Robust Test of the Flatbed Scanner for Air-Void Characterization in Hardened Concrete

Abstract

AbstractIn the province of Ontario, Canada, the Ministry of Transportation administers a quality assurance testing program for all new concrete construction where air content and spacing factor are measured. Testing is performed by qualified operators in accordance with ASTM C457/C457M-12. In this study, 324 routine samples that had been tested by qualified operators were obtained and analyzed for air-void parameters by an alternative automated flatbed scanner method. A strong correlation was found between the manual and automated methods. This research explored the premise that summary statistics from a small set of training samples could be used to define a set of global threshold levels to measure air-void parameters from a much larger population. Two different methods for setting the global thresholds were tested: one based on arithmetic means, and another based on modal location parameters derived from type I extreme value distributions. The frequency of false-negative errors (failure to detect a defect) was used as a criterion to assess the different threshold methods. Automated test results derived using global thresholds based on arithmetic means minimized the occurrence of false-negative events.