Study of Stainless Steel’s Properties for Radiation Safety using Artificial Neural Networks

Abstract

Several general components of reactor are made from stainless steel. Most of the container used for storing nuclear waste is made from stainless steel. In this work, an artificial neural networks (ANNs) model is used to improve the properties of stainless steel during its manufacturing process. The main job of ANNs is determining the layer’s thickness and predicting the influence of different parameters on the growth kinetic of the process. Experimental data of the borided layer was used on AISI 316 steel in a liquid medium (70% borax + 30% silicon carbide). The results obtained by artificial neural network (ANNs) model with those obtained by the mathematical model based on Fick's law and experimental data are nearly the same. The mean error from the neural network was between 1.2 and 1.35 µm. The mathematical model (second Fick's law) several parameters must be used which are: (time, temperature, concentration and the concerned interface, the boron surface concentration and the diffusion coefficient of boron in each phase). On the other hand, for the second model (artificial neural network) only the temperature, the time and the learning base. In other words, implementing artificial neural network for this case save time, effort and cost.

Keywords: Thermochemical treatment, Boriding, Fe2B, radiation safety, artificial neural network

Cite this Article

Ahmed Yousef El-Haseib, Mazhar Mahmoud Hefnawi. Study of Stainless Steel’s Properties for Radiation Safety using Artificial Neural Networks. Journal of Nuclear Engineering & Technology. 2018; 8(3): 31–39p.