A Study on Damage Diagnosis and Prognosis of Lap Joint using Features

Purpose: This paper proposes a feature-based model to estimate lap joint crack.
Methods: First, the piezoelectric sensor data of the aluminum lap joint was subjected to signal processing to filter noise and extract the target area data. Second, three features were extracted from the filtered data. Moreover, a crack length estimation model was derived as a function of loading cycles for each feature. Finally, a crack length model as a function of loading cycles was proposed by combining the features. The performance of the proposed model is verified by comparing the result of the crack length estimation with that of the Paris-Erdogan model.
Results: The error of the proposed model against the measured values was between 0.03% and 1.8%, proving that the combined model outperforms the individual feature-based model. Additionally, the Paris-Erdogan model showed less accurate prediction than that of the proposed model.
Conclusion: This study confirms an excellent result by proposing a model with combined features for estimating the crack length of the lap joint.