PIPE PRODUCTIONSteel Making for Pipe FabricationIn the 1950's the steel for pipe manufacture was produced by the acid open-hearth process.This was the most efficient method then available for producing moderately low carbonsteels. The molten steel was treated with manganese to remove the bulk of the sulphur andoxygen and then killed with calcium, silicon and aluminium to remove the residual oxygenand nitrogen before being poured into ingots. The ingots would then be transferred to theshaping plant to be cut into billets for seamless pipe manufacture or hot rolled into plate forwelded pipe manufacture. Steel is still made by this process but should be avoided for pipeproduction except for pipes of the lower strength grades in non-severe service.Quality pipe is fabricated from slabs of steel continuously cast from molten metal producedby the basic oxygen process (Basic Oxygen Steel), illustrated in the Figures, or by electricarc smelting. The production process must produce steel that is low in carbon and sulphur,phosphorous, oxygen, hydrogen and nitrogen (SPOHN) and additions need to be made ofmanganese and micro-alloying elements. The sequence to produce the slabs involvesprimary steel production, secondary or ladle steel production and casting. Intermediatesteps include magnetic stirring and degassing.The primary furnace to produce the primary steel is lined with a basic material, usuallydolomite. The charge of graded ore, pig iron and scrap is treated in the furnace to removeexcess carbon by blowing oxygen through the steel either using a lance inserted deep intothe molten metal or injecting the oxygen by bottom entry. The other additives are alsointroduced by lancing as this ensures a good dispersion. Sulphur and dissolved oxygen areremoved by sequential additions of manganese, silicon and aluminium.During production the steel can absorb various gases. These reduce the fatigue propertiesof the steel because the small voids in the steel increase localised stresses and maycoalesce under cyclic loading to produce micro-voids. For pipelines the steel must be fully"killed". Oxygen and nitrogen dissolved in the molten steel may be removed by purging withargon, which also stirs the steel and aids in the decarburisation and de-phosphorisation ofthe steel, aids flotation of the slag and reduced the volume of nitrogen absorbed in the steel.Various additives are used to absorb the residual gases, e.g. silicon, aluminium andcalcium, and vacuum degassing is also routinely used.

The treated molten steel is poured into ladles for secondary steel making. The first pour isstirred magnetically to aid removal of the residual slag and is then poured to a second ladleusually from the bottom of the first ladle to reduce carry over of floating slag. After electricalarc reheating, and whilst being magnetically stirred to ensure dispersion, the steel is injectedwith the relevant micro alloying additives and the other major alloying elements adjusted tothe appropriate level. The steel is then degassed and immediately continuously cast(concast) into slabs.

Continuous slab casting is used because it provides a more consistent product and alsohelps to reduce the segregation of phases within the slab which, when rolled, can produceplate with discontinuous mechanical properties. A typical continuous casting system isillustrated in the Figures. The slab caster comprises a tundish from which the steel ispoured from the bottom at a controlled rate into a mould. The outer surface of the steel ischilled in the mould. Water jets continuously cool the strand of liquid steel within the viscouscase of solidifying steel. The strand then flows through a sequence of water cooled rollerswhich gently direct the strand from the vertical to the horizontal where it is hot reduced by asequence of rollers to the final slab geometry. The deleterious materials in the steel are lowmelting point and congregate at the centre of the strand. To reduce the volume of theseunwanted materials the strand of steel from the mould is squeezed by pinch rollers, whichretain the unwanted material as a molten, floating core within the liquid steel. This pinchresults in a hot reduction of about 12 mm in the slab thickness. When the pour is completed

