CAUSES OF ROOF FAILURE AND MODELLING OF PITCHED ROOF BLOW -OFF IN SOUTHWESTERN NIGERIA

Abstract

The increasing incidences of roof failures especially blow-off in recent times in Southwestern Nigeria has become worrisome in view of the damage done to adjacent structures and danger posed to building occupants and owners. There is the urgent need to devise methods to curtail failures and minimize the incidences of blow-off. This study was designed to investigate the causes and patterns of roof failures, recommend curtailment measures and develop a model to predict roof blow-off. Using purposive sampling technique, a survey of 3,780 roofs spread across Ekiti (450), Lagos (450), Ondo (360), Ogun (570), Osun (780) and Oyo (1,170) states was undertaken to establish the causes and patterns of roof failure. During the survey, timber samples at the point of roof construction (780), and those from failed (2000) and unaffected roofs (1000) were collected for moisture content determination in accordance with American Standard for Testing Materials (ASTM) D442 while the common nails used in construction were subjected to corrosion test in accordance with ASTM 1977. The integrity of nail joints was tested in accordance with ASTM 1761. Physical measurements of attic space and ambient temperatures, roof slopes, building dimensions and orientation were taken and combined with topography and courtyard effect to develop aerodynamic model to predict roof blow-off. The model was validated using post-model survey captured data. Data were analysed using descriptive statistics and regression analysis. Causes of roof failures included poor workmanship (30.5%), materials inadequacies (18.6%), design errors (14.8%), roof geometry (14.0%), topographical location (11.8%), age and environment (10.3%). Timber’s moisture contents were 12.0% to 24.0% during construction. Natural seasoning of these moisture contents to 7.0% in service, induced stresses on roof members. Temperature fluctuations between 20.0oC and 40.0oC promoted moisture condensation and dimensional changes in roofs’ wooden members. Poorly fitted joints reduced joint load from 103.1 ± 8.3 kg to 82.6 ± 5.1 kg. Nail diameter reduced from 21 ± 0.2 mm to 14.7 ± 0.3 mm within 90 days of exposure to water indicating potential reduction in joint strength. Blow-off occurred when ≤ 0; where MR and MO are resisting and overturning moments respectively. The model revealed that while gable roof could be adequate at the plain, hip roof with pitch angle between 40o and 60o would be appropriate on 5o and 10o slope hills respectively, with coefficient of multiple regression of 0.91 (p < 0.05). The model also revealed that optimum pitch angle was 55o and presence of courtyard reduced the wake and drag effects on roof. There were no statistical differences between the roof blow-off model predictions and post model survey data. The overturning moments for the rest were also greater than the resisting moment but they did not experience blow-off because of adequate anchorage of the sill. Roof failures in Southwestern Nigeria were caused by weakened joints resulting from corroded nails, interface gaps and wind effect. Remedial measures could include appropriate building orientation, proper anchorage, high pitch and adequate openings.