Abstract

Ground penetrating radar (GPR) is a powerful tool for characterisation of the subsurface in a range of applications. Finite-difference time-domain (FDTD) forward modelling is often used to gain a more quantitative understanding of the interaction between GPR systems and the region of interest. This can be undertaken in 2D, where simulations fail to model multiple polarisations and require a number of simplifying assumptions for both electromagnetic fields and for modelling the environment. Alternatively full 3D modelling may be used, but this can be very computationally expensive. Here we present an idea, in contrast to the more formal 2.5D FDTD GPR modelling approach, which is based on using a thin slice of a full 3D model in cases that we want full 3D fields but we have a 2D modelling environment. The key issue with this approach is minimising, if possible, the error introduced by the very close proximity of the perfectly matched layer (PML) absorbing boundaries of the model in the invariant direction of the modelled geometry. We use gprMax, an open source FDTD GPR modelling package, to check the viability of this idea. An assessment of domain size required is made considering the error produced and the computational demands. Optimising the PMLs will be key in making this approach viable for future GPR modelling that provides a full 3D field solution in cases where a 2D geometry is a reasonable assumption.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call