this plug of material is cut off the end of the final slab and recycled to the primary steelproduction unit.Typical Compositions of Steel for Line PipeAlloying elements are added to pipeline steels to alter the mechanical properties of the steelbut also have an impact on the corrosion behaviour. Table 5 lists the typical alloyingelements and their effect on steels. The API Specification 5L allows for a wide range ofcompositions, specifying only the maximum levels of a limited number of alloying elements.This tolerance allows the steel producer to achieve the specified strength, toughness andweldability by a range of fabrication routes.slabs cut in two then one side of the longitudinal weld of the pipe will be made from steelfrom the centre of the slab where the highest levels of inclusions and tramp compounds arepresent. When formed into pipe the inclusions would be adjacent to the longitudinal weld;this type of pipe may be unsuitable for sour service. With all pipe made from strip it is usualto reject pipe made from 'jointers", these jointers are the welds that connect the individualstrips together to make the long coils.Thermo-Mechanical Controlled Processes for Plate ProductionHot rolling can roll the slab to the correct dimensions and the slab then subsequently heat treatedto obtain the final mechanical properties. In the traditional hat rolling process thesteel slabs are heated well into the austenite region, 1200 - 1250 OC, and held at thistemperature (soaked) where re-crystallisation produces coarse austenitic grains. The plateis hot rolled starting at this temperature and the rolling is continued as the plate cools toabout 1000 C . The rolling results in the formation of large elongated grains. After rollingthe steel is allowed to cool in air. Recrystallisation of the austenite grains occurs and somegrain refining. To refine the grains further the steel is normalised by re-heating into the austenitic region tore-crystallised grains and then cooling at a controlled rate to obtain a fine ferrite grainsize. An alternative is to quench and temper the hot rolled steel plate. In this process thesteel is heated into the austenite region to re-crystallise the grains and then cooled rapidly tofreeze the grains. By cooling rapidly the non-equilibrium structure, martensite, is formed.The steel is fine grained, very strong but brittle and requires tempering which reduces thestrength and improves the toughness. The Q&T procedure gives more control of the finalproperties than normalising. The final strength of the plate material produced by hot rollingand heat treatment is achieved by a relatively high alloying content, micro-alloying andlimited grain refining. In consequence the upper limit of strength is around X60 thoughhigher grades can be achieved with careful control of the heat treatment. The carbonequivalent of these steels is high however, which has an impact on the weldability. Typicallythese steels have CE's in the range 0.35 - 0.43. The blend of strength and toughness thatcan be produced by hot rolling is limited. The higher grade, higher quality plates for pipelinefabrication is increasingly being produced by the controlled rolling processes.There are several variants of the controlled rolling process, often termed thermo-mechanicalcontrol process (TMCP) that involve rolling the plate in several stages at carefully controlledtemperatures. Schematics of the conventional hot rolling production procedures and thevarious TMCP procedures are given in the Figures.Option 1 is the simplest TMCP and is the original controlled rolling process developed toproduce large sheets of high strength steel for construction of ships and super tankers. Theslab is heated to 1200 - 1250 OC and the steel is rolled to the roughing stage. The plate isallowed to cool to a temperature where it remains austenitic but reworking the grains will notre-crystallise as a result of the work input during the second phase of rolling. The steel iscooled further until it is on the borderline of the transition zone of austenite to ferrite and isfinished by re-rolling at this temperature to form elongated austenite grains. On furthercooling the steel re-crystallizes into fine ferrite and pearlite grains. The energy input in

working of the austenite (y) grains at the transition temperature results in a relatively fineferrite grain structure though the steel retains a marked banded appearance, rather alikegrain in wood. All the working of the plate occurs in the austenite region and this option istermed TM(y).Option 2 is to roll the plate for the third time whilst it is in the austenite + ferrite zone beforecomplete conversion to ferrite has occurred; this results in a very fine ferrite grain structurethough with residual structural banding. Some dislocation strengthening also occurs, asthere can be no relaxation of the imposed stress by re-crystallisation. Since the working ofthe steel occurs in both the austenite region and the ferrite region this option is termed TM(y + a).Option 3 is to roll the plate as in Option 2 and to cool the steel using water-coolingimmediately the third rolling is completed. This allows the conversion to ferrite grains butfreezes the growth of these grains that would occur if the plate were allowed to cool slowlyin the air. This option involves working the plate in the austenitic and ferritic regions andaccelerated cooling and is termed TM(y + a) + ACC. The rate of cooling is restricted toavoid quenching the structure, which would result in the formation of bainite and martensite,and these would require subsequent tempering.Option 4 is to roll the plate at the high temperature and then allow some cooling and rerollingof the steel in the temperature region where re-crystallisation of the worked austenitegrains will not take place. The steel is then water cooled until the temperature is in theferrite region; the steel is essentially quenched. This produces a very strong material with afine but disordered grain structure and requires to be tempered. Because the workingoccurs in the austenitic region but grain growth is restricted it is termed TM(a) + Q&T.Option 5 is Option 4 with the steel quenched to a sufficiently high temperature such that ithas sufficient residual heat to self temper. This avoids the need to reheat but can only beapplied to relatively thick material. This option is termed TM(a) + QST.Option 6 is to hot roll the steel as with all previous options, allow the steel to cool to the zonewhere re-crystallisation does not occur and to re-roll and then cool the material into theferrite zone. Ferrite formation starts to occur and the material is re-heated slightly to raise itinto the austenite zone where it is re-worked as it cools through the austenite-ferritetransformation temperature. The material is then acceleration cooled. This option is termedTM(7)IC + (a)ACC.The initial slab heating stage for hot rolling is important as it controls the micro-alloyingprocess and the starting size of the grains. Each of the micro-alloying elements requires adifferent temperature for solubilisation and precipitation. During the rolling process thegrains are deformed and, if this is at the re-heating temperature between the recrystallisationtemperature and the transition temperature, or in the dual phase austenite ferritetemperature band, then the grain size is refined by re-crystallisation between therolling passes when the imposed strain is above the critical deformation level. Becausegrain growth is restricted by the precipitation strengthening precipitates formed at the grainboundaries the grains cannot grow thermally and are increasingly refined by eachrecrystallisation step.Accelerated cooling is increasingly being used because the rate of cooling from the dual

phase (y + a) temperature band has a marked effect on steel strength. The more rapid isthe cooling, the greater is the increase in the strength of the steel even with a low alloyingcontent. Because a low alloying content reduces the carbon equivalent, and hence theweldability, these steels provide very high strength and toughness with excellent weldability.The rate of cooling is restricted to avoid quenching the structure as this would result in theformation of bainite and martensite and these would require subsequent tempering.After forming the plate will contain residual hydrogen that takes time to diffuse out. Most

steel manufacturers stack plate to maintain the temperature high for longer, whichencourages hydrogen degassing. In the pipe fabrication specification provided by the pipeproducer it may be necessary to check that stacking is allowed especially for plate of 20 mmthickness or above and for plate which will be used for fabrication of pipe for sour service.

