Orbital Welding of Super Duplex Header
Orbital Welding of Super Duplex Header
The completed EMBLA project test header shown at Cameron’s facility in Scotland. Twenty-two orbital butt welds were required to complete the manifold and two small connections were done by hand.
Cameron Forged Products Division of Cooper (Great Britain) Ltd, of Livingston (West Lothian, near Edinburgh, Scotland) has recently acquired a state-of-the-art orbital pipe welding system consisting of a power supply and two full-function orbital welding heads manufactured by Arc Machines, Inc. of Pacoima (California, U.S.A.). As a first project for this new venture, they undertook the fabrication of a specially forged super duplex test manifold for Phillips Petroleum to be used for a high-pressure application on an offshore platform in an oil field of the Norwegian sector of the North Sea.
TEST MANIFOLD. The oil/gas output from the oil well is diverted through one of two manifolds. The test manifold tests the flow directly from the well and thus receives the full unregulated pressure, whereas flow to the larger-diameter production manifold has a throttle to regulate the flow through the valve and is not subjected to the same high pressure. The test manifold has a relatively small diameter (8" or 203.2mm) and is designed to have a heavy wall thickness (31mm) in order to withstand the extreme pressures. The product flowing through the test manifold, the high operating temperatures and pressure, in addition to the seawater and moist salt air in the operating environment provide an exceptionally severe service environment. Because of the harshness of this type of environment, there is an increasing demand for the use of more corrosion-resistant materials in offshore applications.
MATERIAL. The suitability and comparatively low costs of duplex and, more recently, super duplex stainless steels for petroleum industry applications such as this, have been recognized and these materials have been specified with increasing frequency in recent years. One of the new generation super duplex alloys, UNS S32760, was selected for the EMBLA project test manifold to be constructed by Cameron. According to Cameron metallurgist Jim Bogie, the high corrosion-resistance and favourable mechanical properties of duplex stainless steels are achieved through a combination of alloying chemistry and careful control of temperature, cleanliness and other factors prior to and during forging. Careful control of these parameters results in a material with a balanced structure of austenite and ferrite in the annealed condition. Welding of the standard duplex material (UNS S31803) often produces a phase imbalance with a resulting loss of corrosion resistance unless annealing is done to restore the ferrite/austenite balance. It is impractical to anneal large weldments such as a manifold, and annealing is not feasible for most field applications.
The second-generation duplexes have made it possible to achieve a favourable phase ratio after welding without having to anneal by increasing the chromium (Cr) content compared to standard grades and alloying with nitrogen (N) which restores the austenite/ferrite phase balance more rapidly after welding and minimized chromium and molybdenum segregation (table 1). Nitrogen has a very beneficial effect on corrosion-resistance as indicated by its increase of the pitting resistance equivalent (PREn = %Cr + 3.3% Mo + 16% N) (ref. 1 and 2). The PREn is used as an index of a material's resistance to corrosion, with higher numbers indicative of greater resistance to pitting corrosion. One criterion for a super duplex is a duplex material with a PREn greater than 40. The PREn of alloy UNS S32760 is 43.3 according to this formula. There is some indication that tungsten should be included in the PRE formula, and the addition of tungsten to alloy UNS S32760 is intended to improve the resistance to localized corrosion (ref. 3). Alloy UNS S32760 also contains trace amounts of copper (Cu). Cu has the beneficial effect of reducing the amount of diffusible hydrogen and therefore limiting the amount of hydrogen-induced cracking (HIC) that occurs in sour service environments (ref. 4).
Figure 1. Forged tees positioned on a V block in preparation for welding. Arc Machines, Inc. Model 15 weld head is shown mounted on an oversized guide ring to accommodate slight modification of the usual torch position.
Table 1. Chemical composition of duplex stainless steels W%.
