Wind Pressure Characteristics Study Of A Flat Low-rise Building By Field Measurement And Wind Tunnel Testing

Windstorm disaster, as one of the most serious natural disaster in the world,causes a large number of casualties and property losses every year. It has been widelyrecognized that most of the losses are not induced by the wind directly, but due to thelocal damage or overall collapse of building structures under the action of strongwinds. Low-rise buildings take over the vast majority of the damaged buildings andstructures, whose main failure forms are the breakdown of cladding components alongthe windward edges of the roofs and the overall collapse of the roof structure. In thisthesis, the data of velocity and pressure were obtained by the field measurementsystem including an instrumented experimental low-rise building during a strongtropical storm. By analyzing these field measured data, characteristics of wind fieldnear the ground and wind pressure distribution on the roof of a flat low-rise buildingunder strong wind environment are presented and discussed. By means of wind tunnelexperiments, essential features of characteristic turbulences on the roof are discussed.Furthermore, the wind-induced internal pressures are investigated via the sets ofdifferent opening sizes and shapes at the windward corner of the roof. The mainresearch results are summarized as follows:The applicability of the hypothesis that the natural wind can be treated as astationary process is proved. The probability density function of longitudinalfluctuating velocity matches well with the Gaussian distribution, while the lateralfluctuating velocity diverges from the Gaussian distribution, deviating to the left atthe negative side and to the right at the positive side. The longitudinal fluctuatingvelocity contains more large-scale vortices while high-frequency energy is moreprominent in the lateral fluctuating velocity.At the height of10m, the field measured longitudinal wind turbulence intensityfor B terrain category changes from0.12to0.17, with a mean value of0.14, and thelateral wind turbulence intensity ranges from0.09to0.13, with a mean value of0.11.For A terrain category, the longitudinal wind turbulence intensity varies from0.08to0.15, with a mean value of0.11, and the lateral wind turbulence intensity changesfrom0.07to0.12, with a mean value of0.09. The measured mean values oflongitudinal turbulence intensity don·t appear to be much different from thespecification in Chinese load code, but the field results violently fluctuate because of the variations of the surrounding terrain conditions in reality.Under the action of the incident oblique flow, the windward corner of the roofexperiences serious suction as the results of the formation of conical vortices. Themeasured max shape coefficient comes to seven times bigger than that of thespecification in the design code, which obviously illustrates that the load codeunderestimates the real wind pressures. The fluctuation of pressures on the roof isclosely related to the incident flow, but the change of wind direction plays a main roleon the variation of the wind pressure coefficients.When the separate bubble is formed, the mean pressure coefficients androot-mean-square pressure coefficients both show symmetrical distribution about theridge line of the building. The absolute values of the pressure coefficients arerelatively large near the windward edge and get a little smaller as going far away fromthe edge. With the increase of the roof ridge height, the turbulence intensity isreduced, the length scale between the stagnation on the windward wall and thewindward edge of the roof is increased, which can account for the increasing tendencyof the absolute values of mean pressure coefficients and the decrease of theroot-mean-square pressure coefficients when the building gets high. Only in a certainrange of wind directions, the conical vortices can appear. In the range of20°to70°,there are two conical vortices simultaneously along the windward edges on the roof.Under the effect of conical vortices, the mean pressure coefficients androot-mean-square pressure coefficients exhibit a cone-shaped distribution. There is nomutual attraction or interaction of other forms between the paired conical vortices andthey move independently, influencing the pressure distribution within their respectivescope.As the wind angle increases, vortex core position and the attachment position ofthe conical vortex along the short windward edge get closer to the edge, while thechanges of the conical vortex along the long edge are just opposite. What is more,with the change of wind direction, the rangeability of attachment position issignificantly greater than that of vortex core position. The quadratic curves fit thevortex core position and attachment position well. Point vortex model underestimatesthe test results of these areas. The modified formulas based on the point vortex modelmatches well with the experiment data of wind pressure profiles. The height of vortexcore is in good agreement with a quadratic curve without the constant term, whichcommendably presents the relationship between the height of vortex core and thecross section position of conical vortices and departs less from the testing results. Internal wind pressures of the building are highly strong and can be consideredto be evenly distributed. Due to the superimposed effect of inside and outside windpressures, the mean values of wind suction on roof are weakened sharply, evenpositive pressures come out. However, the fluctuating net wind pressure displays asmall wane. The internal pressures possess high relevance both in time and frequencydomain.Two apparent spectral peaks turn up in the power spectrum of internal windpressure, one corresponds to the vortex shedding frequency and the other relates tothe Helmholtz frequency resulting from the opening at the windward roof corner.Because of the serious associative relationship between the external pressure in thescope of the conical vortex and the internal pressure, the superimposition effectoffsets the influence of conical vortices, which results in the disappearance of vortexshedding frequency spectral peak. However, Helmholtz resonance phenomenon onlyexists in internal pressure, so the spectral peak at Helmholtz frequency cannot beoffset. It·s worth noting that the Helmholtz frequency will not alter with the winddirection.