Simulation via artificial neural networks and mechanic behavior analysis of steel AISI 430

The tensile test is one of the main techniques of mechanical characterization of materials. It is a destructive test and, considering the preparation of the proof corps, it has high cost and is time consuming. This work intends to use Artificial Neural Networks to simulate the behavior of the tensile testing machine at different annealing temperatures of ferritic stainless steel AISI 430, stabilized with niobium. The samples were annealed at 14 different temperature values for one hour and cooled in the furnace. The results showed that the network correctly simulated the stress-strain curves generated during the tensile tests getting low average squared errors and hardening rates with even minor errors. This characterization indicates the possibility of refining the microstructure in the samples annealed below 700°C, this temperature recrystallization and secondary recrystallization at 900°C.