Identification of Plastic Properties of Metallic Structures by Artificial Neural Networks Based on Plane Strain Small Punch Test

In order to assess the strength of aged and in service components, Small Punch Test (SPT) has emerged. However, it has two disadvantages, firstly using of the hemispherical punch which is difficult to manufacture in most conventional workshops and secondly the known difficulties in obtaining the flat disk samples. This paper discusses a novel approach, the Plane Strain Small Punch Test to identify the plastic properties metallic structures. To do so, a new apparatus was designed and manufactured to perform a series of plane strain SPT in room temperature. Based on the plane strain SPT, finite element was established for simulation of the specimen. The corresponding load displacement responses obtained from the FE simulation were implemented to establish database for an artificial neural network and, hence by training the network a function was obtained to predict the plastic properties of Stainless Steel 304L.

Keywords:

Plane Strain SPT , Stainless Steel , Artificial Neural Network , Plastic Properties