The first quarter century - TWI's sour service test facility
TWI Bulletin, July/August 1998
Trevor Gooch graduated in industrial metallurgy at the University of Birmingham, and continued to obtain the degrees of MSc(Eng) and PhD from the University of London. Dr Gooch joined TWI in 1965. In 1980, Dr Gooch became Head of the Materials Department, embracing the welding characteristics of virtually all metallic materials. In 1994, he was made Materials Technology Manager for TWI, involved with joining characteristics of all materials used for construction.
Testing in environments containing H 2 S has been going on at TWI for nearly 30 years, and 25 years ago, a dedicated laboratory was built at Abington. Trevor Gooch traces the history of the development of this facility, and shows how the data generated and expertise developed have influenced industry and industry standards.
Some of the earliest studies undertaken by BWRA, the forerunner of TWI, were involved with the problem of stress corrosion cracking of welded materials. In the 1950s and 60s, a range of investigations was carried out, especially on the behaviour of carbon steels and high strength aluminium alloys in various media.
During this time, it became clear that, while high strength steels (yield stress above 1000MPa) could be of significant advantage in offering excellent strength/weight ratio for aerospace applications, the materials were sensitive to cracking in a range of aqueous conditions. By 1960, two schools of thought had emerged, namely that cracking was either due to the presence of a readily corrodible 'active path' within the material or that it stemmed from hydrogen embrittlement, the hydrogen being picked up by the steel following a cathodic discharge reaction.
Extensive studies were carried out at Abington, both to define the mechanism of cracking and to generate quantitative data on material behaviour and the effects of welding. It was shown conclusively that, in the general case, cracking was a consequence of hydrogen pick-up. This could occur even under conditions of nominal anodic polarisation, presumably in consequence of metal corrosion leading to hydrolysis and acidification immediately adjacent to the material surface.
Fig. 1
Relationship between SCC defect tolerance parameter and steel hardness, derived from tests on precracked samples: a) Parent materials, HAZs and weld metals, excluding precipitation hardened and twinned martensites; b) Parentmaterials and HAZs having a twinned martensite microstructure. [1]
Given that cracking was a result of hydrogen, it was not surprising that a clear correlation was found between the material hardness and cracking sensitivity as expressed in terms of a critical stress or stress intensity. Further, the precise relationship depended
[1] both on the microstructure developed and on the severity of the service environment in the sense of its propensity to induce hydrogen pick-up
(Fig.1). By this time, it had been known for many years that exposure to sour H 2 S containing environments could promote appreciable hydrogen pick-up and cause 'sulphide stress cracking' (SSC) in a range of transformable steels. The significance of hardness had been well recognised, and various limits were already being applied within the oil industry.
This work led to the significant event of the MR0175 standard being published in 1975. Because TWI had available test points for tensile and bend stress corrosion samples, investigations had been carried out for Member Companies to assess behaviour of different materials and welds, notably for pipeline applications in sour conditions. Up until about 1970, the requirement for welded joints to be tested under sour conditions was only sporadic, and the work that was necessary for TWI members was carried out on a small scale within the main laboratory, using only localised extraction to handle the H 2 S. However, this was about to change, and over the last 25 years the increasing need to evaluate weldment behaviour in sour media has led to progressive and significant expansion in the facilities available at Abington for the use of Member companies.
Sour service laboratory
In 1972, a failure occurred in the Arabian Gulf that had a major impact on oil industry thinking. The problem concerned a subsea steel pipeline in which through-thickness wall penetration occurred by a mechanism that has become variously known as 'hydrogen pressure cracking', 'step-wise cracking', 'hydrogen induced cracking' or, perhaps most appropriately, 'hydrogen pressure induced cracking' (HPIC).
Fig. 2
Hydrogen pressure cracking in linepipe steel, showing the stepped appearance: unetched.
