Comparative Evaluations of Frequency Response Analysis Methods for Fast and Precise Point-to-point Position Control

This paper examines frequency response analysis (FRA) methods using transient response signals in actual operation, for the fast and precise point-to-point position control of industrial servo systems. In industrial processing/manufacturing machines, the enormous fast and precise point-to-point trials are conducted with varieties of interval times, and parameter fluctuations of plants occur due to the self-heating, environment temperature variation, and aged deterioration. Therefore, FRA methods using the transient response signals in operation (e.g. force and position) are required to adaptation of controllers, stability analyses, and abnormal diagnoses. However, since initial and final values of the transient response signals are not the same in general in the point-to-point motion, it is difficult to analyze the accurate frequency characteristic of plant directly through the discrete Fourier transform (DFT)-based FRA. In this study, three practical FRA methods (windowing function, bandpass filter, and differentiator) are comparatively evaluated to examine the appropriate FRA method for the fast and precise servo systems. The comparative evaluations are conducted by experiments using a laboratory prototype galvano scanner.