Constitution diagrams such as the Schaeffler and DeLong diagrams are commonly used to predict the ferrite/austenite ratio of welded stainless based on the chemical composition of the base material (and consumables). The DeLong diagram is considered to be more accurate since it includes the effects of nitrogen. However, the phase balance after welding is influenced to a large extent by the weld parameters and conditions of welding, and chemistry alone is not a reliable indicator of weld results (ref. 5). Although the improved chemistry of the super duplex material lends itself to a favourable ferrite/austenite balance following welding, this still requires careful control of welding parameters affecting the heat input and cooling rates in order to achieve consistently good welds. Duplex weld metal solidifies as ferrite (delta ferrite) and the austenite is formed through a solid-state phase transformation during cooling. The transformation temperature range shows some variation depending on the alloy content but was reported as 1250-1150° C for a 22Cr-6Ni-3Mo-0.21N weld metal (ref. 2).
Another complicating factor in the welding of duplex materials is that they are subject to the formation of sigma phase and the formation of chromium nitrites which impart undesirable mechanical properties. Sigma, which is particularly deleterious to impact strength, forms on cooling after welding in the temperature range of 1050-650° C. Although a sufficiently slow cooling rate is needed to complete the ferritic-austenitic transformation, it is essential to cool quickly enough to prevent the formation of sigma phase. Thus it is critical to establish welding parameters that produce a favourable phase ratio without developing deleterious phases. This requires precise control of welding parameters and cooling rates to achieve consistent results. Although it is often associated with welding, sigma can also form during forging of the base material or in service if subjected to temperatures in the critical range. To arrive at welding parameters that will promote favourable metallurgical properties in the welded material, the manufacturers of the various alloys generally specify permissible heat input during welding in kJoules/mm.
DESIGN AND FABRICATION. Cameron's method of fabrication of the test manifold is unique. Cameron has been making duplex pipe for ten years and have had extensive experience in forging, so they knew what they were looking for in a forged product. Cameron constructed the test manifold from seventeen individual tees which they forged from super duplex material supplied by Acciaieria Foroni of Italy. The material is melted in an AOD (argon oxygen decarburization) furnace which provides excellent control of trace elements such as carbon and sulfur. The solid blocks of material had the hollow main arms forged by reverse extrusion with the outlet machined and drilled to form an integral hub. Cameron's specialized forging equipment allowed them to closely control the metallurgy of the tees and resulted in favourable grain structure with good grain flow which nicely follows the crotch radius of the piece. This is in contrast to the standard construction method of making the manifold from pipe and pulling the tees. Since it is difficult to pull tees on such heavy-walled material, Cameron's technology permits the use of the heavier walls (31 mm) required to resist the very high operating pressures of 621 Bar g. Cameron's manifold also has fewer welds per tee than standard designs which increases the safety factor for high-pressure applications.
CONSIDERATIONS. After consideration of all factors involved, the decision was made to use orbital welding technology for the fabrication of the test manifold. Equipment costs were factored with the need for weld quality, estimated productivity, personnel requiremeets for orbital compared to manual welding, operator training costs, and shop conditions. It is important to recognize that duplex and super duplex stainless steels present a high degree of difficulty for manual welding. The weld puddle is somewhat viscous and hard to control: it is difficult to master the technique without pushing through on the weld roots (ref. 6). Thus skilled manual welders with the capability of making successful welds on these materials are a scarce commodity. Quality control of duplex welds can be a problem since manual welds can have good appearance and pass the required bend and tensile tests, but still have an unfavourable phase balance that makes them vulnerable to corrosion.
Although a rough estimate of the ferrite content of finished welds is possible with instrumentation based on the magnetic properties of ferrite, this method is neither accurate nor reliable. The only accurate methods are destructive tests which require cutting out of the welds. The best assurance of consistently achieving desirable weld metal features is the application of closely held and monitored weld parameters and conditions, repeated with a high degree of accuracy from weld to weld. Orbital welding technology makes this practical, since once successful weld parameters have been developed with the orbital equipment the same parameters can be used over and over to get repeatable, consistently good welds.
The customer, Phillips Petroleum, specified that welders be qualified to do hand welds in the 6G position prior to being trained on the orbital equipment. Cameron selected three of their manual welders for training on the orbital equipment. Initial welding parameters were developed at the Arc Machines, Inc. European Office in Gland, Switzerland, by pipe welding specialist, Francois Exelmans. Pipe welder training takes a week or more to train welders to operate the equipment and to develop weld parameters for their application.
