Introduction

Heat exchangers have an exceptional role in various industrial applications for effective heat transfer process. Heat exchangers are found in different sizes, configurations, designs, and are manufactured using various types of materials to serve the intended purpose. The most prevalent type of heat exchangers utilized in industries are shell-and-tube heat exchangers, due to their simple design, reliability, and incredibly effective heat exchange rate1. A set of tubes with their ends fastened in tubesheets makes up a shell and tube heat exchanger. The transfer fluids are separated by the tube set encased by a cylindrical shell as they pass through and around2. Prevention of the mixing of transfer fluids is utmost important and therefore, tube-to-tubesheet joint (TTJ) is crucial for the shell-and-tube heat exchanger durability3. Manufacturing methods and processing parameters have a substantial impact on TTJ integrity. The primary techniques for joining the tube and tubesheet are welding, tube expansion, or combining the welding and tube expansion processes. Each technique has its own advantages and disadvantages and preferred over the others in specific industries or applications. For instant, expansion process offers a tight contact between the tube and the tubesheet and can be done with a relatively low cost to uphold the axial load, but it produces a high residual stress at the surface of contact, and fluids may pass through the clearance between the tube and the tubesheet due to unsealed joint tip4. Welding is also widely used due to its cost effectiveness and ease to perform. However, welded joints are more vulnerable to mechanical failures such as fatigue and stress corrosion cracking (SCC)5. Combining welding and expansion can enhance the performance of the joint significantly, where the sequences of hybrid welding-expansion process either involves welding followed by expansion or vice versa, yet the residual stresses from heat dissipation during welding, combined with the stresses from the plastic deformation of the tube, greatly alter the strength of the tube-to-tube sheet joints6. It is worth mentioning that an extensive increase of the residual stresses beyond an optimum value may decrease the life span of the joint. Therefore, using the proper amount of pressure and conducting welding in a professional manner are essential. The combined process is mainly preferred to increase the joint strength of heat exchangers exposed to a high operating pressure or used in critical operations as in power plants. It’s commonly preferred to perform welding prior to expansion, because the residues of lubricant used in roller expansion process may lead to the formation of weld porosities at the surface7. Figure 1 illustrates an engineering drawing for a roller-expanded joint with 2 tube hole grooves configuration.

Fig. 1
figure 1

Schematic drawing for an expanded TTJ with double grooves.

Full size image

In many applications, unplanned shutdown due to failure in heat exchangers joints may lead to a huge financial loss in industries8. Failure analysis done by Xu et al.9 showed that heat exchanger tube cracking was caused by stress corrosion cracking, with residual tensile strains due to seal expansion being the stress source. Also, it was found that corrosion or pitting reduces the thickness of the tube wall abruptly, resulting in leakage10. The principal cause of a nuclear power plant heat exchanger failing to work, as reported by Yang et al.11 was an inadequacy of seal expansion in the heat exchanger. Adullah et al.12 suggested that the reasons of the early collapse of heat exchanger in a gas plant include stress corrosion cracking and inappropriate tube-to-tube sheet welding. As per Thekkuden et al.13 there have been several instances of premature heat exchanger failure as a result of stresses developed at tube- to-tubesheet joints. Residual stresses are highly likely to damage welded joints, where propagation of cracks is expected at the tube-to-tubesheet (TT) welded joints due to overload mechanisms14,15. Throughout recent years, a large number of TT welded joint failures in shell-tube heat exchangers have been researched16,17,18. Considering literature, several factors such as the selection of manufacturing process, inappropriate welding parameters, under expansion, over expansion, weld defects, inexpert welder or labor, residual stress, corrosivity of transfer fluids, corrosion resistance of tube and tubesheet materials, inappropriate selection of seal or strength welding are found the causes for the heat exchanger failure at the joint. These conclusions emphasize the necessity of reliable manufacturing of TTJs to improve heat exchanger structural performance. Even though the manufacturing processes were established a decades ago, premature failures of tube and tubesheet joints draw attention to need for careful evaluation of geometrical conditions, manufacturing processes and process parameters.

