How Accurate Is Ground Penetrating Radar (GPR) for Bridge Deck Condition Assessment?

As Ground Penetrating Radar (GPR) is increasingly employed for assessing the condition of bridge decks, quantifying and controlling the quality of GPR data becomes an important challenge. This paper uses a methodology to assess the accuracy of deck condition measures, and demonstrates the methodology in a case study involving real data. The latter are generated during GPR applications on a large bridge deck, and are processed using a commercial image-processing algorithm. The measures extracted from the processed GPR data are the rebar reflection amplitude and the dielectric constant of the deck material. The accuracy of the GPR assessments is evaluated by comparing the core data (‘ground truth’) to the GPR measures. The methodology uses appropriate statistical characteristic curves for quality control. It is based on a use of data to plot the probabilities of true detection vs. false detection. Image interpretation requires the use of a threshold value (typically established from experience) selected as to optimize true and false detection rates. The results of the case study show that rebar reflection data detect defects of the bridge decks at 75 percent true detection rate with 15 percent false detection rate. However, the results using the dielectric constant data fail to discriminate defects from the decks. The reason for this is attributed to the presence of latex modified concrete (LMC) overlay on the bridge deck, and the fact that values of the dielectric constant are derived from surface reflection data. Findings from the analysis and important conclusions are presented and discussed.