One limitation of the use of super duplex has been the difficulty of obtaining suitable consumables. Filler material with correct chemistry is often hard to obtain. Sandvik alloy 25.10.4-L was selected because it is slightly over-alloyed in Ni compared to the base material to counteract the effects of Ni segregation during welding (see table 1).
The weld joint end preparation requirements are more stringent for orbital than for manual welding. End preparations must be done to very tight tolerances to assure a consistent uniform joint, so selectivity must be exercised in choosing a sub-contractor for this purpose. The preparation used for orbital welding was a compound bevel with a closed root, whereas an open root is typically specified as the preparation for manual welds on this material.
ORBITAL PIPE WELDING EQUIPMENT. The Arc Machines, Inc. Model 215 Power Supply has a library for the storage of welding programs or schedules for various sizes of pipe or special application welds. Weld parameters for the control of welding current, pulse times, arc voltage control to maintain a constant arc gap during welding, travel speed, wire feed speed, torch oscillation, etc. are entered via a Model 215 Program Operator Pendant (POP) and stored in memory. Separate schedule numbers may be used for each pass of a weld and changing from one program to another is simply a matter of entering the program number and initiating the weld sequence.
While the weld is in progress, slight steering adjustments to the torch position may be made via the POP or the smaller auxiliary pendant. The Model 15 full-function pipe weld head with an 'A' type torch is mounted on a guide ring which is clamped to the pipe or manifold. The weld head travels around the guide ring to complete the weld. Because of limited axial clearance for welding to one side of the hubs on the test manifold, one of the weld heads was fitted with a bracket to reposition the torch within a narrower envelope. This head was mounted on a 14" guide ring instead of the standard 10" guide ring used for the other weld head to permit positioning of the torch (figure 1).
WELD TEST RESULTS. Weld Procedure Specifications and Procedure Qualification Records were written according to the requirements of ASME Section IX of the Boiler and Pressure Vessel Code and ANSI/ASME B31.3 Code for Pressure Piping. These documents describe the welding procedures and conditions based on the weld parameters arrived at by welding of the test coupons in Switzerland. Ronnie Crawford, a welding engineer for Cameron, submitted a finished test coupon for testing. The sample weld (figure 2) was sectioned and examined at 2x (macrosection) and at lOOOx and 200x (microsection) for determination of ferrite and austenite ratios in the weld root (figure 3a) and heat-affected zone (HAZ) (figure 3b). The ferrite counts were carried out by a computerized counting process on a 16 point grid with 30 randomly selected fields after etching in a solution of ferrofluid EMG 909. The optimum corrosion resistance of the product is obtained with base material and weld metal having 50% ferrite and 50% austenite. The specification permitted a minimum of 35% ferrite and a maximum value of 65% ferrite with the balance being austenite. For the root the ferrite count was 48.75% ferrite, while the ferrite count for the HAZ was 55.75% ferrite, which is well within the specified limits. The welded samples were further examined after etching in 10% oxalic acid followed by a swab etch with 10% potassium hydroxide. No deleterious phases could be discerned.
Figure 2. Macrosection (2x) of qualification weld.
Figure 3a. Microsection of weld root (1000x).
Figure 3b. Microsection of HAZ (200x).
An estimation of the corrosion resistance of the manifold in service is made by testing welded samples in an accelerated corrosion test under conditions simulating the service environment. Samples were tested according to ASTM G-48 practice A, to determine the critical pitting temperature (CPT) of the welded samples. The CPT is a measure of the resistance to pitting in a chloride environment. Samples are immersed in a 6.0% ferric chloride solution for 24 hours at a starting temperature of 35° C, which is the minimum CPT specified for the weld metal. If no pitting occurs, the sample is put back in the solution with a 5° C increase in temperature. The process is repeated until failure occurs. The samples passed at 35° C, they passed at 40° C, but showed signs of corrosion at 45° C with pitting on the weld face and corrosion of the HAZ.
Bend and tensile tests were done to determine the ductility and ultimate tensile strength of the weldment. Four side bends were done through 180° with acceptable results. The ultimate tensile strength of the weld, determined by pulling the welded sample on either side of the weld until it breaks, was found to range from 839.31 to 855.61 N/mm2 which far exceeded the minimal 750 N/mm2 required to pass the test. The breaks occurred in the parent material and not in the weld.
