Estimation of Bending Strength of CFRP Cross-Ply Laminates from Damping Capacity Using by Neural Network

摘要

For the wider use of CFRP, the mechanical properties of CFRP laminates having damage in their interior should be investigated and it is also desirable to develop a technique to estimate their strengths from other mechanical properties by a nondestructive method. In this paper, the damping capacities of the CFRP cross-ply laminated beams having the different dmaged areas are measured by using an impulse hammer test and their static and impact bending strengths are also obtained by a four point static bending and a three point impact bending test. As a result, the relations among the damaged area, the damping capacity and the static or the impact bending strength of the CFRP cross-ply laminated beam can be experimentally obtained. Next, by using some of the expermental data as tutor data, neural networks are developed for estimating the static and impact bending strengths of CFRP crossply laminated beams from their damping capacity. Finally, this neural network is shown to be a quite useful method to estimate bending strength non-destructively because the other static and impact bending strengths as output data can be obtained from the damping capacity as input data.