Evaluation of buffeting predictions of a cable-stayed bridge from full-scale measurements

Over 40 years ago, Alan Davenport introduced statistical methods for the analysis of wind buffeting of slender structures, which still form the basis of most buffeting calculations today. However, there remains some uncertainty over some of the functions used in the method, and there have been surprisingly few detailed comparisons of buffeting predictions with measurements from full-scale bridges. This paper presents results from long-term full-scale monitoring of the Second Severn Crossing cable-stayed bridge and compares them with the design predictions and other variations of Davenportʹs method. ntributions of aerodynamic and structural damping of numerous modes are separated, believed to be the first time this has been achieved from site measurements of a long-span bridge with a bluff cross-section. The aerodynamic damping of low-frequency vertical modes was found to be approximately half of the values predicted by quasi-steady theory, while for low-frequency lateral modes it was somewhat higher, causing reduced lateral response amplitudes. gh the amplitudes of vertical and torsional vibrations predicted in design appeared to be reasonable, re-calculation using the measured wind turbulence and total damping revealed significant discrepancies from the measured responses. Finally, the three-dimensional aerodynamic admittance is estimated from the full-scale data, and comparisons made with various functions that have been suggested.