Modelling structure cladding interaction in large single-storey steel-framed buildings

  • Michael Roberts

Student thesis: Phd

Abstract

As single-storey steel framed structures, colloquially termed “sheds”, have grown and modern cladding systems have proliferated, the assumption that buildings may be designed independently of the cladding must be challenged. A thorough literature review of stressed skin theory highlights the fundamental shortcomings with the theory: the component properties of modern cladding systems are unknown, additional failure modes are not accounted for within the codes, the simple equilibrium models upon which the code expressions are derived do not necessarily scale and the compatibility model used to combine frames and cladding is only applicable in a very limited number of situations. Sway-frame structures are typically not designed taking the global effect of the cladding into account, however regardless of the designers’ intentions the cladding has a high in-plane stiffness, and “parasitic diaphragm action” (unwanted stressed skin action) will occur; this may lead to premature failure of the cladding. There are currently no accepted methods to accurately assess how different cladding systems influence the load path of vertical or horizontal loads through a structure, except in the simplest of cases, nor is there any way to evaluate the load factors at which the cladding, or the interface between the cladding and the structure, is likely to fail. A modelling framework is proposed which allows complete models of very large steel-framed structures, including detailed consideration of the cladding, and the interaction between the cladding and the primary frames, to be quickly generated and analysed. No tests on such structures exist in the public domain, therefore validation was performed in three phases. Initially, individual models of the primary frames were validated against frame results in the public domain. Next, novel finite element models of the four common cladding systems (trapezoidal sheeting, sandwich panels, cassettes and two skin built-up systems) were developed using the “Component Method” which aims to minimise the number of required elements by encapsulating complex, often nonlinear, component behaviour within simple elements. Due to the large number of elements required, and complex connectivities, custom data generators were considered mandatory. These generators automatically produce detailed cladding models of any dimension in a short length of time. The results from several diaphragm panels, which were tested to destruction and whose results are available in the public domain, were used to validate these models. This comparison served to validate not only the model of the tested diaphragm, but also the underlying assumptions and simplifications made, as well as the implementation of the data generator. Finally, to validate the interfaced models, models of two semi-full-scale structures, both of which have been tested previously, were automatically generated. The results were found to compare very well with the measured results, further increasing the confidence in the generated models. Extrapolation of the structures showed how effects of scale influence diaphragm action. These structures together with a much larger multi-bay frame building, were then analysed with both the simplified methods (including the current Eurocode procedure) and using the developed framework; the inadequacy of the simplified methods was clear, particularly on larger, multi-bay structures. The framework is novel because the generated models contain the primary structure together with detailed consideration of the cladding; every component (i.e., every single fastener, individual roof sheet, cleat etc) is included in the model allowing for assessment of both the global stiffening effect on the primary frames, and assessment of damage to the cladding caused by parasitic forces as load is applied and redistributed. The framework is applicable to the three cladding systems in common usage: profiled steel sheeting, sandwich panel
Date of Award31 Dec 2023
Original languageEnglish
Awarding Institution
  • The University of Manchester
SupervisorYong Wang (Supervisor)

Keywords

  • Portal frame design
  • Finite element modelling
  • Sandwich panels
  • Elastic-plastic analysis
  • Structure cladding interaction
  • Stability
  • Metal cladding
  • Stressed skin design
  • Steel structures
  • Parasitic diaphragm action

Cite this

'