كاربرد شبكه عصبي مصنوعي در پيش بيني رفتار پيري و پخت سختي فولاد ST12

Artificial neural network in prediction of aging and bake hardening behavior of ST 12 steel

In this work, an artificial neural network (Ann) model was established for prediction of mechanical properties of a baked steel(ST12) using the results of experimental data. For simulation the industrial process of forming and painting of autobody panel, the specimens first were heated to the temperature in the range of 500 to 1000°C and followed by the room temperature at different rates. Then, pre- strained 8% and aged for 1 hour at 100°C. At last they were pulled to fracture at room temperature. Comparing the predicted amounts of bake hardening by Ann model with the experimental ones indicates that well-trained neural network model has more capability in prediction of bake hardening amount and find yield stress for ST 12.