Paul Hilton, Martin Ogle, David Taylor*, TWI Ltd. * now at JCB;
Paul Lurie and Roger D Howard, BP Exploration Operating Company Ltd
Paper presented at ICALEO 2003 Conference, 13-16 October 2003, Jacksonville, Florida, USA
In oil and gas exploration, there are advantages to be gained from exploiting new pockets of hydrocarbons by branching laterally from existing casings, in addition to drilling new wells. These so-called multilateral wells require a junction to be constructed and sealed. Thus there is an interest in examining the feasibility of using a fibre delivered focussed laser beam for cutting and welding steel pipe materials in the branching operation. In production, any such process would be required to function in an environment which comprises high density mud, a temperature possibly in excess of 100°C and pressures up to 500bar. This paper reports on a series of trials to examine the use of fibre delivered Nd:YAG laser power for welding in a high pressure gaseous environment, to simulate the downhole conditions. The work involved the design and manufacture of a pressure chamber equipped with access, via a sapphire window, for the laser beam. To simulate the welding process, material samples in the form of small diameter discs, were rotated inside a chamber under the focussed laser beam, to perform a series of melt runs in high pressure gaseous atmospheres of helium and argon. The results of these trials are presented, which include what is believed to be the first laser melt run produced under a pressure of 500bar.
In the so-called Upstream or Exploration sector of the Oil and Gas Industry, a major effort is extended in finding new large hydrocarbon reserves, proving the size and productivity of the field by drilling a number of exploration wells, and then, if judged to be economic, developing the infrastructure to produce the hydrocarbons to sell onto the international market. In the drilling and completions segment, large sums are spent on drilling holes from the surface into the oil or gas strata of the reservoirs, running individual lengths of casing, with threaded connections, into the well, to support the well during production and then sealing the annulus between the casing and the wall of the hole with cement.  Individual wells may cost from US$1-50 million depending on location, i.e. offshore, deep or shallow water, or onshore, and specific difficulties associated with local geology. Profitability in the oil and gas sector clearly depends on maximising the productivity of oil and gas wells.
Initial reservoir pressure usually ensures that the well productivity is high for the first few years, but thereafter the flow rates decline as the well pressure reduces. The flow rates may also decline for many other reasons, such as, water ingress or more rapid pressure reductions due to reservoir complexity. When the well production becomes un-economic, a new well may be drilled from the bottom of the existing well (this is often called a side-track well) into another part of the reservoir where oil or gas remain. In the last few years, it has become more common to drill a number of wells 'laterally' from the mainbore, and these new wells are termed multi-lateral wells.  These multilateral wells are drilled because they are often cheaper than drilling a completely new wellbore, and the productivity of these wells may be higher than a single well.
One of the critical components of the multilateral well is the junction. It must be created in such a way that there is some mechanism of joining the steel liner in new lateral wells back to the steel casing of the mainbore. This can be done, for new wells, either by building a special section of casing with a pre-installed exit window mechanism, or by designing a special joint that can be slightly collapsed while running it into the new wellbore and subsequently expanded to create the new lateral connection. Both of these approaches are used,  but they cannot be utilized for the many hundreds of existing wells. For these wells a number of mechanical systems have been designed which give varying degrees of seal. 
The ideal situation would be to create the window in the casing of an existing well with relatively simple cutting technology, drill the new lateral well, run the steel liner into the lateral, and then simply weld up the connections between the new liner and the existing well casing. However, the ambient conditions at depth in the well bore i.e. hydrostatic pressures of, say, 350bar (~5,000psi) together with hydrocarbon fluids, possibly mixed with water, are not conducive to welding. The problem is compounded by the need to weld remotely at distances of perhaps 15,000ft (~4,500m) from the energy source and the operator.
The work reported here, is part of a programme undertaken to investigate the possibility of utilizing fibre delivered Nd:YAG laser light, to perform the operations of welding and cutting mentioned above. It was envisaged that cutting and welding would take place in a section of the well bore which had been temporarily isolated by the creation of a pocket of suitable welding gas. The programme also investigated aspects of 'remote' operation of the laser source, and delivery of laser power, via optical fibre over the large distances involved. The part of the work described in this paper, involved the design and manufacture of a pressure chamber, which could provide access for an Nd:YAGlaser beam via a suitable pressure window, with the objectives of performing welding trials in gaseous atmospheres at pressures up to 520bar.