Shape ControlIn the 1960's there was an increase in production of oil and gas and a concomitant demandfor large diameter pipe that meant that very wide steel sheets of strong, tough steel wereneeded from which to form the pipe. For example to produce a 48-inch diameter pipe by theU-0-E process requires a sheet of steel -4 m by 12.5 m. Much of the new production wasalso sour. It was known that pipelines able to tolerate sour product had to be constructed ina manner that avoided hard spots at welds in the material to avoid risk of sulphide stresscracking (SCC) as defined in NACE MR-0176. This requirement is best met by the use ofsteel with a low alloying content, but as discussed above, to obtain high strength steels withlow alloy contents required steel with a fine grain size to provide both strength andtoughness. The plate for these large diameter pipes was obtained by use of the controlledrolling process, at that time restricted to Option 1.Pipe fabricated from controlled rolled steel was used for transmission of sour crude oil andthe pipe failed by internal cracking, as a result of the formation of hydrogen gas within thesteel. Some of the atomic hydrogen evolved during internal corrosion migrates into the steeland is reformed as hydrogen gas at manganese sulphide inclusion platelets in the steel. Thebuilcl up of hydrogen gas develops large pressures that raise sufficient internal stress tocause the internal cracking of the steel, a similar process to the blistering of tank steel atlaminations. The working of the steel had resulted in a marked banding of the steelmicrostructure as the grains were elongated in the rolling direction. Such steel has markedanisotropic properties between the longitudinal and transverse directions, for examplesignificant differences in tensile strengths and a factor of 2 - 3 in absorbed impact energy.More importantly for sour service inclusions of manganese sulphide are also flattened andelongated and thus presented a large surface area for absorption of the atomic hydrogenmigrating through the steel. Here comment is restricted to the methods used to alterinclusion morphology to reduce steel susceptibility to HIC and also reduce the anisotropy.The first attempt to produce steels resistant to HIC was by incorporation of copper into thesteel. Subsequent to some corrosion the copper forms an enriched copper sulphide layeron the inner surface and, by catalysing the combination of hydrogen atom to molecularhydrogen, the quantity of atomic hydrogen entering the steel is reduced. However crackingstill occurred when copper-containing pipe formed from controlled rolled steel was used forproducer it may be necessary to check that stacking is allowed especially for plate of 20 mmthickness or above and for plate which will be used for fabrication of pipe for sour service.

Inclusion Shape ControlIn the 1960's there was an increase in production of oil and gas and a concomitant demandfor large diameter pipe that meant that very wide steel sheets of strong, tough steel wereneeded from which to form the pipe. For example to produce a 48-inch diameter pipe by theU-0-E process requires a sheet of steel -4 m by 12.5 m. Much of the new production wasalso sour. It was known that pipelines able to tolerate sour product had to be constructed ina manner that avoided hard spots at welds in the material to avoid risk of sulphide stresscracking (SCC) as defined in NACE MR-0176. This requirement is best met by the use ofsteel with a low alloying content, but as discussed above, to obtain high strength steels withlow alloy contents required steel with a fine grain size to provide both strength andtoughness. The plate for these large diameter pipes was obtained by use of the controlledrolling process, at that time restricted to Option 1.Pipe fabricated from controlled rolled steel was used for transmission of sour crude oil and

the pipe failed by internal cracking, as a result of the formation of hydrogen gas within thesteel. Some of the atomic hydrogen evolved during internal corrosion migrates into the steeland is reformed as hydrogen gas at manganese sulphide inclusion platelets in the steel. Thebuild up of hydrogen gas develops large pressures that raise sufficient internal stress tocause the internal cracking of the steel, a similar process to the blistering of tank steel atlaminations. The working of the steel had resulted in a marked banding of the steelmicrostructure as the grains were elongated in the rolling direction. Such steel has markedanisotropic properties between the longitudinal and transverse directions, for examplesignificant differences in tensile strengths and a factor of 2 - 3 in absorbed impact energy.More importantly for sour service inclusions of manganese sulphide are also flattened andelongated and thus presented a large surface area for absorption of the atomic hydrogenmigrating through the steel. Here comment is restricted to the methods used to alterinclusion morphology to reduce steel susceptibility to HIC and also reduce the anisotropy.The first attempt to produce steels resistant to HIC was by incorporation of copper into thesteel. Subsequent to some corrosion the copper forms an enriched copper sulphide layeron the inner surface and, by catalysing the combination of hydrogen atom to molecularhydrogen, the quantity of atomic hydrogen entering the steel is reduced. However crackingstill occurred when copper-containing pipe formed from controlled rolled steel was used fortransmission of wet sour gas. Steel composition had to be reformulated again to reduce thesulphide inclusions and to alter their shape.The primary method of improving resistance to HIC was reduction of residual sulphur in thesteel. The small quantity of sulphide remaining, as manganese sulphide, could be furthermodified by shape inclusion. Primary calcium treatment of steel is used after the aluminiumadditions to reduce the concentration of sulphur. The calcium is added after the aluminiumaddition as calcium silicide or carbide into the molten steel. As a vapour it combines withsulphur and oxygen in the steel and the reaction products are removed in the floating slag.Very low levels of sulphur are achievable typically in the range 0.001 - 0.003%.Magnesium has the same effect but is more expensive than calcium, though magnesium isstill used to form ductile iron pipe from cast iron. A secondary addition of calcium wasintroduced at the ladle stage to shape the residual sulphides as spheroids. With calciumtreatment the residual sulphide is associated with the calcium principally as calcium sulphiderather than as manganese sulphide. These particles tend to be fine and globular and,because they are hard, retain their shape through the hot rolling process. The areapresented to the atomic hydrogen is markedly reduced resulting in less internal hydrogenaccumulation and a marked improvement in resistance to HIC. The ratio of Ca to S shouldbe restricted with a minimum of 1.5:1 to 2:l at the ladle analysis. A lower calcium ratioprovides inadequate calcium to form the spheroids whilst excess calcium formsunfavourable compounds that affect the mechanical properties in the HAZ. A typicalcomposition for internal crack resistant steel is given in Table 2.

