A NEURAL NETWORK FOR DEPTH DETERMINATION OF SEPARATIONS BETWEEN A RUBBER MATRIX AND REINFORCING STEEL BELTS

摘要

A neural network for depth determination of separations between a rubber matrix and reinforcing steel belts has been developed. To evaluate a separation depth up to 10 mm, the neural network has been designed and constructed of four sub-neural networks. Each sub-network has been trained by using simulated time-domain signals reflected by the structure containing separations at various depths. A classifier that employs a cross-correlation algorithm and a gate are used to preprocess input data and to send the signal to the desired sub-network. The neural network has been tested on both simulated and measured signals. The estimated depths of separations agree well with the actual ones.