Strengthening of hot-rolled S355 steel I-section beams using WAAM high-strength steel

Contents lists available at ScienceDirect Thin-Walled Structures journal homepage: www.elsevier.com/locate/tws Full length article Strengthening of hot-rolled S355 steel I-section beams using WAAM high strength steel Jiachi Yang ∗, M. Ahmer Wadee , Leroy Gardner Department of Civil and Environmental Engineering, Imperial College London, London, SW7 2AZ, UK A R T I C L E I N F O Keywords: Directed energy deposition (DED) Experiments Hybrid manufacturing Strengthened steel I-section beams Testing Wire arc additive manufacturing (WAAM) 3D printing A B S T R A C T An experimental investigation to assess the major-axis flexural behaviour of 12 hot-rolled S355 steel I-section beams strengthened by the addition of high strength steel (HSS) through wire arc additive manufacturing (WAAM) is presented. The geometry of the beam specimens was obtained by means of 3D laser scanning. The mechanical properties of both the hot-rolled and the WAAM steel were determined through monotonic tensile testing. Physical testing of the strengthened beam specimens was conducted. The results showed that the WAAM strengthening led to dramatic increases in bending resistances of between 35% and 80% under fourpoint bending, and of between 30% and 85% under three-point bending, for increases in mass of between just 5% and 15% respectively. At the same time, the specimens exhibited good ductility, despite the high strength of the WAAM additions. The presented experimental results, which are the first of their kind, successfully demonstrate the applicability of the proposed strengthening approach for both new and retrofitted steel beams, and the game-changing potential for enhancements in structural efficiency and reductions in embodied carbon in the construction industry. 1. Introduction Wire arc additive manufacturing (WAAM), or directed energy deposition-arc (DED-arc) additive manufacturing (AM), has been shown to offer a range of opportunities for the construction sector. Extensive work has been undertaken to investigate the material mechanical properties of additively manufactured steel [1–3], while recently, WAAM has been explored as a strategy for hybrid construction and as a viable means of strengthening steel structures [4,5]. The structural performance of a series of hot-rolled steel I-section columns and beams, strengthened by the addition of WAAM steel, were examined in [6,7] through physical testing; the strength of the WAAM steel approximately matched that of the underlying steel I-sections, which was grade S355. The experimental results demonstrated that a significantly enhanced load-carrying capacity was achievable for the strengthened members with a relatively small percentage increase in mass. WAAM has also been explored recently to strengthen steel plates under fatigue loading [8,9], and to produce a variety of structural elements [10–14] and connections [15–19]. Another directed energy deposition (DED) AM method, laser cladding, has been used recently to repair damaged steel elements [20–23]. The residual stresses and distortions in the WAAM-strengthened steel structures have been examined through experimental methods in [24,25] and through numerical simulations in [9,25,26]. ∗Corresponding author. E-mail address: [email protected] (J. Yang). Hitherto, the strengthening of normal strength hot-rolled I-section beams with WAAM high strength steel (HSS) (i.e., with a nominal yield strength greater than 460 MPa [27]) has yet to be investigated; this, therefore, forms the focus of the present work. A total of 12 WAAM HSS-strengthened I-section beams, designed and manufactured with optimised WAAM additions, are introduced. The geometry of the strengthened specimens, obtained by means of 3D laser scanning, and the mechanical properties of both the hot-rolled I-sections and the added WAAM steel, obtained through monotonic tensile testing, are then described. Physical testing was carried out on the beam specimens to assess their major-axis flexural behaviour and the results are presented herein. Analyses of the effect and efficiency of the proposed strengthening approach are then carried out. The presented experimental results successfully demonstrate the applicability and considerable performance benefits of strengthening normal strength steel beams with high strength steel added through WAAM. 2. Specimen design and manufacture The design and manufacture of the examined WAAM HSS-strengthened beam specimens are presented in the current section. A total of 12 two-metre long (i.e., 1.8 m span 𝐿 with two 0.1 m overhangs) steel I-section beams with hot-rolled IPE160 profiles in grade S355 https://doi.org/10.1016/j.tws.2025.113437 Received 19 February 2025; Received in revised form 9 May 2025; Accepted 9 May 2025 Thin–Walled Structures 215 (2025) 113437 Available online 27 May 2025 0263-8231/© 2025 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ). J. Yang et al. Fig. 1. Schematic illustrations of (a) four-point bending tests and (b) three-point bending tests, with dimensions in mm. Table 1 Nominal cross-sectional dimensions of IPE160 profile. 