Real-Time Error Prediction for High-Precision Operation of Parallel Kinematic Machines

Closed-loop parallel kinematic machines (PKMs) have been proposed to improve precision and operation speed over conventional machine tools and robots. However, an embarrassing dilemma is that most of the existing PKMs achieve very lower precision in contrast to equivalent serial machine tools or robots, which are competitive to same tasks. Limited works have been conducted to evaluate errors thus improve precision of machine in real-time control. It becomes necessary to explore the relation of the motion error with the dynamics of a PKM. In this paper, the new model of the error evaluation has been proposed; three major sources of error under consideration are the deformations of the components under dynamic loads, the deformations at joint contacts, and the clearances of passive joints. To illustrate the modeling procedure, the dynamic model of machine is developed to determine internal forces among components and locations of joint contacts. Errors caused by machine dynamics are evaluated analytically in real time; in particular, the errors happened at the contacts of passive joints are estimated based on Hertz theory. The developed error models can be applied to compensate the motion errors of tool tip in real-time. The Exechon parallel kinematic machine is used as a case study, the results from simulation has been compared with the test data.