Using Artificial Neural Network for Predict Tensile Strength of AISI 304 Stainless Steel

Abstract

The objective of this research is to Ultimate tensile strength of AISI 304 stainless steel. The test speed is identified into 4 difference levels and 12 difference forms of stress concentration which are tested by tensile Testing machine, Zwick (Z020). Artificial Neural network is used to predict the error. The result of the experiment finds that the specimen that able to stand the highest tensile member is the specimen with single hole which the diameter of the its hole is 1 mm. The specimen with offset hole is the specimen that able to stand the lowest tensile member, which the diameter of its hole is 3 mm. From using artificial Neural network for predicting the ultimate tensile strength of AISI 304 finds it can predict properly which the average training error is 1.63% and the average error of the predicted tensile strength is 1.15% that can reduce that experiment time and able to predict the untested experiment correctly.