Self-heating and acoustic emission guided fatigue damage evolution analysis and rapid life prediction for thermoplastic composite ultrasonic welds

With the increasing usage of thermoplastic composites, ultrasonic welding has become a critical joining technique owing to its superior performance. However, due to the heterogeneity of the welded joint, the damage evolution mechanism is complex and hard to predict its fatigue performance. In this paper, the fatigue damage evolution behavior of CF/PEEK ultrasonic welded joints is systematically investigated by combining Infrared Thermography (IRT) and Acoustic Emission (AE) data. The results reveal that both the frequency and amplitude of AE signals significantly increase with the loading amplitude, accompanied by a noticeable temperature rise on the specimen surface. By applying K-means clustering to AE amplitude data, three damage modes during the fatigue failure process are identified. Furthermore, the fatigue damage evolution process of the joints is analyzed in conjunction with optical microscopy observations. Based on K-means clustering and Miner's Rule, a fatigue life prediction model utilizing AE data is developed. Compared with traditional fatigue test results, the predicted fatigue limit exhibits an error of only 1.07%, and the predicted S-N curve falls within the 95% confidence band of the experimental S-N curve. This study provides both theoretical and experimental support for the fatigue life prediction of CF/PEEK ultrasonic welded joints.