Utilization of Wilcoxon-Mann-Whitney statistics in assessing the reliability of nondestructive evaluation technologies

Abstract

Nondestructive evaluation (NDE) technologies are increasingly being used for the condition assessment of existing structures. In order to ensure best funding allocations and most effective maintenance and repair activities, the reliability of NDE technologies for a particular application needs to be carefully assessed to determine the margin of error (MOE) of NDE data. This research aims to investigate the utilization of “Wilcoxon-Mann-Whitney (WMW) Statistics” for the reliability assessment of NDE technologies. The reliability of three of the most widely used NDE technologies for the detection of subsurface cracking in concrete was assessed using WMW analysis. The NDE technologies assessed were: Impact Echo (IE), Infrared Thermography (IRT), and Ground Penetrating Radar (GPR). Two in-service bridge decks and two full-scale fabricated deck slabs were nondestructively tested to provide suitable data for analysis. The in-service bridge decks contained areas of mild and severe subsurface cracking as determined from previous testing performed by the Missouri Department of Transportation (MoDOT). The fabricated decks, on the other hand, contained areas of synthetic cracking of different sizes and depths. The reliability indices determined from WMW analysis for the three NDE technologies discussed in this manuscript are exactly equivalent to the indices determined previously by the same authors using Receiver Operating Characteristics (ROC) analysis. Results of WMW analysis, by the way, indicated that IE and IRT technologies were more reliable than GPR for the detection of concrete subsurface cracking. The final WMW indices of IE and IRT data were much higher than the corresponding indices of GPR data for both the in-service bridge decks and the fabricated slabs. The interesting equivalence between WMW and ROC analyses may encourage other researchers to utilize the former method in order to investigate the reliability of other NDE technologies in other applications. The findings of this research may also assist decision-makers considering the use of these NDE technologies in order to form decisions regarding maintenance and repair needs for bridges.