Nano-coated composite fastener: Structure health monitoring based on magneto-mechanical effect

Abstract

The heavy metallic fasteners and undetectable defects in the joining area have long posed challenges for composite structures. In this study, we propose a composite fastener that incorporates a nano-coating on the thread and embeds magnet in the shank. Based on magneto-mechanical effect, a physics-guided neural network (PGNN) is proposed to establish a correlation between magnetic density and bearing stress. For the strengthening effect of nano-coating, the bearing strength of composite joint reinforced by nano-coated composite fastener reaches 196, representing a 26.5% increase compared to the composite joint reinforced by an un-coated fastener. When all-composite joint fails, the nano-coated fastener shears off at the shank, whereas the un-coated fastener shears off at the thread. The predicted bearing stress from PGNN can reasonably quantify the real stress situation of testing joints. The maximum prediction error of the PGNN is merely 18.9 MPa, which is 51.7% lower than that of a neural network (NN) without physical guidance. Compared with the NN, the PGNN shows the abilities of reducing prediction randomness, better convergence, and higher prediction accuracy. This research presents a possible substitute for metallic fasteners to reduce the composite structure weight and monitor structure health.