Evaluation of superalloy heavy-duty grinding based on multivariate tests

Abstract The quality and economy of grinding depend on proper selection of grinding conditions for the materials to be ground. In order to evaluate the effect of heavy-duty grinding, a new performance index, which includes specific material removal rate, size accuracy, and grinding forces, was proposed. Robust design of experiment, including orthogonal arrays, the signal-to-noise ratio (SNR) method, and analysis of variance (ANOVA) for multivariate data, was employed to estimate the effect of uniform experimental design and to optimize grinding parameters. Empirical models of grinding force were investigated for finite element analysis of new fixture design. These empirical models, based on robust design of experiments and multiple regression methodology, have been confirmed through further verification experiments. Correlation coefficients from 0.87 to 0.96 were achieved.