To give an outline about the tube expansion process, the three categories of expansion method can be classified as mechanical rolling, hydraulic expansion, and explosive expansion5. Tube expansion by rolling is a process of cold working the end of the tube to fit to the tubesheet using rollers. The outer diameter, inner diameter, and length of a tube increases as the tube is extended, but the tube wall thickness tends to reduce19. In explosive forming, a controlled explosion of specified and correctly positioned explosive charges generates substantial pressures in the tube hole, resulting in nearly immediate expansion of the appropriate sections. Despite the surface polish of the tubesheet hole, the outside of the extended lengths of tube are forced into the tubesheet holes, resulting in a high integrity connection20. In the case of hydraulic expansion, tubes are plastically deformed by a pressurized liquid. The tube material deforms till the tube hits the tube hole wall due to the pressure generated by the liquid. By raising the pressure, the tube is forced against the receiving tube hole’s wall and tube wall thinning occurs21. However, expanded joints only are vulnerable to axial load, which makes them unsuitable for applications requiring great axial load persistence. The tubes came out from the tubesheet for all expansion percentages and groove conditions in the study conducted by Thekkuden et al., because the strength of the expanded joint was lower than the axial strength of the tube. These findings emphasize the importance of the use of welding in addition to tube expansion process for enhancing the strength of the TTJs22. Another common technique for fabricating TTJS is hybrid welding- expansion method23. The impact of TIG welding and expansion percentages on the strength of carbon steel-based TTJs was investigated by Thekkuden et al.24 and concluded that the hybrid welding and 3% expansion percentage in roller expansion process togethor was adequate for constructing TTJs with a strength greater than tube strength.

Tubes and tubesheets can be made from different metals and are joined in a range of ways demanding the necessity for more research in the area of TTJS. Table 1 lists some of the materials used to manufacture tubes and tubesheets, in addition to the joining method as mentioned in literature.

Table 1 Commonly used material to manufacture TTJ.
Full size table

On the other hand, in certain industries where fouling and system purity are a concern, polymer heat exchangers are found to be suitable for those situations31, in such heat exchangers, Iftikhar et al. recommended32 to use a novel technique called friction stir welding to fabricate TTJS in thermoplastic heat exchangers. Friction stir welded thermoplastic TTJS are proven to have joints with excellent strength and leak path compared to using commercial adhesive in the conventional welding technique.

Many FEM models have been put forth and used in the past to predict residual stress in TTJS. Using an axisymmetric FE model, Xu et al. predicted the thermo-mechanical stresses, including welding residual stress, in a tube-to-tube-sheet weld. A numerical analysis was also performed on the thermo-mechanical stress distribution, with and without welding residual stress. It was discovered that the stress state in the tube-to-tubesheet joint was greatly impacted by residual stresses from welding33. Merah et al. investigated the influence of initial clearance on contact pressure and percent tube wall reduction using finite element analysis. The results revealed that the initial clearing impact is insignificant for low strain hardening materials26. Yu Wan et al. used X-ray diffraction and the finite element method to investigate the residual stress distribution in 316 L stainless steel-based tube-to-tubesheet welded joints. It was determined that substantial tensile residual stresses are formed in TTJS, and it was suggested that duplex welding be used to reduce residual stress34.

Several factors call for the use of Titanium-based TTJ in shell and tube heat exchanger applications over other materials in the market. For instance, its excellent corrosion resistance makes it ideal for power condensers and desalination plants applications in seawater environments35. This resistance includes a variety of industrial conditions, leading to reduced maintenance and extension of the life of heat exchangers36. Furthermore, Titanium has excellent tensile strength, making it very desirable for high stress applications. In Cooper et al. work, Ti Gr2 based TTJs showed an exceptional tensile strength of 403.8 ± 4.2 MPa, a higher value than the parent material’s tensile strength37. In addition, it is less dense than metals like stainless steel, another high corrosion resistance metal commonly used for fabricating TTJs, which is particularly useful in some situations where weight is a concern37. On the other hand, Titanium has its own drawbacks, starting with lower thermal conductivity when compared to other metals like copper-nickel, which can reduce heat transfer efficiency unless thin walled-tubes are used35,38. Also, the cost of Titanium compared to other metals like copper can be a limiting factor, especially when utilized in large scale applications35. The welding process can also be challenging especially in thin-walled components, and for this, special methods like friction welding or gas tungsten constricted arc welding are recommended to be used to avoid defects39. Moreover, while titanium is very corrosion resistant, it can suffer from pitting corrosion and transgranular cracking under certain conditions, such as contact with certain chemicals at elevated temperatures36. Despite these drawbacks, titanium remains a strong candidate for applications requiring specific properties such as corrosion resistance, strength, and light weight. On the other hand, when comparing titanium grade 2 to titanium grade 5, another typical grade of titanium used to make TTJs, the two have differing weldability and mechanical characteristics. However, the major distinction has to do with their resistance to crevice corrosion at different temperatures. Both grades are resistant to stress corrosion cracking and pitting corrosion, but Grade 2 is susceptible to crevice corrosion at about 80 °C, while Grade 5 becomes susceptible at 200 °C, especially in a CO2 rich environment, as found in Pang et al. work40. These temperature dependent corrosion behaviors are very important in choosing the right alloy for a particular application, particularly when the service conditions are aggressive.