At that time, the mechanism was not clearly established, but the oil company, BP, and pipe supplier, Sumitomo Metals, carried out the initial failure investigation and thought it likely that the unusual cracking morphology observed (Fig.2) was in fact a consequence of hydrogen pick-up by the steel. TWI was consulted at this stage, and a number of the fracture faces were found to have very similar morphologies to those induced in the previous studies at Abington on aqueous stress corrosion of welded steels. This was very strong evidence that hydrogen pick-up was responsible for the subsea failure, a view which was confirmed by hydrogen analyses on failure material.
To examine material susceptibility to the problem, BP and Sumitomo had developed a direct exposure test in which pipeline steel samples were exposed to seawater saturated with H 2 S with no application of external stress. The initial aim was essentially to sort materials and the procedure was fairly simple, but with lacquering of through-thickness faces of the sample on the basis that hydrogen entry should develop only on the as-received surface of a linepipe.
A tripartite interlaboratory study was carried out involving Sumitomo Metals, TWI and the Fulmer Research Institute in which sample cracking behaviour was examined and the specimens were analysed for hydrogen pick-up during the test exposure. Lacquering was found to give inconsistent results and was therefore dropped from the test method.
The first 'standard' sampling and test procedure was drafted by TWI, BP and Sumitomo Metals in 1973. This method was adopted by other oil companies, and eventually formed the basis of NACE TM0284.
Following the introduction of this formalised test procedure, a number of oil companies began to request routine testing of materials for pipelines, and pressure vessels, to demonstrate resistance to the problem. Paralleling this effort on HPIC, member companies increasingly required stress corrosion testing to be carried out in support of projects world wide, albeit with some emphasis on assessing susceptibility of pipeline welds to sulphide stress cracking. The demand to handle H 2 S at Abington therefore increased dramatically, and in 1973 the decision was swiftly taken to build a dedicated H 2 S laboratory on site, incorporating the available test machines and HPIC equipment.
The design of the facility fully recognised the singular hazards of toxicity and fire potentially associated with handling H 2 S. A very great deal of effort went into ensuring that a safe working environment could be maintained. Positive ventilation was installed on every test point (Fig.3) and in the general laboratory, with appropriate sensors linked to trips on H 2 S supplies should the level of H 2 S in the working atmosphere rise too high. Rigorous and continuing staff training was undertaken, and a working system established in which only qualified operators may enter the laboratory and use the equipment installed.
Following the opening of the laboratory, it became apparent that demand for the facility was still increasing, and an expansion was undertaken in 1980. TWI installed its first slow-strain rate machine in 1985, the unit being equipped with a Hastelloy autoclave (Fig.4). The usefulness of the pressurised facility was soon manifest. Accordingly, a range of autoclaves was installed to handle both direct exposure coupons and bent-beam stress corrosion specimens (Fig.5). Facilities were set up so that constant gas flow through the autoclaves could be achieved rather than working with a sealed unit, enabling necessary H 2 S and CO 2 levels to be maintained throughout the test duration. Further expansion became necessary and the laboratory was refurbished and extended in 1990 - and a further room is being added at present.
Fig.3 The laboratory in 1978, showing dead weight
load machines with constant temperature baths and
local extraction vents.
Fig.4 Slow strain rate test machine.
Fig.5 Some of the autoclaves available in the H 2 S laboratory.
Fig. 6
SSC tests on internally stressed pipe ring sections.
In consequence, TWI can now offer a wide range of H
2 S test facilities. Constant load, constant deflection and slow strain rate tests can be carried out both under normal atmospheric pressure and at high pressure, and at temperatures up to 250°C, with, if necessary pH measurement under pressure. The equipment available is flexible, and in many cases tests have been carried out on full size welded samples, both pipe rings and plate
(Fig.6). Major investigations at TWI
Many of the early data generated on step-wise cracking were incorporated in the 'standard' reference paper produced by Moore and Warga in 1976.
