Application of Hybrid Laser Arc Welding for Construction of LNG Tanks Made of Thick Cryogenic 9% Ni Steel Plates

Quenched and tempered 9% Ni steel is preferably supplied as X8Ni9 + QT according to DIN EN 10028-4 or ASTM/ASME A553 Type 1. Due to its favorable combination of production costs and high mechanical-technological properties [1‐3], X8Ni9 + QT steel is best suited for use in cryogenic environments down to -196 °C, where high impact strength is primarily required. The high cold resistance of this ferrite steel alloyed with nickel is comparable to that of austenitic steels and is achieved in the course of heat treatment and due to complex microstructure, which includes fine tempered-martensite with up to approx. 10% of the thermally stable retained austenite [4‐6].

The large volume LNG storage tanks and containers as well as components of modern LNG fueled ships exhibit representative examples for the use of 9% Ni steel [7‐9]. The LNG market continues to grow and, according to a forecast by the International Energy Agency (IEA), the global demand for LNG volumes will increase more than double by 2040 [10]. For steelmaking, this means that more and more sheets made of 9% Ni steel are needed to expand the LNG infrastructure. From the LNG plant manufacturer’s side, this large tonnage of sheet metal needs to be processed by welding into a finished product. In modern LNG facilities, steel sheets made of 9% Ni steel with a thickness in the range between approx. 10 and 60 mm are used for the construction of the inner wall as well as flat bottom, which withstand the cryogenic medium [11, 12]. The design of the large above-ground LNG tanks often requires that the dome roof is also made of 9% Ni steel [13]. In LNG tank construction, the steel plates are usually welded together in multi-run technique using conventional arc welding methods. These are submerged arc welding (SAW), gas metal arc welding (GMAW), shielded metal arc welding (SMAW) or gas tungsten arc welding (GTAW). Welding can be performed in different positions. A novel vertical SAW process has been developed for this purpose [14]. It uses a small-diameter filler wire and a waving technique to control the weld pool. This results in a relatively low deposition rate of less than 100 g/min, which does not appear to be particularly effective in LNG production.

A major problem that frequently occurs with nickel alloyed steels is the residual magnetism that causes arc blow. This is a particular problem with 9% Ni steel, which can be easily and very strongly magnetized, making it impossible to perform the weld [15]. The alternating welding current (AC) is therefore recommended to prevent the arc blow effect. Other welding techniques such as friction stir welding (FSW) or electron beam welding (EBW) for joining thick sheets of 9% Ni steel are currently being researched [16, 17]. Nevertheless, according to current regulations in the LNG sector, only arc-based welding processes are used for the construction of LNG facilities.

Austenitic Cr-Ni and Ni-based filler metals are the standard welding consumables for joining 9% Ni steel as this ensures good fatigue crack resistance and high fracture toughness of welds at low temperatures [18, 19]. The regulations of the International Classification Society Det Norske Veritas and Germanischer Lloyd (DNV GL) set the required average energy of the Charpy V-notch impact test of 34 J for the welds on Ni-alloyed steels for LNG applications [20]. Projects carried out under the scope of German Technical Inspectorate (TÜV) sometimes specify a minimum Charpy toughness of 0.40 J/mm², corresponding to 32 J on a standard Charpy impact specimen [21]. Depending on the grade of 9% Ni steel used, other LNG projects have even higher requirements, with a minimum impact energy of 60 J as an acceptance criterion [22].

Satisfied notched impact strength of austenitic WM at -196 °C on quenched and tempered 9% Ni base metal (BM) can be achieved if the welding procedure is carried out in compliance with the welding regulations. The welding process has to be carried out with the lowest heat input in order to control the grain size in the HAZ and the degree of mixing between the austenitic filler material and the ferritic base material, so as not to reduce the toughness of the welded joint [23, 24]. It is therefore recommended to limit the heat input during welding to about 3.5 kJ/mm for SAW and 2.5 kJ/mm for SMAW [25]. In addition to high toughness, maintaining the required tensile strength is an essential task when using austenitic filler metals. This is particularly the case for conventional Ni-based welding consumables, as the tensile strength of the austenitic WM is lower than that of the quenched and tempered 9% Ni steel [26]. To balance this effect, the wall thicknesses should be over-designed in order to ensure the strength of the welded construction. As a consequence, there is a greater need for materials, higher processing costs and additional construction costs, e.g. for stronger foundations.

