Noncontact Ultrasonic and Computer Vision Assessment for Sawcut Initiation Time

Abstract

Sawcut timing of contraction joints in concrete pavements is affected by interactions between the time-dependent concrete properties, notch depth, base friction, climate, and type of saw. Even with an understanding of these variables, the experience of the sawcut operator is equally important. A novel framework is introduced to predict the sawcut initiation time (ts) by integrating a one-sided noncontact, ultrasonic testing system (UTS) to estimate the concrete’s final setting time (tf), and a computer vision (CV) algorithm to determine an acceptable level of joint raveling. Six distinct concrete mixtures were cast into slab specimens (500×500×150 mm3) in the laboratory. Dry sawcuts were then introduced to the specimens at every 30–60-min intervals. The noncontact UTS estimated the final setting time (tf) by transmission of leaky Rayleigh waves from a 12° incident angle. The CV-based technique quantified the joint raveling damage of each sawcut through a newly defined raveling damage index (RDI) and provided a magnitude of RDI over time for each concrete mixture. By surveying ten different contraction joints on two field projects in Urbana, Illinois (US), an acceptable damage index (RDIa) was statistically defined to be 3%. Given the time-dependent RDI for each concrete mixture in the lab and RDIa threshold of 3%, the sawcut initiation time (ts) was estimated and then correlated to the final concrete setting time (tf). For the six concrete mixtures, ts was estimated as 1.3×tf plus 39 min in order to achieve sawcut joint raveling levels at RDIa. The proposed UTS and CV-based techniques provide a standardized method to predict sawcut initiation time (ts) and assess the quality of a sawn joint by any field construction personnel.