Elsevier

Structures

Volume 27, October 2020, Pages 141-150
Structures

Effect of roof to wall connection stiffness variations on the load sharing and hold-down forces of Australian timber-framed houses

https://doi.org/10.1016/j.istruc.2020.05.040Get rights and content

Abstract

Roof to wall connections (RWCs) play a major role in wind load sharing and load transfer of timber-framed houses. The lack of coherence in the load sharing and load transfer can cause premature failure of timber-framed houses. Thus, it is essential to provide RWCs with adequate strength and stiffness, to ensure the dynamic roof wind loads are adequately shared and transferred through the structural system to the foundations. Typical construction defects in RWCs (i.e. missing nails) reduce the stiffness of the RWCs, introducing discontinuities in load paths, contributing to total roof failure under windstorms. This study combines realistic spatially and temporally varying wind loads, captured from model-scale wind tunnel-studies, with a validated finite element model of the roof structure, that accurately describes the load sharing through the structural system and evaluate the effect of RWC stiffness variations and realistic wind loading variation on the load sharing and hold-down forces of Australian contemporary timber-framed houses. The combined effect of RWC stiffness variations and realistic wind loading variation are compared to the traditional design method approach (i.e. tributary area method and use of simplified design roof pressures from wind loading standard AS/NZS 1170.2 (2011)). Results show that the maximum RWC hold-down force from AS/NZS 1170.2 (2011) is about 10% greater relative than the maximum RWC hold-down force from realistic dynamic wind loads, and overestimates wind loads for most wind angles. The maximum RWC hold-down force derived by the load sharing method with defective connections (i.e. single and double nails missed on the RWCs), induced hold-down forces 10% higher than that of the traditional method at neighbouring RWCs. Further, hold-down force variation between traditional and load sharing methods can increase up to 50% in a real house roof structure, as the stiffness of their RWCs vary throughout the system due to the material non-linearity and construction defects/human errors.

Introduction

Timber-framed houses are the most common residential building in Australia and the roof of these structures experience high wind loads during windstorms. Thus, the fundamental design objective and task of a house’s timber-framed structural system is to securely transfer the wind loads from roof and walls to the foundations, through their inter-component connections. A better understanding of the load path is essential to evaluate the structural response of a house to windstorms. Uplift or vertical load transmission and lateral load transmission are the two basic types of wind load transmission in a timber-framed house structure. The vertical load transmission typically depends on the roof loads and structural response of the roof structure (i.e. roof cladding, batten, truss, ceiling, ceiling cornice, etc.) and their connections (i.e. batten to cladding, batten to truss, and roof to wall connections). Discontinuity in the load sharing and load transfer will increase loads to structural elements, potentially leading to premature damage to the structural system.

Several research studies associated with load paths have been conducted worldwide [e.g., [1], [2], [3], [4], [5], [6], [7], [8], which have provided some qualitative information of the load sharing of a house structural system. However, there is a lack of quantitative data on load sharing through timber-framed house construction as it is very complex. There are several factors affecting the load sharing in timber-framed construction, such as spacing of members and fasteners (i.e. roof trusses, battens, cladding fixings, etc.), stiffness of members and their connections, fascia beams, roof sheathing type and orientation, etc. [9].

Damage observed after recent windstorms, such as in Cyclone Yasi (2011), Cyclone Marcia (2015), Cyclone Olwyn (2015), Brisbane Severe Storm (2014), and the Tornado in Oklahoma (2015) reported roof structure failures due to combination of high wind speed, high internal pressure and construction defects/human errors, which initiated a cascading structural failure. These failures indicate that timber-framed houses remain vulnerable in cyclonic and non-cyclonic regions. After these events, research and investigations were conducted in order to assess design and structural strength of timber-framed houses [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34]. The windstorm damage investigations show that the loss of roof cladding resulted in extreme damage, leading to a significant loss of strength, which resulted in progressive collapse [10], [14], [16], [17], [23], [30].

Research into full-scale house tests has shown the complexity of actual load transmission was difficult to incorporate in design, without full-scale testing of individual house designs [11], [12], [13], [18], [19], [20], [21]. Recently, several other full-scale house tests have been carried out to evaluate the structural response and load transmission of the North American and Canadian residential houses [15], [25], [26], [27]. The studies by Datin et al. [26], Zisis and Stathopoulos [35], Morrison [15] and Doudak et al. [27] have evaluated the wind load sharing and load path network through a typical timber-framed house in North America and Canada. The results from these studies cannot be used to evaluate the structural stability of Australian timber framed houses as their structural systems, construction methods, RWCs and design wind speeds are different to houses from North America and Canada. These research studies and damage investigations have shown that the roof, roofing components and connections are the most vulnerable structural elements in the timber-framed house structural system [10], [14], [16], [17], [23], [30]. Furthermore, these previous researches have highlighted that the construction defects on the inter-component connections reduce their stiffness and create discontinuity in the load sharing and load transfer to the foundations, which contributes to overloading of connections and cascading house failures to windstorms.