Another factor that increases the construction costs of LNG tanks is a large amount of expensive austenitic welding consumables, which are needed to fill the edge preparations on thick plates. A more suitable approach for joining 9% Ni sheets would be the use of less expensive ferritic welding consumables of the same type as the BM. Such fillers could offer matching properties of the WM and BM as well as 100% efficiency of the weld joint. Intensive research into the development of ferritic filler materials for welding the 9% Ni-alloy was already conducted in the 1980s [27]. The latest studies show that with the use of matching ferritic filler materials and appropriate post weld heat treatment (PWHT), the required mechanical properties of cryogenic 9% Ni steel sheets can be achieved with confidence [28, 29].

Despite the existing recipes for the manufacture of ferritic consumables for cryogenic service, the use of these materials has only become established in practice for specific applications. In addition, the market availability of these fillers is still very limited. The reason for this is considered to be a possible impairment of the weld quality at operating temperature. The question of whether these filler metals represent an alternative to austenitic filler metals has not yet been clarified. Given the current state of the art, the implementation of highly efficient welding processes such as HLAW appears to be very relevant to the LNG sector, not only from an economic point of view, but also in terms of productivity and quality. The great economic potential of the HLAW process compared to the conventional welding processes results from the significant advantages of this technology, such as the high achievable penetration depth, high welding speed, low heat input during welding and ability to automate the process [30, 31]. In particular, when welding thick-walled LNG constructions, the advantages of the hybrid welding process becomes noticeable. The thick-walled parts can be welded faster and with less amount of filler material. Numerous studies showed the applicability of HLAW for single-pass welding of flat steel plates up to thicknesses of 25 mm [32‐34]. An electromagnetic weld pool support allows steel plates with a thickness of 28 mm to be joined in single layer in butt joint using the HLAW process [35]. 45 mm thick high strength steel sheets could be joined by HLAW with double-sided technique and application of metal cored wire [36]. The current state of development thus shows that existing HLAW technologies can cover the range of wall thicknesses presented in the LNG sector.

A number of studies have been carried out by research groups worldwide to investigate the application of laser-based welding techniques for joining thick plates of nickel alloyed steels [37‐39]. Laser beam welding (LBW) tests have shown V-notch impact strength of up to 1.0 J/mm² (80 J) for 7.2 mm thick sheets of X8Ni9 [40]. The results could not be extrapolated to thicker plates. In particular, an average V-notch toughness of 0.26 J/mm² (21 J) was achieved for the 14.5 mm thick 9% Ni steel, which was below the acceptance level [41]. It has to be mentioned that the autogenous LBW process of 14.5 mm plate thickness was characterized by an unfavorable welding thermal cycle with a high cooling rate of up to 300 °C/s and no metallurgical influence on the WM was possible because no filler material was added. In [42] it is discussed that impact strength is influenced not only by the WM microstructure but also by the dimension of the heat affected zone (HAZ) and the heterogeneity of the mechanical properties, which is strongly dependent on a preheating parameter. The 14.5 mm thick plates of 9% Ni steel were preheated before LBW experiments. Preheating resulted in the formation of both WM and HAZ without significant hardness differences. Impact toughness values up to 2.28 J/mm² (182 J) were obtained for the preheat temperature range of 160 to 210 °C and cooling times t_8/5 of 2.1 and 2.3 s, respectively. The obstacle is that preheating is an additional technological step that requires special equipment and, in addition, cannot always be carried out on the construction site.

In contrast to LBW, the HLAW process offers more flexibility. Additional energy of the arc allows obtaining a more favorable welding thermal cycle. Satisfactory metallurgical or mechanical properties of the HLAW welds can be achieved by using suitable filler metals. A uniform distribution of the filler material in the narrow and deep melting zone is of importance for achieving the required mechanical properties of the hybrid welds [43‐45]. Furthermore, it is of great practical interest to determine the extent to which the dilution of the filler material in the hybrid welds of cryogenic 9% Ni steels can have a positive influence on the notched impact strength of the hybrid welded joints.

