3D step frequency GPR Asphalt pavement stripping detection: Case study evaluating filtering approaches

Stripping between hot mix asphalt layers can cause premature failure of pavements. Knowledge of the lateral location, depth and extent of stripping can affect the timing and type of pavement preservation, maintenance, or rehabilitation. Since subsurface stripping extent is important for maintaining the road network, Minnesota Department of Transportation (MnDOT) uses forensic measures such as coring to make this evaluation typically. The destructive nature and lack of coverage provided by this accepted method makes finding nondestructive methods for detecting stripping of in-situ asphalt pavements an important need for MnDOT and many other agencies in the National Road Research Alliance (NRRA). The case study presented in this paper uses a DX1821 antenna to collect data at a frequency step of 20 MHz ranging from 50 MHz to 3050 MHz with a dwell time of 7.52 µs using a GeoScope Mk IV control unit on a full-scale asphalt pavement constructed on a National Center for Asphalt Technology (NCAT) test track with built-in stripping. Signals from the known stripped and non-stripped locations were used to evaluate statistical and energy based approaches. It is shown that a maximum energy ratio method, an approach that has been successfully applied to acoustic emission applications in the past, can improve signal clarity for stripping detection using ground penetrating radar. The results are demonstrated in the ability to enhance the stripping indicators in both traditional impulse time-history data and C-scan depth slices when applied to step-frequency ground penetrating radar data. The effectiveness of the energy ratio-based method in detecting fabricated stripping makes it a promising method for scoping pavement projects for construction activities. By identifying the presence and extent of stripping, road owners will have information that can allow them to avoid misapplication of preservation techniques that do not address the root stripping problem thus minimizing premature failures, or avoid major rehabilitation in situations where the evaluated lack of stripping suggests a preservation or less costly alternative can be applied.