PIPE FABRICATIONPipe for the oil and gas industry is largely restricted to four fabrication routes: seamless,longitudinally welded by electrical resistance welding, helical or spiral welded andlongitudinally welded using submerged arc welding. Production of pipe by furnace buttweldingof hot plate, though permitted, cannot produce the large diameters required.Nowadays many pipe fabricators are independent and not part of an integrated steelcompany. Fabricators also tend to specialise in certain pipe fabrication techniques and fewproduce pipe material over the full range of diameters and wall thicknesses. The Oil andGas Journal make a regular survey of pipe fabricators but this listing may not be

comprehensive as some East European and Far Eastern fabricators are omitted.Independent pipe mills form pipe from plate or coiled plate that is bought from whateversource can provide suitable material, at an acceptable price. For the production of a largequantity of pipe the plate may be sourced from several steel suppliers. The mechanicalproperties of the pipe will be consistent with specification but caution is necessary regardingthe weld procedures as small variations in composition can affect the quality of the weldachieved.

Seamless PipeSeamless pipe is formed by hot working steel to form a pipe without a welded seam. Theprocess is shown schematically in the Figures. The initially formed pipe may besubsequently cold worked to obtain the required diameter and wall thickness and heattreated to modify the mechanical properties. A solid bar of steel, termed a billet, is cut froma slab and is heated and formed by rollers around a piercer to produce a length of pipe.The Mannesmann mill is perhaps the best known type of piercing mill. In this mill the steelbillet is driven between rotating, barrel-shaped rolls set at a slight angle to each other. Therolls rotate at about 100 - 150 rpm and the billet also rotates. The piercer is placed justbeyond the point where the billet is squeezed by the rolls so that as the formed billet passesthrough the "pinch" zone between the two rollers the reducing stress tend to open the metalover the piercer.

The piercing mill produces the primary tube that requires finishing to form the pipe. The wallthickness is further reduced and the pipe finished in plug rolling mills that drive the pipe overlong mandrels fitted with plugs of the correct internal diameter between rollers that extrudethe tube to the required external diameter. An older process is the Pilger process. Thisprocess uses eccentric rolls to form the pipe in discrete stages. A mandrel is inserted intothe partly formed pipe from the piercing mill. The assembly is driven into the open rolls and,as the rolls rotate back and forth, sequential sections of the pipe are drawn into theeccentric rolls and the outer diameter formed to the required dimension set by the rollereccentricity. Other methods use multiple conical or conventional horizontal rolls or offsetrollers. Pipe is finished in a reeling process in which it is driven between slightly conicalrollers followed by passage through a sizing mill that ensures circularity.After forming, the pipe may be delivered as produced or, more usual for the oil industry, aseither normalised or quenched and tempered. These heat treatments homogenise themechanical properties and further improve strength and toughness. After forming iscompleted the pipe is inspected for internal laminations and pressure tested hydraulically.The hydraulic test is at high pressure and for up to 20 seconds.This type of pipe is generally available in diameters up to 16-inch but can be obtained insizes up to a maximum of 28-inch from a restricted number of suppliers. The largerdiameter pipe is made by expansion of smaller diameter pipe. Seamless pipe is thepreferred material of several operators for small to modest diameter pipelines. Its principaladvantages are its good track record in service and that there are no welds in the pipesections.

The larger diameter seamless pipe may be more expensive than pipe fabricated by thealternative processes. Disadvantages of seamless pipe are a fairly wide variation of wallthickness, typically +5I1 -12.5% and out-of-roundness and straightness. The premier pipefabrication mills can produce seamless pipe to closer tolerances. The outer surface of thepipe may also be highly distorted such that when it is grit-blasted, prior to coating, tinyslivers of steel rise up. This can be a serious drawback when the pipe is to be coated with athin anti-corrosion coating such as fusion bonded epoxy (FBE) and it is prudent for thepipeline engineer to check for such effects when pre-qualifying pipe suppliers.