The state of the art and technology shows that the application of laser-based welding technology for the LNG service is an interesting topic from an economic and technological point of view. Unfortunately, this aspect has not been sufficiently researched up to now.

This paper discusses the results of welding tests on the relationship between welding process parameters, filler wire type and achievable mechanical properties of HLAW joints of thick 9% Ni steel plates. Furthermore, attempts were made to clarify the performance of the hybrid welding process for prototype manufacturing. A segment of the inner wall of large size LNG storage tank was hybrid welded in both horizontal (2G) and vertical (3G) positions as a simulation to the actual assembly onsite.

The low carbon 9% Ni quenched and tempered steel X8Ni9 + QT640 with a minimum yield strength of 640 MPa was used for the welding tests. The sheet thickness was 14.5 mm. According to the manufacturer’s specifications, the steel has a tensile strength of 717 MPa and an absorbed impact energy of 197 J at a cryogenic test temperature of -196 °C. The Vickers hardness is 250 HV. Austenitic solid wire ERNiCrMo-3 according to AWS A5.11 and experimentally produced ferritic wire ERNi11 with 11% Ni were selected as filler metals for comparative welding tests. The chemical compositions of the materials used are given in Table 1.

A gas mixture of Ar with 18% CO₂ as well as Ar with 30% He with a flow rate of 25 l/min served as shielding gases for HLAW tests. The back shielding was Argon 4.6 with a purity grade of at least 99.996%.

A 20 kW IPG YLR-20,000 fibre laser with an emission wavelength of 1070 nm and a beam parameter product of 11 mm x mrad was used as the laser radiation source. A HIGHYAG BIMO HP laser processing head with a focal length of 350 mm provided a focused spot diameter of the laser beam of 560 μm. The arc was powered by a Cloos Qineo Pulse 600 welding machine with a maximum current of 600 A.

The HLAW experiments were performed with an arc leading orientation were the GMAW torch was tilted at an angle (γ) of 25° to the laser beam axis. The process distance (a) between the wire tip and the laser incidence point on the workpiece surface was adjusted to 4 mm. With a negative focus position (Δz) of -3 mm, the laser beam was slightly defocused on the workpiece surface. The contact tip to work distance (S) was adjusted to 16 mm. The configuration of the HLAW process is shown schematically in Fig. 1.

The welding head was mounted on a 6-axis industrial robot. For butt welds with a technical 0-gap, sheets of 9% Ni steel with dimensions 300 mm x 200 mm x 14.5 mm were precisely aligned and clamped on the welding table. The welding setup used allowed welding in flat position 1G/PA (Fig. 2a) as well as horizontal position 2G/PC (Fig. 2b) and vertical down position 3G/PG (Fig. 2c).

After welding, the hybrid welds were visually inspected for surface irregularities in accordance with ASME IX and DIN EN ISO 13919-1 qualification standards. Selected welds were radiographed (X-rayed) in accordance with DIN EN ISO 17636-1 for internal quality assessment. Metallographic examinations and hardness measurements were carried out in accordance with DIN EN ISO 17,639 and DIN EN ISO 6507-1 respectively.

Cross sections were obtained at the middle part of welded specimens. After metallographic polishing, the samples were etched in Nital reagent (Alcoholic nitric acid 4%). The macro morphology of the welds was then examined using a Zeiss Stemi 508 stereo microscope. The microstructure of the HAZ and the WM of the welded specimens was investigated by optical microscope Zeiss Imager A1m. The X-ray diffractometry (XRD) was used for quantification of residual austenite in the WM of hybrid welds.

Tensile and impact tests of welded joints were carried out at room temperature and − 196 °C, respectively according to concerned standards. The Charpy V-notch impact test was done according DIN EN ISO 50125. The V-notch impact test specimens were processed into size of 10 × 10 × 55 mm³ with a 2 mm depth V-type 45° notch based on ISO 148-1. The V-notch was milled in the centerline of the laser-hybrid WM. The fracture surface morphology was characterized using a Phenom XL desktop scanning electron microscope (SEM).