In this current work, Pure titanium grade 2 is used as the tube and tubesheet materials. Due to all mentioned advantages, Titanium grade 2 may be used in various heat exchange applications such as desalination equipment or submarine equipment, aerospace, chemical processing facilities, oil refineries, and power plants41. Nevertheless, there is a dearth of information in the literature regarding the impact of process parameters and manufacturing techniques on the performance of costly pure titanium grade 2. Moreover, a study on the effect of doubled tube hole groove inclusions on the expanded tubes at various expansion percentages is necessary for evaluating whether the tubes are under expanded or over expanded based on the tube pull-out strength. In addition, studies comparing the hybrid welded-expanded, welded only and expanded only are essential to compare the manufacturing methods based on tube pull-out strength. In many of the known articles, strength welding on protruded tubes has been given importance whereas in the current work, influence of seal welding on flush joint configured tube is investigated. In nutshell, the integrity of pure titanium grade 2 TTJS is proposed rather than the widely reported carbon steels. The proposed work is expected to be beneficial for engineers and researchers for understanding the behavior of joint characterizations with respect to manufacturing techniques and geometrical parameters. In addition, this work also provides in-depth analysis of the physical metallurgy associated with titanium tubes and tubesheets under various conditions of expansion percentages and welds.

Experimental procedure

Fifteen TTJS were fabricated of commercially pure Titanium grade 2 (CP Ti Gr2) tubes and tubesheets. Mock-up tubesheets with a size of 40 × 40 × 25 mm, double tube hole grooved configuration and 19.25 mm drilled tube hole are fabricated as shown in Fig. 2. Double groove on the tube hole is considered over single grooved and without any grooved tubesheets for the current experimentation because the bulging of the tube to the tube hole grooves by tube expansion is expected to increase the strength of the joint42. Improving the strength of the tube to tubsesheet joint is given utmost importance for the fabrication of shell and tube heat exchangers. Another way of improving the strength of the joint apart from the inclusion of tube hole grooves would be to consider U groove, V groove and J groove at the tubesheets, so that the molten electrode would be deposited more causing more fusion between tube and tubesheet. Each groove has a width of 3 mm and separated by a distance of 6 mm, where the first groove is fabricated 6 mm away from the chamfer side of the tubesheet. The chamfer at the outer edge of the tubesheet is 1.5 mm long at 45°. The sectional view with dimensions of the tubesheet is shown in Fig. 3.

Fig. 2
figure 2

Tubesheet mockup block.

Full size image
Fig. 3
figure 3

Sectional view of mock-up tubesheet (All dimensions are in mm).

Full size image

Table 2 provides the chemical composition of CP Ti Gr2. Three different cases of manufacturing process were taken into consideration when fabricating the joints, namely expansion only, welding only, and hybrid welding-expansion (welding followed by expansion) processes. Tubes with 19.05 mm outer diameter and 1.53 mm of thickness were placed to the tubesheet using flush joint configuration. The TTJ welds were fabricated by Tungsten inert gas welding process using INVERTEC 300TPX powersource and DRATEC DT-ErTi2 consumable. The chemical composition of DRATEC DT-ErTi2 filler rod is shown in Table 3. Two welding passes—one for the root and one for the cap—were made during the welding process at room temperature. The reasons for welding using TIG are less heat input, controlled weld metal deposition and less weld spatter. SPOTCHECK SKC-S and CRC Lectra Clean was used to clean the spots of welding. The roller expansion is performed using 99 series electric rolling motors, tube expanders and ELC110220 electric torque controller. Figure 4 shows the fabrication setup of the joints. In the expansion process, tubes were expanded for three different wall thinning reduction percentages, 4%, 6%, and 8%. Table 4 lists the various configurations of the fabricated joints for testing purposes. Figure 5 shows the fabricated joints in the order mentioned in Table 4. Table 5 provides pre-and-post rolling expansion dimensions for expansion percentages related to expansion only, welding only, and hybrid welding-expansion procedures. The mechanical and metallurgical evaluations of TTJS are done by inspecting minimum leak path (MLP), pull-out load, hardness, microstructures and phase determination. The macro examination is conducted in compliance with the requirements outlined in ASME Section (IX) standard. Validation criterions specified in ASME SEC. IX Ed. 2019-QW-193.1.1 and 2019-QW-193.1.3 are used to evaluate the qualification of welded TTJS by measuring the weld throat to determine the MLP43. To assess the pull-out strength capacity of the joint, a pull-out test was performed on all joints including welded only cases, expanded only cases, and hybrid welded-expanded cases following Appendix A criterion in ASME SEC. VIII DIV. I standard44. The Vickers hardness of base material, weld zone, and heat affected zone are measured under the conditions of 10 kgf test load and square-based diamond pyramid indenter and, the results are compared with the acceptance criteria in NACE MR0175 standard. For macroscopic analysis, a specimen from the welded TTJ was grinded with abrasive silicon carbide (SiC) paper with various grit sizes (P180, P220, P400, P800, and P1200) as a preparation for MLP test. For microscopic analysis, multiple specimens from tube base material, tubesheet base material, and welded TTJ were prepared by undergoing the same griding steps, then followed by cloth polishing process using Aluminum Oxide solution. Subsequently, they were thoroughly washed with distilled water and dried with compressed air to accomplish a mirror like finish as shown in Fig. 6. To gain insight into the microstructural features, such as grain size, crystallographic structure, and phase composition of the specimens, Kroll’s reagent (1.5% vol/vol HF + 4% vol/vol HNO3) was used as an etchant, where the polished surfaces were rubbed by the reagent then left for 3 min etching time in the case of tubesheet base material and welded TTJ specimens, and 30 s in the case of the tube base material specimen. mean lineal intercept technique was used to measure ASTM grain sizes for all specimens.

