Abstract
It is well known that the crack propagation in a structure is a serious problem from the structural integrity point of view. In order to prevent serious crack propagation, a circular hole is using as a crack arrester. This paper describes the application of a hierarchical neural network to the estimation of a crack arrester position in a solid with the FEM. We solve the 2D crack propagation problem in three point bending model. NN is used as a method of asking for a crack arrester position. It was shown that NN is very effective in prediction of crack arrester position.