𝐻 𝐵 𝑡w𝑡f𝑟 𝐴 𝐼y (mm) (mm) (mm) (mm) (mm) (mm2) (mm4) 160 82 5.0 7.4 9.0 2009 8.693 × 106 Table 2 Chemical composition of tested hot-rolled steel profiles (% by weight, balance Fe). C Si Mn P S Cr Mo 0.07 0.18 1.42 0.017 0.024 0.09 0.03 Al Cu Nb Ti N V Ni 0.013 0.28 0.041 0.018 0.009 0.008 0.12 Table 3 Typical chemical composition of ER110S-G and ER120S-G low-alloy carbon steel welding wire from wire certificates (% by weight, balance Fe). Wire grade C Si Mn Cr Ni Mo ER110S-G 0.09 0.7 1.70 0.30 1.85 0.60 ER120S-G 0.10 0.80 1.80 0.35 2.25 0.60 steel, strengthened with WAAM HSS, were designed for either fourpoint bending (4PB) or three-point bending (3PB) tests, as illustrated in Fig. 1, showing the designated loading conditions. The WAAM high strength steel was produced using low-alloy carbon steel solid welding wire in grade ER110S-G or ER120S-G (as per the specification from the American Welding Society AWS A5.28 [28]), which are the two highest-strength carbon steel welding wires currently available on the market. For each loading condition, a bare I-section specimen was also tested as a benchmark, and labelled as ‘B-4PB/3PB’. The WAAM-strengthened specimens were fabricated through metal inert gas (MIG) welding by the Dutch manufacturer MX3D using a 6-axis ABB robotic arm system. The nominal cross-sectional properties of the IPE160 cross-section are given in Table 1, where 𝐻, 𝐵, 𝑡w, 𝑡f and 𝑟 are the cross-sectional height, cross-sectional width, thickness of web, thickness of flange and root radius respectively; 𝐴 is the cross-sectional area and 𝐼y is the second moment of area about the major axis. The chemical composition of the hot-rolled I-section, as provided by the supplier, is given in Table 2, while the typical chemical composition and key nominal as-welded mechanical properties of the welding wire, as specified in the wire certificates, are provided in Tables 3and 4 respectively. The same process parameters were adopted for the two types of welding wire during the WAAM operations, as listed in Table 5, where the nominal dimensions of a single weld bead are also given. For each combination of wire grade and test type (i.e., ER110S-G and 4PB), three specimens were designed with nominally 5%, 10% and Table 4 Key nominal as-welded mechanical properties of ER110S-G and ER120S-G low-alloy carbon steel welding wire from wire certificates. Wire grade ER110S-G ER120S-G Yield strength (MPa) 790 915 Tensile strength (MPa) 880 960 Percentage elongation (%) (𝐿0= 5𝑑0) 16 20 Impact energy (J) (ISO-V KV 20 ◦C) 90 130 Table 5 Process parameters and nominal weld bead dimensions adopted in WAAM. Process parameters Value Wire diameter (mm) 1.2 Voltage (V) 18.5 Current (A) 165 Travel speed (mm/s) 7 Inter-pass temperature (◦C) <150 Metal transfer mode Short-circuiting transfer Shielding gas 98% Ar, 2% CO2 Gas flow rate (l/min) 17–20 Nominal weld bead height (mm) 1.5 Nominal weld bead width (mm) 7.5 Fig. 2. Illustrations of the strengthened cross-section and sectional dimensions (CF — compression flange, TF — tension flange). 15% percentage increases in volume, (𝑉W∕𝑉I)n, respectively, where 𝑉W is the volume of the WAAM addition, 𝑉I is the volume of the bare Isection within the 1.8 m span, and the subscript ‘n’ denotes ‘nominal’. The specimens were hence labelled in the sequence of ‘wire gradetest type-(𝑉W∕𝑉I)n’ (e.g., Specimen ER110-3PB-5%). Each specimen was designed with WAAM additions optimally deposited on the Isection profile to seek a moment resistance diagram closely matching the bending moment diagram corresponding to the applied loading within the deposition region, thereby making material utilisation more uniform and therefore more efficient. At the cross-sectional level, the WAAM material was added onto the outer surface of the tension flange in a distributed manner, whereas on the compression flange, it was deposited in the form of tip stiffeners, see Fig. 2 with the cross-sectional dimensions illustrated, where 𝑏w and ℎw are the width and height of a single weld bead respectively. This material arrangement has been shown to be capable of enhancing the bending moment resistance of I-section beams while largely preserving their ductility, as reported in [7]. The design information of two typical WAAM HSS-strengthened beam specimens, ER110-4PB-10% and ER120-3PB-5%, is provided in Tables 6and 7 respectively. The moment resistance diagram, 𝑀Rd, was established based on the cross-sectional plastic bending moment resistances 𝑀pl that vary along the length of the specimen according to the strengthening arrangement, in accordance with the specifications provided in EN 1993-1-1 [27] for Class 1 cross-sections such as the IPE160 examined herein. The anticipated vertical load(s) 𝑃 and bending moment diagram 𝑀Ed (see Fig. 1) were determined correspondingly according to 𝑀Rd. Note that 𝑀pl was determined using the nominal Thin-Walled Structures 