Resistance-welded thermoset composites: A Bayesian approach to process optimisation for improved fracture toughness

• Published in: Composites. Part A, Applied science and manufacturing Applied science and manufacturing, 2024-02, Vol.177, p.107894, Article 107894
• Main Authors: Maierhofer, Thomas, Loukaides, Evripides G., Carr, Craig, Bisagni, Chiara, Butler, Richard
• Format: Article
• Language: English
• Subjects: Bayesian theory; normal distribution; thermoplastics
• ISSN: 1359-835X
• DOI: 10.1016/j.compositesa.2023.107894

Joining thermoset composites via resistance welding offers a novel highly efficient assembly method for next-generation aerospace structures. Resistance-welded joints combine the benefits of bonding with the capacity for high-volume manufacturing rates and eliminate the need for complex surface preparation. The influence of key welding parameters on the joint performance is investigated by assessing the Mode I fracture toughness. Double Cantilever Beam specimens with different welding parameter combinations are manufactured, tested and compared with each other. Thermoset laminates are made weldable by co-curing a chemically compatible thermoplastic film with an uncured thermoset laminate. A Bayesian approach is used to study the correlation between processing parameters and to select parameters yielding high performance by training a Gaussian process emulator. Observed Mode I fracture toughness values are comparable to high-performance thermoplastic composites. This is equivalent to an improvement of approximately 290 % in Mode I fracture toughness when compared to a co-cured thermoset joint.