Satheeskumar et al. [32] highlighted that every timber-framed house has construction defects (i.e. missing nails) on the RWCs, reducing the stiffness and uplift capacity up to 40% of the standard specified capacity. The safety factor specified in Australian standards and building codes (AS 1649 [36], NCC [37], AS 1684.3 [38], AS 1720.1 [39], AS 4055 [40], AS/NZS 1170.2 [41]) cover up to 25% of variability. Satheeskumar et al. [33] found that these poorly constructed RWCs, structural and non-structural elements in the roof structure could create variations of the vertical reaction forces at RWCs. This can increase the vulnerability of the timber-framed house structural performance under windstorms. The Australian standards and National Construction Codes [37] have requirements that every structure and structural element should be designed to satisfy the “performance-based design and deemed-to-Satisfy condition”. To satisfy these conditions, the timber framed house and their inter-component connections stiffness, structural response and load sharing must be investigated.

In standard design practice, the hold down forces derived from a single truss analysis does not account for the loads shared to and from adjacent trusses and the connections stiffness variation. This standard design practice could underestimate or overestimate the hold down forces, where underestimation of hold-down forces my overload connections, contributing to the whole-roof failure under windstorm. Thus, this paper aims to investigates the effect of RWC stiffness variation and how its influences the load sharing through the structure and the RWC vertical reaction force of a timber-framed house. This paper further investigates the hold-down forces at the RWCs from realistic wind loads across the roof during a windstorm by combining the wind loads from wind tunnel tests with the load sharing through the structural system with a finite element model (FEM) analysis. Results are compared with RWC hold down forces derived from standard design practice (i.e. basic tributary area analysis using wind loads derived from the wind loading standard AS/NZS1170.2 [41]).

Section snippets

Contemporary house structural system

A field survey of contemporary houses under construction in non-cyclonic regions of Australia, near Brisbane and Melbourne, was conducted by a team from the Cyclone Testing Station (CTS), to determine representative houses and their structural systems [34]. Based on the analysed survey data, the houses in Melbourne are of similar size, shape and construction type to houses in the Brisbane region. The houses’ construction types in both Melbourne and Brisbane are brick veneer with timber trusses

Influence function βi

The βi can be determined from the single truss analysis (i.e. traditional method) and full or part of house structure analysis (i.e. load sharing method).

Results and discussion

Truss hold-down forces were derived for Trusses A, B, C, D and E using the time varying pressure distributions for all 36 wind angles obtained from the wind tunnel study and the βi values obtained from the traditional and load sharing methods. For the discussion in this section, a dimensionless hold-down force coefficient (CN) is introduced below,CNt=X(t)/12ρU-h2ANwhere AN is the total tributary area of single truss (i.e. 5.6 m2).

Conclusions

This study investigated the effects of stiffness variation in the RWCs on load sharing in the roof structure and hold-down forces of roof trusses by modelling the structure of the timber-framed house. Wind loads on the RWCs were obtained by using vertical reaction influence functions obtained from the single truss analysis (i.e. traditional design method) and FEM analysis of part of a contemporary timber-framed house (i.e. load sharing method). Realistic dynamic wind loads were obtained via a

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This study is part of the research conducted by Climate Adaptation Engineering for Extreme Events Cluster funded by CSIRO Land and Water Business Unit. The authors gratefully acknowledge the funding support of CSIRO and the assistance provided by Dr David Henderson and Prof. John Ginger from Cyclone Testing Station, James Cook University, Townsville, Australia.

References (49)

  • D.V. Rosowsky et al.

    Establishing design values for uplift connections in light-frame construction

    J. Testing Evaluation, ASTM

    (1998)
  • D.V. Rosowsky, S.D. Schiff, T.A. Reinhold, P.R. Sparks, B.L. Sill, Performance of Low-Rise Structures Subject to High...
  • D.V. Rosowsky et al.

    Reliability of light-frame wall systems subject to combined axial and transverse loads

    J. Struct. Eng.

    (2005)
  • T.D. Reed et al.

    Uplift capacity of light-frame rafter to top plate connections

    J. Architectural Eng. ASCE

    (1997)
  • R. Shivarudrappa et al.

    Sensitivity of load distribution in light-framed wood roof systems due to typical modeling parameters

    J. Perform. Constr. Facil

    (2013)
  • G.R. Walker, “Report on Cyclone Tracy-Effect on buildings-Dec 1974”, Australian Department of Housing and...
  • G.N. Boughton, G.F. Reardon, Simulated Wind Test on a House: Part 1. Cyclone Testing Station, James Cook University,...
  • G.N. Boughton, G.F. Reardon, Simulated Wind Test on a House: Part 2. Cyclone Testing Station, James Cook University,...
  • G.N. Boughton, G.F. Reardon, Simulated Wind load Test on the Tongan Hurricane House. Cyclone Testing Station, James...
  • G.N. Boughton, D.J. Henderson, J.D. Ginger, J.D. Holmes, G.R. Walker, C. Leitch, L.R. Somerville, U. Frye, N.C....
  • M.J. Morrison, Response of a Two-Story Residential House under Realistic Fluctuating Wind Loads. PhD Thesis, Department...
  • E. Keith et al.

    Hurricane Andrew—structural performance of buildings in South Florida

    J. Perform. Constr. Facil

    (1994)
  • C. Leitch, J.D. Ginger, B. Harper, L.R. Somerville, N.C. Jayasinghe, P. Kim, Investigation of performance of housing in...
  • G.F. Reardon, Simulated Cyclone Wind loading of a Nu-Steel House. Cyclone Testing Station, James Cook University,...
  • Cited by (0)

    View full text