Table 2 Chemical composition of CP-Ti (Grade 2)45.
Full size table
Table 3 Chemical structure of DRATEC DT-ErTi2 welding rod.
Full size table
Table 4 Configurations of the fabricated joints.
Full size table
Fig. 4
figure 4

(a) Welding setup (b) Welding power source (c) Tube roller expansion setup (d) Fixture (e) Samples of welded coupons.

Full size image
Fig. 5
figure 5

Fabricated TTJS in the experimental order given in Table 4.

Full size image
Table 5 Dimensions before and after fabricating the joints.
Full size table
Fig. 6
figure 6

Surface finish after grinding and polishing tubesheet base material for metallurgical test.

Full size image

Results and discussion

Microstructures of CP Ti gr. 2 tube and tubesheet base materials

The microstructure of the parent materials of the tube and the tubesheet observed using optical microscope are shown in Figs. 7 and 8, respectively. Typical hexagonal closed pack (HCP) structure with equiaxed grains is present in both parent materials46. The pitting spots evident on the surface of the tubesheet etched specimen formed due to the breakdown of the passivity film on the surface47. The average grain size for tube and the tubesheet base materials were measured to be 19 μm and 44 μm respectively.

Fig. 7
figure 7

Microstructure of tube base material. (a) ×100 magnification. (b) ×200 magnification. (c) ×500 magnification.

Full size image
Fig. 8
figure 8

Microstructure of tubesheet base material. (a) ×100 magnification. (b) ×200 magnification. (c) ×500 magnification.

Full size image

Microstructures of welded joints

A coupon was sectioned from the welded joint at the weld region to examine the effect of welding on the microstructure of the tube-to-tubesheet joint. Four regions of interest were examined, namely, the welded zone (WZ), the heat affected zone (HAZ), and the base material (BM) of the tube and tubesheet. Figures 9, 10, 11, 12, 13 and 14 show the optical microscopy images of weld zone, HAZ, and base material zone formed between the tube and the tubesheet. The microstructure evident in the base material is HCP α-phase equiaxed grains. Moving from BM toward WZ through HAZ, the grains become coarser due to complex thermal cycling occurred during welding, forming noticeably larger grains as shown in Fig. 1048. The microstructure of the HAZ shows a different characteristic when compared to the microstructure of BM. The polyhedral grains in BM are replaced with coarser polyhedral α′-grains with acicular morphology in HAZ. This observation is consistent with the findings of the research conducted by BÁRTA et al.49. This transition in morphology occurred due to the allotropic transformation of β -phase grains to α-phase grains as it cools down at a slow rate50. Cao et al. found in their work that heating and cooling rate play crucial part in the formation of the metastable α′ martensite phase51. Also, studies confirmed that α′ phase exists in greater amounts in fusion zone (FZ) compared to HAZ in titanium welds, which is a typical result of welding processes involved in Titanium alloys52,53,54. When comparing the microstructure of the WZ to HAZ, both exhibit the same α′ microstructure. However, the grain size in the WZ is greater than the grain size in the HAZ. This difference in size can be attributed to the different cooling process experienced by both zones, as the molten metal must cool down from a greater temperature in the WZ compared to the HAZ55,56. According to the literature, cooling rate has a significant impact on the microstructure of titanium welds57,58. Gupta et al. established in their work that heat input due to welding technique used has a significant effect on the martensitic microstructure and mechanical properties of commercially pure Titanium. It was found that the strength increases as the temperature drops from room temperature to 20 K, elongation percentage has been determined to be maintained or increased as well. Furthermore, in the case of CP Ti Grade 2, TEM images suggested that weldment microstructure shows acicular martensitic α′ microstructure along with twinning59. An XRD analysis was conducted on specimens taken from the tube, tubesheet, and the weld to verify the match in findings. It showed that all peaks obtained for all specimens form at the same angles, aligning with the peaks formed by Titanium’s α-phase. Since α′-phase and α-phase grains are known to have similar diffraction patterns, it can be inferred that the observed polyhedral grains in the WZ and HAZ are α′-phase grains.

