This paper presents the investigation of cold-formed Q1100 ultra-high strength steel unequal-leg angle section stub columns through experimental testing and numerical simulations. A total of 18 angle section stub columns were tested, including 13 unequal-leg angel specimens and 5 equal-leg angle specimens. Tensile coupon material properties and initial local geometric imperfections were measured. Test results indicated that all angle section stub columns occurred interacted failure of local buckling and torsional buckling before attaining the yielding strength. For each series of specimens with the same width of the longer leg, unequal-leg angle sections had lower ultimate load capacities due to the smaller cross-sectional areas, compared to the equal-leg angle section counterpart. Slenderness of the shorter leg decreased with the reduce of its width, which also led to the delayed buckling of the specimen and increased utilisation of material strength. Finite element models were validated by the test results and then used to implement parametric studies to generate supplementary data. With the obtained test and numerical data, design provisions in existing codes and specifications were evaluated to examine their applicability on predicting the cross-sectional resistance of the cold-formed Q1100 ultra-high strength steel unequal-leg angle section stub columns. It is found that Eurocode, the North American specification AISI S100 and Australian/New Zealand standard AS/NZS 4600 can generally provide safe yet conservative predictions. Direct Strength Method can provide conservative predictions for non-slender sections, while slightly overpredicts the cross-sectional resistance for slender sections. Modified design recommendations based on Eurocode, AISI S100 and DSM have also been proposed, which present more accurate predictions than the existing design methods.

Structural machine learning (ML) is increasingly used for rapid performance prediction and parametric design, but its reliability is often limited by scarce structural data and by the weak formalisation of engineering knowledge embedded in design specifications. This study proposes a specification-guided ML framework that converts specification knowledge into learnable, generative and deployable components for parametric structural design. The framework is demonstrated on cold-formed stainless steel RHS/SHS beams, where AISC 370 provisions define the code baseline, feasibility constraints and expert partitioning rules. A two-stage generative augmentation strategy is developed to mitigate the adverse effects of limited data. Feasible design variables are first conditionally generated, followed by synthesising the structural capacity through residual learning anchored to specification-based resistance predictions. To reflect regime-dependent behaviour, a mixture-of-experts (MoE) predictor is constructed, in which a slenderness-based gate blends non-slender and slender experts. The augmented MoE models substantially improve test accuracy, reduce prediction scatter and mitigate overfitting compared with code calculations and single global learners trained on limited real data. Finally, the trained model is deployed in Rhino–Grasshopper, enabling real-time capacity assessment, geometry visualisation and interactive exploration. The proposed framework demonstrates how design specifications can be formalised as computable priors for data generation, model specialisation and design automation.

Accurate modeling of stainless steel behavior in cold-formed corner regions requires reliable input material parameters, which can be obtained through corner coupon tests. However, the implementation of existing stress-strain models requires key material parameters that are often unavailable due to the challenges of conducting corner coupon tests. To support more generalized applications of stress-strain models and reduce reliance on physical testing, a comprehensive database comprising 321 corner coupon test results and 93 full-range stress-strain curves was collected from 58 published sources, covering the most widely used structural stainless steel types, including austenitic, duplex, and ferritic grades. Subsequently, the applicability of existing predictive models for estimating essential material parameters was assessed and modified or newly developed equations were proposed where improved accuracy was needed. Based on either measured or predicted material inputs, stress-strain curves were generated by adopting a one-stage Ma-Chan-Young model and a modified two-stage Ramberg-Osgood model, and were validated against experimental data. The evaluation results confirm that the proposed predictive framework can reliably capture the nonlinear stainless steel material behavior in cold-formed corner regions, even in the absence of direct test data, and is suitable for use in numerical simulations and structural design involving cold-formed stainless steel structures.