215 (2025) 113437 2 J. Yang et al. Table 6 Design information of Specimen ER110-4PB-10% (unit for dimensions: mm). Specimen ID (𝑉W∕𝑉I)nLoad (s) 𝑃 (kN) ER110-4PB-10% 10.97% 120.4 Profile and bending moment diagrams dimensions of the IPE160 profile (see Table 1) and those of the weld bead (see Table 5), along with the assumed yield strengths of the hot-rolled and the WAAM steel. The assumed yield strengths were adopted to be 410 MPa for the S355 hot-rolled steel and 630 MPa for the ER110S-G WAAM steel, based on previous studies [7,29]; these values were considered as closer representations of the likely properties than nominal values. However, owing to a lack of available data from previous studies, the typical yield strength equal to 915 MPa from Table 4 was adopted as the assumed yield strength of the ER120S-G WAAM steel. Some additional design rules were applied to maintain consistency among all the specimens: (a) the WAAM material was deliberately arranged on each strengthened cross-section to maintain the cross-sectional plastic neutral axis at approximately the mid-height of the I-section (i.e., approximately the same area of WAAM material was added to the compression flange and the tension flange at each strengthened cross-section); (b) the weld lines were symmetrically distributed across the tension flange, with up to seven weld lines spaced 11 mm apart in each layer (row); for instance, the Section 3 of Specimen ER110-4PB-10% (see Table 6) contains 17 weld lines on the tension flange, which form two full layers with 7 weld lines in each and one partial layer with 3 weld lines. The arrangement of the weld lines for all specimens was automatically implemented using a Python script. All strengthened specimens were designed consistently in accordance with the rules outlined above; therefore, their detailed design information is not presented herein for brevity. The specimens were manufactured in a similar manner to that reported in [7]. Weld lines were deposited alternately onto the two flange outstands, with those on the tension flange deposited first, followed by those on the compression flange. This deposition strategy was intended to reduce lateral welding distortions during WAAM, thereby improving deposition accuracy and manufacturing quality, and is therefore recommended in future applications. 3. Material properties The mechanical properties of the materials examined in the present study were obtained through monotonic tensile testing and are presented in this section. A 400 mm long IPE160 section was fabricated with both flanges extended through WAAM, as shown in Fig. 3(a). Those WAAM extensions were manufactured using the aforementioned two types of welding wire, and adopting the same process parameters given in Table 5. For each type of WAAM steel (i.e., ER110S-G or ER120S-G), four dog bone-shaped tensile coupons designed in accordance with EN ISO 6892-1 [30] were extracted from the WAAM extensions (see Fig. 3(b)), where two of these four coupons were left in the as-built condition and two were machined/milled to flat surfaces to assess the potential influence of the WAAM surface undulations on the mechanical behaviour. These coupons, with their parallel regions spanning approximately the first nine layers of the WAAM extensions, were considered capable of reflecting the properties of the WAAM deposits in the strengthened beam specimens. Additionally, four tensile coupons were extracted from the examined hot-rolled IPE160 section (i.e., two from the flanges and two from the web). All tensile coupons were extracted from the parent materials by means of waterjet cutting. The dimensions of the tensile coupons are presented in Fig. 4. Prior to testing, gauge lengths equal to 5.65√𝐴, where 𝐴 is the cross-sectional area of the parallel region, were marked on each side of the coupons for the later calculation of fracture strains, 𝜀f. Note that for as-built coupons, 𝐴 was determined as the average cross-sectional area within the parallel region, using the method described in [1,7], where the geometry of each coupon was reconstructed from the laser-scanned point clouds of its surface; a series of contour slices was then used to section the parallel region of the reconstructed geometry, allowing analysis of the cross-sectional properties. A 250 kN Instron 8802 testing machine was used to apply tensile loads to the coupons under displacement control. During the tests, Thin-Walled Structures 215 (2025) 113437 3 J. Yang et al. Table 7 Design information of Specimen ER120-3PB-5% (unit for dimensions: mm). Specimen ID (𝑉W∕𝑉I)nLoad (s) 𝑃 (kN) ER120-3PB-5% 5.52% 182.1 Profile and bending moment diagrams Fig. 3. (a) I-section with extended flanges fabricated through WAAM and (b) schematic illustration of coupon extraction. Fig. 4. Dimensions of tensile coupons (unit: mm). strain rates of approximately 7 × 10−5/s and 2.5 × 10−4/s were employed up to and beyond the yield strength 𝑓y respectively. A fourcamera LaVision digital image correlation (DIC) system was used to