Development of Regression Model to Predicting Yield Strength for Different Steel Grades

Abstract

Yield strength of steel is an important property in designing a steel component. The maximum load that material can sustain without undergoing permanent deformation is represented by yield strength. A robust model is purposed, which can accurately predict the yield strength of different steel grades, based on the varying chemical composition of steel from the employed database. The dataset consisting of chemical composition and experimental yield strength of steel rods from a collapsed building sites. MATLAB® was used to generate a regression model such as multivariate regression and linear regression, to predict the yield strength values for different steel grades. A comparative study among the different regression models for RMS error, Max error, and highest accuracy is discussed in this paper. [copyright information to be updated in production process]