Remote monitoring and prognosis of fatigue cracking in steel bridges with acoustic emission

Abstract

Acoustic emission (AE) monitoring is desirable to nondestructively detect fatigue damage in steel bridges. Investigations of the relationship between AE signals and crack growth behavior are of paramount importance prior to the widespread application of passive piezoelectric sensing for monitoring of fatigue crack propagation in steel bridges. Tests have been performed to detect AE from fatigue cracks in A572G50 steel. Noise induced AE signals were filtered based on friction emission tests, loading pattern, and a combined approach involving Swansong II filters and investigation of waveforms. The filtering methods based on friction emission tests and load pattern are of interest to the field evaluation using sparse datasets. The combined approach is suitable for data filtering and interpretation of actual field tests. The pattern recognition program NOESIS (Envirocoustics) was utilized for the evaluation of AE data quality. AE parameters are associated with crack length, crack growth rate, maximum stress intensity and stress intensity range. It is shown that AE hits, counts, absolute energy, and signal strength are able to provide warnings at the critical cracking level where cracking progresses from stage II (stable propagation) to stage III (unstable propagation which may result in failure). Absolute energy rate and signal strength rate may be better than count rate to assess the remaining fatigue life of inservice steel bridges.