Fig. 9
figure 9

Microstructure of the weld zone. (a) ×100 magnification. (b) ×200 magnification.

Full size image
Fig. 10
figure 10

Microstructure of the HAZ near the weld. (a) ×100 magnification. (b) ×200 magnification.

Full size image
Fig. 11
figure 11

Microstructure of the weld zone near the tube. (a) ×100 magnification. (b) ×200 magnification.

Full size image
Fig. 12
figure 12

Microstructure of the HAZ near the tube. (a) ×100 magnification. (b) ×200 magnification.

Full size image
Fig. 13
figure 13

Microstructure of the transition zone from weld zone to tube base material through HAZ. (a) ×100 magnification. (b) ×200 magnification.

Full size image
Fig. 14
figure 14

Microstructure of the tube base material. (a) ×100 magnification. (b) ×200 magnification.

Full size image

Microscopic analysis of expanded tube

One of the most important aspects of joint integrity is using an appropriate expansion percentage. According to Farrahi et al.‘s investigation, one crucial factor affecting the tube-to-tubesheet joint’s performance is the residual stress introduced through expansion60. In contrast, Haneklaus et al. stressed in their work how crucial it is to apply an initial contact pressure in order to help ensure subsequent partial diffusion bonding between the tube and the tubesheet61. In order to establish sufficient contact between the tube’s exterior and the tubesheet, tubes expand plastically during the roller expansion process. A number of factors, including friction between the tube and tubesheet, tube thickness, residual stress, and initial clearance, affect the strength and integrity of the joint. Al-Aboodi et al.‘s study demonstrated that decreasing cutoff clearances and increasing residual contact stress between tube and tubesheet are the consequences of increasing friction. The residual contact stress also increased linearly with increasing tube material strain hardening level for all friction coefficients examined in their study62. In addition, a FE analysis was conducted by Merah A et al. to assess the combined effects of large initial clearance and strain hardening of tube material on interfacial pressure and tube deformation. Results showed that excessive expansion due large initial clearance between the tube and the tubesheet compromises the robustness of the joints, as residual contact stress progressively rises until it drops at lower cut-off clearances63. The impact of expansion percentage on the degree of plastic deformation of the tube-to-tubesheet joint has been investigated by comparing the microstructure of 4%, 6%, and 8% expanded joints to the base metal. To achieve the aforementioned expansion percentages, the tubes were expanded to achieve inner diameter − 16.64 mm, 16.69 mm, 16.75 mm respectively. The effect of expansion percentage on the average grain size at the tube cross-sections close to inner and outer regions of the expanded/unexpanded tube at three distinct zones—the expanded, the transition, and the unexpanded zones—was determined by performing microscopic analysis. Figure 15 depict the microstructure of the 4%, 6%, and 8% expanded joints at the expansion zone respectively. As the inner sides of the tubes in all cases were in direct contact with the rollers, finer grains were observed at those regions compared to the side in contact with the tubesheet. For the three expansion percentages, 4%, 6%, and 8%, the average grain sizes were measured to be 23.72 μm, 21.81 μm, and 18.57 μm with an extent of expansion affecting the radial depth of the tubes by 307 μm, 333 μm, 365 μm respectively. Another notable observation is that the 8% expansion caused the greatest amount of twinning in the grains. Accordingly, the greater the expansion percentage is, the smaller the grains in the expanded region, and the more occurrences of grains’ twinning. For the outer edge of the tubes, the average grains sizes were measured to be 25.26 μm, 24.74 μm for the 4% and 6%, and 24.72 μm for the 8% expansion percentages, indicating that the expansion process has a minor effect on the outer side of the tube. Moving towards the transition zone, the average grain sizes were measured to be 23.72 μm, 22.13 μm, and 20.71 μm using Fig. 16 for the three consecutive expansion rates, indicating that expansion process has little to no effect at the transition zone. Thus, based on the findings of the microstructure analysis, 4% expansion rate causes the least refinement in the grains at the inner surface of the tube, whereas the 6% and 8% expansion rates have more pronounced effect on the grains size, formation of twinning, and the radial expansion extension from the inner surface of the tube towards the centre of the tube. In contrast, all three cases showed little to no effect on the outer surface of the tube at the expansion zone, as well as a negligible effect away from the roller and in the transition zone.

Fig. 15
figure 15

Microstructures of the tube at the expanded zone at the inner and outer tube cross-sections (×200). Red mark shows the influence of roller expansion on the inner surfaces of the tube.