monitor the displacement/deformation development within the tensile coupons, with images recorded at a rate of 1 Hz. The test setup is shown in Fig. 5. Longitudinal engineering stresses within the parallel region of the coupons were determined as the load divided by the original cross-sectional area, where, again, the average area was used for asbuilt coupons; the longitudinal engineering strains were determined using the software DaVis [31] based on the recorded images of the speckle pattern. A similar testing procedure was adopted in [32,33]. The engineering stress–strain relationships of the tested coupons are given in Fig. 6, where the as-built and the machined coupons are labelled with ‘A’ and ‘M’ respectively. The key mechanical properties are summarised in Table 8, where 𝑓y is the yield strength taken as the 0.2% proof strength, 𝑓u is the ultimate tensile strength, 𝜀u is the strain at 𝑓u, 𝐸 is the Young’s modulus and 𝜀f is the fracture strain determined as the percentage elongation over the marked gauge length. The fractured WAAM tensile coupons are shown in Fig. 7. Some previous studies have reported reduced mechanical properties for as-built coupons compared to machined coupons, primarily owing to surface undulations that lead to localised reductions in crosssectional areas [1–3]. However, no significant difference can be observed between the stress–strain curves of the as-built coupons and the corresponding machined coupons in the present study, see Fig. 6. This Thin-Walled Structures 215 (2025) 113437 4 J. Yang et al. Fig. 5. Experimental setup for tensile coupon tests. Fig. 6. Engineering stress–strain curves of tested tensile coupons. Table 8 Key mechanical properties of tested tensile coupons. Coupon ID 𝑓y𝑓u𝜀u𝜀f𝐸 𝑓u∕𝑓y (MPa) (MPa) (%) (%) (MPa) ER110-A1 589 787 13.67 24.58 204 000 1.34 ER110-A2 585 791 13.04 23.83 209 000 1.35 ER110-M1 593 788 11.53 16.40 204 000 1.33 ER110-M2 589 799 12.65 21.49 205 000 1.36 Mean (ER110) 589 791 12.72 21.58 205 500 1.34 ER120-A1 654 931 12.10 21.77 203 000 1.42 ER120-A2 639 907 11.12 19.90 207 000 1.42 ER120-M1 644 915 12.17 23.26 202 000 1.42 ER120-M2 647 950 11.74 21.11 203 000 1.47 Mean (ER120) 646 926 11.78 21.51 203 000 1.43 Flange-1 407 541 13.79 28.50 203 000 1.33 Flange-2 408 541 14.21 31.40 205 000 1.33 Web-1 413 544 14.99 31.87 203 000 1.32 Web-2 414 542 13.74 29.94 200 000 1.31 Mean (Hot-rolled) 411 542 14.18 30.43 202 750 1.32 Fig. 7. WAAM tensile coupons after fracture. Fig. 8. Porosity within the fractured cross-section of Coupon ER110-M1. is attributed to the loading direction being aligned with the welding direction, along which surface undulations have minimal influence on the cross-sectional area of the coupons. More specifically, for the machined coupons, the smallest cross-sectional area was 99% of the average cross-sectional area within the parallel region, while for the asbuilt coupons, this only reduced slightly to 98%. The hot-rolled coupons and the ER110S-G coupons exhibited a similar form of stress–strain curve, featuring a sharply defined yield point, a yield plateau, and strain-hardening [34]. In contrast, the ER120S-G coupons showed no distinct yield plateau. Note that Coupon ER110-M1 exhibited particularly low ductility compared to the other ER110S-G coupons, which can be attributed to the large porosity observed within the cross-section where failure occurred, as shown in Fig. 8. It should also be mentioned that the yield strengths of the WAAM steel are typically lower than the corresponding as-welded values provided by manufacturers, as given in Table 4. In the present study, the average measured yield strength of the ER110S-G WAAM steel was 6% below the manufacturer value, while for the ER120S-G material, it was 30% lower. This is most likely owing to the much slower cooling rate in the WAAM process compared to that in conventional welding scenarios [1,29]. 4. Geometric and mass analyses The surface geometry of the examined beam specimens was obtained by means of 3D laser scanning using a FARO scan arm with the capability of capturing 600,000 points per second with a precision of ±0.075 mm — see Fig. 9, showing the scanning setup and the scanned point cloud of a typical specimen, ER110-3PB-15%. The crosssectional dimensions of the examined I-sections were measured from Thin-Walled Structures 215 (2025) 113437 5 J. Yang et al. Fig. 9. (a) 3D laser scanning of the tested specimens and (b) scanned point cloud of Specimen ER110-3PB-15%. Fig. 10. Cross-sectional stress distribution at (a) elastic bending moment resistance 𝑀el and (b) plastic bending moment resistance 𝑀pl (+ – tensile, - – compressive). Fig. 11. Illustration of the bending distortion measurements. the scanned point clouds. Laser scanning was also performed on the ‘flange-extended’ I-section (see Fig. 3(a)) to determine the average height and width of a single weld bead (i.e., ℎw and 𝑏w, see Fig. 2), which measured 1.495 mm and 7.70 mm respectively for the ER110S-G WAAM steel, and 1.450 mm and 7.94 mm respectively for the ER120S-G