Inverse identification of pipeline steel mechanical state by means of coupled magnetic measurements and artificial neural networks

Artificial neural networks are widely used to develop models able predict properties of interest by learning and establishing relationships between inputs outputs a system. It is particularly relevant for the inversion process in framework Non-Destructive Testing (NDT) involving strongly non-linear non-monotonic behavior and/or saturations. The case-study considered this work magnetic material subjected uniaxial mechanical stress, plastic strain field. An Neural Networks (ANN) model proposed corresponding remanent magnetization, coercive field maximum magnetization as target properties. A series experimental data made various magneto-mechanical measurements train, evaluate validate ANN model. suitably predicts second specimen same field, stress ranges first specimen. inverse then loading from signature. Unique accurate solutions found that proves relevance machine approach such NDT application.