Effect of positional errors on the accuracy of multi-probe roundness measurement methods

Abstract

Multi-probe roundness measurement methods can be used to measure cross section roundness profiles and dynamic behaviour of large flexible rotors such as paper machine rolls. Other roundness measurement methods are not suitable for such measurements, since the rotors are too large to be measured on precision spindles and the center point of the measured profile can move in an unpredictable and unrepeatable way during the measurement. Multi-probe roundness measurement methods can, to a limited extent, separate the center point movement (commonly also called error motion) and the roundness profile of a cross section of a rotating workpiece. This study compares the effect of positional errors and center point movement on the accuracy of three different multi-probe roundness measurement methods. The research included quantification of the effects of probe noise, positional errors and center point movement on the accuracy of the roundness profiles produced by the different methods. A novel method for generating continuous random center point movement is presented. Signals of a rotating workpiece with center point motion were simulated, and following GUM handbook supplement 1 guidelines, the Monte Carlo method was used to obtain distributions for the harmonic components of the methods. Distributions for different errors and their effects on roundness parameters are presented separately for each roundness measurement method. The results of this research only show minor differences in lower order harmonic components between the methods. The results also show that diameter sampling is immune to horizontal positional errors although it is particularly sensitive to angular positional errors of the probes, which lead to more errors accumulating in the phases, not the amplitudes of the harmonic components. With higher order harmonic components, the obtained error distributions were observed to correspond well with theoretical error propagation rates for harmonic sensitivity.