Stress State Identification of Critical Bridge Components using Nonlinear Acoustics. Final Report for Highway Innovations Deserving Exploratory Analysis (IDEA) Project 158.

摘要

This project is developing and demonstrating the application of a nonlinear acoustics-based technique for identifying the stress state of critical highway bridge components through laboratory-scale and field testing. The project involved laboratory-scale and field demonstration of the proposed approach. Theoretical and numerical models were developed to identify the most sensitive ultrasonic waves to the level of stress on structural steel. The selected ultrasonic waves were tested on an L profile loaded uniaxially and a thick plate loaded uniaxially and bi-axially. The load value was increased incrementally, and ultrasonic measurement was taken at each loading step to develop a sensitivity curve of the stress-ultrasonic velocity relationship. Plates with different thicknesses and with and without holes as well as with and without paint were numerically modeled and analyzed under dynamic loading in order to identify variables affecting the ultrasonic waves. The plates were tested under stress-free state for comparing with the numerical results. It was demonstrated both numerically and experimentally that the Rayleigh wave arrival for frequencies above 700 kHz was not influenced by the plate waves for plates thicker than 0.5 inch. A special hand-holder was designed and fabricated that allowed adjusting transmitter-receiver spacing and varying sensor angle to control the propagating waves. The holder keeps the transmitter-receiver distance fixed during stress measurements, preventing any change in distance between the transmitter and the receiver when the structural material is under stress. The effect of bolt holes on the ultrasonic signature was identified. A change in arrival time of 0.03 μsec was identified as the error range to be considered in stress measurement. The structural shapes were tested under uniaxial and biaxial loading to identify the frequency stress calibration curve. The tests were conducted using 1 MHz ultrasonic transducers. The limitations of the approach due to the presence of paint and surface roughness were identified. The peak frequency for a painted sample was 0.66 MHz while that for a similar unpainted sample 0.89 MHz. An approach to overcome the paint influence on the stress-free reference point was developed through taking measurement from low stress region at a gusset plate under investigation in the field. The effect of recoupling on stress measurement was studied and an approach to reduce the coupling error was developed through a surface preparation process. If the coupling fluid was wiped over the testing surface and then wiped smooth, the variation decreased. A laboratory test was performed by placing the sensor wedges parallel to the application of loading. The purpose of this test was to measure the biaxial loading coefficients K1 (-4.25x10-5 MPa-1) and K2 (4.01x10-4 MPa-1) required for complex loading cases.