Stainless steel structural members demonstrate great promise in supporting the green and sustainable transformation of the construction sector. This study experimentally and numerically investigated the performance of cold-formed stainless steel built-up open section beams. In the experimental programme, eight beam specimens were manufactured by assembling two edge-stiffened channel components, which were brake-pressed from EN 1.4162 lean duplex and EN 1.4301 austenitic stainless steel plates. M5.5 self-drilling screws were adopted to connect the built-up section members. Four-point bending tests were conducted on the eight beam specimens with two nominal web heights and three nominal plate thicknesses. The experimental results of moment-curvature curves, failure modes, and moment capacities of all test specimens were obtained. In the numerical investigation, a nonlinear finite element model was constructed and calibrated using the four-point bending experimental results. Using the validated finite element model, a parametric study was carried out to generate additional 90 numerical data on lean duplex and austenitic stainless steel edge-stiffened built-up open section beams. Underpinned by the dataset of tests and finite element analyses, the appropriateness of the codified direct strength method (DSM) was evaluated. Modified DSM design formulae were proposed for cold-formed stainless steel built-up open section members under bending.

The structural performance of concrete-filled double skin tubular (CFDST) beams is examined in this study. Stainless steel rectangular hollow sections (RHS) were used as the outer tubes, while steel circular, square and rectangular hollow sections (CHS, SHS and RHS) were used as the inner tubes. The inner tubes were eccentrically positioned in the tensile zone of the cross-sections to enhance the bending moment capacity. A total of 30 CFDST specimens was tested under three-point bending. Three levels of inner tube eccentricity were investigated. The nominal compressive strength of the infilled-concrete ranged from 40 to 120 MPa. The full moment-deflection responses, bending moment capacities and failure modes of the investigated CFDST beams are reported and discussed herein. It was observed that increasing the inner tube eccentricity enhanced the bending moment capacity by up to 18% and increasing the concrete compressive strength from C40 to C120 led to a maximum improvement of 50% in bending moment capacity. Additionally, the experimental bending moment capacities were compared against predictions determined from current European, American and Chinese standards for composite carbon steel members. It was shown that the existing design provisions yielded conservative and scattered predictions for the investigated CFDST beams. Developments required in the current design provisions in relation to their application to the studied cross-sections include improved allowance for the strain hardening of stainless steel, concrete confinement and shear-bending interaction. These factors are to be addressed in future research.

Steel modular construction technology has been increasingly adopted in high-rise buildings, where a concrete core has been typically employed to ensure the lateral stability of the overall structure. Within this type of steel modular system, the connections between steel modules and concrete core walls are critical for effective load transfer, especially under severe horizontal loads. However, up to now, only a few experimental tests and systematic numerical investigations have been conducted on such module-to-core wall (M2C) connections, limiting their application in real-life engineering cases. Therefore, this study aimed to develop an innovative M2C connection for high-rise steel modular buildings and systematically examine its tensile behaviour through experimental tests and numerical simulations. The innovative M2C connection was developed based on the principles established from examining existing connection forms. The experimental tests were performed on ten specimens, and numerical simulations were conducted on 253 models with due consideration on various critical influencing factors. The analysis results demonstrate that the failure mode of the developed connection is the fracture of welds in the T-shaped plate or the failure of the bolts, and that the steel grade, thickness of the T-shaped plate flange, and thickness of the constraint plates are the most critical influencing factors. The findings provide valuable insights into the tensile behaviours of the newly developed M2C connection, and practical guidelines for the design and application of this M2C connection in future high-rise steel modular buildings.

