BAE Systems has been building submarines in the United Kingdom since 1888 and has developed many welding processes, including those used on Astute class submarines.

An Astute class submarine in its berth.

Astute class boats wil comprise the largest nuclear-powered attack submarines Britain's Royal Navy has commissioned.

Vickers Armstrong – now BAE Systems – built its first submarine in 1888 and has been in the forefront of submarine construction ever since. This is a review of the uses of welding in building submarines at BAE Systems, and shows that many of the problems of fabricating heavy, high-strength welds encountered and often solved, in advance of the rest of commercial fabrication.

The Beginnings
The first all-welded ship built in the United Kingdom was completed at Cammell Laird's shipyard, in Birkenhead, in 1923, but it was not until the late 1920's that electric welding was permitted in United Kingdom naval construction, and not until 1933 was electric welding used for naval construction.

Riveting remained the principal method of plate joining until the early part of World War II, despite the fact that Vickers' Production Manager Sir Len Redshaw had seen extensive welding in German submarine building in the late 1930's, and was desperate to increase it in United Kingdom.

Although ESAB, marketing Oscar Kjellberg's patented coated electrodes, had begun the manufacture of coated electrodes before the First World War, it appears that the first arc welding process used in the shipyard involved dipping bare wire, mild steel rods in a slurry bucket. The dipped rods then were deposited using DC power controlled by resistance boxes. This primitive process was used for the fabrication of deck and angle brackets that previously had been hot forged by blacksmiths, and to seal minor joints that had been difficult to make watertight.

In the early 1930's British-made electrodes pre-coated with flux were adopted. The flux was reinforced with asbestos string to provide some strength and adhesion to the flux coating during drying and subsequent handling, but welding processes continued to use, primarily, dipping and drying. It was still necessary to use DC power, because the flux formulations were not capable of sustaining the arc through the voltage reversals inherent with AC.

The Coming of the Welder
Welding originally was carried out by electricians, but electricians soon relinquished this "dirty and unpleasant" process and caulkers took it up as it was replacing their traditional work of riveting and caulking. Welding as a separate trade/department evolved around 1935. The new caulker/welders soon extended their skills to include positional welding.

By 1938, a number of companies were selling extruded, machinemade rods; and developments in power sources improved weldability characteristics to suit specific applications, including deep penetration, touch welding, ease of slag removal and welding on the open berth.

The Second World War
At the start of the WWII, welding was being used to join the circumferential butts of pressure hull plates in submarines, but seams still were riveted. In submarine building terminology, a "circumferential butt" is a plate butt joint aligned 90 degrees to the longitudinal axis of the ship, a "seam" is a butt joint running parallel to that axis.

By 1942, submarines were more or less completely welded, mainly using extruded, "Ironex" electrodes, an electrode that was still helically wrapped with asbestos string and iron wire, giving good tolerance to rough handling, external atmospheric conditions, dirty plate surfaces and, inadvertently, slightly higher deposition rates. Electrode diameters were available in the range of 3.2 mm to 6 mm (0.126 in. to 0.236 in.), and AC current could be used, even positionally.

However, in 1943 the British Admiralty – now known as MOD(N) – began to take random X-ray shots of various butt welds using "portable" X-ray machines and there were some experiments with the use of such techniques as preheating rods to reduce cracking.

Mechanised Welding
Also around 1942, the "Fusarc" process was introduced. This was a process that had been under development in France and was "rescued" by an RAF raid just before the Germans arrived at the factory. It was developed by "Quasi-Arc," a welding equipment manufacturer in Gateshead, and introduced to the shipyard in 1942. It used a continuous coated electrode, supplied as a large coil, comprising a core wire, with helically wound binder wires embedded in the flux coating, thus providing strength to help the flux remain on the core wire as it was straightened and fed by drive rollers through long contact jaws into the arc, The binding wires also provided the electrical contact path between the jaws and the core wire to enable the arc to be drawn at the electrode tip.

Prefabrication
Fusarc soon was used to weld submarine hull units set up on rotator rollers. An initial, manually deposited root run was laid internally in the butts and seams. Tractor mounted Fusarc machines were set externally at the top dead center of hull units, where they could weld the seams, by driving the tractor along the joint, or for circumferential butts, by driving the tractor while the unit was rotated on the rollers underneath. At the same time the internal ring frames that had been positioned "toast rack" fashion in the unit were manually fillet welded to the hull plating by a team of welders.

