NONDESTRUCTIVE TESTING OF DRILLED SHAFTS USING PATTERN RECOGNITION

AVAILABLE NDT TECHNIQUES CAN DETECT RELATIVELY LARGE DEFECTS IN DRILLED SHAFTS. TO PROVIDE INFORMATION ABOUT THE SIZE OR SHAPE OF THE DEFECT, SOME TECHNIQUES REQUIRE EMBEDDING SENSORS INSIDE OF HAVING ACCESS ALONG THE LENGTH OF THE SHAFTS. SUCH REQUIREMENTS LIMIT THEIR APPLICABILITY TO NEWLY CONSTRUCTED SHAFTS AND ALSO COMPLICATE THE DESIGN AND INSTALLATION PROCESSES. IN THE STRESS WAVE APPROACH, A MECHANICAL WAVE IS INTRODUCED AT THE TOP OF A SHAFT. FROM THE VIBRATION SIGNAL, THE TIME REQUIRED FOR THE WAVE TO TRAVEL DOWN AND BACK UP IS AN INDICATION OF THE EXISTENCE OF A DEFECT. DUE TO THE LACK OF ACCURATE MATHEMATICAL MODELS, ONLY A SMALL PORTION OF THE VIBRATION SIGNAL IS USED; THE REST IS OVERLOOKED. RECENT ADVANCES IN THE AREAS OF ARTIFICIAL INTELLIGENCE AND PATTERN RECOGNITION MADE POSSIBLE IN THE USE OF SUCH NEGLECTED INFORMATION FOR DIAGNOSTICS IN MEDICINE, MACHINERY MONITORING, AND SPEECH ANALYSIS. IN THIS PROPOSAL, WE WOULD INVESTIGATE THE FEASIBILITY OF UTILIZING THE PATTERN RECOGNITION TECHNOLOGY TO CHARACTERIZE DEFECTS IN DRILLED SHAFTS. LABORATORY MODELS WILL BE USED TO VERIFY THE FEASIBILITY OF THE TECHNIQUES.