head assembly, thereby producing a corresponding change in the vertical position of the welding head assembly to provide the commanded arc voltage.

Where a particular welding station such as stations A or B herein performs separate kinds of weld passes in succession, such as root pass followed by hot pass, the predetermined weld parameters for each kind of weld pass are programmed as aforementioned and then those weld parameters for a specific weld pass are selected to control that weld pass. Weld pass selection, which determines the set of preprogrammed weld parameters to be used, may be operator-controlled or semiautomatic (for example, with a counter which advances to select the programmed weld parameters for the next weld pass). It follows that a single welding station can perform each weld pass for a complete welded joint, by appropriate programming for the sets of weld parameters for each weld pass, so that a multiple-station pipe laying facility can remain in operation although one of the welding stations may be inoperative.

It should be understood that the disclosed use of the present invention to weld a vertical-plane joint about horizontal pipe is by way of example only, since the present invention is readily adaptable for automatic welding along any path in response to preprogrammed weld parameters chosen for the weld path.

It will be understood that the foregoing relates only to a preferred embodiment of the present invention, and that numerous changes and modifications may be made therein without departing from the spirit and the scope of the invention as set forth in the following claims.

Claims (10)

1. Apparatus for welding a joint on a workpiece, said apparatus comprising guide means defining a fixed path of movement in proximately parallel relation to a joint to be welded, and means for carrying at least one welding torch along a fixed path which lies adjacent to said joint, and characterized by position sensing means responsive to the location of said welding torch on said fixed path to provide location signals which denote said torch location, signal storage means operative to receive and retain selectably variable parameter command signals that correspond to desired values of a selected welding parameter at each torch location which said position sensing means senses along said fixed path, and further operative in response to said location signals from said position sensing means to provide the previously retained parameter command signals which correspond to said desired value of the parameter for the sensed location of said torch, and means responsive to said parameter command signal to adjust said welding parameter so that said parameter responsive signal maintains a predetermined relation to said parameter command signal.

2. Apparatus as in Claim 1, wherein said welding torch includes a filler wire feed means and means operative to supply a selectably variable amount of preheating current to said filler wire, and wherein said apparatus is further characterized in that said selected welding parameter comprises said preheating current, and said means responsive to said parameter command signal comprises power supply means operative to supply a selectably variable heating current to said filler wire in response to signals from said signal storage means, so that the amount of preheating of said filler wire for each location of said torch along said path is determined by said stored parameter command signals for each such location.

3. Apparatus as in Claim 2, further characterized in that the rate at which said preheated filler metal is supplied to said welding torch comprises a separate welding parameter, and comprising filler wire feed means operative in response to signals from said signal storage means to adjust the selected rate at which filler wire is supplied, so that the rate of filler wire feed is predetermined by said stored parameter command signals for each location of said torch along said path.

4. Apparatus as in Claim 3, further characterized by means responsive to selected heating current and selected feed rate of said filler wire, said means being operatively associated with said filler wire feed means and said power supply means to maintain a selected deposition rate of filler wire by adjusting said heating current in response to a change in said selected feed rate of filler wire, and by adjusting said feed rate in response to a change in said selected heating current.

5. The process of welding a joint along a path while selectably controlling the size of the weld puddle in the joint, comprising the steps of establishing an electric arc with the material being welded so as to create a weld puddle of molten metal at a location on the joint, moving the arc to traverse the path being welded, and supplying preheated filler metal to said weld puddle while moving said arc; and characterized by controlling the preheating of said filler metal in response to a selectably variable function of the location of said puddle along said path, so that the size of said weld puddle does not exceed a size previously determined to be appropriate for said puddle location along said path.

6. The process as in Claim 5, further characterized in that the rate at which said preheated filler metal is supplied to said weld puddle is also controlled in response to a selectably variable function of said puddle position, so that the size of said weld puddle does not exceed said appropriate size.

7. The process as in Claim 6, further characterized in that said filler metal rate of supply is additionally variable in response to a change in the selected preheating of said filler metal, and said preheating is additionally variable in response to a change in the selected rate of supply, so as to maintain a selected deposition rate of said filler metal.

8. The process as in Claim 5, wherein at least part of said path is displaced from a horizontal plane, and further characterized in that said preheating of said filler metal is controlled in response to location on said path so that the size of said weld puddle does not exceed a size whereat natural forces acting on said puddle retain the puddle in the joint in opposition to gravitational attraction tending to pull the puddle out of the joint at nonhorizontal locations along said path.

9. The process as in Claim 8, said path is circumferential to a joint of a pipe to be welded, and further characterized in that said preheating is controlled so that the size of said weld puddle does not permit gravitational dropout of said puddle while at nonupright locations long said path.

10. The process as in Claim 5, further characterized in that said electric arc is a nonpulsed electric arc.