Compared with code-based design, data-driven prediction learns context-specific response patterns from tests and simulations, offering finer-grained and less conservative estimates. Reliable data-driven prediction in structural engineering is often hampered by limited and imbalanced datasets as well as the weak coupling between learned models and practical tools. This work presents an integrated framework that combines physics-informed generative data augmentation, Bayesian-optimised machine learning, model interpretability, and parametric design for stainless steel tubular columns. The generative stage is two-fold: a conditional model synthesizes feasible feature vectors under design labels, after which a physics-aware residual is generated under the joint context of the features and the Eurocode baseline. A post-generation rebalancing enforces adequate representation across cross-section classes and non-dimensional slenderness bands. Tree-structured Parzen Estimator optimisation is used to tune both the physics-informed CTGAN/CTVAE generators and the gradient-boosted learners (XGBoost, LightGBM). Across multiple train/test settings, the augmented learners consistently improve predictive fidelity and robustness relative to Eurocode formulas and to models trained without augmentation. SHAP analyses confirm that learned effects are congruent with mechanics and remain stable across datasets and algorithms. The tuned predictor is embedded in a Rhino-Grasshopper component that provides real-time what-if exploration and side-by-side comparison with the Eurocode baseline inside a familiar parametric workflow. The study establishes a reproducible pathway for combining mechanics-aware augmentation, explainable learning, and CAD-centric deployment to support reliable design under data scarcity.

This study is focused on the numerical investigation of stainless steel tubular T-joints with chord preload. The developed finite element models were validated against experimental results. The tests previously reported by the authors comprised ten square hollow section T-joints with equal brace and chord widths (/3 = 1): five joints with chord's height-to-thickness ratio (2 gamma) = 40 fabricated from 80 & times; 80 mm sections and five joints with 2 gamma = 20 fabricated from 40 & times; 40 mm sections, all with a wall thickness of 2 mm. Each joint configuration was subjected to tensile and compressive chord loading corresponding to 75%, 40%, and 0% of the chord cross-section resistance. The parametric study comprising 900 numerical simulations using the validated models, and the resulting data were compared with predictions from current design standards and formulations available in the literature. Based on the observations of the obtained results, a revised calculation methodology is proposed to reproduce the numerical response better. The results show that joint resistance decreases under chord compression, whereas under chord tension, a stress-relief effect occurs, leading to a smaller reduction in resistance. Furthermore, the design standards generally provide conservative resistance predictions compared with numerical results, whereas alternative formulations from the literature yield closer mean predictions but with greater relative scatter.

This study aimed to develop an innovative module-to-module (M2M) connection for high-rise steel modular buildings and examine its structural behaviour under shear loads. The M2M connection was developed based on the examination of the special characteristics of connections in high-rise steel modular buildings from the perspectives of construction efficiency and structural safety. Experimental tests with eight specimens were performed to investigate the shear behaviour of the developed connection. Numerical models were then established and validated base on the test results to accurately simulate the shear behaviour of the developed connection. Subsequent parametric numerical simulations with 153 models were performed to examine the effects of critical influencing factors on the shear behaviour of the connection. The analysis results indicated that the middle sleeve in the developed M2M connection can provide effective protection for the vertical connectors from horizontal loads. Furthermore, the identified critical influencing factors, such as the steel grade of different components and axial compression ratio, can facilitate the protection of the vertical connector across all loading stages, thereby ensuring the disassembly and reassembly ability of the developed connection. The axial compression ratio had the most significant effects on the shear behaviour of the connection, as it not only increased the yield and ultimate load of the connection by over 100%, but also exhibited significant interactive effects with the thickness of the connection plate and the steel grade of the middle sleeve and the connection plate.

Wire-laser additive manufacturing (WLAM) is an additive manufacturing technique to fabricate metal components, using metal wires as the raw material and laser as the energy source. It is capable of producing sections with complex shapes. However, knowledge about WLAM structural components is lacking, and investigation on their geometric characteristics and structural behavior is limited. This paper presents an experimental study on the compressive behavior of WLAM ER110S I-section stub columns with various stiffener configurations; this concept makes full use of the characteristic of high geometric freedom provided by additive manufacturing. The experimental program comprised 4 tensile coupon tests and 13 stub column tests, with the geometric characteristics of all specimens captured via 3D laser scanning. Their failure modes and ultimate loads are comprehensively discussed in this paper. The obtained test data were then used to evaluate the applicability of relevant design rules in EN 1993-1-1, the Direct Strength Method, and the Continuous Strength Method for WLAM ER110S I-section stub columns. The findings revealed that all design methods provided reasonably accurate strength predictions for WLAM ER110S I-section stub columns, although all three methods provided unsafe unfactored predictions for some slender I-section specimens. Moreover, compared to WLAM ER110S conventional I-section stub columns, the stiffened I-section specimens exhibit clearly enhanced structural efficiency when the flange-tip stiffener length exceeds one-fifth of the flange outstand width, achieving a disproportionate increase in both load-bearing capacity and deformability relative to the added mass. These findings offer insights into structural strengthening and retrofitting via metal additive manufacturing.

