Abstract:
High internal pressure induced through openings in buildings is one of the major causes of building related damage during strong wind conditions. In combination with external pressure, it can exacerbate the net load on the building envelope leading to its failure, either in whole or in part. Although, various aspects of the “internal pressure problem” have received attention since the 1970’s, scope for further understanding of its effects continues to be repeatedly felt during major post-cyclone forensic investigations around the world even today. This thesis attempts to investigate some of the issues related to this problem in detail, using a combination of analytical, numerical, wind tunnel and full-scale studies, with a goal of proposing realistic design provisions of internal pressure for a range of building volumes and opening sizes. The thesis begins with a discussion on the wind loading chain of internal pressure inductionand goes on to explain the load transfer mechanism from the turbulent approach flow on to internal pressures, induced through dominant openings in buildings, using a multi-stage admittance process and mechanical admittance approach. A detailed investigation of the effect of opening size, location, terrain roughness and wind direction on internal pressure in rigid, non-porous buildings using volume-scaled wind tunnel experiments is carried out in combination with complementary numerical analysis. The numerical predictions of mean and fluctuating internal pressures are found to conform well to the experimental results. In particular, the internal pressures are found to exhibit a dynamic response to external pressures near the opening resulting in Root Mean Square (RMS) and peak internal pressure values in excess of those predicted by the quasi-steady analysis. The size and location of the opening are found to have a considerable influence on the mean and fluctuating internal pressures induced inside the building. Calibration of loss coefficients by the way of a spectral match between the measured and predicted spectra resulted in values ranging from 6 to 36 with increase in opening size. Investigations into the effects of envelope flexibility in a building with an opening are conducted using a generalized two-degree-of-freedom dynamic model and its simplified counterpart, the single-degree-of-freedom quasi-static model, available in literature. The models predict a reduction in the Helmholtz resonance frequency with increase in envelope flexibility along with an associated increase in the damping of internal and net envelope pressures. This is indicated by the lowering of the resonant peak in the internal and net pressure linearized admittance functions, consequently resulting in a somewhat smaller Root Mean Square (RMS) as well as peak internal and net roof pressures relative to the case of a rigid building. Some wind tunnel experiments are carried out to validate the model predictions. Treatment of the influence of background porosity on internal pressure is based on the development of a simplified and computationally efficient analytical model based on certain simplifications, such as lumping the leakages in the leeward side and usage of a temporospatial leeward mean external pressure coefficient, shown to be adequate for all practical purposes. The presented model adequately linearized is used to develop closed form nondimensional design solutions for estimation of the ratio of internal to external pressure fluctuations and gust factors for a range of building volumes, opening sizes and porosity ratios. Experiments carried out with different combination of leakages, lumped and uniformly distributed on the roof and leeward wall, in the presence of dominant openings of various sizes are used to justify the simplifications adopted in the derivation of the model and to validate the model predictions. A performance analysis of the model carried out in conjunction with some approaches, found in the literature on benchmarking data, exhibits the relative advantage of the model in terms of accuracy, practical usage and computational efficiency. Significant research efforts are directed towards developing a generic analytical model of internal pressure by combining the effects of skin flexibility and background leakage in the building envelope. Two variants of the model, which includes the effect of roof external pressure characteristics, are developed; that for a building with a dynamically flexible and leaky envelope and the more simplified quasi-statically flexible and leaky envelope. The linearized model and the analytical transfer function estimates are used to numerically quantify the combined damping effects of flexibility and leakage for a range of roof flexibility and leakage combinations representative of real building envelopes, and compared with the quasi-steady predictions. The model predictions of internal pressure and the damping due to flexibility and building porosity are validated using wind tunnel experiments involving a flexible Styrofoam roof and uniformly distributed background leakage. A generalized theoretical model of internal pressure dynamics in a building with multiple openings on a single wall with highly correlated external pressure is developed. Analytical and wind tunnel studies on a model building for the case of two closely spaced dominant openings in