Finally, a prototype was constructed where a segment of the inner wall of large size LNG storage tank was hybrid welded in both horizontal (2G) and vertical (3G) positions as a simulation to the onsite manufacturing. The prototype was constructed twice where its quality was confirmed using non-destructive and destructive examinations.

The type of filler metal is an important factor in achieving satisfactory HLAW joints on 14.5 mm thick sheets of 9% Ni steel. The single pass HLAW process with its low heat input and high welding speed has resulted in a narrow width for both WM and HAZ, which is of remarkable importance for maintaining the mechanical properties. A stable and reproducible welding process with good weld formation was achieved in a relatively wide parameter window for both filler materials tested, the austenitic ERNiCrMo-3 filler wire and the less expensive matching ferritic 11% Ni filler wire.

In the case of the austenitic filler, a drop in the strength of the laser hybrid weld was to be expected due to the relatively soft austenitic WM indicated by the hardness measurement in Fig. 8a. Nevertheless, the tensile test showed reproducibly good results for all joints made with the austenitic filler material ENiCrMo-3. The ultimate strength achieved was equal to that of the BM and therefore met the requirements of the ASME code. In fact, the tensile specimens were fractured in the BM far from the WM and HAZ. Apparently, the local width of the undermatched austenitic WM is small compared to the plate thickness and has a negligible negative effect on the global weld strength [51].

The use of experimentally produced 11% Ni ferritic filler wire played a significant role in maintaining not only the tensile strength but also the impact toughness of the HLAW joint, despite the martensitic structure of its weld zone. This is attributed to the fact that the chemical composition of the weld zone, particularly the Ni content, is similar to that of the BM and the weld zone is relatively small. It is noteworthy that the impact energy correlates with the heat input of the welding process, which is controlled by the wire feed speed or welding current. The heat input of the HLAW process in the range 0.45 kJ/mm (v_w 14 m/min) to 0.50 kJ/mm (v_w 18 m/min) proved to be good for maintaining sufficient impact energy. At a low wire feeding speed v_w 8 m/min, a relatively low heat input of 0.41 kJ/mm is achieved. The amount of filler material added is low and the cooling rate of the weld zone is closer to that of laser beam welding, where insufficient impact energy was found for the 14.5 mm thick X8Ni9 sheets [41].

An acceptable impact toughness of hybrid laser-arc welded joint could not be maintained using the commercially available austenitic filler wire ERNiCrMo-3 since its major volume is deposited at the weld zone top and it is difficult to be penetrated toward the weld root as a result of narrow weld root width. The ductile austenitic structure is obtained mainly at the weld top zone while less ductile structure is formed at the weld lower zone. The problem with insufficient penetration of the austenitic filler metal in the HLAW of a ferritic BM has a high potential for optimization. There is a great practical interest in carrying out the HLAW process in such a way that better mixing is achieved. The research work of Muhammad et al. [52] proves experimentally that the direction of wire feeding affects the mixing behavior. The study shows that in GMA leading configuration is an unfavorable solidification front in the weld pool present that may obstruct the circulation of the melt flow between the weld surface and root and reduce the mixing of weld metal. In the GMA trailing configuration, the flow recirculation is more pronounced and mixing is more efficient than in the GMA leading case. The study emphasizes that the absence of such recirculation in the GMA leading configuration can be seen as the main reason for having low mixing at the bottom part of the weld. This observation results in the recommendation to perform the HLAW process in a GMA trailing configuration. However, experience shows that such process configuration is only feasible up to a sheet thickness of approx. 10 mm. With higher wall thicknesses of welded parts, problems with process instabilities such as increased droplet formation will occur [53]. The use of thin nickel foil strips in the joint can promote the austenite formation in the complete cross section of the HLAW weld [54]. In [55] a TIG welding process is described in which a 0.5 mm thick Ni interlayer is introduced into the groove to improve the cryogenic toughness of the 9% Ni welded joint. Another method to realize a homogeneous distribution of the filler metal in the weld seam for thick plates is a so-called “two-step process”, in which the austenitic filler metal is applied to the edges of the weld partners by means of laser metal deposition (LMD) in the first step. In a second step, the parts coated in this way are HLAW welded [56].

