Feasibility Assessment of Support Vector Regression Models with Immune Algorithms in Predicting Fatigue Life of Composites

Abstract

Predicting fatigue life of composite materials is essential to increase reliability of manufacturing systems. The predicting techniques for fatigue life of composite materials are not widely investigated. The support vector regression (SVR) is an emerging forecasting technique and has been applied in many areas successfully. Therefore, this study attempts to examine the feasibility of SVR in predicting the fatigue life of composite materials. Additionally, immune algorithms (IA) are used to select three parameters of SVR models. An experimental data set from a laboratory was employed to depict the feasibility of develpoed SVRIA (support vector regression with immune algorithms) approach in predicting fatigue life of composite materials. Empirical results indicate that the SVRIA is a valid way in predicting fatigue life of composite materials.