Electrical Resistance Welded (ERW) PipeERW pipe is formed from coiled plate steel. The plate is uncoiled and sheared to aconvenient workable length, flattened and the edges dressed. The plate is passed througha sequence of rolls to form the pipe. The sequence of rolls crimps the edges of the plateand then progressively bends the body of the plate into a circular form ready for welding ofthe longitudinal seam. The longitudinal seam weld is made by electricai resistance welding(hence the name of the pipe). No consumable or weld metal is used to make theautogenous weld, which is produced by pressure similar to a weld produced by ablacksmith. The process is illustrated schematically in the Figures. When a new coil ofplate is started it is welded to the end of the previous coil to allow it to be pulled through therolling mill. The pipe formed with a joint in the middle (a jointer) is generally not accepted foruse for pipelines.An electrical current is passed across the interface to heat the steel pipe faces that are to beERW welded. Once molten the faces are pressed together to produce the longitudinalseam weld. The heating may be by low frequency AC current, typically 60 - 360 Hz,introduced directly into the pipe by rolling contacts or induced into the steel with inductioncoils operating at high frequencies of above 400,000 Hz. The later process of producingthe pipe is term high frequency induced (HFI) ERW pipe.Pipe for oil and gas pipelines is almost exclusively produced using high frequency inductionwelding. Two American, three European and four Japanese manufacturers presentlyproduce the majority of this type of ERW pipe. High frequency induction is effectivebecause of the skin heating effect that results from the use of the very high frequency ACcurrent. As the frequency of an AC current in a conductor increases the electrons that carrythe current tend to flow increasingly at the outer surface of the conductor. At very highfrequencies the electron movement is exclusively in the outer 1 mm or less of the conductor.When applied to steel pipe this results in a very high rate of heating of the steel facessufficient to melt the metal in the weld region.The pressure exerted on the faces during the weld forming process result in the moltenmetal at the faces being squeezed outward to form stubs of metal above and below theweld. Any debris or oxides on the steel faces is discharged in the stubs of metal. The stubsof metal are trimmed off and the weld is inspected using ultrasonic probes. The weld is thenlocally heat treated to anneal the weld and heat-affected zone. Depending on the diameter,the pipe is re-heated and stretch reduced slightly or is passed through sizing andstraightening rollers to ensure exact dimensions and circularity. The weld is re-inspectedand the pipe cut to length. The pipe lengths may be further heat treated. A final inspectionsequence is hydraulic testing. In the premier mills there will also be a full body inspectionafter the hydrotesting.The weld is extremely fine as the bulk of the molten metal is squeezed out. It is not possibleto detect the weld by eye and it is prudent to specify that a paint line mark the weld line.Samples of weld are cut from the ends of pipe for metallographic inspection, analysis,tensile, ductile and toughness testing. The invisibility of the weld can lead to ERW pipebeing mistaken for seamless pipe.ERW pipe is the main competitor to seamless pipe. It is cheaper than seamless and it canhave considerably tighter tolerances on wall thickness. Pipe lengths are typically standardlength + 50 mm and the pipe can be produced in lengths up to 27 m. Though APISpecification 5L permits a wide tolerance on wall thickness, +19.5/-8°/o, typical modern wallthickness tolerances are + 5 %. It is also claimed that pipe wall thickness can be specifiedto 0.1 mm and non-AP1 Specification 5L sizes are possible. These tight tolerances can havea cost benefit as the smaller tolerance in wall thickness and circularity permit a more rapidset up and lend themselves to semi-automatic and automatic welding processes.The original ERW pipe was fabricated with low frequency welding and resulted in a poortrack record for ERW. In response the HFI-ERW pipe fabricators provide a high quality

inspection to restore confidence in the product. Conventional, low-frequency ERW doesneed a high level of inspection because the welds are prone to a form of failure termed"stitching". Dirt or oxide on the weld seam surfaces, arcing or insufficient pressure duringthe welding causes a lack of through-wall-thickness welding. The weld formed is minute andfailures of the heat treatment may leave the weld area vulnerable to enhanced corrosion andthe toughness values of the HAZ may also be affected. Most of these problems have beenovercome in the HFI production process but because of problems with the early ERW pipethere is reluctance to use this type of pipe for high pressure pipeline service though thematerial is specified routinely for moderate service pipelines.U-0-E (or SAW) Longitudinally Welded PipeU-0-E pipe is formed from individual plates of steel by firstly forming plate into a U, then intoa tube (0). After longitudinally welding, the pipe is then expanded (E) to ensure circularity.The process is illustrated schematically in the Figures. An alternative method uses sets ofrollers that gently shape the plate into pipe. The shaped plate is joined by a longitudinalweld normally produced using the submerged arc welding process, and for this reason thepipe is sometimes termed SAW pipe. The submerged arc welding procedure is described inthe section on Welding.Tab plates are fixed to the steel plate and the plate cut to exact size and the edges dressed.The edges are then crimped and the complete plate is progressively bent into a U-shapeand then into a tube in presses. The 0-press leaves a residual 0.2 - 0.4 O/O compression inthe pipe. A higher compression is provided for pipe for sour service. The butting edges aretack welded at the tab plates to prevent movement during the main welding. The buttingedges of the tube are then welded using submerged arc welding with multiple head weldingdevices. At least two welding passes are made: first the internal weld is formed and thenthe pipe is rotated through 180' and a second external weld pass made. The tab plates areprovided to allow the weld to start and finish beyond the end of the pipe to ensure a qualityweld at the ends of the pipe.Some pipe fabricators use gas welding to produce a complete first weld pass to joint the twoedges of the plate. The weld is then completed using SAW. The weld is inspected byultrasonic examination and X-ray radiography. If the weld is acceptable then the pipe isexpanded using an hydraulic head that is inserted at one end of the pipe and, after eachexpansion, is moved incrementally along the pipe. At each halt the head expands to form