The new pre-fabrication techniques made possible by welding and the demands of the general war effort enabled production to reach the stage where one submarine was being launched each month from the same slipway. Subsequent outfitting of the hull took two months, so that a boat could be in commission within three months of launch. This rate was possible because the boats were being ordered in batches with fixed designs, and the material was made available.

By the end of the war, production was high, and welding had proved its worth in terms of strength, reliability, speed, economy, and flexibility. Radiography was being used regularly, but welders still worked strict "piecework." A ‘clear picture' bonus was introduced whereby welders who produced welds proven defect-free by radiography, were paid a bonus. Empirical, mutually agreed acceptance standards were developed in-house. There was also an elaborate system of visual checking by the foreman; with signed statements that he had accepted the weld preparation prior to welding, and records of actual joint dimensions, gaps and angles etc to prove that he had inspected the root run on both the first and second sides.

improved Steel and Weld Properties
By the late 1940's, higher tensile steels were introduced. There were extensive investigations by the Engineering Department of Cambridge University, from which began the British Welding Research Association now known as TWI, to discover the causes of the dramatic cracking failures of large welded ships, notably the Liberty ships and tankers, such as the Schenectady, which had been mass produced in America.

Research focused attention on, and increased knowledge of, the significance of welding design, steel chemistry and hydrogen. Controls for carbon, silicon, manganese and sulphur levels in steels for welding were introduced, and the significance of hydrogen and carbon became understood.

Consequently, low hydrogen, which was then termed "lime ferritic," MMA electrodes were developed which had to be dried or baked. These required a very short arc length, and hence, greater welder skill to control the arc, but produced joints with better tensile, elongation and toughness properties, in both weld metal and heat affected zone.

Various "Admiralty steels" and corresponding electrodes were developed for higher strengths through the ranges of A quality, B quality, D quality: DW and UXW, being carbon-manganese low alloy steels with higher carbon equivalents that gave rise to some weldability problems. In order to alleviate these problems, QT28, an electrode with a lower carbon equivalent, was introduced in 1956. This, the first fully quenched and tempered low alloy steel, was used for pressure hull plating for the successful "Oberon" class of boats.

The nuclear Submarines
In 1959, the QT28 electrode was replaced by the QT35 electrode, which had a 35-ton-per-square-in. 0.2 percent proof stress. That was attained by stringent heat treatment from an open hearth steel making process. This steel was intended to match the properties required for Dreadnought, Britain's first nuclear submarine that was based on an American design that used HY80, also a quenched and tempered steel.

Because the properties of these steels were obtained by heat treatment, the very process of welding degraded the properties of the heat affected zone, so that very strict controls of pre-heat, heat input (run out length) and temper bead sequences were developed, to minimise the size, hardness and brittleness of the HAZ.

The as-deposited weld metal could not be heat-treated, and therefore had to have a higher alloy content to match the strength of the plate. That, in turn, increased the risks of weld cracking. Therefore the weld metal analysis/ strength was increased even more, to over-match the plate strength, to try to compensate for the lower toughness of the weld metal.

Whole joint testing was carried out by the MOD(N) Research Establishment using a specially devised "Pellini Bulge Explosion" test, in which whole test weld plates were subjected to severe mechanical deformation by explosions close to their surface.

Automatic hull butt welding still was carried out with the development of the Fusarc process, where alloying elements were added to the continuous flux coating to improve the deposited weld strength. This reduced the welding characteristics of the flux coating, but was offset by the use of an extra granulated arc flux, as in then-common submerged arc process.

This hybrid process was called "Fusemelt," and provided by BOC who had taken over Quasi-Arc. Low basicity powdered fluxes had not been developed, so that the standard submerged arc process was not capable of giving the required joint properties

Ultrasonic testing
In 1959, for the first time anywhere in large scale, heavy fabrication, the quality of (pressure hull) welding was examined not only by gamma radiography, but also ultrasonic testing. At first, normal probes scans were used.

