Determination of elastoplastic properties of in-service pipeline steel based on backpropagation neural network and small punch test

Abstract

Evaluating the mechanical properties of materials of in-service pipelines without shutting down transportation has been always a challenge. A semi-destructive method for determining the true stress-strain curve of in-service natural gas pipeline steel based on backpropagation (BP) artificial neural network and small punch test (SPT) was proposed in this study. The load-displacement curves of 457 groups of different hypothetical materials were obtained by the verified finite element model of SPT within Gurson-Tvergaard-Needleman (GTN) damage parameters and used to train the neural network. The relationship between the load-displacement curve of the SPT and the true stress-strain curve of the conventional tensile test was established based on the trained neural network. The elastoplastic properties of in-service natural gas X80 pipeline steel were obtained by this method. The accuracy and wide applicability of the trained neural network were verified by the experimental results of four types of materials obtained by the SPT and conventional tensile test. This work demonstrates a semi-destructive method, which can be applied to derive the true stress-strain curve of the in-service pipeline steels to determine the elastoplastic properties only by the load-displacement curve of the SPT without performing conventional tensile test.