Design of block experiments applied to the mechanical properties of steel wires for springs

Abstract

The purpose of this article is to demonstrate an application of the design of block experiments via analysis and multiple linear regression in the investigation of a steel thermal treatment process with multiple responses. The study aimed to design statistical models to predict the mechanical properties in SAE 9254 draw steel wires, with diameters of 2.00 mm and 6.50 mm, after quench hardening and tempering. For this purpose, process input variables (wire diameter, processing speed, tempering temperature and polymer concentration) were investigated regarding their influence on the material tensile strength, yield point and hardness. The results revealed that the mechanical properties of the steel wire are significantly influenced by the selected variables, and analysis of variance (ANOVA) was employed to validate the design of the statistical models. Multiple linear regression allowed for an appropriate representation of the process, and graphical analysis was found to be very useful in displaying the behavior of the multiple responses.