Finite element analysis and field observation of a residential roof subjected to hurricane winds

Thumbnail Image
Jungmann, Matz
Major Professor
Fouad Fanous Sivalingam Sritharan
Committee Member
Journal Title
Journal ISSN
Volume Title
Research Projects
Organizational Units
Journal Issue
Is Version Of
Civil, Construction, and Environmental Engineering

Every year, natural disasters take a great toll on the economy and infrastructure of the United States. When specifically considering the insured property losses of U.S. natural disasters, seven out of ten highest natural disasters are caused by a hurricane event. Hurricane seasons have increased in activity over the past few years, with the 2005 Atlantic storm season seeing 14 hurricanes, 7 of which were major hurricanes. This thesis focuses on an analysis performed under the winds of Hurricane Katrina on an instrumented roof structure of a house built in Pensacola, Florida. The data recorded is compared to results obtained from a theoretical finite element (FE) model as well as those stipulated from a design standard.;The instrumented test structure was a low-rise residential house with 76 pressure cells attached to the hip roof sheathing and 68 load cells placed between the rafters/trusses and the top of the wall. A wind anemometer was located near the test structure to record the wind's angle of attack and speeds during a hurricane event. Although Hurricane Katrina data is the focus of this report, the test structure, which was built in 2002, has experienced several other hurricanes. Data recorded during Hurricane Katrina data was chosen to be analyzed and compared against the FE analysis results as this storm event produced one of the highest peak wind speed of 56.2 mph.;The results from the FE model was compared to Hurricane Katrina field data by performing a gravity analysis, analyzing equivalent loads from three wind speeds, and comparing the field data to the requirements of ASCE 7-02 Standard. The gravity analysis was important to verify that the assumptions used in the FE model were satisfactory and produce the load distribution similar to that observed for gravity loads determined from the field data at very low wind speeds during Hurricane Katrina.;The three static wind pressures were applied to the FE model to determine if the FE model was comparable to the wind loading at the particular times during Hurricane Katrina. The model showed that the FE model was significantly higher in wind loading than the field data because all sheathing elements in the FE model had wind pressures applied normal to the test roof. Also, the linear interpolation used for the applied wind pressures to the FE model could have been different from what was measured in the field over the test roof.;Thirdly, the ASCE 7-02 Standard wind loads were analyzed for both applied pressures and total wind loading on the test roof for a comparative analysis to the field data. This was done to investigate the adequate representation of wind loads in the ASCE 7-02 Standard. The results showed that the components and cladding (C&C) wind pressures were much greater than the main wind force-resistance system (MWFRS) since C&C accounts for the localized effect that could occur. The conservativeness of the ASCE 7-02 Standard could not be addressed for the MWFRS applied pressures at the peak wind speed of 56.2 mph during Hurricane Katrina because the averaged load cell reactions of the field data fluctuated greatly in adjacent load cells. However the Standard wasn't conservative when comparing the C&C design pressures to the field data.;Overall the loads induced by wind effects on the test roof during Hurricane Katrina were small in comparison to gravity effects, if the peak wind speed of a hurricane event been greater, then the field data may provide a better comparison to FE analysis results. This leaves many areas of future studies for finite element analysis such as dynamic loads, fatigue, or cyclic loading. A field test of the test roof is greatly needed to get a satisfactory feel for correct gravity loading to be established for the future studies of the test structure. Also, the precipitation during the hurricane event should be studied as this is suspected to have influenced the load cell data.

Subject Categories
Mon Jan 01 00:00:00 UTC 2007