Charpy impact tests were done at -40° C. The presence of sigma phase in the weld or HAZ would result in failure to pass this test. The minimum required average energy absorption specified was 42 Joules. The average impact value obtained from the specimens comfortably met this 42 Joules requirement.
The customer specified a Vickers hardness test to determine the root hardness of the weldment, but the testing agency could find no ASTM standard for this material; values for standard duplex were the closest available against which to compare the super duplex. The numbers were higher than expected and it was thought at first that this was due to root hardening during specimen preparation, but then it was found that the numbers were expected to be higher for super duplex than for standard duplex. The NACE (National Association of Corrosion Engineers) standard calls for a maximum value of 28 for standard duplex and NACE are currently being balloted to include a value of 32 for super duplex on the Rockwell C hardness scale. The customer accepted this procedure and the test results for various locations on the parent material, weld and HAZ were well within this specification.
After the initial weld passed all the required tests, the welding operators trained by Arc Machines performed additional test welds for Performance Qualification Tests according to ASME Section IX. These welds were subjected to the same test procedures as above, and since the test results for these welds were also satisfactory, the welding personnel were then certified to perform the orbital welding of the test manifold.
PRODUCTION WELDING. The production environment is critical to the success of welding duplex materials so Cameron constructed a special clean area which was a tent-like enclosure to protect the portion of the manifold being welded. Contamination of the weld joint with metals through dust or direct contact could upset the delicate phase balance of the weldment.
Tees arrived at the Cameron weld shop after end-preparation with the ends capped to keep the joints and inside surface (ID) clean. Prior to welding, the tees must be inspected visually for surface finish, and dye penetrant tests done to test for cracks. Prepared tees were mounted on specially constructed tables equipped with stainless V blocks to protect the material and hold the tees in position in preparation for welding (see figure 1). The alignment for welding was critical since very tight specifications were in effect for dimensional distortion. The finished manifold is 56' long (17.1 metres) with an overall length tolerance of ± 6 mm. A centering line was lightly inscribed on the tee to indicate the precise alignment of the tees. Final alignment was done with a laser beam. The components were then manually tacked in place to hold the position during welding. Even with these precautions, there was a marked tendency for the tacks to pull apart during the welding process.
It took about fifteen minutes to complete a single pass of the weld compared to about an hour with manual welding. However, it was necessary to wait between passes to observe the specified 150°C interpass temperature. The welders took the interpass temperature down to 125° C to be on the safe side. The oxygen level in the ID purge was measured with an oxygen meter at 0.3% maximum between welds to prevent oxidation of the weld ID. During the time it took to reach the interpass temperature, the welder, Davy Watt, connected a second Model 15 weld head to the power supply and made a pass on another weld. By alternating weld heads they were able to achieve a duty cycle of about 80%. While they would have preferred a 95% duty cycle, they were reasonably pleased with the 80% figure. Manual welding would have required two men welding from both sides of the joint at once to balance the weld. A total of 64 passes were required to complete an orbital weld with a total arc time of about fourteen hours. The direction of the weld was the same for each pass (see figure 4).
Although the root pass is considered difficult on this material with manual welding techniques, the root pass did not present any particular problems with the orbital welds. With a closed root and a precise land preparation, smooth, even penetration was easily achieved. The welds were cosmetically very pleasing and did not require any grinding. Brushing with a stainless steel brush produced a clean weld. This is a significant advantage when working with duplex materials since grinding is a potential source of contamination.
The sequence of the work was to join three tees together as a section and then to join the sections together to complete the manifold. There was some difficulty in meeting the original schedule since the finished tees supplied by the subcontractor did not arrive in the sequence in which they were intended to be welded. Although the welding itself was within the projected time-frame, delays in these areas meant that welding had to be done in two 12-hour shifts a day, seven days a week. Even with this demanding schedule, the welders experienced much less fatigue using the orbital equipment since there was no need for tiring out-of-position welding. By choosing welders such as Davy Watt, who had a keen interest in the project and was highly motivated to succeed with the orbital equipment, Cameron was on the right track to success with their orbital welding project. Mr. Watt learned quickly and exercised the 'tender loving care' during welding that assured the success of welding duplex. Still, Cameron have a continual program of welder training in duplex, super duplex and other high-alloy materials.
