Detection of fatigue crack under temperature changes using stress waves and statistical analysis
Abstract »In aerospace engineering, one of the critical problems threatening the structural integrity especially for aging aircraft is fatigue crack. This paper presents a study on monitoring and detecting of fatigue crack in a Titanium compact test (CT) specimen under temperature changes using stress waves and statistical analysis. The CT specimen is installed on a MTS testing machine for generating fatigue crack under cyclic load and an electrical oven is employed to simulate temperature changes around the CT specimen. Two PZT transducers are surface mounted on the CT specimen to excite and receive stress waves at different intervals. The stress wave signals under different temperatures are processed by continuous wavelet transform (CWT) to extract multiple features to characterize the existence and growth of the fatigue crack. To consider the temperature effect on the stress wave signals, a statistical multi-variable outlier analysis is adopted and the multiple features are fused by Mahalanobis distance, which is used to determine whether fatigue crack exist in the CT specimen from a probabilistic and statistical prospective. The experimental results have demonstrate the applicability and effectiveness of the proposed method.