Localization of surface dent deformation and inter-laminated damage in CFRP laminates under low-velocity impact behavior based on multi-channel one-dimensional convolutional gated recurrent unit

Abstract

Lamb wave has become one of the most promising methods for carbon fiber-reinforced plastics (CFRP) damage detection. However, the wave velocity difference caused by CFRP anisotropy makes the time information-based localization methods lack accuracy. Obtaining high resolution and accuracy with limited excitations and sensors is difficult, which leads to difficulties in localizing small-size impact damage. Therefore, this paper proposed multi-channel one-dimensional convolutional gated recurrent unit (MC1-DCGRU) for surface dent deformation and inter-laminated damage caused by low-velocity impact behavior localization combining the feature extraction capability of the multi-channel one-dimensional convolutional neural network (MC1-DCNN) and the temporal information capturing ability of gated recurrent unit (GRU). The localization mean Euclidean distance (MED) of single-point and two-point cumulative damage is 4.29 mm and 5.78 mm, respectively. The proposed method overcomes high-resolution requirements for damage localization. It achieves higher accuracy without increasing excitations and sensors, which also achieves multi-point cumulative damage localization.