Figure 4. A completed weld on the test manifold. 64 passes were required to complete the weld, which took about fourteen hours.
Welding Engineer Ronnie Crawford tried welding with 5% nitrogen added to the shield gas because it has been reported that this helps to retain the benefits of alloying with nitrogen. Since the addition of nitrogen resulted in no discernable advantage, pure argon gas was used instead. The heat input into the weld was specified as 1.5 kJoules/mm as a maximum and 0.5 kJoules/mm as minimum. Mr. Crawford claims that with the orbital power supply they could accurately control the heat input into the weld in kJoules/mm to several decimal points. The finished welds were inspected visually, with dye penetrants, and by radiography. No defects were found in any of the production welds. They are cosmetically better than manual, with a very nice appearance. The manifold has 22 butt welds altogether and in addition two small connections were done by hand. The equipment was flexible enough to be adapted to allow the welding of hubs directly to manifold branches using orbital techniques. This was done by placing the second hub on top of the one being welded in order to mount the head.
FINISHED PRODUCT. An unforeseen advantage of the orbital welding process was the ability to predict with accuracy the precise amount of shrinkage that occurred during the welding of each tee. Super duplex has a very high rate of shrinkage which presents a potential problem with manual welding since, with that process, the amount of shrinkage is unpredictable. Shrinkage and distortion were held to a minimum with the orbital welding process. The first two tees, welded to a pup piece were only 0.25mm out of a built-in allowance of 6mm for shrinkage. When the customer came to Cameron to inspect the progress of the manifold construction they were amazed by the quality of the welds and the overall dimensional precision of the nearly completed manifold. The customer was very pleased with the results.
Production Supervisor Ernie Ross was delighted with Cameron's involvement with orbital welding and is convinced that it is a real breakthrough for them. They are seeing an increase in the use of so-called exotic materials including duplex and super duplex, and feel that orbital welding will allow them to take full advantage of the favourable properties of these materials. Cameron now has the ability to make consistent high-quality welds with uniformly good metallurgical properties that will retain the inherent corrosion resistance and mechanical properties of the base material under harsh environmental conditions. They expect that the use of these materials in offshore and subsea applications will extend the pipe service life from 5-10 years currently seen with carbon steel systems, to 25 years before requiring maintenance. This should effectively reduce the cost of expensive hyperbaric welding and subsea repairs in general. Cameron's Product Manager, Douglas Armet, foresees that their recently acquired orbital welding capabilities will provide new business opportunities in offshore manifold and piping fabrication.
By Barbara K. Henon, Ph.D., Arc Machines, Inc.
1. R. Dohnke, C. Gillessen, T. Ladwein and U. Reichel: 'Nitrogen makes duplex better', Stainless Steel Europe, pp. 22-27, Dec. 1991.
2. T. Ogowa and T. Koseki: 'Effect of composition profiles on metallurgy and corrosion behaviour of duplex stainless steel weld metals', Welding J., Res. Suppl. 181-s - 191-s, May 1989.
3. D. Fruytier: 'Industrial experiences with duplex stainless steels', Stainless Steel Europe, pp. 28-39, Dec. 1991.
4. R.D. Kane, J.P. Ribble and M.J. Schofield: 'What's behind the corrosion of micro-alloy steel weldments?', Welding J., 56-64, May 1991.
5. D.V. Smith: 'A practical approach to ferrite in stainless steel weld metal', Practical welder, Welding J., 57-60, May 1989.
6. Mike Hayes, Senior Metallurgist with Acute Technologies, Inc., Houston, Texas, U.S.A. Personal communication on his experiences with orbital welding of super duplex stainless steel pipe for a critical petroleum piping system application.
ABOUT THE AUTHOR
Barbara K. Henon has a Ph.D. and works with Arc Machines, Inc. in Pacoima, California, U.S.A.