Simulation of the virtual shape meter

To improve the accuracy of shape control, shape detection and recognition is essential. In this paper, it established model of virtual instruments and a variety of defect shape model. It calculated the residual stress distribution along the width direction of the variable defect shapes used the finite element method and identified them through the SelPSO-BP shape pattern recognition model. It obtained that all the penetration defect shape can be used for neural network, which solved the problem of sample collection and can be obtained the residual stress distribution of the arbitrarily complex wave-shape.