a single wall showed that internal pressure in such configurations increase with increase in the ratio of opening sizes, and become almost equal to, but slightly less than that for the most critical single opening configuration under normal onset flow, when the combined area of the two openings become double the critical single opening size. For oblique wind angles between ±40-60º, the RMS and the peak ratio internal pressure coefficients for the two-opening configuration of area ratio unity are found to be much higher than the most critical single opening configuration. This is thought to be due to a comparatively stronger “tangential flow excitation” compared to single opening configurations. A theoretical and experimental investigation of the net dynamic load on the partition wall resulting from the internal pressure in a compartmentalized building with and without an internal opening is undertaken. The fluctuating net load on the wall, predicted using a spectral model, is validated using wind tunnel experiments. The net dynamic load on the partition wall is found to decrease with an increase in the size of the internal opening. While the mean load on the partition wall with an internal opening is found to be predictably low due to pressure equalization on either side of the wall, gust load factors in the range of 2-3 in combination with Helmholtz resonance observed in the study can have severe implications on the direct and fatigue wind loads on partition walls, not usually designed to withstand such loads. Large mean and fluctuating loads due to Helmholtz resonance of internal pressure is also observed for the limiting case of a sealed partition wall without an internal opening. A comprehensive investigation of the transient response of internal pressure and the likely level of overshoot experienced by buildings, following suddenly created openings during severe storms, is studied by numerically simulating the analytical models of internal pressure at different temporal proximities to the occurrence of a peak instantaneous external pressure near the opening. Different combinations of envelope flexibility and background porosity for the Texas Technical University (TTU) test building are used for the study. Computational Fluid Dynamics (CFD) modelling is used to calibrate the ill-defined coefficients in the internal pressure analytical models. The level of overshoot is predicted to be higher than the pre-sequent and subsequent steady state resonant response when the openings are created zero to one oscillation period preceding the occurrence of that peak external pressure. Nondimensional overshoot factors of internal pressure for a range of building volumes and opening sizes, obtained by numerically simulating the non-dimensional governing equations for different envelope flexibility-leakage combinations, are presented in the form of design curves. The results show that while the overshoot phenomena may be significant for buildings and structures with very small internal volumes and comparatively large openings (like small garages and garden sheds), it gets damped out in large industrial structures with flexible porous envelopes, even when a sudden dominant opening is created rapidly during the occurrence of a strong gust. Efforts into addressing the codification issues related to the steady state resonant response of internal pressure involves estimation of influence factors for a range of building volumes and opening sizes. The approach requires determination of the correlated fluctuating external pressure field around the opening as well as the internal pressure fluctuations from the measured data in order to evaluate the influence factors using the covariance integration approach coupled with Eigenvalue analysis technique. The results are used to empirically develop non-dimensional design equations of internal pressure influence factors, for a normal onset wind flow, that are sensitive to the opening size and location as exhibited in the measurements. Research into the different factors influencing internal pressure finally culminates in a field study involving internal pressure measurements in an industrial warehouse with a single dominant opening of different size. A range of wind directions are investigated at moderate wind speeds. Evidence for resonance of internal pressure for a windward opening situation is found. Resonance, if any, is found to be hardly discernable at other opening configurations. This lack of resonance is attributed to the inherent leakage and flexibility in the envelope of the building. Numerical validation of the measured data and particularly, the importance of correctly parameterising the different dampers are exhibited, absence of which is shown to result in an unrealistically high value of the opening loss coefficient. Ratios of the fluctuating and the peak internal to opening external pressure obtained in the study are found to compare well with other reported full scale measurements and adequately covered by the nondimensional design equation available in the literature. The results and conclusions of this study are expected to be useful in further recognizing the importance of internal pressure as a significant parameter in building design considerations and pave the way for more comprehensive design provisions of internal pressure in wind loading standards of the future.