Investigation to the Effect of Processing Parameters on Impact Resistance of Pip-Line Micro Alloyed Steels using Neural Network

Micro alloyed steels are widely used in several industries including crude oil and natural gas for which they called API steels. These steels are produced using hot rolling after steel making and continuous casting and impact resistance is one of the main mechanical properties that are measured in the final products. Impact resistance is affected by several processing parameters. In the present study, the effect of 33 processing parameters such as chemical composition, strain, strain rate and temperature of hot rolling were studied using a neural network model for 122 independent samples. Two model were studied to investigate the error range and increasing the accuracy based on the difference between neural network and experimental results. Two neural network models with one and two hidden layers were used to train the system. It has been shown that two hidden layer model is more accurate with less errors. The results showed that the model can be used as a quantitative guide to control the final impact properties of API steels. According to the prediction by the neural network, the maximum values of impact properties were affected by Carbon and Niobium content of the billet, billet temperature in the final rolling stage, final rolling temperature and Coiling temperature.