Little previous work on laser welding under pressure has been reported. Shannon et al  have used CO 2 lasers to perform hyperbaric welding on steel, and Durchame et al  have investigated hyperbaric laser welding using a point and line source model. High pressure arc welding studies have been undertaken by Richardson. 
2. Equipment and experimental procedures
The prime requirements for the high pressure test facility were as follows:
- The welding environment would be an inert gas (helium or argon at working pressures from -1 (vacuum) to +517bar(g) (7500 psi), at ambient temperature.
- The laser focussing head, would be mounted outside the pressure chamber and the beam admitted to the work piece via an anti-reflection coated sapphire window.
- The test sample would be rotated to give an effective welding speed between 4 and 14mm/sec and large enough to produce three welds of 35mm minimum length each.
- The test sample could be immersed in a liquid environment (mud, water, etc).
- The chamber would need to be quickly opened and closed to ensure rapid exchange of test samples.
The general construction of the pressure chamber, which was designed and manufactured by TWI, is shown schematically in Figure 1. A photograph of the actual chamber can be seen in Figure 2. The high pressure gas was contained by the stainless steel barrel. The ends of the pressure vessel were closed by a stainless steel lid at the top and bed at the bottom. The whole assembly was held closed by a 2500kNhydraulic jack equipped with a locking ring. The barrel of the chamber was sealed at both ends with 'O' rings. In order to facilitate rapid opening of the chamber and access to samples, the barrel was mounted in a sprung loaded plate to lift it out of the bed recess. The barrel, including the test sample and its drive mechanism, could then be slid clear of the bed for easy access.
Fig. 1. Schematic arrangement of the high pressure welding chamber
Fig. 2. Photograph of the high pressure welding chamber
The test sample consisted of an 80mm diameter, 8mm thick steel disc screwed to a stainless steel turntable. The turntable was mounted on a ball race and the complete assembly could be lifted out for sample changeover. A friction drive from an electric motor inside the cell rotated the sample. Closure of the chamber engaged an 8-way high pressure electrical plug and socket providing power to the sample drive motor. The anti-reflection coated sapphire window, which was 50mm diameter and 20mm thick, was held in an aluminium housing and sealed with 'O' rings, to the lid of the chamber. Bottled argon and helium were used as the chamber atmospheres. For work above about 150bar, a gas compressor was brought into operation. The internal volume of the chamber was less than 10lt. To insert a sample, close the chamber and pressurize to 500bar took approximately 35 minutes. A beta-site Multiwave Auto TM Nd:YAG laser, originally manufactured by Lumonics Ltd, was used. A step index fibre optic of core diameter 0.6mm and length 30m, transmitted the beam to the output housing, which collimated the laser light and focused it to a 0.9mm diameter spot at a distance of about 300mm from the end of the housing. The head was mounted in the arm of a robot, above the pressure cell, so that it could be easily moved up and down. The converging beam from the output housing was then transmitted through the sapphire window into the chamber of the pressure cell. The 300mm stand off distance was chosen to minimise the possibility of damage to the sapphire window during processing, by ensuring it was a significant distance from the workpiece, and therefore avoiding excessive heat, spatter, etc. A disposable glass coverslide was positioned below the sapphire window to provide further protection for the inside surface of the window. The maximum laser power available at the sample, after transmission through the window and coverslide, was 2.6kW.
During laser processing it was known that two factors could influence the effective position of the laser beam focal plane. The first was the insertion of the sapphire window and the coverslide into the converging laser beam and the second was due to the effect of the refractive index change in the gas used, as a function of pressure. As a result, the effective focal length of the system was increased by 7mm to compensate for the windows and a further 5mm to compensate for the high pressure helium environment (at 500bar). For argon (at 500bar) the effective focal length of the system was increased by a further 41mm. It was calculated that any increase in spherical aberration of the focused spot due to the worst case of the windows plus 500bar of argon was negligible. Depending on the pressure being used, therefore, the focal plane of the beam was positioned on the surface of the sample for all the trials. All the samples were laser cut discs of 8mm thick BS EN 10025: 1993 S355K2G3 structural CMn steel. In practice, two melt runs on each disc were made. The laser power was kept constant at 2.6kW on the workpiece (measured only at atmospheric pressure),the only other variables in the experiment being gas pressure and sample rotational speed.
