Investigation Of The Effect Of Modular Construction Details On The Lateral Behaviour Of Cold-formed Steel Framed Shear Walls

Introduction

In recent years, structural- and cost-efficiency, durability as well as sustainability [1] have ramped-up the use of cold-formed steel (CFS) profiles in many countries as both structural and non-structural elements [2]. Shear walls made of CFS members (studs, tracks and blockings) and sheathed with wood or cement particle (CP) panels are one of the lateral load resisting systems (LLRSs) adopted in lightweight steel construction [3]. The main codes that currently define methodologies for the design of CFS structures are AISI S400 (2015) [4] and AS/NZS 4600 (2018) [5]. However, in the current market, CFS modular buildings can include construction details that could influence their lateral behaviour, and are not covered by the current lateral design provisions and guidelines for CFS structures [4]. Furthermore, the complex analyses and design procedures related to the significant number of thin components, which are locally unstable and show several failure mechanisms, require an advanced lateral behaviour investigation [6]. For the past two decades, full-scale testing has been largely adopted to investigate the behaviour of CFS framed shear walls under lateral loads [7], [8], [9], [10], [11], posing the basis for design and code development.

Virtual testing (i.e., numerical simulation) has also been largely adopted for advancing understating of CFS structural capacity and predict their behaviour at varying of loading conditions and structural components, to the point that now it can be deemed of primary importance for the purpose of optimizing the structural performance of the CFS framed buildings, in particular, at the early stages of the product development process.

Over the last decade, several endeavours have been devoted to the numerical simulation of CFS framed shear walls subjected to monotonic and cyclic (quasi-static and dynamic) lateral load. Stewart model (1987) [12] was deemed suitable for the simulation of the experimental tests carried out by Nisreen Balh (2010) [13] on CFS framed shear walls, however, the strength deterioration observed in the test results was not considered. Martínez and Xu (2010) [14] proposed a simplified, yet accurate, approach for modelling a CFS framed shear wall using a 16-node shell element with equivalent geometric and material properties derived from the actual properties of a CFS framed shear wall. Liu P. et al (2012) [15] adopted Pinching4 model [16] developed by Lowes and Altoontash (2003) [17] in order to characterize the cyclic behaviour of CFS wood-sheathed shear walls; this model was calibrated based on experimental test results and reproduced the hysteretic behaviour with an acceptable accuracy (below 10 % difference). Based on the same model, 2- and 3-dimensional models were established by Leng J. et al. (2017) [18] for nonlinear dynamic response history analyses of full CFS systems (2-storey buildings). Shamim and Rogers (2013) [19] simulated the nonlinear response history of two-storey CFS framed shear walls under seismic load using Pinching4 model that was calibrated based on results of dynamic tests carried out by the same authors. Vigh et al. (2014) [20] developed and calibrated a simplified strut model with the adoption of the Ibarra-Medina-Krawinkler constitutive model [21] to represent the deteriorating hysteretic loops of CFS corrugated steel-sheathed shear walls. Buonopane et al. (2015) [22] developed a computationally efficient screw-based modelling protocol in OpenSees software for CFS OSB-sheathed shear walls. Two hysteretic models that take into account strength and stiffness deterioration as well as pinching, have been developed and implemented in the official OpenSees release (version 2.4.5 and above) by Kechidi and Bourahla (2016) [23] to simulate CFS wood- and steel-sheathed shear walls behaviour under monotonic and cyclic lateral load. It is worth noting that all the above-described numerical simulations adopted beam-column elements to model the CFS frame members. Consequently, local and distortional buckling or their combination were not captured. David Padilla-Llano (2015) [24] proposed a numerical framework for CFS framed shear walls that captures the nonlinear cyclic behaviour of critical components including frame members (chord studs) as well as screws. More advanced modelling techniques have been undertaken by Hung Huy Ngo (2014) [25] through the adoption of SpringA element in ABAQUS to simulate the shear behaviour of the screws connecting OSB sheathing to CFS frame members. Deverni et al. (2021) [26], [27] replicated the same efforts with a simplistic approach of modelling the shear behaviour of the sheathing-to-CFS screws using CONN3D2 element in ABAQUS assuming a constant angle between the screw deformation and the global horizontal axis throughout all levels of lateral demand on the shear wall. Moreover, with no un- and re-loading paths defined, SpringA and CONN3D2 elements can merely be adopted in the simulation of the lateral behaviour of CFS shear walls under monotonic load. The Bouc–Wen–Baber–Noori (BWBN) (1993) [28] model was used by Nithyadharan and Kalyanaraman (2013) [29] to capture the deteriorating behaviour, in terms of strength and stiffness deterioration with severe pinching, that has been observed in the screw fasteners between the sheathing and CFS frame members under cyclic load. Subsequently, the BWBN constitutive model along with a variably oriented spring pair element have been implemented in ABAQUS as a user-element (UEL) to replicate the cyclic behaviour of the screws under shear demand [30]. In all the above-described modelling efforts, the aim was to replicate the results of tests on conventional CFS framed shear walls rather than optimizing the structural performance of CFS framed shear walls with constructional details that are not covered by the current lateral design provisions and guidelines.

The innovation in the study presented in this paper is to uncover the effect of modular construction details on the behaviour of laterally-loaded CFS framed shear walls and to optimize the screws pattern and sheathings layout efficacy in this LLRS. Therefore in this paper, first experimental tests on sheathing-to-CFS screws (Section 2) and tensile tests on CFS frame members (Section 3) are presented to characterize the basic components of the shear walls under investigation. An advanced modelling protocol is proposed in Section 4, which uses radial springs with experimentally-derived backbone curves implemented in UELs, to model the shear behaviour of sheathing-to-CFS screws, while accounting for the deformation of the shear wall frame members. The proposed modelling protocol is validated using results coming from experimental tests carried out by the authors [31], where a good agreement has been achieved. Subsequently, the effect of additional details that are commonly adopted in CFS modular construction and go beyond the scope of the current lateral design provisions is assessed (5 Parametric study, 6 Screw shear demand evaluation, 7 Comparison with design codes). The main details include: (i) presence of floor and ceiling ledger beams on the interior face of the shear wall, (ii) sheathing boards having different sizes from the overall shear wall and thus the presence of both vertical and horizontal seams, (iii) use of cement particle (CP) boards at the bottom stripe of the shear wall and (iv) different screw spacing in the top and bottom stripes from the middle part of the shear wall. Lastly, rules have been established for optimizing the screws pattern and sheathings layout efficacy in the above-described LLRS.