the tube into an exact circular pipe. Typically the cold expansion is limited to below 1 - 2%and the expander is shaped to avoid damage to the weld. The pipe is hydrostatically testedand then re-inspected using automatic ultrasonic testing and radiography. The weld mayalso be MPI tested. Finally the end faces of the pipe are end faced or bevelled, fitted withend protectors and the pipe moved to the storage racks.U-0-E pipe is used for the larger diameter pipelines. It is competitive with seamless pipe forthe intermediate diameters (14-inch to 28-inch). For the smaller diameter pipes the pipefabricator may use cut down plate because producing narrow plate is less economical.Because inclusions and segregation tend to concentrate at the centre of the plate a pipeformed from a split plate may have inclusions and segregation adjacent to the weld. Suchpipe may be unsuitable for sour service. If split plates are to be used then an odd numbershould be cut to avoid the centre line of the original plate abutting the longitudinal weld.U-0-E pipe is the premier pipe material for large diameter high-pressure pipelines. it hassmaller variations in wall thickness and out of roundness than seamless pipe; a typical wallthickness tolerance is +12%, -10% and ovality k lolMoany. p ipe fabricators can producepipe to tighter tolerances, typically k 5% and ovality k %Oh. Specifications may be based onpercentages or dimension as mm. When specifying, the percentage route is better for the

larger diameters and the dimension better for small diameter pipe. Press bent pipe is usedfor very heavy wall pipe and is formed similarly to U-0-E by forming a cylinder from plateand welding longitudinally. The bending uses three rollers, one offset and moved to formthe required diameter.Helical (Spiral) Welded PipeA coil of hot-coiled plate is uncoiled, straightened and flattened and the edges dressed. Theplate is then helical wound to form a pipe. The width of the strip and the angle of coilingdetermine the pipe diameter. As the pipe is formed the helical seam is welded using inertgas welding or submerged arc welding (SAW) first internally and as the seam rotates to thetop position the external weld is made. A continuous length of pipe is produced. Afterforming the pipe is passed through a sequence of rollers to ensure circularity.The pipe weld is tested using radiography or ultrasonic testing and the pipe is then cut to therequired lengths. The pipe joints are hydrostatically tested before being re-inspected. If thepipe passes inspection then it is end-faced or bevelled, the end protector caps are fitted andthe pipe transported to the pipe racks. A schematic of the process is given in the Figures.The end of the coiled plate is welded to the start of the next plate coil and this results in aweld perpendicular to the helical weld forming the pipe. This weld joint should not be lessthan 300 mm from the end of the pipe. This weld may not receive the same degree ofscrutiny as the helical welds and the pipe specification may need to call for an additionalinspection of these welds after the hydrotesting stage.Helical welded pipe can be made in a wider range of diameters and wall thickness thannominal API Specification 5L sizes. It can also be produced in long lengths above thenormal double random length of 12 m. Because the pipe is formed from plate the wallthickness tolerances are good being similar to U-0-E pipe though there may higher out-of roundness.It has been used for large diameter pipelines, both crude oil and gas, but isgenerally considered a less reliable material than U-0-E formed pipe. It is cheaper than U-0-E and is widely used for caissons, sleeves, low-pressure hydrocarbon service, dried gasservice and for water transportation where the service is moderate. The large length ofweldment and the need for some of this weld to be at the bottom of the pipe, where theworst case corrosion occurs, is seen as a potential weakness which tends to preclude thistype of pipe from general high pressure service. The helical welding may reduce the datagained in intelligence pig surveys because the sensors skip at the welds resulting in blindspots just downstream of the welds.

Selection of Pipe Wall ThicknessThe diameter of a pipeline is selected based on hydraulic studies. The wall thickness isdetermined by service pressure including hydraulic heads, the design factor, the corrosionallowance and the fabrication tolerance (wall thickness variation). It is then commonpractice to select the nearest higher API Specification 5L standard thickness pipe. Thelarger negative tolerance in wall thickness for seamless pipe as compared to SAW pipemeans that for a given diameter a seamless pipe will need to be thicker on average than anERW or SAW pipe. As wall thickness increases the complexity of welding may increase.However many modern pipe fabricators can produce pipe to closer tolerances and this mayneed to be reviewed using a cost benefit analysis. Many companies insist on smallertolerances than API Specification 5L tolerances and also specify exactly the wall thicknessthey require. However for small batches of pipe this may not be economical, as standardAPI sizes tend to be cheaper than non-standard.

