Artificial Neural Network (ANN) Modeling for the Energy Absorption of Hot‐Rolled Plates in Charpy Impact Tests

Abstract

In the present paper, a modeling for the energy absorption (CVN) at room temperature of hot‐rolled plates in Charpy V‐notch impact tests was investigated, in which an BP(Back Propagation) ANN (Artificial Neural Network) model with three layers was developed to take into considerations chemical compositions, processing parameters, yield strength, tensile strength and product thickness. The measured or predicted strength values have been used to predict the energy absorption in Charpy impact tests, both showing good agreements with the measured values. In order to compare the precision of the neural‐network methods in predicting CVN, linear regression analysis was performed by using the same data. Also, analyses were made for the effects of alloying elements, microstructure and processing parameters on CVN using ANN model, being consistent with the metallurgical rules. It concluded that the absorbed energy in Charpy impact tests for given steel compositions, processing parameters, strengths and plate thickness can be predicted by using the modeling.

Citation

Tan, W., Liu, Z.Y., Wu, D. and Wang, G.D. (2008), "Artificial Neural Network (ANN) Modeling for the Energy Absorption of Hot‐Rolled Plates in Charpy Impact Tests", Multidiscipline Modeling in Materials and Structures, Vol. 4 No. 1, pp. 37-46. https://doi.org/10.1163/157361108783470423