Algorithms for synthesis and inversion of radar data from concrete bridge decks

There is an urgent need to develop methods for rapid identification of major deterioration in bridge decks and pavements. As traditional methods are slow and cumbersome, the focus has shifted to the use of modem nondestructive techniques such as Ground Penetrating Radar (GPR). GPR is a fast, non-contact technique, but interpretation of the radar datais difficult and requires complex analysis. Recently, great improvements in the analysis of radar data have been made and theoretical models for the prediction of subsurface condition of concrete structures have been developed. This paper describes models for predicting the velocity and attenuation of electromagnetic waves in concrete as a function of frequency, temperature, moisture content, chloride content and concrete mix constituents. The electromagnetic properties of concrete are predicted by aggregating the individual properties of its constituents: water, salt, air, cement paste, and aggregate solids. This mixture model, in conjunction with a rebar model developed to account for the reflection produced from reinforcing bars embedded within the concrete, has been utilized to synthesize radar waveforms for representative reinforced concrete bridge deck geometries. A least squares inversion procedure has been applied to the computer generated synthetic waveforms. This paper demonstrates the use of this inversion procedure to predict the spatial variations in volumetric'water content, salt content, and rebar cover.