3.1 Melt runs produced in a helium environment at elevated pressures
Melt run trials using helium gas were carried out at a range of pressures: 50, 100, 175, 250, 350 and 500bar. All trials were made at a nominal speed of 0.6m/min and with the focus position at the workpiece surface, adjusted for the effect of ambient pressure. The general trend in the results obtained was for an increase in the width of the top bead produced from 2.5mm, at 50bar pressure, to 3.4mm at 500bar pressure. The graph of achieved penetration against pressure, Figure 3, possibly shows a trend of slightly increased penetration as pressure increases. Figures 4 and 5 show macrographs of the melt runs produced at 50 and 500bar respectively. Visual inspection of the top bead and melt penetration did not show any variation over the length of the melt run. The top bead for all trials appeared smooth, even and shiny, as can be seen from Figure 6, which is a melt run at a pressure of 500bar. In addition, there was a visual reduction in heat generated discolouration on the material surface as pressure increased. On opening the chamber for all the trials carried out at pressures greater than or equal to 50bar, there was little or none of the orange/brown 'sooty' deposit, normally seen when Nd:YAG welding CMn steel.
Fig. 3. Graph of penetration against pressure for melt runs in a helium environment
Fig. 4. Transverse section of a melt run produced in helium at a pressure of 50bar at 2.6kW with a welding speed of 0.6m/min
Fig. 5. Transverse section of a melt run produced in helium at a pressure of 500bar at 2.6kW with a welding speed of 0.6m/min
Fig. 6. Melt run produced in helium environment at a pressure of 500bar at 2.6kW with a welding speed of 0.6m/min
3.2 Melt runs produced in an argon environment at elevated pressure
Melt run trials with argon gas were carried out at pressures of 100, 350 and 500bar. All trials were carried out at a nominal speed of 0.6m/min and with the focus position at the workpiece surface, which was adjusted for the effect of ambient pressure. The general trend observed for argon was for a greater increase in the width of the top bead as pressure increased. The graph of penetration against pressure in argon, Figure 7, shows reduced penetration as pressure increases, from 1.4mm at 100bar to 0.6mm at 500bar. The top bead at 100bar was shiny and fairly smooth and even. At 500bar, the melt run remained shiny but was now very uneven, as can be seen in Figure 8. On opening the chamber for all the trials, once again there was little or no evidence of the orange/brown welding 'soot'.
Fig. 7. Graph of penetration against pressure for melt runs in an argon environment
Fig. 8. Melt run produced in an argon environment at a pressure of 500bar at 2.6kW with a welding speed of 0.6m/min
3.3 Melt runs produced in a helium environment at elevated pressure in the presence of contaminants
A nominal pressure of 175bar helium was selected for trials involving the presence of an oil-based mud inside the chamber. On visual inspection of the melt run produced, there was no apparent difference between the top bead in this trial and previous trials with no contaminant. The penetration achieved was 1.3mm with a top bead 3.1mm wide, compared with 1.3mm penetration and 2.7mm top bead for the trial with no oil-based mud.
As can be seen in Figures 4 and 5, all the melt runs produced at elevated pressure inside the confines of the chamber produced conduction limited type weld profiles. In the helium gas environment, the weld penetration remained approximately constant at about1.4mm, and in the nitrogen gas environment, the penetration appeared to decrease from 1.4mm to 0.6mm, over the pressure range from 50-500bar. No evidence was seen with either gas at elevated pressure, for the keyhole mode of laser welding. In addition, at pressures of 50bar and above, little or no evidence could be found inside the chamber, or on the samples, for the type of fine 'sooty' deposit generally associated with keyhole Nd:YAG laser welding of CMn steels. The latter results contrasted strongly with reference trials undertaken at atmospheric pressure in the closed chamber with atmospheres of both air and helium. All the welds made at atmospheric pressure, in the closed chamber, generated large amounts of fume, which appeared to condense rapidly as fine 'soot' on the sample, the walls of the chamber, the coverslides and the inside surface of the sapphire pressure window. With an atmosphere of air in the chamber, the initial part of the weld produced exhibited a keyhole type profile, with a penetration of over 3mm. By the end of the weld (approximately 50mm further around the circumference of the sample), penetration had dropped to~2.0mm. This reduced penetration is thought to be due to significant fume build up in the confined volume of the closed cell. (Trials in an 'open' air environment, while welding through the sapphire window and coverslide, did not show this effect). With one atmosphere of helium inside the closed chamber, significant fume generation and 'sooting' again occurred, but none of the melt runs produced showed evidence of keyhole type behaviour and weld penetration was reduced to about 1.1mm. On every occasion that a weld was made with the chamber closed, at atmospheric pressure, with either air or helium as the environment, the window protection coverslide was broken, notwithstanding the relatively short welding time. During all the work at pressures between 50 and 500bar, no coverslide damage was noted.