[2] Since that time, much of the effort on H
2 S related phenomena has involved carbon and carbon-manganese steels, but recent years have seen extensive testing also of corrosion resistant alloys (CRAs).
Critical weld area hardness levels
The emergence of NACE standard MR0175 in 1975 formalised the requirement of material hardness control to avoid SSC. Because of the hardening which inevitably takes place at welded steel joints, the question arose as to what criteria should be applied to arc welds in line-pipe, pressure vessel steels etc. The provisions of NACE MR0175 require plain carbon and allied steels to be below Rockwell 22C hardness for service under sour conditions. This limit was, however, derived from studies on materials of entirely different compositions and microstructure from those expected in the weld area of linepipe steels for example.
Fig. 7
SSC in hardened HAZ of a weld in a C-Mn steel.
Moreover it was obvious, although not always recognised, that Rockwell testing was not suitable for welded joints since the diameter of the hardness impression was of the same order of size as the HAZ for example. Hence, even though Rockwell/Vickers hardness conversions existed, there was uncertainty as to what level of Vickers hardness should be prescribed for welds in typical linepipe and pressure vessel steels to avoid SSC (Fig.7). A major group sponsored investigation was therefore undertaken. The main aim was to define a suitable maximum hardness for welded pipeline and similar steels under conditions defined as sour in MR0175, but the project sought further to examine the dependence of any critical hardness on the specific environmental conditions experienced.
Fig. 8
Relationship between maximum weld area hardness and threshold stress in NACE TM0177 solution: below 250HV10 (equivalent to Rockwell 22HRC), SSC occurs only with applied stress above base metal yield.
The study successfully showed that, under sour conditions as exemplified by the NACE TM0177 Method A solution, weld area hardness gave a good indication of SSC behaviour. In particular, it was found that the existing limit of 22HRC taken as equivalent to 250HV was applicable to welded carbon-manganese and similar steels, despite the wide range of transformed weld area microstructure that could develop in commercial heats of varying composition (Fig.8). The results also indicated that weld metals might show somewhat inferior behaviour to base steel or transformed heat affected zones: although it is reasonable to associate this with a higher inclusion content, the reason for the effect has not been fully defined and further study is needed.
The project further demonstrated the effect of varying environmental conditions, and associated propensity to form hydrogen, on cracking behaviour. Limiting hardness levels were defined for both low H 2 S levels (ca 80-100ppm) and for cathodic polarisation in seawater to highly negative potentials.
The demonstration of the applicability of the NACE 22HRC/250HV limit to welds in pipeline and structural steel has reached general acceptance. Such a limit is commonly specified in welding procedure qualification for sour service, and is incorporated in BS 4515.
External hard zones
When welding pipelines intended to carry sour products, it may be difficult to maintain all areas of the weld below 250HV, especially given the requirement for high joint completion rates and the use of rapid, low heat input welding procedures. High hardness areas are most likely on the outside of the pipe, which are not subject to tempering by further weld passes, yet these areas are furthest from the hydrogen entry at the internal surface, and nearest to a free surface at which hydrogen can be evolved.
They should, therefore be subject to lower peak hydrogen level than the internal surface, and hence should be able to tolerate higher hardness. This was demonstrated at TWI in two studies, one sponsored by a number of companies and using nominally unstressed specimens, and one funded by the American Gas Association and using internally stressed full ring specimens. Both programmes gave consistent results, with a clear indication that external hardness limits on welded linepipe could be relaxed relative to the inner surface. Again, this work has been recognised in the maximum hardness levels specified in BS 4515, and latterly in EFC Document 16.
Effect of nickel
Following various testing programmes in the 1960s, the NACE MR0175 standard stipulates a maximum Ni content of 1%Ni in carbon steels for sour service. This limit has sometimes rendered it difficult to obtain required low temperature toughness levels, especially for weld metal. Accordingly, the necessity for such a compositional restriction was examined, using manual metal arc electrodes with a common flux formulation so that any effect of Ni could be isolated. In the event, no significant influence of some 2%Ni on SSC behaviour was found. This work was noted by the European Federation of Corrosion, and EFC Document 16 specifically permits Ni-containing materials to be used for sour service applications necessitating good toughness.
