Predicting SCC Behaviour of Austenitic Stainless Steels in High Temperature Water by Empirical Learning Methods(EAC2)

概要

論文の詳細を見る
Two kinds of empirical learning methods based on artificial neural network (ANN), i.e., Double Layer Perceptron (DLP) model and Elman Feedback (EF) model, were used to analyze SCC data and predict the SCC susceptibility of austenitic stainless steels in high temperature water(HTW). DLP model could not converge after long training epochs while EF model could reach a steady value within limited training epochs for two sets of SCC data of 304SS and 316SS respectively. Threshold value (ThV) used in EF model had obvious effect on prediction ratios. In the processes of EF model, method I of including sample to be predicted generally had higher prediction ratio than that of method II of excluding the sample to be predicted. The data of the SCC susceptibility of 304SS and 316SS in HTW related to environmental factors such as temperature (T), dissolved O_2 content (DO), chloride ion content ([Cl^<-1>]), electrode potential (E) were used. In the case of ThV<0.6, the ranges of prediction ratio were ca.0.70〜0.95 with method I and ca.0.6〜0.85 with method II for 304SS, ca.0.75〜0.95 with method I and 0.75〜0.90 with method II for 316SS. Results showed that Elman Feedback model was a useful tool for qualitatively predicting the SCC behaviour of austenitic stainless steels in high temperature water.
一般社団法人日本機械学会の論文
2001-10-20