As with the introduction of X-radiography in the 1940s, there were very few objective acceptance standards at first, and no truly standardised ultrasonic testing procedure. No significant problems were identified for some time, until the absence of a properly standardized ultrasonic testing calibration system lead to considerable confusion and widely differing defect reports by various ultrasonic testing teams in the shipyard and subsequent in-service monitoring inspections. Sequential in-service surveys appeared to show defects in the frame to pressure hull welds growing at alarming rates, causing consternation.

Eventually, a section of hull was actually cut out and tested by all the teams who each found different defect lengths. The sample then was cut up to prove that the "indications" were merely surface profile effects that had been found during the post welding inspection in the yard, but not recorded.

A more significant problem resulted from the change to the connection of the stiffener frame web to the pressure hull from simple fillet welds to full penetration tee-butts. This may have been introduced to improve and simplify the inspection and interpretation by the ultrasonic testing process, and perhaps to improve resistance to the possibility of fatigue toe cracking in subsequent service.

These welds were deposited by a semi-automatic MIG process using a 1 percent argon/oxygen shield gas and 1.6-mm-diameter Airco A632 low alloy bare wire (at 380-420 amps).

The tensile strength of this deposit was 25 percent greater than the plate strength (overmatched), to compensate for the lower toughness of as-deposited weld metal compared with the plate toughness.

It is interesting to note that, even today, submarine hulls designed with a maximum working stress close to the plate yield stress, and subjected to external pressures with extreme modes of failure, are not, and cannot be subjected to post weld stress relief. Compare this with almost universal code requirements for post weld treatments for commercial pressure vessels designed with much greater factors of safety.

Lamellar Tearing
The resulting very high tensile residual stresses induced in the tee butt welds, from the over–matched weld metals and the pre-heating regime used, exceeded the "through the thickness" or short-transverse strength and ductility of the QT35. Soon, the incidence of lamellar tearing (cracking) in the pressure hull plating below the frame webs was revealed by the ultrasonic testing. Many strange theories for the cause of the mainly sub-surface cracking were postulated, together with procedures to both avoid its incidence and repair it when found.

In the end, it was the shipyard welding engineers who deduced that the problem lay with the previously unrecognised poor through-the-thickness properties of the QT35. It was clear that the residual stresses of the tee-butt connections had to be reduced. Lower strength electrodes, with pro-active control of preheating and deposition techniques were developed that effectively overcame the problem for the frame welds.

New repair techniques were devised involving management of the application of preheat and buttering using under-matched electrodes. More research showed that it was the open hearth steel making process for QT35 which gave rise to distribution of lamellar, brittle, manganese-silicates which caused the poor throughthickness, which now is known as "Z" quality, properties.

It was realised that these were not a feature of the comparable HY80 steel, which was produced by electric furnace and vacuum de-gassed. Thus, by 1969, NQ1 steel, a modified version of HY80 steel, was introduced. This steel still is used in current submarine fabrication, and the experience gained has lead to the development of "Z" quality steels for susceptible heavy joints in commercial use in the United Kingdom.

Welding process developments
In 1967, the first commercial lowhydrogen basic flux was tested and approved using a low alloy wire, as a conventional submerged arc process. The old Fusemelt process was abandoned, leading to significantly improved applications of automatic welding for hulls.

The older MMA electrodes were replaced by slightly higher strength rods, but with improved weld metal toughness and reduced hydrogen levels. Unfortunately, careful and extensive developments of pulsed arc positional processes in the late 1960's failed to produce a viable semi-automatic process to replace positional MMA. During the early 1970's, ultrasonic testing began to be carried out using both shear and normal probes, giving much better defect location and representation. In the mid 1980's, a flux-cored wire, with acceptable mechanical properties, was introduced, which was much more productive than MMA for positional work.

Current Processes
Current welding practice for submarine manufacture involves twin tandem submerged arc for rotated sub-unit circumferential butts, and for frame to hull and web-to-table tee butts. Pressure hull static circumferential butts and sub-unit vertical seams are welded by a mechanised (positional) FCAW process, and semi-auto FCAW is used for all other welding. The use of MMA is currently limited to very few applications, mainly where access is very restricted. Current NDT practice promotes the use of digitised ultrasonics (time-of-flightdiffraction), replacing radiography for butt welds wherever practical.