The utilisation of metal additive manufacturing in civil engineering is at an exploratory stage. Aluminium alloys, which are commonly used in building fa & ccedil;ades and offshore structures due to their excellent corrosion resistance and lightness, may find new opportunities in the construction sector with advances in metal additive manufacturing. In this study, an experimental programme and numerical investigation were performed on the structural performance of AlSi10Mg aluminium alloy angle stub columns additively manufactured by selective laser melting (SLM). Taking advantage of the flexibility in geometry design provided by SLM process, the aluminium angle specimens were devised with varying leg widths and thicknesses. The experimental programme involved tensile coupon tests, Vickers hardness tests, microstructural characterization, measurements of initial geometric imperfections and residual stresses, as well as stub column tests. The mechanical properties, microstructural morphology and compression resistances of the SLM-fabricated AlSi10Mg aluminium alloy angle specimens were extensively examined. It was revealed in the stub column tests that, when the width-to-thickness ratio of the slender legs was smaller than 17.8, the ultimate stresses of the SLM-fabricated AlSi10Mg angle specimens could exceed the static 0.2% proof stresses measured from the tensile coupon tests. Subsequently, a finite element model was developed using ABAQUS to simulate the structural behaviour of the additively manufactured aluminium alloy angles under axial compression, which was verified against the obtained stub column test results. A parametric study was conducted utilising the validated finite element model, aiming to generate additional numerical data of SLM-fabricated angle stub columns over an extended scope of geometric dimensions. Moreover, based on the established experimental and numerical database, the suitability of design provisions specified in the European Code for conventional aluminium structures and a novel direct strength method (DSM) developed in the literature for steel angle columns was assessed for the additively manufactured AlSi10Mg angles under axial compression. In the light of shortcomings in the existing design rules, a modified DSM-based approach was proposed in this study for the accurate and reliable design of additively manufactured aluminium alloy angle stub columns.

This study examined the structural performance of cold-formed stainless steel built-up open and closed section beams through experimental and numerical investigation. Three kinds of built-up section configurations, consisting of two edge-stiffened channels back-to-back, two unstiffened channels face-to-face and two web-stiffened channels face-to-face were assembled using discretely distributed self-drilling screws. The channel components were brake-pressed from stainless steel plates of austenitic grade EN 1.4301 and lean duplex grade EN 1.4162. Firstly, 21 built-up open and closed section specimens were tested under four-point bending about the major axis. The experimental results in terms of failure mode, moment versus curvature response and moment capacity were obtained and discussed. Subsequently, by validating against the results obtained from the experimental programme, finite element models were established using the ABAQUS to simulate the behaviour of cold-formed stainless steel built-up open and closed section members subjected to major axis bending. An extensive parametric study was conducted utilising the verified finite element models to predict the moment capacities of additional 256 austenitic and lean duplex stainless steel built-up section specimens over a wide range of geometric dimensions. It was found that the investigated built-up open and closed section beams were able to develop strength beyond first yield and even exceeding the member plastic moment. Furthermore, based on the experimental and numerical data pool, the applicability of the direct strength method (DSM) design curves adopted in the ASCE/SEI 8-22 was assessed for the cold-formed stainless steel built-up section beams. The evaluation results indicated that the codified design rules exhibited shortcomings in predicting the flexural strengths of austenitic and lean duplex stainless steel members with the built-up sections investigated in this study. Hence, a series of modified DSM formulae was recommended to offer accurate, consistent and reliable strength predictions for cold-formed stainless steel built-up open and closed section beams bent about the major axis.