The corrosion allowance is determined by calculation and a method is given in the Sectionon Internal Corrosion. For a non-corrosive service (e.g. a refined product pipeline) there isno need for a corrosion allowance. In a mildly corrosive environment a corrosion allowancealone may be adequate, though it is usual to use a corrosion inhibitor to suppress pitting

corrosion. More aggressive environments may demand a corrosion allowance andcontinuous corrosion inhibition. For very corrosive environments carbon steel may not be asuitable material and an alternative corrosion resistant material may have to be used.There are no easily applicable criteria for these changeovers. The design life expectancy ofthe pipeline, the pressure and flow profiles over the design life and the expectations ofchanges in production fluid composition all need to be taken into account. Other factorsrelevant include the criticality of the line, the ease of inspection, the availability of inhibitor,the level of skill of operatives and, increasingly, the environmental aspects. These arediscussed elsewhere.

Pipeline Joint LengthsWhen fabricating land pipelines the welders move along the pipeline, welding the pipe jointsto the end of the growing pipeline. This allows considerable flexibility in the lengths ofindividual joints. Manual welding remains the most flexible welding procedure. If semi automatic welding techniques are used, as the complexity of the welding proceduresincreases, so the need to reduce the tolerance on pipe length generally increases toaccommodate the machine limitations.BendsIt is important to check that the material can be hot formed and retain its strength. The hotbend fabrication process may excessively reduce the strength of very low carbon steels. Ifthis is the case then special steel joints may be required from which the bends can befabricated.If a pipeline is to be constructed to allow inspection by magnetic flux leakage pigging toolsthen there are restrictions on the acuteness of the bends. The bend size decreases withincreasing pipeline diameter. For 6 and 8-inch pipes the pigs may require 5-D bendsminimum whilst for 24-inch pipes the more advanced pigs are designed for 1 %-D or 2-0bends.StraightnessUsually the out of straightness of the pipe follows a log-normal distribution with diameterand, as would be expected, the larger the diameter of the pipe the straighter is the pipe.API Specification 5L allows 24 mm in a 12 m joint. Usually it is possible to obtain pipe to 12mm in a joint length of 12 m. To reduce total cost over specification of straightness should be avoided.Residual MagnetismThe final step in pipe fabrication is to bevel the ends of the pipe to ensure that they aresquare to the longitudinal axis of the pipe and are ready for welding. The bevels may beinspected for laminations using magnetic particle inspection. Pipe should be de-magnetisedas residual magnetism can cause problems during welding, as control of the plasma arc islost. The plasma is an electrical current that is distorted by the magnetism. Note thatproblems with magnetisms may also arise when attempting to repair a pipeline that hasbeen inspected using a magnetic flux inspection pig.

IMPROVING CORROSION RESISTANCECorrosive service pipelines may be fabricated in solid corrosion resistant alloy (CRA) orcarbon steel internally clad with CRA, or non-metallic pipe. This option would only beselected if carbon-manganese steel pipe were unsuitable for the service. Selection is clearif the fluid to be transported is too corrosive for carbon manganese steel even with inhibition.Borderline cases arise where carbon steel with inhibition is theoretically feasible but the useof corrosion inhibition may not be actually economically or technically feasible. Examplesthat may arise are remote production facilities, production facilities in hostile environmentsor to minimise risk of environmental damage. Under these circumstances it may be costeffective to use a corrosion resistant alloy for the pipeline compared to carbon steel plusinhibition.

In conjunction with the technical reason for selection of the CRA there will need to be ajustification of the particular CRA selected on strictly economic grounds. In all cases thedesign life is of critical importance. If the cost benefit analysis is between a CRA andcarbon steel it is important to include hidden costs of operating a carbon steel pipeline, e.g.the regular monitoring and inspection, transport and application of inhibitor, and the higherrisk of failure of the pipeline. If the analysis is to distinguish between different CRAs thenthe costs associated with installation and potential failure risk tend to dominate. Severalstudies are reported which indicate that the more expensive material option can be thecheapest overall for modest diameter (to 10-inch), short length (to 15 km) pipelines withdesign lives of fifteen or more years. Outside this diameter-length domain the overalloperation of the Field may need to be re-evaluated taking account of the processing costs toalter the corrosiveness of the produced fluids.

Alloying ElementsThe corrosion resistance of carbon steel is markedly improved by the addition of nickelandlor chromium. The separate effects of nickel and chromium are shown in the Figures.Other elements, for example molybdenum, are also be added to improve resistance tospecific types of corrosion attack. Relatively small additions of chromium result in markedreduction of corrosion in sweet systems (fluids containing carbon dioxide). The bulk ofcorrosion in sweet systems is suppressed and reaches a minimum when the nickel contentis 9% or greater or the chromium content equals or exceeds 12%. Unfortunately, thoughgeneral corrosion is reduced, other corrosion mechanisms become dominant, for examplechloride stress corrosion cracking, crevice corrosion and pitting. The external corrosionissues must also be addressed. The selection of the CRA is largely determined by the needto avoid these potential internal and external failure mechanisms.Use of a single alloying element would minimise material cost. Chromium alone would bethe preferred alloying element because it is cheaper than nickel but the simple chromiumalloys have poor weldability, showing delayed cracking and a marked reduction in themechanical properties of the heat affected zone (HAZ). To overcome these limitations,blends of chromium, nickel and iron are used.In addition to nickel and chromium other alloying elements may be used to improve thematerial resistance to pitting and crevice attack. Molybdenum and nitrogen are twoelements that have the most significant effect.The chromium and nickel steels are corrosion resistant because, subsequent to a smallamount of corrosion, the alloys form a continuous and tenacious surface film enriched in thecorrosion resistant element. This is particularly the case with chromium where thechromium forms a chromium oxide film. This film is normally invisible but it can be grown byanodising andlor oxidising and then become sufficiently thick to appear tinted by lightdiffraction. The film has a high electrochemical potential compared to the base material andis less likely to corrode.The corrosion resistance resulting from the presence of the surface film is termed passivityand the film is termed the passive film. The passive film is not a static entity but isconstantly being reformed but at such a low rate that the material appears to be uncorroded.If conditions are such that the surface film is preserved and can reform correctly then thematerial will give long service. High levels of chloride and the other halides can preventlocal reformation of the film and corrosion occurs at these defective areas. Because thenormal chromium oxide film is cathodic to the anodic (corroding) areas a galvanic couple isestablished.The rate of corrosion at the anode will depend on the ratio of the cathodic and anodic areas.If this area ratio is large then the continued corrosion may appear as pitting attack. Theeffective area ratio of cathode to anode depends on the local conductivity of the water.Pitting Resistance Number, PRENThe pitting resistance number is a convenient method of comparing the corrosion resistance

