Frequency Domain Analysis and Precision Realization in Deterministic Figuring of Ultra-Precision Shaft Parts

An aerostatic spindle is a core component in ultra-precision machine tools. The rotor of the spindle has extremely high manufacturing accuracy, which cannot be directly achieved via traditional machining, but always via manual grinding. The deterministic figuring theory is introduced into the machining of shaft parts, which overcomes many shortcomings of manual grinding. The manufacturing error of the shaft’s surface contains different frequency components, which have different effects on its working performance and the figuring process. Because the deterministic figuring method can only correct the error within a limited frequency range, in order to ensure high efficiency and high precision of the figuring process, we need to use reasonable filtering parameters to filter out the error with unnecessary frequencies. In this paper, the influence of contour error with different frequencies and amplitudes on the air film are analyzed using computational fluid dynamics (CFD) software, and the amplitude–frequency analysis as a function of the power spectral density (PSD) characteristic curve is used to study the filtering parameters of the measured data. After the figuring experiment using the filtering parameters obtained from the analysis, the average roundness of the shaft converged from 0.419 μm to 0.101 μm, and the cylindricity converged from 0.76 μm to 0.35 μm. The precision reached the level of manual grinding, which proves the rationality of the analysis using filtering parameters in a shaft’s deterministic figuring.