Abstract A novel concrete‐filled steel tubular (CFST) column with inner semi‐circular stiffeners is proposed in this study. The semi‐circular stiffeners are developed based on the confinement mechanism in square CFST columns and aim to strengthen the confinement effect on concrete in the whole cross‐section area, therefore improving the overall structural performance of the composite column. The section is also flexible to use different grades of concrete in the core and cell regions. Experimental investigations on axial compressive behaviour of the semi‐circle stiffened concrete‐filled steel tubular (SCS‐CFST) columns as well as the hollow section columns were carried out. High strength steel Q690 and two concrete grades C40 and C90 were adopted for the test specimens. Various parameters including steel tube thickness, cross‐section slenderness, combination of different concrete grades as well as semicircular stiffener size were investigated. The test results showed that the stiffened hollow sections achieved significantly higher compressive strength with delayed local buckling failure compared to the unstiffened counterparts. SCS‐CFST columns presented significantly enhanced compressive strength and ductility compared to the corresponding unstiffened CFST columns. The applicability of existing design equations was evaluated for the novel SCS‐CFST columns. A new design approach considering the confinement effect in different regions was proposed and exhibited good predictions.

Experimental and numerical investigations of cold-formed austenitic stainless steel (CFASS) tubular T-joints are presented in this paper. The CFASS tubular joints were fabricated by braces and chords with circular hollow sections (CHS). The joint specimens were designed by varying the ratios of geometric parameters (beta, 2 gamma and tau). A total of sixteen CFASS CHS T-joints were tested by applying axial compression through the brace without preloading in the chord. Numerical investigation was then performed including the successful validation of the nonlinear finite element model and the subsequent parametric studies. The CFASS CHS T-joints mainly failed by chord plastification. The effects of the beta, 2 gamma and tau parameters on the T-joint strengths were investigated. The design equations for carbon steel CHS T-joints provided in CIDECT, EN-1993-1-8 and the literature were assessed by comparing their predictions with the test and numerical results of the CFASS CHS T-joints. It is shown that these predictions are generally unconservative. New design equations that consider the influences of geometric parameters on joint strengths were proposed. Overall, it was found that the newly proposed equations provided more accurate, reliable and less scattered predictions compared to the design equations in CIDECT and EN-1993-1-8. Specially, the unconservative predictions of 11 % by CIDECT and the unreliable predictions by both CIDECT and EN-1993-1-8 were improved.

This paper presents a comprehensive study on the design of cold-formed stainless steel (CFSS) plain channels under concentrated bearing loads via experimental and numerical investigations. Two loading cases were considered in the study: end loading (EL) and interior loading (IL), reflecting floor joist members seated on solid foundation subjected to localised bearing loads. A total of 20 tests was conducted, and the tested plain channels were press-braked from stainless steel sheets of austenitic (EN 1.4404) and lean duplex (EN 1.4162) grades. Numerical models to simulate web crippling of plain channels under the floor joist loading cases were developed and verified against the test results. The verification analysis showed that the outcomes of the numerical modelling were close to the test results, and an extensive parametric study, comprised of 144 numerical simulations, was subsequently performed based on the verified model. The experimental and numerical results were utilised to assess the applicability of existing web crippling provisions codified in American specification, European code and Australian/New Zealand standard for stainless steel structures. New design rules were proposed for CFSS plain channels under the EL and IL cases. The reliability levels of the existing and proposed design rules were evaluated using the reliability analysis from the current American specification for cold-formed stainless steel structures.

