CFRP composite fatigue damage pattern recognition using ultrasonic guided waves

The damage patterns of carbon fiber reinforced polymer (CFRP) composites are various and evolutionary with the increase of service life. With the development of condition-based maintenance and digital twins for aerospace composite structures, it is particularly important to recognize damage patterns online. The traditional acoustic emission technology commonly used for damage pattern recognition has some limitations, such as passive monitoring, and large amounts of data. In this paper, a new ultrasonic guided wave-based two-stage method integrated with similarity measurement and improved Fuzzy C-means clustering is presented to recognize fatigue damage patterns of CFRP laminates. The entire damage pattern recognition process was conducted in two stages. In Stage I, the dominant fatigue damage modes are firstly identified by the sequence similarity of guided wave envelope signals measured in different fatigue periods. The similarity of waveform shapes is analyzed and obtained by dynamic time warping. By exploring similarity curves under different measurement strategies, the evolution trend of dominant fatigue damage was characterized. In Stage II, an improved Fuzzy C-means clustering method is proposed to solve the mixing characteristics of guided waves under the coexistence of multiple damage modes. Furthermore, a cumulative membership degree indicator is presented to inversely predict the evolution trend of fatigue damage patterns. Compared with the pattern recognition results by acoustic emission, it is shown that the proposed method can effectively identify the fatigue damage mode of CFRP.