Low H 2 S levels
While the general problem of progressive well souring following water injection for secondary recovery has been long recognised, it has become increasingly important over the last 10 or 15 years. This is certainly the case in the North Sea where many originally sweet formations can be expected to develop appreciable H
2 S levels in the product as time goes by. Accordingly a further group sponsored project has been in progress at Abington studying particularly hardness limits to avoid sulphide stress cracking at low H
2 S levels, recognising also the effects of pH. The aim is to provide data that can be used to evaluate whether or not existing plant designed for sweet operation can be safely used as souring develops. Most existing data on the effects of varying H
2 S and pH have been derived from tests undertaken at 1bar total pressure in view of the experimental convenience. Trials at Abington indicated that sulph
Corrosion resistant alloys
The introduction of duplex stainless steel to avoid CO
2 attack in linepipes in the early 1980s represented a major shift in material type used in much of the oil and gas sector. It was soon recognised that these corrosion resistant alloys (CRAs) could suffer SSC, especially when work hardened and employed in downhole applications. A major review was carried out, sponsored by over 20 companies, to examine published literature and field experience on CRAs in sour conditions. Particular attention was paid to defining environmental limits for different types of material, recognising H
2 S, pH, chloride and temperature levels.
Fig. 9
SSCC at a weld in duplex stainless steel.
Testing at TWI for Member companies has been carried out on different CRAs in a range of media, primarily to assess the suitability of individual alloys for specific applications (Fig.9).
A Group Sponsored Project has examined the effects of welding on duplex steels under sour conditions, with particular reference to the influence of ferrite/austenite balance, of local hardening, and of intermetallic phases.
At present, there is a high level of interest in martensitic stainless steels, developed from traditional chromium alloys by reduction in carbon and increase in alloying elements such as nickel and molybdenum to obtain improved properties. In principle, they can offer economic alternative to duplex stainless steels for many applications. It is vital that the behaviour of welded joints in sour media be defined, and such environmental testing is presently in hand with this aim in a group sponsored project on the weldability of these new alloys.
Work on CRAs has inevitably required testing under elevated pressure and temperature using the range of autoclave facilities installed in the H 2 S laboratory. Most testing has used constant deflection samples, but increasing use is being made of the slow strain rate test points available, since the approach can give a rapid indication of possible material cracking sensitivity.
Present position
Over the years, TWI has acquired suitable expertise in the testing of ferrous and non-ferrous materials in sour media. Facilities have been continually upgraded and extended, and a wide range of test points is now available to meet the needs of Member companies. Constant load and constant deflection testing can be undertaken under both ambient and high temperature/high pressure conditions, and work has been carried out on the effect of fluctuating stress during test.
TWI staff members are active on Committees and Work Groups, especially of NACE and the European Federation Corrosion, and the total effort expended on problems associated with sour service represents a significant proportion of the Materials department activity. Certainly, the overall testing capability has increased considerably over the last 25 years; the requirements to optimise alloy selection and welding procedures for more aggressive situations, and at the same time to define material behaviour more precisely and economically, mean that this trend can be expected to continue.
Acknowledgements
The author thanks colleagues at TWI for assistance and advice over many years. Appreciation is given also to TWI Research Member companies who have supported the sour service work undertaken.
References
| N° | Author | Title | |
| 1 | Gooch T G: | 'Hardness and stress corrosion cracking of ferritic steel'. The Research Bulletin 1982 23 (8) 241-246. | |
| 2 | Moore E M and Warga J J: | 'Factors influencing the hydrogen cracking sensitivity of pipeline steels'. Materials Performance 1976 15 (6) 17-23. | Return to text |