WAAM steel; both dimensions are relatively consistent with the nominal dimensions adopted in design (see Table 5). The densities of the hot-rolled and the WAAM steel were measured based on Archimedes’s principle. The measured densities were 7.86 g/cm3 for the hot-rolled steel and the ER110S-G WAAM steel, and 7.83 g/cm3 for the ER120S-G WAAM steel. In Table 9, the measured cross-sectional dimensions of the examined specimens, together with some key section properties of the mid-span cross-section, including the second moment of area 𝐼y, the elastic bending moment resistance Thin-Walled Structures 215 (2025) 113437 6 J. Yang et al. Table 9 Cross-sectional properties and mass of examined specimens. Specimen ID 𝐻 𝐵 𝑡w𝑡f𝑟 𝐼y𝑀el 𝑀pl 𝑚I𝑚W𝑚W∕𝑚I (mm) (mm) (mm) (mm) (mm) (×104 mm4) (kN m) (kN m) (kg) (kg) B-4PB 161.82 83.67 5.37 6.96 8.95 879 44.66 51.17 28.68 – – ER110-4PB-5% 162.14 83.84 5.26 7.05 8.94 1042 52.21 62.17 28.73 1.67 5.80% ER110-4PB-10% 162.36 84.11 5.32 7.13 8.92 1172 57.78 72.18 29.10 3.19 10.97% ER110-4PB-15% 162.06 83.92 5.37 6.94 8.94 1224 61.15 79.22 28.70 4.42 15.38% ER120-4PB-5% 162.42 83.92 5.25 7.02 8.92 1044 52.44 63.56 28.66 1.73 6.05% ER120-4PB-10% 162.25 83.84 5.31 6.90 8.92 1150 56.80 72.39 28.50 3.34 11.72% ER120-4PB-15% 162.24 83.91 5.29 7.02 8.90 1208 61.08 81.36 28.73 4.34 15.09% B-3PB 162.29 84.10 5.36 7.04 8.96 895 45.35 51.92 28.96 – – ER110-3PB-5% 162.28 84.31 5.42 7.04 8.94 1209 59.16 76.13 29.13 1.64 5.64% ER110-3PB-10% 162.68 84.07 5.27 7.10 8.87 1338 65.34 86.52 28.92 3.10 10.71% ER110-3PB-15% 162.10 83.78 5.35 7.02 8.92 1482 69.78 98.49 28.82 4.46 15.46% ER120-3PB-5% 162.42 84.28 5.42 7.08 8.96 1174 58.08 75.35 29.22 1.61 5.52% ER120-3PB-10% 162.17 83.88 5.34 6.98 8.93 1269 62.26 84.33 28.73 2.97 10.33% ER120-3PB-15% 162.20 83.92 5.40 7.01 8.91 1410 67.88 96.80 28.92 4.45 15.37% Fig. 12. Flexural distortions of four-point bending WAAM HSS-strengthened beam specimens in comparison with the original underlying imperfection of the benchmark specimen. 𝑀el and the plastic bending moment resistance 𝑀pl about the major axis, are provided for all specimens; the mass of the I-section portion (𝑚I), and that of the WAAM additions (𝑚W), are also included. Note that 𝑀el was calculated as the bending moment at which either the top or the bottom extreme fibre of the base I-section reached its yield strength 𝑓y,I, while 𝑀pl was determined as the bending moment at which the entire base I-section and the WAAM regions reach their corresponding yield strengths (𝑓y,I and 𝑓y,W for the I-section and WAAM material, respectively), as illustrated in Fig. 10. The section properties and masses were determined based on the measured values of the crosssectional dimensions, densities and the average material properties (see Table 8). The calculated total mass 𝑚=𝑚I+𝑚W of the specimens was compared to the measured total mass, where the error margin was below ±1%. Previous studies [7,25] have investigated the flexural distortions in WAAM-strengthened steel I-section beams caused by the heat input from WAAM through physical measurements and numerical methods. The in-plane flexural distortion in the 𝑥-𝑧 plane where the loading is applied, see Fig. 2, may be referred to as pre-cambering if it opposes the loading direction, enabling the pre-cambered specimens to counteract some of the deflection caused by the external load. Conversely, the outof-plane flexural distortion in the 𝑥-𝑦 plane induces initial imperfections to the specimens. The flexural distortions of the examined specimens were extracted from the scanned 3D point clouds, as illustrated in Fig. 11; the 𝑧 coordinate of the mid-point on the outer surface of the compression flanges was used for the determination of the in-plane flexural distortions, while the out-of-plane distortion was evaluated by averaging the 𝑦 coordinates of the mid-points on the web surfaces. In Figs. 12 and 13, the in-plane and out-of-plane flexural distortions of the WAAM HSS-strengthened specimens, denoted as 𝛿0,z and 𝛿0,y, respectively, are plotted alongside the original underlying imperfections of the two benchmark specimens. It can be observed that all WAAM HSS-strengthened specimens exhibit some slight pre-cambering up to approximately 𝐿∕450, while showing a greater out-of-plane imperfection/out-of-straightness compared to the benchmark specimens. Thin-Walled Structures 215 (2025) 113437 7 J. Yang et al. Fig. 13. Flexural distortions of three-point bending WAAM HSS-strengthened beam specimens in comparison with the original underlying imperfection of the benchmark specimen. Fig. 14. Some misalignment of material deposition in Specimen ER120-4PB-10%. This amplification effect on the out-of-plane out-of-straightness can be attributed to the asymmetric material deposition sequence, as described in [7] and Section 2, and some positional inaccuracies during specimen manufacture. For instance, Fig. 14 shows the tension flange of Specimen ER120-4PB-10% where the deposited weld lines were eccentric about the centreline, resulting in