Towards the wider applicability of the Pultruded fibre-reinforced polymer (PFRP) structures, this study focused on developing an optimum connection geometry. PFRP multi-bolt staggered connection behaviour is investigated in this study. A total of sixty connection tests were carried out by varying the design parameters, such as the angle of loading, thickness of the PFRP plate, diameter of the bolts, and different patterns of staggered alignments. The test results are comprehensively analysed to determine the optimum connection geometry for achieving the bearing failure with progressive deformation. The appropriateness of the existing design standards for determining the bearing and shear-out strength of the bolted connections is verified. Based on the observations from the experimental results, minimum required bolt connection geometry dimensions are proposed. The comparison of test and design results shows that the strength predictions by design specifications CNR 2008, ASCE Pre-standard 2010, EU 2016, CEN/TC 19101, and FprCEN/TS 19101 are largely conservative but still not accurate. Therefore, suggestions to improve the design equations are proposed.

Abstract Additive manufacturing, commonly known as 3D printing, is an innovative technique that allows the easy production of structural members. This paper presents an experimental investigation on the performance of 15‐5 PH stainless steel square and circular hollow section tubular stub columns manufactured by selective laser melting (SLM). An initial investigation was conducted to identify appropriate process parameters for the fabrication of the 15‐5 PH stainless steel specimens. Material properties of coupon specimens extracted from the printed stub columns using three different scanning strategies were examined through tensile coupon tests. The initial local geometric imperfections of the stub column test specimens were measured. In addition, a total of thirteen tubular stub column specimens were tested under axial compression. The ultimate loads, load‐axial shortening curves and failure modes were obtained for all the square and circular hollow section specimens. It should be noted that among the three scanning strategies adopted in this investigation, the scanning strategy S2 resulted in higher ultimate strengths of the SLM‐fabricated 15‐5 PH stainless steel tubular stub columns compared to those printed using the scanning strategies S1 and S3. Furthermore, the ultimate loads of thirteen stub column specimens were used to assess the applicability of European Code to additively manufactured stainless steel tubular stub columns.

The design of pultruded fiber-reinforced polymer (PFRP) structures can be governed by the beam-to-column joints because they exhibit brittle behavior. The objectives of this study were to understand the load path in PFRP structural joints, improve the failure mode, and delay the brittle mode of failure by modifying the joint configuration. The components in the PFRP beam-to-column joints were made from E-glass pultruded structural shapes. Ten beam-to-column joint tests were carried out, including parameters such as three different end distances (e1), cleat thicknesses (ta), and additional T-stiffeners. The conventional beam-to-column joints failed in a brittle mode, with cracks initiated at flange cleats followed by progressive stiffness reduction leading to ultimate failure. The analysis using strain data confirmed that there is a need for an additional load transfer component in the top flange to delay the first brittle failure. The use of T-stiffeners significantly increased the initial stiffness of the beam-to-column joint and delayed the first failure. The overall rotational stiffness of the PFRP beam-to-column joint was determined using the joint component method in the Eurocode. It is shown that the Eurocode method is conservative for connection components with higher end distances. The appropriateness of the stiffness prediction method in the Eurocode was demonstrated with a design example.

A comprehensive numerical investigation looking into the static structural behaviour and design of cold-formed steel elliptical hollow section X-joints (CFS-EHS-XJs) is reported. Detailed finite element (FE) models were developed for three distinct brace-chord configurations, incorporating cold-forming effects and heat-affected zones. The numerical models were validated against experimental results of 30 specimens through comparisons of load-deformation responses, joint resistance and failure mode. Parametric analyses encompassing 199 CFS-EHS-XJs were performed to examine the influence of key geometric parameters on joint static behaviour. The combined experimental and numerical dataset of 229 joints was used to evaluate the existing design provisions, including CIDECT and Eurocode 3 methods with equivalent circular and rectangular section approaches, and Wardenier's EHS joint design method. These comparisons revealed significant scatter and inaccuracy in joint resistance predictions using the current design methods. A new design methodology based on the ring model theory has been proposed and calibrated against the developed dataset. The proposed approach demonstrates superior accuracy and consistency in predicting CFS-EHS-XJ resistances compared to existing design methods.