A Fast Algorithm for On-Line Machining Process Modeling and Adaptive Control

Abstract

A fast, on-line algorithm for machining process modeling and control is proposed. The modeling is accomplished via a new recursive estimator that offers good accuracy at a minimal computational load. Its Fast Kalman-type version, that further reduces its computational complexity, is also presented. The adaptive controller, which is based on on-line identification and closed-loop pole assignment, is characterized by a low computational load and no need for a priori process information. The analytical results are supplemented by numerical simulations, where the proposed scheme is used for the control of a turning operation and shown to offer very good performance under noisy conditions and suddenly changing machining dynamics.