significant out-of-plane flexural distortion, see Fig. 12. This issue could be mitigated against with the simultaneous deposition of two weld beads maintaining symmetry about the vertical axis through the cross-section. With the mechanical and the geometric properties established in Section 3 and the current section, the discrepancy between the true/measured yield strength of the WAAM material and the corresponding design yield strength used in Section 2 should be addressed. The moment resistance diagrams, calculated using both the design and the measured properties, are shown in Fig. 15 for Specimen ER110-3PB-10% and in Fig. 16 for Specimen ER120-4PB-10%. For Specimen ER120-4PB-10%, the lower yield strength of the WAAM material – approximately 30% lower than the design value – resulted in a significant reduction in the bending moment resistance. Additionally, although the bending moment diagram remains within the envelope of the moment resistance diagram, the match between the two becomes less close. In contrast, the discrepancy for Specimen ER110-3PB-10% is minimal. In structural design applications, to ensure that the WAAMstrengthened beams can achieve their intended design resistances, suitable design values for the material properties of WAAM steel are needed. 5. Experiments on WAAM HSS-strengthened beam specimens Physical testing was carried out on the WAAM HSS-strengthened beam specimens to assess their major-axis flexural behaviour under either four-point bending or three-point bending conditions. The experimental setup and test results are detailed in the current section. 5.1. Test setup A 500 kN Instron hydraulic jack was used to apply the vertical loads (see Fig. 1) to the beams under displacement control. Prior to testing, web stiffeners were welded onto the specimens at the supports and the loading points to prevent web crippling under the concentrated loads. The schematic illustration and physical test setups for the four-point bending tests and the three-point bending tests are shown in Figs. 17 and 18 respectively. Roller supports were used to enable end rotations, under which load cells were placed to measure end reactions. Bearing plates were placed between the specimens and the roller supports to avoid stress concentrations. In the four-point bending tests, an HPE180 spreader beam was used to achieve a pair of equal vertical loads acting at the third points of the beam span. Inclinometers were mounted onto the web stiffeners beneath the loading points to measure the in-plane rotation at these locations. String potentiometers were mounted onto the outer surface of the tension flange under the loading points and at the mid-span to measure the vertical displacements. In the threepoint bending tests, inclinometers were mounted on the web stiffeners Thin-Walled Structures 215 (2025) 113437 8 J. Yang et al. Fig. 15. Moment resistance diagram and anticipated bending moment diagram of Specimen ER110-3PB-10% determined using (a) design/nominal yield strength and geometric properties and (b) measured yield strength and geometric properties. Fig. 16. Moment resistance diagram and anticipated bending moment diagram of Specimen ER120-4PB-10% determined using (a) design/nominal yield strength and geometric properties and (b) measured yield strength and geometric properties. above the supports, while a single string potentiometer was installed at the mid-span. A series of top-flange lateral restraints were employed at multiple locations along the length of the specimens to prevent lateral–torsional buckling. A lateral restraint system similar to that reported in [7] was employed herein, where the lateral restraints were pre-loaded with 1.5 kN axial tension prior to testing to stabilise the testing system. Data from the instrumentations were logged with the DATASCAN system at a frequency of 3 Hz. Additionally, a four-camera LaVision DIC system was used to monitor the displacement/strain development within the specimens at a frequency of 2 Hz. All specimens were loaded vertically at a rate of 3 mm/min at the loading points. 5.2. Test results All four-point bending and three-point bending specimens failed by inelastic local buckling or/and distortional buckling in the compression flange and web, with some lateral displacements occurring between the lateral restraints. No visible or audible signs of failure in the WAAM material or at the interface between the WAAM material and the steel I-section material were observed. The deformed specimens after testing are shown in Figs. 19 and 20. The bending moment 𝑀 within the purely flexural region of the four-point bending specimens, and the bending moment 𝑀 at the mid-span of the three-point bending specimens, calculated from the end reactions, are plotted with respect to the mid-span vertical displacement 𝛿m in Fig. 21(a) and (b) respectively. Additionally, curves of the bending moment 𝑀 versus the curvature 𝜅 in the purely flexural region of the four-point bending specimens are presented in Fig. 22(a), while curves of the