Parameter identification when designing a solid-modeling-based grinding-machine controller

Abstract

Modern computer-aided design and dynamics simulation tools allow design and testing to be performed in a virtual-reality environment. The integration of controller synthesis programs also makes it possible to perform servo tuning without actually having to build a prototype. However, most servo controllers are based on integer processors for machine tools, and the integer numbers entering into a servo tuning interface often do not directly reflect the magnitudes of the actual controller parameters, impairing the virtual tuning. This study addressed this problem by employing a gain parameter to each control parameter, and developing an identification algorithm for these parameter gains. The identification results were used to tune the servo, which was first carried out within the solid-modeling simulation environment. This procedure makes it easy to implement the controller parameters in the actual machine. The actual implementation of the resultant controller verified that the proposed method improved the servo performance, including eliminating the overshoot in both position and velocity responses of the grinding machine.