Full size image
Fig. 16
figure 16

Microstructures of the tube at the transition zone at the inner and outer tube cross-sections (×200).

Full size image
Fig. 17
figure 17

Microstructures of the unexpanded tube at the inner and outer tube cross-sections (×200).

Full size image

Minimum leak path

MLP is the shortest path along which the transfer medium can leak, and this value can be estimated by measuring the distance in any direction from the root of the tube-to-tubesheet weld to the nearest surface. MLP for both welded and welded-expanded joints were measured from macro images of a magnification factor of 10× as shown in Fig. 18. Table 6 presents the measured values of MLP for welded joints, as well as welded-expanded joints at expansion rates of 4%, 6%, and 8% respectively. To assess the welds in this test, the criteria stated in ASME Boiler and Vessel Pressure Code, 2004 edition is implemented to qualify the welded joints based on clause QW-193.1.343. Weld qualification is accomplished when the weld throats equal or surpass two by three of the tube wall thickness, and no faults are seen during macro inspection at a magnification of 10X. The thickness of the tube is 1.44 mm, thus the weld throat or MLP should be greater than two-thirds of this value, which is 0.96 mm. Examination of the welded joints revealed that all measured weld throats are greater in magnitude than two-third of the thickness of the tube, indicating that all joints are qualified based on QW-193.1.3 clause in ASME Section IX. In addition to the MLP, the macrostructure clearly indicated the excellent fusion of titanium grade 2 tube to titanium grade 2 tubesheet without any presence of porosities, blow holes and burnthrough defects. Lack of weld defects along with satisfactory MLP indicated the appropriateness of the welding parameters.

Fig. 18
figure 18

Macro images of sectioned TTJ at 10× magnification. (a) Welded only joint. (b) Welded-expanded joint with 4% expansion percentage. (c) Welded-expanded joint with 6% expansion percentage. (d) Welded-expanded joint with 8% expansion percentage.

Full size image
Table 6 MLP of TTJS.
Full size table

Tube pull out strength

For qualifying the strength of TTJS, tube pull-out test was conducted according to ASME Sec-VIII, DIV-1, Non-Mandatory Appendix A, standards to determine the maximum allowable loads for tube-to-tube-sheet joints44. The test was done on seven different specimens to assess the strength of the various scenarios of the study, including welded-only joints, joints expanded only by 4%, 6%, and 8%, and welded-expanded joints at 4%, 6%, and 8% expansion percentages, as shown in the corresponding Fig. 19. Figures 20, 21 and 22 show a top view of the fractured joints after conducting the test. Figure 23 depicts the load-displacement data produced from the tube pull-out test. The results show that hybrid welded-expanded joints have greater pull-out capacity than expanded-only joints for the same expansion percentages. The pull-out peak loads for welded-expanded joints with 4%, 6%, and 8% expansion percentages were determined to be 36.080 kN, 36.680 kN, and 36.200 kN, respectively, compared to significantly lower values of 11.310 kN, 15.910 kN, and 14.240 kN for expanded-only joints with 4%, 6%, and 8% expansion percentages in the same order. On the other hand, the results of welded-only joints test were very close to those of welded-expanded joints, achieving a peak load of 34.020 kN. Nevertheless, failure has occurred at the weld joint region. The results of 4% and 8% welded-expanded joints were very similar. This similarity is attributed to the fact that fracture occurred within the tube itself for both 4% and 8% expansion rates, away from the joint area, implying that the obtained results are for the tube rather than the tube-to-tubesheet joint for these two cases. Furthermore, no slippage or cracking in the joint region was visually observed, providing a strong support for this conclusion indicating that the two joints are qualified according to ASME BPVC, Section VIII, Division 1 standards since the joint strength is greater than the tensile strength of the tube. The welded-expanded joint with an expansion rate of 6%, on the other hand, failed at the joint region, yet the peak load bearing capacity was found slightly greater than the bearing capacity of the other two welded-expanded specimens. In this case, the failure at the weld may be due to the flush joint configuration causing limitation in the deposition of the molten electrode over the thin tube region. Moreover, welding of the thin titanium tubes needs careful attention while welding because of the chance for burnthrough defect, however the formation of burnthrough defect is avoided by proper welding parameters. Despite the limitation of the flush joint configuration, flush configuration is preferred over protrusion configuration as the chance for melting of thin titanium tubes and burthrough are more while using protrusion configuration. The failure at the weld zone in the case of welding followed by 6% expansion indicate that the weld strength was close to the tube strength. To assess the integrity of this joint, the maximum allowable axial load was calculated in accordance with ASME VIII - Basis for establishing allowable loads for TTJS, using Eq. (2), and subsequently compared to the obtained peak load from the tube pull-out test. The calculations showed a maximum allowable axial load of 5.92 kN, which is a much lower value than the peak load of the pull-out test of 36.680 kN at which the joint experienced failure. Accordingly, the welded-expanded joint at an expansion percentage of 6% is qualified. Although seal welding was employed in this study, it had a substantial effect on the pull-out load capacity of the TTJS. On the contrary, the ones that were expanded without welding had a lower pull-out load capacity. Considering that both scenarios experienced failure at the joint region in all cases, it was essential to obtain the maximum allowable axial loads using ASME VIII, Nonmandatory Appendix A, to compare them with the measured peak loads by the test to assess whether the joints were qualified or not, regardless of the slippage that occurred at the joint region. Equation (1) was used to determine the maximum allowable axial load for welded only joint, which was found to be 4.32 kN, a much lower value than the 34.02 kN peak load measured during the test. In the same fashion, for 4%, 6% and 8% expanded-only joints, Eq. (3) was used to determine the maximum allowable axial load, which was found to be 5.10 kN, also a considerably smaller value than the measured peak loads of the three cases, which were 11.310 kN, 15.910 kN, and 14.240 kN respectively. Since peak loads at which slippage occurred for all trials were found to be greater than the maximum allowable axial load, welded-only joints and expanded-only joints with expansion percentages 4%, 6%, and 8% were all found to be qualified in compliance with ASME BPVC VIII standards.

