Application of forward modeling and appropriate processing algorithm to locate aqueduct by GPR

As a high resolution, nondestructive and rapid geophysical tool, Ground Penetrating Radar (GPR) provides the best method that can be used for locating superficial targets. A gallery of the aqueduct is well locatable by the GPR method due to distinct permittivity contrast between the target and the surrounding media. In this research work, as a 2D study, the GPR data acquisition was carried out using 250 MHz shielded antenna in Mashkan area near Kashan city. The target is not well visible in any processed radargram obtained from data acquisition. This paper aims at achieving the best processed section that is strongly similar to real conditions of the subsurface. In this circumstance, the proficiency of suitable processing algorithm will be very useful. To evaluate the ability of GPR, in addition to the suitable processing and appropriate interpretation of actual data, a model that is required based on the conditions of the subterranean features is designed and implemented. What is more prominent in this paper than that in previous similar studies carried out in Iran is the manipulation of modeling in GPR method and comparison of the results with the interpretation of actual data. In modeling, we generally tried to make the best simulation of wave reflection received from the underground layers obtained through the propagation of electromagnetic waves according to the targets and anomalies. Therefore the purpose of this article, according to the properties of the mentioned aqueduct, is to apply forward modeling and designing of model based on the expected conditions of the study at first. Then we want to investigate responses of electromagnetic waves when the processed section has been made by appropriate processing parameters. In this paper, modeling has been done based on a finite difference method. We can simulate complex consequential answers from complicated cases inasmuch as it has benefits such as flexibility, ability to simulate and model complex environments. Further, this method has acceptable responses as seen in cases when compared to other numerical methods, and it is more practical. The finite difference time domain (FDTD) method can model with high accuracy, when the properties of the underground layers are concerned. The final congenial model is acquired based on the characters of the layers, several parts and boundaries of this case study. The mentioned areas have been divided into cube-shaped sections named orthogonal grid cells. For each cell, the intensity of the electric and magnetic fields are considered according to the vertical and horizontal components of the invariable electromagnetic fields on the basis of the measured values of physical properties and the soil type. The resulting radargram is achieved through processing parameters such as Dewow filter, Spreading and Exponential Compensating (SEC), band-pass filtering and background removal. The soil of the study area contains sand, gravel and silt with an average electrical resistivity of about 200 Ωm. The analysis of traveling velocity of the electromagnetic wave is implemented using average electrical permittivity according to components of the soil. Our results show a close agreement between the model and the processed radargram, which consequently implies that the proposed processing procedure indeed provides reasonable results.