Reference-free crack detection under varying temperature

Abstract

A new paradigm for damage diagnosis is proposed by developing a damage detection technique that eliminates the need for initially collected baseline data. Traditional techniques often identify damage by comparing the current data set with the reference data collected from the pristine condition of the structure being monitored. However, this conventional pattern comparison approach is shown to be vulnerable to other changes such as temperature variation that may not be relevant to defects of interest. One of potential advantages of the proposed reference-free approach is that false-alarm due to these undesirable variations could be minimized particularly for field applications where varying structural and environmental conditions impose significant challenges for damage diagnosis. In this paper, the effect of varying temperature on the previously developed reference-free crack detection technique is investigated using an aluminum plate with an increasing crack depth.