bending moment 𝑀 at the mid-span versus the mid-span rotation 𝜃 of the three-point bending specimens are shown in Fig. 22(b). The curvature 𝜅 is calculated as 𝜅= (𝜃1+𝜃2)∕𝐿f, and the mid-span rotation 𝜃 is determined as 𝜃=𝜃1+𝜃2, where 𝐿f is the length of the pure flexural region, while 𝜃1 and 𝜃2 are the rotations measured by the inclinometers. The structural behaviour of the examined specimens is characterised with an equivalent yield point Y, the ultimate point U, the elastic stiffness 𝐾e (i.e., the slope of the initial linear portion of the 𝑀-𝜅 (or 𝜃) curve) and the tangent stiffness at the equivalent yield point, 𝐾y, as depicted in Fig. 23(a) and (b) for the two loading conditions. The equivalent yield point Y is determined based on elasto-plastic energy absorption [35] by seeking two equal areas, 𝐴1 and 𝐴2 on the 𝑀-𝜅 (or 𝜃) curve (see Fig. 23). Additionally, the ductility of the beam specimens is evaluated using the rotation capacity 𝑅, defined as 𝑅=𝜅rot ∕𝜅y−1 for the four-point bending specimens and 𝑅=𝜃rot ∕𝜃y−1 for the three-point bending specimens [36]. Here, 𝜃rot is taken as 𝜃85 corresponding to the point of 85% 𝑀u on the descending range of the 𝑀-𝜃 curve, while 𝜅rot is taken as 𝜅u instead, since the 𝑀85 point was not reached during testing for the benchmark specimen B-4PB. A summary of the key four-point bending and three-point bending test results is provided in Tables 10 and 11 respectively, where 𝐾e is obtained to be the best linear fit to the initial linear portion of the 𝑀-𝜅 (or 𝜃) curves. Colour maps of the longitudinal surface strains in two typical specimens – ER110-4PB-15% and ER110-3PB-5% – at different load levels, obtained through DIC, are presented in Fig. 24(a) and (b) respectively. 6. Discussion and analyses of experimental results In the present section, the test results of the WAAM HSS-strengthened beam specimens, as described in Section 5, are discussed, and the levels of structural performance enhancement are analysed using different metrics. All tested specimens exceeded their calculated plastic moment capacities — see Tables 9, 10 and 11. In Figs. 21 and 22, it can be seen that the benchmark specimens exhibited a relatively well-defined elasto-plastic structural response with strain hardening, featuring a small transition zone following the initial elastic range, where plastic strains spread throughout the cross-sections and the stiffness degraded sharply. The strengthened specimens showed more gradual stiffness degradation, beginning at relatively early stages during testing without a sharply-defined yield point. This gradual-yielding behaviour shares similarities with stainless steel [37,38], but is attributed in the case of the studied WAAM-strengthened beams to the use of two different steel grades (for the I-section and WAAM material) and the high residual stresses within the specimens caused by the WAAM heat input [25]. Therefore, the equivalent yield moment 𝑀y and the corresponding Thin-Walled Structures 215 (2025) 113437 9 J. Yang et al. Data availability Data will be made available on request. References [1] C. Huang, P. Kyvelou, R. Zhang, T.B. Britton, L. Gardner, Mechanical testing and microstructural analysis of wire arc additively manufactured steels, Mater. Des. 216 (2022) 110544. [2] M.T. Chen, T. Zhang, Z. Gong, W. Zuo, Z. Wang, L. Zong, O. Zhao, L. Hu, Mechanical properties and microstructure characteristics of wire arc additively manufactured high-strength steels, Eng. Struct. 300 (2024) 117092. [3] M.T. Chen, Z. Gong, T. Zhang, W. Zuo, Y. Zhao, O. Zhao, G. Zhang, Z. Wang, Mechanical behavior of austenitic stainless steels produced by wire arc additive manufacturing, Thin-Walled Struct. 196 (2024) 111455. [4] L. Gardner, Metal additive manufacturing in structural engineering – review, advances, opportunities and outlook, Struct. 47 (2023) 2178–2193. [5] X. Meng, L. Gardner, Hybrid construction featuring wire arc additive manufacturing: review, concepts, challenges and opportunities, Eng. Struct. 326 (2025) 119337. [6] L. Gardner, J. Li, X. Meng, C. Huang, P. Kyvelou, I-section steel columns strengthened by wire arc additive manufacturing – concept and experiments, Eng. Struct. 306 (2024) 117763. [7] J. Yang, M.A. Wadee, L. Gardner, Strengthening of steel I-section beams by wire arc additive manufacturing – concept and experiments, Eng. Struct. 322 (2025) 119113. [8] E. Ghafoori, H. Dahaghin, C. Diao, N. Pichler, L. Li, M. Mohri, J. Ding, S. Ganguly, S. Williams, Fatigue strengthening of damaged steel members using wire arc additive manufacturing, Eng. Struct. 284 (2023) 115911. [9] H. Dahaghin, M. Motavalli, H. Moshayedi, S.M. Zahrai, E. Ghafoori, Wire and arc additive manufacturing for strengthening of metallic components, Thin-Walled Struct. 203 (2024) 112074. [10] X. Meng, B. Weber, M. Nitawaki, L. Gardner, Optimisation and testing of wire arc additively manufactured steel stub columns, Thin-Walled Struct. 