of the CRAs. The number has no fully developed theoretical basis but is empirically proven.The PREN is sometimes also called the pitting index and is calculated from:where the symbols represent the weight percentages (%) of chromium, molybdenum, tungsten andnitrogen.Solid CRA PipeThough there are many CRAs available only a limited number can be used as solid material

for pipelines that are to be fabricated by welding. Once the CRA is welded part of the heat affected zone of the weldment is in the solution-annealed condition. Many CRAs have low

yield strength in this condition compared to their strength in the wrought condition.Regaining strength by cold working, heat treatment or ageing cannot be applied to thefabricated pipeline. Overcoming the low strength would necessitate thick wall pipe andconsequent severe economic impact. The range of materials that can be used for internalcladding is wider but there are limitations on the melting point of the cladding material.Studies have been made on the use of screwed connectors as an alternative to welding forsmall diameter pipelines; this would permit a much wider range of materials to be used.As a consequence of these limitations the most commonly used CRA at present is solidduplex stainless steel. The duplex stainless steels have yield strengths of X65 or above andthe operational temperature range is -50 to +150 degrees C. Duplex stainless steel has goodcorrosion resistance to sweet and mildly soured fluids and can also tolerate externalexposure to saline waters (in case the cathodic protection system is inoperative).Recently developed alloys that are intermediate in price between carbon steel and duplexsteels are the low nickel 130h Cr alloys. These are weldable and suitable for hightemperature sweet and sour service though limited in their tolerance to the chloride contentof the product.

Clad PipeClad pipelines are formed from a carbon manganese steel outer pipe lined internally with athin layer (2 - 3 mm thick) of corrosion resistant material. The internal cladding is notconsidered to contribute to pressure retention. The clad pipe may be formed with the liningeither metallurgical-bonded to the steel pipe or tight fit into the steel pipe. The clad piperepresents an economic solution of mechanical and corrosion requirements. The clad pipeconfers the high strength of carbon steel (typically API 5L X70 or X80) with the corrosionresistance of the liner.Commonly used CRAs are austenitic stainless steels (mainly type 316L), and the high nickelalloys type 825 and 625. 13 Cr materials have also been used and there is interest in usingsuper grade stainless steels, e.g. nickel alloy 904L or 25-6M0. Clad pipe has beenextensively reviewed in a two-year cross-industry study released in 1995.Possible Future MaterialsThe modified 13% Cr materials, often termed weldable 13Cr, will be more widely used asconfidence is gained. To date about 600 km of pipe has been installed.The super-austenitic materials are possible materials for use either as solid pipelines or,more likely, as a cladding material. The lower nickel content would reduce the cost of thesteels to mid-way between the type 300 austenitic steels and the high nickel alloys. Superaustenitic steels have high PREN values and hence good resistance to pitting, crevicing and stress corrosion cracking. These materials are also readily weldable compared to the duplex stainless steels.Candidates are 904L and 25-6M0. The 904L requires to be modified by addition of moremolybdenum to a minimum content of 6.1%. The multiplier of 3.3 for Mo moves the PRENfor the 904L into the super-austenitic domain. 25-6M0 is a generic type but is presentlylimited in availability.

CHECKLISTType of pipeSteel making routeIdentification of pipeline service conditionsLimitations on pipe steel chemistryLimitations on longitudinal weld chemistryMinimum yield strength (SMYS)Upper strength limitationNon-ductile transition temperature (NDTT)Minimum acceptable average and minimum toughnessToughness test temperatureNon-destructive testing, acceptance criteria, procedures for out-of-specification productHydrostatic testing, acceptance criteria, procedures for out-of-specification productWall thickness toleranceLength toleranceOvalityStraightnessEnd finish: square, standard or special bevelsInternal cleanliness of the pipeProvision of caps at pipe endsAvoidance of jointers, split skelps, residual magnetisminspection requirements, frequencies, reporting proceduresPerformance tests required, special tests, acceptance criteriaTrack record of fabricatorWarrantiesPre-qualification requirements, mill visits, complaints procedures

TABLES