When welding at elevated pressure in an argon environment, the reduced penetration described above was coupled with a large increase in the width and irregularly of the weld top bead. This effect could be caused by a 'smearing' of the heat source, produced by movement of the very dense gas during welding. Argon at 500bar, has a density close to that of gasoline at atmospheric pressure.
The results indicated that at a pressure somewhere between 1 and 50bar, a mechanism takes place which possibly inhibits generation of vapour from the molten weld pool. This mechanism certainly prevents the condensation of any vapour close to the sample or inside the pressure chamber which kept the coverslides clean and intact during all high pressure experiments. It is also possible that this same mechanism prevents the formation of a keyhole, as no keyhole behaviour was seen at any pressure above 50bar. The existence or not of a keyhole in these trials is however, confused by the fact that inside the small confines of the chamber, even in an environment of one atmosphere of helium, only conduction limited welding was apparent.
It is recommended that further experiments are performed at pressures between 50 and 1bar, with higher laser powers/power densities, to investigate the transition from keyhole to conduction limited welding, at elevated pressure.
- Melt runs on CMn steel samples were made at pressures up to 500bar in an inert gas environment using 2.6kW of Nd:YAG laser power.
In addition, from an analysis of the melt runs produced, the following more detailed conclusions can be drawn.
- Little change could be seen in the penetration observed in the helium high pressure environment over the range from 50-500bar.
- The generation of melt runs in a small, confined chamber at low pressures, would appear to cause a build up of fume which can affect the transmission of the laser beam through the chamber, causing a reduction in observed penetration.
- Between 1 and 50bar pressure, in a helium (or argon) environment, a mechanism which restricts the formation of or prevents the condensation of, metal vapour inside the chamber, is operative.
The authors would like to thank BP Exploration Operating Company Ltd for permission to publish this work, which was undertaken together with Mobil, Texaco and Chevron under the MoBPTeCh charter.
- Economides M J, Watters L T, Dunn-Norman S. 'Petroleum Well Construction'. Published by J Wiley & Sons, 1998. ISBN 0471969389.
- Shannon, G J, McNaught W, Deans W F and Watson J, 1997. 'High Power Laser Welding in Hyperbaric Gas and Water Environment'. Journal of Laser Applications, 9 (3):129.
- Durchame R, Kapadia P, Lampa C. 'A Point and Line Source Analysis of the Laser Material Interaction in Hyperbaric Keyhole Laser Welding Proceedings'. ICALEO 1995.
- Richardson I M. 'The Influence of Ambient Pressure on Arc Welding Processes - A Review'. Physical Aspects of Arc Welding. Proceedings, Study group 212. Seminar in honour of J F Lancaster, Glasgow, UK, September 1993.
Meet the authors:
Paul Hilton is Technology Manager - Lasers, at TWI in the UK, where he has specific responsibility for TWI's strategic development of laser materials processing. He is also the current President Elect of the UK's Association of Industrial Laser Users.
Martin Ogle is Head of Structures and is Principal Design Consultant and Technology Manager (Design) at TWI. He is currently a member of British, European and International Standards Committees for design and construction of steel and aluminium structures. He has had over 20 years of experience in drafting British and foreign design and manufacturing standards. Major research projects have included incremental collapse of building frames, residual stresses in bridges and nuclear plant and fatigue loading for traffic on highway bridges.
Paul Lurie is a Novel Drilling Engineer with the BP Exploration and Production Technology Group in Sunbury-on-Thames, UK. He holds a PhD degree in Applied Surface Physics from Cambridge University. He had been involved in developing and implementing novel technologies for more than 30 years, initially in diamond mineral processing and diamond exploration technologies, and in novel drilling for the Oil and Gas Industry for the last 15 years.
Roger Howard graduated in metallurgy from Imperial College, London, in 1969 and worked for The Welding Institute in the metallurgical laboratories until 1974. He then moved to Clarke Chapman Power Engineering and worked on welding and metallurgical aspects of automated welding related to the fabrication of nuclear and fossil fired steam generators. In 1980 he moved to BP and is currently employed as a 'materials and welding engineer' in the Exploration and Production Technology Group at Sunbury-on-Thames.