$$\:{L}_{max}={A}_{t}{S}_{a}{f}_{r}$$
(1)
$$\:{L}_{max}=MIN\left[{A}_{t}{S}_{a}{f}_{re},\:{A}_{t}{S}_{a}\right]\:$$
(2)
$$\:{L}_{max}=MIN\left[{A}_{t}{S}_{a}{f}_{e}{f}_{r}{f}_{y}{f}_{T},\:{A}_{t}{S}_{a}\right]\:\:$$
(3)

where At is the tube cross-sectional area, Sa is the allowable stress for tube material. fre is the factor for the overall efficiency of welded and expanded joints, fe is the factor for the length of the expanded portion of the tube, fr is the tube joint efficiency, fy is the factor for differences in the mechanical properties of tubesheet and tube materials, fT is the factor to account for the increase or decrease of tube joint strength due to radial differential thermal expansion at the tube-to-tubesheet joint.

Fig. 19
figure 19

TTJS used for tube pull-out test. (a) 4%, 6%, and 8% wall reduction expanded joints. (b) 4%, 6%, and 8% wall reduction welded-expanded joints. (c) Welded only joints.

Full size image
Fig. 20
figure 20

Top view of welded only joints post tube pull-out test, demonstrating a slippage at the joint region. (a) First welded only specimen. (b) Second welded only specimen.

Full size image
Fig. 21
figure 21

Top view of expanded only joints post tube pull-out test, demonstrating a slippage at the joint region. (a) 4% expansion percentage. (b) 6% expansion percentage. (c) 8% expansion percentage.

Full size image
Fig. 22
figure 22

Top view of welded-expanded joints post tube pull-out test, demonstrating a slippage at the joint in joint (b), and fracture at tube in joints (a). Expansion rates for the specimens are 4%, 6%, and 8% respectively.

Full size image
Fig. 23
figure 23

Load versus displacement graphs for 4%, 6%, and 8% expanded, welded-expanded, and welded only joint.

Full size image

X-ray diffraction analysis

XRD test was conducted to examine the crystal structure and the present phases in the welded TTJ. Three specimens from the welded joint were tested including the tube base material, tubsheet base material, and weld material. All peaks plotted in Fig. 24 formed around angles 2θ = 35.09° (010), 2θ = 38.41° (002), 2θ = 40.16° (011), 2θ = 52.99° (012), 2θ = 62.94° (110), 2θ = 70.64° (013), 2θ = 76.20° (112), 2θ = 77.34° (021). Results indicate that all specimens are composed solely of hexagonal a-phase and match the Alpha CP Ti coded in ICDD. No other peaks were formed at other angles, implying the absence of the β-phase. The absence of β-phase is mainly attributed to Titanium’s temperature below 882 °C, as well as the low cooling rate64,65,66. The absence of peaks at any other angles supports the conclusions drawn from the microscopic analysis of the welded joint, which suggested that the phases present there consist entirely of α and α′ phases.

Fig. 24
figure 24

XRD analysis results.

