Evaluation of bending fatigue damage for FRP with AE (2nd report - application o neural network and fractal dimension of wavelet)

摘要

Carbon-fiber reinforced plastic (CFRP) has become an important structural material in various fields. Therefore, it is important to evaluate the fracture modes and the fatigue damage of CFRP laminates. In the previous paper, the acoustic emission (AE) signals of CFRP laminates (i.e. [0°]{sub}11, [0°/90°]{sub}11 and [±45°]{sub}11 subjected to cyclic bending loads were recorded at each cycle during the fatigue test, and were analyzed by wavelet transform (WT) to evaluate the damage and fracture modes. By observing the resultant mapping of wavelet coefficients in the time-frequency coordinate plane at each cycle, it was proved possible to recognize the damage and the modes of CFRP laminates using nondestructive inspection. The purpose of this paper is to develop a methodology for evaluating fatigue damage and fracture modes using the characteristic features of the mapping. This system consists of an AF measuring device and a neural network. The network has learned the pattern sets dealing with the interaction between the features of the mapping and the fatigue damage. The character of the mapping expressed by wavelet coefficients in the time-frequency coordinate plane is bracketed wording clear scientifically by fractal dimensions that were led by the box-counting method in fractal theory. The effectiveness of this system is demonstrated by comparing results of the neural network with experimental data obtained from the fatigue tests.