GPR detection of subsurface voids and its validation based on similarity principle

In this paper we adopt the similarity principle to study GPR response to voids under road surface or in the levees and dams by cross-examination of GPR data acquired using both high-frequency antennas on laboratory bench-size models and low-frequency antennas in the field. In physical world the similarity principle is widely used in theoretical study and engineering practice. In short, by the similarity principle, the following statement can be true: the experiment result obtained from a smaller scale can be extrapolated to a larger scale as long as the ratio of the wavelength of the signal to the size of the target keeps the same. Complete similarity includes three aspects: geometric similarity, kinematic similarity, and dynamic similarity. For GPR detection of road voids, the geometric similarity (i.e., depth, size, and shape of the voids) and kinematic similarity (i.e., radar wave propagation velocity) are more critical. The dynamic similarity (i.e., the reflectivity or the contrast of the materials) can be automatically met when we use the similar material in the model and the field. For example, if we can detect a 10-cm void by using 2-GHz antenna, we sure should detect a 1-m void by using 200-MHz (0.2-GHz) GPR antenna in the same medium. That implies that all the phenomena observed in the high-frequency test can be extrapolated to the low-frequency situation. Meanwhile, experiments with high-frequency GPR on smaller-scale models are more economical and easier to be carried out. We conducted the high-frequency experiment by using 2.6-GHz antennas over a sandbox model in the lab. We also examine the validity of the similarity principle to GPR void detection by comparing the high-frequency, lab-scale data to low-frequency, field-scale GPR data acquired from GPR surveys on highways and levees.