Damage detection in composite laminated beams based on multi-scale enhancement of mode shape curvature using stationary wavelet transform

This paper proposes a damage detection method enhanced by a multi-scale product feature, which combines mode shape curvature and stationary wavelet transform (SWT) to address the difficulty in locating damage in composite laminated beams using traditional mode shape identification. The effectiveness of this method has been validated through numerical modelling and experimental validation. The primary innovation lies in the construction of an integrated damage identification framework that fuses ‘Mode Shape Curvature-SWT-Multi-Scale Product’. This framework effectively amplifies the local singularity induced by damage by using mode shape curvature as the initial input. Furthermore, it employs a multi-scale product strategy to achieve synergistic enhancement of damage features and nonlinear suppression of background noise. By using mode shape curvature as the SWT input, low-frequency damage features are more effectively elicited, causing the D3 frequency band to exhibit distinct damage peaks. Implementing the multiplication strategy further improves damage localization accuracy by 33%. In addition, a scanning laser vibrometer (PSV-400) was employed to measure the mode shapes of a carbon fibre laminated beam containing a pre-set crack, and damage detection on mode shape curvature. The results demonstrate that the proposed method can accurately identify damage locations, effectively validating its efficacy and practicality for detecting damage in composite laminated beams.