Robust sub-gridding schemes for the finite-difference time-domain (FDTD) analysis of ground penetrating radar (GPR) scenarios

点击翻译按钮获取中文摘要

Enabling technologies for the detection of buried objects, such as land mines, power cables and water pipes, cracks and oil reservoirs, are extremely valuable for many areas of the Canadian economy, from construction and manufacturing to the oil industry. Among those, ground penetrating radar (GPR) stands out as a portable, economic and effective means of subsurface exploration. The operation of GPR is based on post-processing the reflections of a short, electromagnetic pulse emitted by the radar towards the ground. These reflections are caused by the interaction of the pulse with the air-ground interface and, most importantly, by the scattering of the pulse from subsurface layers and buried objects. Hence, an accurate interpretation of these reflections reveals useful information about the composition of the subsurface, as well as the position and condition of buried objects. This project is focused on dramatically improving the efficiency and accuracy of numerical electromagnetic simulators of GPR pulse propagation scenarios. Such simulations are used to build a "dictionary" of pulse reflections from buried objects, which is necessary for the interpretation of the reflected signals collected by GPR. Evidently, the reliability of GPR depends on the quality of this "dictionary", which is typically limited by numerical errors. This work will build on the applicant's extensive experience with multi-scale electromagnetic modeling (i.e. modeling of problems that span multiple scales in space and time), combined with the state-of-the art expertise in current GPR techologies offered by Sensors and Software, a Canadian champion in the GPR market worldwide.