Volumetric water content estimation of concrete by particle swarm optimization of GPR data

Abstract

This study investigates the Volumetric Water Content (VWC) of concrete block by inversion of Ground-penetrating radar (GPR) traces by Particle Swarm Optimization (PSO) method. PSO method was applied to GPR data set, which contains 24 days of data. PSO was used as a tool to estimate relative dielectric permittivity, relative magnetic permeability and electrical conductivity values of concrete. The relative dielectric permittivity has a direct relationship with the water content. Generally, the relative dielectric permittivity value is estimated from velocity estimation by picking reflections on radargrams. However, the picking of reflection may be erroneous and subjective. In this case, this study was aimed on PSO method as an inversion tool to estimate relative dielectric permittivity to calculate VWC of concrete instead of velocity calculation approach to avoid man-made error during first arrivals picking. Firstly, PSO application on the 3rd day data of the fresh concrete was introduced in detail. Afterwards, PSO results on all data set were presented. In order to show the accuracy of PSO calculations, the percentages of relative error (%RE) values, which indicate the fit between observed and calculated GPR traces, were estimated. %RE values were varied between 3.11% and 6.05%. After the PSO process, estimated dielectric permittivity values were used to calculate the VWC. So, the PSO application was utilized to monitor water content change of concrete. Nevertheless, besides the traditional methods, the information on the water content can be obtained with the proposed method non-destructively.