Machine learning (ML) methods for the structural health monitoring (SHM) of composite structures rely on sufficient domain knowledge as they typically demand to extract damage-sensitive features from raw data before training the ML model. In practice, prior knowledge is not available in most cases. Deep learning (DL) methods, on the other hand, can obtain higher-level features from raw input data and have proven superior in several applications. This paper proposes a Convolutional Neural Network (CNN) based approach for the delamination prediction in CFRP double cantilever beam (DCB) specimens using raw local array strain measurements via distributed optical fiber sensors. The conventional CNN architecture is modified to perform regression, as the delamination size is a continuous value. 1D and 2D CNN architectures are deployed and compared and different techniques are exploited to encode 1D spatial strain pattern series as 2D images. Raw strain patterns collected during static testing are used to train the CNNs, while testing is performed on unseen raw fatigue strain patterns, showing the CNN ability to automatically extract discriminative features from the non-pre-processed static strain pattern-based signals that generalize to raw fatigue signals as well. This strategy has the potential to reduce fatigue testing expenditures while also shortening the time required to gather training data.@en

This work addresses the effect of the test temperature on the cohesive model parameters for the 3M Scotch-Weld™ 7260 B/A epoxy adhesive. It extends a previous experimental work done at room temperature and further develops a previously proposed parameter identification method based on optimization. Double-Cantilever Beam (DCB) and End-Notched Flexure (ENF) tests were conducted at four different temperatures: 20°C, 40°C, 55°C and 70°C. Moreover, Single Lap Joints (SLJ), and bulk specimens’ tensile tests were carried out. Finite element analyses of DCB and ENF tests were performed and optimization algorithms were used to calculate constitutive cohesive model parameters. Three approaches were followed. First, cohesive parameters were derived from bulk tests. Then, optimization with two variables for each mode was conducted, to identify the parameters ei and σu,i (i = I, II) of a triangular traction separation law. In the third, only σu,i, was taken as variable, while ei was derived from the elastic modulus of the bulk adhesive. Finally, the cohesive models were applied to simulate the response of the SLJ, and numerical results were compared with experiments. The two free variables optimization method allowed to obtain the most accurate predictions of the maximum load and SLJ displacement at failure.@en