Dynamic turning force prediction and feature parameters extraction of machine tool based on ARMA and HHT

Abstract

The dynamic prediction of turning force is an effective means to reflect the changing course and characteristics of the force in the whole cutting process. It is very important to model and predict the turning force and analyze its spectrum characteristics. In order to obtain the dynamic turning force, the circular cutting experiment of 12Cr18Ni9 rotary piece was carried out on the numerical control machine ETC1625P. The KISTLER sensor was mounted on the tool head of the machine tool, and the real-time turning forces in three cutting directions were measured. The experimental data show that the turning force fluctuates with the change of displacement in the feed direction. In order to study the complex nonlinear relationship between turning force and cutting parameters, the dynamic turning force was predicted by autoregressive-moving average modeling. The time–frequency analysis of the main turning force was carried out by using Hilbert–Huang transform. The local time–frequency characteristics of the signal were obtained by analyzing the Hilbert amplitude spectrum of the signal. When only one cutting parameter was changed, the maximum amplitude of Hilbert marginal spectrum of turning force signal changed with the change of cutting parameters. The results show that the high-precision modeling of dynamic turning force and the extraction of cutting features can be effectively realized by using autoregressive-moving average and Hilbert–Huang transform.