189 (2023) 110857. [11] I.H. Shah, N. Hadjipantelis, L. Walter, R.J. Myers, L. Gardner, Environmental life cycle assessment of wire arc additively manufactured steel structural components, J. Clean. Prod. 389 (2023) 136071. [12] J. Chen, S.S. Song, J. Ye, G. Quan, P. Kyvelou, L. Gardner, Axial compressive behaviour and design of concrete-filled wire arc additively manufactured steel tubes, Structures 70 (2024) 107495. [13] P. Kyvelou, A. Spinasa, L. Gardner, Testing and analysis of optimized wire arc additively manufactured steel trusses, J. Struct. Eng. 150 (3) (2024) 04024008. [14] V. Laghi, M. Palermo, G. Gasparini, V.A. Girelli, T. Trombetti, Experimental results for structural design of wire-and-arc additive manufactured stainless steel members, J. Constr. Steel Res. 167 (2020) 105858. [15] X. Guo, P. Kyvelou, J. Ye, L.H. Teh, L. Gardner, Experimental study of DED-arc additively manufactured steel double-lap shear bolted connections, Eng. Struct. 281 (2023) 115736. [16] X. Guo, P. Kyvelou, J. Ye, L.H. Teh, L. Gardner, Experimental investigation of wire arc additively manufactured steel single-lap shear bolted connections, Thin-Walled Struct. 181 (2022) 110029. [17] W. Zuo, O. M. Chen, L. Gardner, Testing of wire arc additively manufactured duplex stainless steel double-lap shear bolted connections, J. Struct. Eng. 151 (2) (2025) 04024205. [18] W. Zuo, O. M. Chen, A. Su, S. Liu, X. Yun, F. Xu, Structural performance of wire arc additively manufactured duplex stainless steel single-lap shear bolted connections, Eng. Struct. 319 (2024) 118706. [19] W. Zuo, M.T. Chen, O. Zhao, A. Su, S. Liu, F. Xu, Y. Huang, B. Cheng, Behavior of wire arc additively manufactured 316L austenitic stainless steel single shear bolted connections, Thin-Walled Struct. 202 (2024) 112075. [20] L. Kang, C. Zhang, M.A. Bradford, X. Liu, Axial compressive behaviour of corroded circular steel tube columns retrofitted by laser-cladding additive manufacturing, Thin-Walled Struct. 192 (2023) 111129. [21] L. Kang, C. Zhang, M.A. Bradford, X. Liu, Residual stresses in circular steel tubular columns repaired by laser-cladding additive manufacturing, Thin-Walled Struct. 193 (2023) 111275. [22] L. Kang, P. Song, Bending behaviour of surface corroded and perforated corroded steel tubes repaired by laser cladding additive manufacturing, Thin-Walled Struct. 203 (2024) 112213. [23] Y. Yuan, X. Zhang, J. Ji, L. Zhang, B. Zeng, C.L. Wang, Axial compression performance of circular steel tubes with localized corrosion repaired by laser-cladding additive manufacturing, Eng. Struct. 318 (2024) 118811. [24] P. Kyvelou, C. Huang, J. Li, L. Gardner, Residual stresses in steel I-sections strengthened by wire arc additive manufacturing, Structures 60 (2024) 105828. [25] J. Yang, M.A. Wadee, L. Gardner, Residual stresses in steel I-section beams strengthened by wire arc additive manufacturing, J. Constr. Steel Res. 232 (2025) 109606. [26] J. Yang, P. Kyvelou, M.A. Wadee, L. Gardner, Simulation and prediction of residual stresses in WAAM-strengthened I-sections, Structures 69 (2024) 107248. [27] Eurocode 3: Design of Steel Structures. Part 1-1: General Rules and Rules for Buildings, European Committee for Standardization, Brussels, 2022. [28] AWS A5.28: Specification for Low-Alloy Steel Electrodes and Rods for Gas Shielded Arc Welding, American Welding Society, 2022. [29] B. Weber, X. Meng, R. Zhang, M. Nitawaki, T. Sagawa, L. Gardner, Tensile behaviour of WAAM high strength steel material and members, Mater. Des. 237 (2024) 112517. [30] EN ISO 6892-1: Metallic Materials – Tensile Testing – Part 1: Method of Test At Room Temperature, European Committee for Standardization (CEN), Brussels, 2016. [31] LaVision GmbH, Davis 10, version 10.2.1, 2017. [32] X. Meng, L. Gardner, Behavior and design of normal and high strength steel SHS and RHS columns, J. Struct. Eng. 146 (11) (2020) 04020227. [33] X. Meng, L. Gardner, Cross-sectional behaviour of cold-formed high strength steel circular hollow sections, Thin-Walled Struct. 156 (2020) 106822. [34] C. Huang, P. Kyvelou, L. Gardner, Stress–strain curves for wire arc additively manufactured steels, Eng. Struct. 279 (2023) 115628. [35] R. Park, Evaluation of ductility of structures and structural assemblages from laboratory testing, Bull. N. Z. Soc. Earthq. Eng. 22 (3) (1989) 155–166. [36] C. Huang, X. Meng, L. Gardner, Cross-sectional behaviour of wire arc additively manufactured tubular beams, Eng. Struct. 272 (2022) 114922. [37] M. Theofanous, T.M. Chan, L. Gardner, Flexural behaviour of stainless steel oval hollow sections, Thin-Walled Struct. 47 (6–7) (2009) 776–787. [38] O. Zhao, L. Gardner, B. Young, Behaviour and design of stainless steel SHS and RHS beam–columns, Thin-Walled Struct. 106 (2016) 330–345. [39] A.I. Osofero, M.A. Wadee, L. Gardner, Experimental study of critical and postbuckling behaviour of prestressed stayed columns, J. Constr. Steel Res. 79 (2012) 226–241. [40] H.X. Yuan, Y.Q. Wang, L. Gardner, Y.J. Shi, Local-overall interactive buckling of welded stainless steel box section compression members, Eng. Struct. 67 (2014) 62–76. Thin-Walled Structures 215 (2025) 113437 16