Orbital Welding Solves Critical Metallurgical Problems in Furnace Shutdown
Orbital Welding Solves Critical Metallurgical Problems in Furnace Shutdown
Very little clearance was available for the operation of the GTA welding machine.
Gas tungsten arc welding has often been moved far afield to solve critical metallurgical problems. A good example is how Dow Chemical Canada, Inc., Western Canada division, in Fort Saskatchewan, Canada, near Edmonton, Alberta, made excellent use of automatic orbital GTA welding equipment during its June 1991 furnace shutdown. According to Dow maintenance personnel, the June shutdown was the largest major furnace shutdown since the plant was built more than 11 years ago. The shutdown was planned in order to replace the furnace tubes in its four VCM (vinyl chloride monomer) furnaces.
The VCM cracking furnaces at Dow are used to heat liquid ethylene dichloride (EDC) inside the coils to a temperature of 500 degrees C (932 degrees F) until the liquid changes to a vapor producing a vinyl chloride monomer. The intense heating and the hydrochloric acid (HCI) creates a very severe service environment subjecting the metal coils to extreme thermal as well as chemical stresses.
Reasons for Using Orbital Welding
Al Schell, supervisor of Dow maintenance services, along with Gilles Benoit, the company's welding specialist, elected to use orbital pipe welding equipment because of the very tight space restrictions for welding the upper convection coils. Previous manual welds on these coils had to be done using mirrors in order for the welder to see the weld joint, and this situation often resulted in welds having to be redone in order to meet Dow's stringent quality control requirements.
Fig. 1 - Welding operators at Dow Chemical were trained by Frank York (in dark glasses) to run the new GTA equipment.
Fig. 2 - Close-up of the inside of a weld made by the GTA orbital welding equipment during training and qualification, in April 1991. The weld is fully penetrated, even and smooth, and is free from concavity or excessive penetration.
The success of this job depended upon cooperation and planning among Dow's maintenance services, purchasing and engineering, as well as between Dow and its suppliers. A Model 215 microprocessor-controlled pipe welding power supply and a water-cooled Model 81 small diameter pipe weld head were purchased from Arc Machines, Inc. Training on the operation of this equipment was provided to Dow welding personnel long in advance of the planned shutdown. Frank York (Fig. 1), pipe welding product manager and welding specialist for Arc Machines, trained Benoit, Norm Potvin and three orbital weld welding operators for seven days in April. This was sufficient time for the operators to become comfortable in the operation of the equipment and for the development and testing of welding procedures.
Solving One Problem
In the course of developing the welding parameters, a problem was encountered initially with a slight concavity of the root on the weld inside diameter (ID) at the 6 o'clock position. This difficulty was overcome by making a split pass starting at 5 o'clock and welding clockwise to the 12 o'clock position, and then starting at 7 o'clock and welding counterclockwise to the 12 o'clock position. This meant that wire would feed into the weld pool from the leading edge of the pool in one direction and from the trailing edge in the other direction. The resulting weld had a smooth, uniform inner surface with no concavity on the ID and a desirable weld profile.
Because of the severe service environment, which, in addition to the stresses mentioned above, would include very high thermal stresses brought about by steam snuffing of the system in case of an emergency situation, the welds had to meet the requirements of ANSI/ASME B31.3 code for severe cyclic chemical piping as well as ASME Section IX of the Boiler and Pressure Vessel Code for the weld procedure used. Guided bend, tensile tests and sections of the welds were done by Hanson Materials Engineering, a local testing facility. The weld procedures and the welding operators were certified at this time and the results of all tests registered with the Alberta Boiler Branch which monitors the codes for Alberta Province. The metallurgical engineers were very impressed by the uniformity of the welds. Although 100% radiography was not required for these welds, Larry Young, Dow inspection services, elected to work to even tighter than required tolerances in the interest of safety and quality. No linear indications were permitted and 100% radiography was employed to attain this goal.
During the June shutdown, 60 or nearly half of the out-convection furnace tubes were replaced making a total of 138 butt joints. Included in this total were 120 return bend-to-tube welds (Fig. 3) and tube-to-tube welds required to tie into the existing system and welding of a pup piece to one end of the tubing and a large flange to another end. The tubes were 3.5-in. (89-mm) OD Inconel(1) 600 with a wall thickness of 0.320 in. (8 mm). Inconel 600 consists of 76%Ni, 15.5%Cr, and 8%Fe. The weld joint was a 45 deg-25 deg compound bevel with 1/8-in. (3.2-mm) OD Inconel 600 insert rings, fusion-tacked in place prior to welding. Inconel 600 wire (ERNICR-3), 0.035 in. (0.9 mm) OD, was used as the filler metal. An argon purge was provided to the inside diameter of the weld joint prior to and during the tacking. This step is an important part of the weld procedure, since the arc tends not to consume an oxidized tack and this could result in incomplete fusion at this point on the finished weld. The shielding gas was a mixture of 95% argon and 5% hydrogen. The hydrogen was added to produce a hotter weld in order to effectively consume the insert ring. The resulting weld was also cleaner appearing than a similar weld using pure argon.
