Some conclusions from the measurements of temperatures and their gradients on eigenfrequencies of bridges

摘要

The ambient air temperature and solar radiation are affecting the soil and asphalts' stiffness and hence the eigenfrequencies of a bridge. Very often eigenfrequencies are automatically determined by special algorithms from structural response data generated by ambient excitation and measured by permanently installed sensors that is sometimes called "output only analysis" or "operating modal analysis". Additionally the ambient air temperature is registered and finally the eigenfrequencies are analyzed versus the air temperature. The graph is normally a scatter diagram and each point is one measurement. In general the eigenfrequencies decrease with increasing temperature, whereas often linear regression is used to determine the line of best fit. But the slope of this straight line and the width of the scatter around the regression line differ from bridge to bridge. Especially this scatter field around the mean value at a determined temperature complicates the detection of damage, which is also often based on eigenfrequencies' reduction. Hence the difficulty among others consists in separating damage from environmental effects. In Luxembourg the eigenfrequencies of a new two field composite bridge with steel girder, concrete slab and a relative thick asphalt layer were monitored over years. Additionally several temperatures at different points of the structure were registered. The eigenfrequencies were determined by Stochastic Subspace Identification (SSI). It turned out that the slope of the first eigenfrequency versus temperature is extremely high with 7‰ per °C and that any part of the bridge has at any moment its own temperature. Hence the bridge is not characterized by only temperature and moreover the temperature difference between steel and concrete is essential for the deviation from the mean value of the eigenfrequency at a given structural temperature. Especially in summer the day and night variations of ambient temperature are high due to the high solar radiation and hence the temperature gradients in the bridge are important, whereas in winter with overcast sky the gradients are small. It can be shown that the temperature gradient between steel base frame and top concrete slab is influencing the measured eigenfrequency. This knowledge is of high importance prior to the analysis of the data for damage detection and a simple reduction of the related uncertainty is possible by using only days with low temperature gradients for damage detection.