Stage 4 Report, July 2004
Development of a Cyclone Wind Damage Model for use in Cairns, Townsville and Mackay
The purpose of the study is to assess the amount of wind damage likely to occur in the major coastal communities of Townsville, Cairns and Mackay, and to obtain a distribution of that damage over each township for a cyclone of given intensity and path. This study focuses on houses and flats (i.e. domestic construction) as they represent the shelter mainly used during cyclone events. For this reason, knowledge of housing performance (resilience) is crucial for agencies involved in disaster mitigation and response, as it serves to target disaster amelioration.
Towns have a mixture of house types. Differences in size, shape, window size, cladding type, roof shape, age, and methods of construction have an effect on the resilience of the house to resist wind forces. Houses also have varying degrees of exposure to wind forces, with those dwellings located in a suburban environment gaining shelter from surrounding structures as distinct from those exposed houses near the sea or in open terrain. Topographical features such as hills can concentrate or divert wind flow. Wind speeds impacting on a community will vary according to a tropical cyclone’s intensity, size and distance from the community. Therefore an assessment of the wind resistance of housing requires knowledge of house types and their distribution throughout the community.
The concept of assessing the vulnerability of housing in the community is to combine these effects of the varying wind forces on the population of houses of different shape, size, construction and age that comprise the township. All of these factors make it difficult to accurately predict the likely damage to a community’s housing from the cyclonic winds.
The study is not about trying to assess the performance of individual houses or even small groups of houses but more a general estimation or indication of potential damage from a particular cyclone event.
As the different geometries and house elements react to the impact wind in different ways, the distribution of the pertinent housing parameters across the community is required. Comprehensive data surveys of the external features of housing were carried out for Cairns, Townsville and Mackay as part of the Tropical Cyclone Coastal Impacts Program (TCCIP). The data collection for Cairns and Mackay was conducted by AGSO Cities project for the TCCIP. Mackay City Council and Cairns City Council provided the TCCIP housing survey and spatial data to the Cyclone Testing Station (CTS). The CTS conducted the physical attribute housing survey for Townsville along with detailed structural inspections of houses.
From knowledge of the development of the towns forming this study, a review of current and superseded building regulations, detailed house inspections, and an overall survey of the housing stock, the methodology was to categorise the various house styles into six general classes based on overall geometry and construction technique, covering houses from the 1860’s through to present day forms.
Because a house frame is a very complex structure it does not lend itself to a straightforward structural analysis, as there are a multitude of building elements providing load sharing and in some cases redundancy. The CTS developed housing wind resistance models used here are to give an estimate of the likely failure mode and failure load for a representative proportion of houses. Findings from full scale house testing, and component joint tests, have been incorporated in the estimation of failure capacities.
The CTS housing wind resistance models focus on the chain of connections from the roof cladding fixings down to the wall tie-down. Findings from damage surveys and full scale house testing results conclude that the predominant mode of failure is associated with the lower load capacity of the joints in the house structure. Five failure modes (at cladding, at cladding to batten connection, at batten to truss/rafter connection, at truss/rafter to wall connection, and of wall tie down connection) were derived for each of the six house models for the two primary wind orientations for both full and partial internal pressure.
When sudden internal pressurisation of the house occurs, such as that caused by a breach in the building envelope by wind driven debris breaking a window or by a blown in door, theload can be dramatically increased on the structure. Failures of house elements add to the debris field impacting on down wind houses that increases their risk of failure leading to higher probabilities of further failures down wind (eg a snowball effect). The report demonstrates the sensitivity of the model output to the estimation of houses subjected to full internal pressure and also highlights the importance of mitigating the risk of full internal pressure (sound structure, window protection, securing potential debris prior to the cyclone, etc).
The deterministic wind field model SEACATd was amalgamated with the CTS housing wind resistance models to provide a user-friendly software interface to estimate the number of houses suffering wind induced damage from a cyclone of given parameters (track, intensity, radius etc). In the wind model framework, topographic factors such as ground height, surrounding terrain and neighbouring structures were represented in accordance with the boundary layer coefficients as detailed in the Australian Wind Load Standard AS1170.2- 1989.
As a verification check, the output of the combined SEACATd and CTS housing wind resistance model has been calibrated against real data. This is not an easy task, because there is a dearth of damage survey data available since, thankfully, severe cyclones are rare events. Tropical Cyclone Althea, which hit Townsville in 1971, is the one event that can be used to calibrate the output from the wind model and pre 1970’s house models. Althea caused overall damage levels in the order of 16 % and a damage level of 68 % for an exposed coastal suburb, which were published in a JCU report on Cyclone Althea. For pre 1970s housing the CTS model estimated the percentage of houses damaged at 15 % for a shielded suburb and 71% for an exposed coastal suburb, which compares favourably with the reported damage. There are no other events in the survey areas that can be used for the calibration of the complete system. However, damage investigations following cyclones Tracy and Vance for example, have been used for validation of specific data points.
The model generated output is intended to give an estimate of the number of houses sustaining wind damage when subjected to a derived wind speed. Limitations, assumptions and variability in the house resistance models, house class distribution, wind speed model, and terrain models should all be considered when assessing the model output. The report and model results are of a sensitive nature and care should be taken in the interpretation of the output and findings.
Variations in model parameters such as the assessed construction quality of the housing stock, the composition of the housing wind resistance models, and the estimated proportion of houses that might be susceptible to full internal pressure can play a large factor in the model’s estimation of overall damage numbers.
Results from the study detail, for defined districts within the modelled community, the estimated numbers of houses suffering damage and give a broad-brush classification of the types of structural failures, using varying winds from a generated cyclone or a blanket wind speed across all districts. In running the model with various cyclone scenarios, areas of housing stock within the modelled districts are shown to have varying relative vulnerability through combinations of topographic and building class features.
The study has shown that the modelled region of Cairns is estimated to have a lower relative overall vulnerability of its housing stock to wind loads than the other modelled regions of Townsville and Mackay, because it has a greater proportion of housing built following the introduction of engineered house building requirements in the early 1980’s.
In varying the house model parameters, the report highlights the potential reduction in damage to houses by protecting the house envelope to reduce the probability of subjecting the structure to full internal pressure.
The project has demonstrated a practical yet advanced methodology for combining multiple CTS house wind resistance models, representing real communities, with the SEA tropical cyclone wind field modelling system. There is much opportunity to extend and improve upon the sophistication of the Housing wind resistance models and the cyclonic wind field model. However, the CTS and SEA are of the view that this work already represents the most advanced assessment of house performance in high winds yet undertaken in Australia.
Download full report
- Full report: high resolution (PDF, 6.1M, last updated 01:38PM, 5 December 2008)*
* Requires Adobe Reader
Last updated 22 February 2016