Fig. 3 - Close-up of the weld of a return bend to upper convection tube was made by the new GTA weld head. The weld bead pattern on the cap pass shows the effects of torch oscillation.
The initial welds of the return bends to the straight sections of the upper convection coils were done in the shop, and installed or "speared" into place as a unit. Then the other end of the return bends were done in the field. A tent-like enclosure (hoarding) was built to enclose the outdoor area where the field welds were done to protect against wind and cold. As an extra precaution against the cool weather, the tubing was preheated with an acetylene torch prior to welding. An interpass temperature of 4000° C (7264° F) was observed. The radial clearance for the field welds was a maximum of 2-1/2 in. (63.5 mm) with some clearances under 2 in. (51 mm). The Model 81 weld head has a nominal radial clearance of 1-3/4 in. (44.5 mm) and in some cases the tubing was spread slightly in order to facilitate mounting of the weld head. The head fit snugly but neatly between the tubes to do easily what would have been extremely difficult welds to do manually (Fig. 4).
Purging of the inside of the coils was done as a unit with argon gas entering from the bottom. An oxygen meter reading of 0% was used as a guideline to indicate when the purge was adequate before welding on additional segments. The weld joint design with the insert rings provided a good fitup which made it easier to get a good ID purge. After the initial purge was established, it took about 6 min. with a flow rate of 40 ft3/hr (1.12 m/h) to purge the next section of two tubes connected at one end by the return bend for a total length of 56 ft (171 m). During the actual weld, the flow rate was turned down to 2 to 3 ft3/h (0.06 to 0.08 m/h), thus eliminating the possibility of pressurization of the weld ID which could result in a concave inner weld bead, or even a blow-out of the liquid weld pool.
Hoisted into Place
The Model 215 power supply had been hoisted to the top of the furnace platform above the location of the upper convection coils. The weld head cables controlling motor drive, wire feed, arc voltage control (AVC), torch oscillation and shielding gas were draped down from above. The field return bend welds were done in five passes with each pass using a single level. This procedure was developed to be very forgiving in the field. Similar welds done in the shop were done with two levels or current changes for each pass. Tube-to-tube welds were done in four passes.
Fig. 4 - Very little clearance was available for the operation of the GTA welding machine. Here the weld head is mounted on tubing of the upper convection coils.
Fig. 5 - Norm Potvin of Dow Chemical employs the auxiliary operator pendant to make minor steering adjustments during welding on the upper convection coils.
Each type of weld had a different program, and the weld programs or schedules specifying weld parameters for welding current, travel speed, time, pulse times, wire feed rate, AVC and oscillation, were entered into the power supply using the hand-held program operator pendant (POP), and stored in the power supply memory (Fig. 5). The AVC automatically maintains a constant arc gap between the tungsten electrode and the weld surface when traveling over an uneven surface. The oscillation controls the tungsten travel from side to side and is programmed as part of the weld program. Minor corrections in steering of the torch may be made by the welding operator as necessary during the weld from the POP or the smaller auxiliary pendant.
There were four tie-ins required to connect the new tubing to existing tubing. These were difficult, but critical welds to make. In these joints there was as much as l/8 or 3/32 in. (3.2 or 2.3 mm) difference in the wall thickness resulting from variable wear on the old tubing. It was necessary to machine away some of the wall of the new tubing to match that of the old tubing. These variations in wall thickness made it virtually impossible to use identical weld parameters for each of these joints. Instead, the welders sealed the tubing ends with a transparent Plexiglas purge plug, with a suitable opening for the ID purge gas to escape, and used the auxiliary operator pendant from the power supply to adjust the welding current based on their observations of penetration while the welds were in progress.
Dow's welding personnel worked 12-hr shifts and were able to complete an average of six welds per shift, with a minimum of five welds and a maximum of eight welds logged per shift. Maintenance supervisor Schell keeps meticulous records of previous repairs including extensive photos neatly arranged in an album. His records show that on previous jobs done with manual techniques, one acceptable x-ray joint not requiring repair per shift was considered good. During the course of this job, only three repairs were required. These occurred on the same day near the end of the project when gusts of wind coming from the tubesheet side of the weld caused the arc to extinguish. These were easily repaired. Radiographic inspection was done by Dow personnel, and zero linear defects were achieved within the scheduled one-month time frame. Thus, not only were Dow's quality requirements met, but the automatic welding equipment provided a bonus of a substantial increase in productivity as well. The cost savings realized in time nearly covered the cost of the equipment on this job alone.
Dow Chemical Canada, Western Canada division, has assumed a leadership role in the application of state-of-the-art orbital welding technology. For example, Dave Thomas, one of the division's engineering employees, is on a Dow committee of U.S. and Canadian personnel that meets several times a year to discuss welding problems and their solutions within the Dow plants. Technology exchange and cooperation among plants on a national and international basis is consistent with Dow's commitment to upgrade the quality, efficiency and safety of its operations.
By Barbara K. Henon, Ph.D., Arc Machines, Inc.
The photographs for this article were taken by Al Schell, maintenance supervisor, Dow Chemical Canada, Inc., Western Canada Division.