Thus, it can be seen that methods for solving the problem of inhomogeneous mixing of filler metal can be proposed, although they are associated with the introduction of additional technological steps in the production procedure. The use of these techniques for welding of thick sheets X8Ni9 for LNG sector is fundamentally possible, but has not been investigated within the framework of the present study.

The results obtained indicate that the controlling factors in producing a welded joint with an acceptable combination of tensile strength and impact toughness are both the chemical composition and the size of the weld zone as a function of the filler wire and the HLAW process parameters.

All the welding experiments discussed above were carried out only in the flat welding position (1G), as shown in Fig. 2a. On the other hand, the actual on-site manufacturing of large LNG storage tanks is carried out using not only flat welding position, but also horizontal (2G) and vertical (3G) positions [14, 57]. In this concern, the optimum welding parameters achieved for flat welding position (Fig. 2a) were properly adapted for both horizontal position (Fig. 2b) as well as vertical-down position (Fig. 2c), which in turn maintained high weld quality. While the laser power (P_L) and welding speed (v) remained constant, the wire feed speed (v_w) had to be decreased in order to reduce the molten pool volume and to prevent the molten metal from flowing out of the weld zone. Proper welding parameters for well processed single-pass HLAW of 14.5 mm thick X8Ni9 plates in different welding positions are summarized in Table 3.

Based on these proper welding parameters, successful attempts were made for prototype manufacturing where a wall segment of large size LNG storage tank was fabricated using HLAW of 14.5 mm thick cryogenic 9% Ni steel plates. Concerning this, welding was conducted in both horizontal (2G) and vertical (3G) positions as a simulation to the actual onsite manufacturing. This prototype, with an overall dimension of approximately 500 mm x 1000 mm, was twice fabricated and its quality was confirmed using non-destructive and destructive tests.

Figure 13 shows schematic illustration of LNG storage tank (a), HLAW in horizontal (2G) and vertical (3G) positions of prototype of a wall segment of LNG storage tank (b), overall view of fabricated prototype as a simulation to the onsite manufacturing (c) and macrograph of a cross section of fabricated prototype showing well processed weld bead with full penetration, narrow and sound weld confirming high quality (d). This successfully manufactured prototype is an important practical step for introducing laser-based welding technology to the LNG industry to replace the currently used conventional arc welding technology.

The type of filler metal as well as the applied process parameters play an important role in the HLAW of 14.5 mm thick plates of 9%Ni steel in order to obtain satisfactory welded joints.

The full penetration single pass HLAW process, with its moderate heat input and high welding speed, resulted in a narrow width for both WM and HAZ, which is of remarkable importance for maintaining the mechanical properties of the welded joints. With appropriate selection of welding parameters, a stable and reproducible HLAW process with good weld formation was achieved for both filler materials tested, the austenitic filler wire ERNiCrMo-3 and the experimentally produced ferritic filler wire ERNi11 with 11% Ni.

The satisfactory tensile properties of the hybrid welds were obtained for both the austenitic filler wire ERNiCrMo-3 and the experimentally produced ferritic filler wire ERNi11. The local width of the undermatched austenitic WM is small compared to the plate thickness and has a negligible negative effect on the global weld strength. All tensile specimens tested fractured in the BM far from the WM and HAZ.

While the use of the ferritic filler wire ERNi11 resulted in satisfactory impact energy values, the austenitic filler wire ERNiCrMo-3 did not result in acceptable impact toughness at the test temperature of -196 °C. This is due to the fact that most of the austenitic filler material was deposited at the top of the weld zone without reliable penetration to the root of the weld. Additional technological steps are required to achieve uniform distribution of the austenitic filler material in the narrow laser hybrid weld.

The results obtained indicate that the controlling factors in producing a welded joint with an acceptable combination of tensile and impact strength are both the chemical composition and the size of the weld zone as a function of the filler wire and the HLAW process parameters.

The efficiency of the HLAW process for the LNG industry has been successfully demonstrated by a prototype construction where a segment of the inner wall of large size LNG storage tank was successfully fabricated. Concerning this, HLAW was performed in both horizontal (2G) and vertical (3G) positions simulating the actual production conditions for the construction site. This is of remarkable importance for introducing HLAW technology to the LNG industry for possible replacing the currently implemented conventional arc welding technology.