Full size image

Vicker’s hardness

Vicker’s hardness were evaluated on welded joint at various zones including the base materials of tube and tubesheet, heat affected zone at the tube and tubesheet and the weld zone with a 10kgf indenting load (HV10) according to ISO 6507-1 standards. Figure 25 depicts the Vickers hardness at the above-mentioned zones. The Vicker’s hardness at the mentioned regions were found to be in the acceptable range since the measured values in all tests at the various regions did not exceed the maximum allowable value of 230 HV listed in NACE MR0175/ISO 15156-367. After welding, the base and weld metal’s hardness cannot be greater than the maximum hardness permitted for that metal and weld alloy respectively. This requirement is important because the vulnerability of Titanium alloy to sulphide-stress cracking (SSC), stress-corrosion cracking (SCC), and stress-oriented hydrogen-induced cracking (SOHIC) may be affected by the metallurgical changes that take place during welding at the weld region. The greatest hardness value, 173.6 HV, was recorded at the weld zone, whereas the greatest recorded values at tubesheet HAZ and tube HAZ near weld were 168.7 HV and 153.4 HV respectively. Additionally, the greatest recorded values at the tube and tubesheet zones were 114.9 HV and 163.8 HV respectively. The hardness spectrum indicates that the weld zone experienced the maximum hardness in comparison to other regions, and the hardness decreased as moving from the weld zone towards the base material through the HAZ. This finding contradicts the Hall-Petch effect, which states that the relationship between grain size and material hardness is inverse68. This phenomenon arises from the fact that metallic deformation requires dislocation movement in the metal, and that movement is impeded by the grain boundary. Given that the size of the grains determines the number of grain boundaries (the finer the grains, the greater the presence of grains’ boundaries), the coarser weld zone should have had the lowest hardness in comparison to other regions, because coarse grain size decreases the possibility of dislocations, which dramatically lowers the material’s hardness value69,70,71. Studies confirm that a noticeable increase in the material’s hardness is caused by the brittle martensite α′ phase, which is dominant at the weld zone. The primary cause of this discrepancy is the formation of the martensitic α′ phase in the weld zone and the heat-affected zone, leading to a substructure strengthening due to abrupt cooling72,73. Also, studies showed that the crystals orientations in the welded region have a significant impact on altering the microhardness of alpha alloys, as crossing boundaries for slip planes becomes more difficult, leading to an increase in the microhardness of the crystals74,75.

Fig. 25
figure 25

Vicker’s hardness test results for welded-only joint.

Full size image

Conclusion

High quality TTJS have been fabricated of commercially pure Ti Gr. 2 tubes and tubesheets using three different techniques common in industry to manufacture TTJS, namely welded joints, roller expanded joints, and hybrid welded-expanded joints. Three different tube wall reduction percentages were selected to fabricate the expanded joints, and welded-expanded joints, which are 4%, 6%, and 8%. The TIG welded joints met all requirements and had a flawless weld fusion without any surface flaws or cavities between the root and cap welds. The expansion rates for the roller expanded tubes were selected to study the effect of the expansion rates on TTJS at microscopic and macroscopic levels. The work’s primary conclusions can be summarized up as follows:

  • The microstructural analysis revealed a coarsening in the grains structure as moving from BM to WZ through HAZ. It was observed that HCP α-phase crystals transformed into α′-phase crystals post rapid cooling in WZ and HAZ. Also, the results of the XRD analysis confirmed the absence of β-phase at HAZ and WZ due to the indistinguishable diffraction pattern for the base material (α-phase dominant) and the weld (α′-phase dominant).

  • The microscopic analysis of the expanded tubes revealed that the extent of grain refinement is directly correlated with the expansion percentage, and the grain sizes decreased at the cross-section close to the inner surface of the tube exposed to the roller expander.

  • The MLPs of the welds were greater than two-third ratio of the tube wall thickness, or 0.96 mm as required by the ASME section IX (Clause QW-193.1.3).

  • Tube pull-out tests, conducted in accordance with ASME Sec-VIII, DIV-1 (Non-Mandatory Appendix A standards), verified that all the joints manufactured using welding, roller expansion and hybrid welding-roller expansion process was qualified with the failure load exceeding the maximum permissible axial loads.

  • The Vicker’s hardness at the weld zone, tube and tubesheet base materials, heat affected zones of tube and tubesheet were restricted successfully to be less than the maximum value of 230 HV according to NACE MR0175. The tubesheet heat affected zone indicated the region of highest hardness with a maximum Vicker’s hardness of 173.6 HV.

  • The cost of the Titanium grade 2 tubes and tubsheets are high compared to the carbon steels and stainless steels, however for heat exchangers working at high critical temperatures and corrosive environment, despite of the high cost, the importance is given to the performance of the material to suit engineering applications. The future scope of this work is to investigate the joint between Titanium grade 2 tube and titanium grade 2 cladded carbon steel tubesheets, wherein the Titanium grade 2 cladded carbon steel tubesheets reduces the cost of production considerably.