A GENERAL METHODOLOGY FOR MACHINE TOOL ACCURACY ENHANCEMENT: THEORY, APPLICATION AND IMPLEMENTATION (COMPENSATION, CNC, SOFTWARE)

M. ALKAN DONMEZ, Purdue University

Abstract

The accuracy of a machine tool is the limiting factor in the accuracy of finished parts. Errors in the machine tool motion produce a one-to-one error correspondence in the final workpiece. These errors in motion are caused by geometric errors of the structural elements and by thermally-induced errors caused by the machining process itself. It is impossible to completely eliminate errors by design and/or manufacturing modifications. Even a small geometric error of one machine element will be amplified at the cutting tool by the long travel ranges of the slides and the tool offsets. Mechanical design cannot eliminate thermally-induced errors because of continuous heat generation by drive motors, friction in slideways, spindle and leadscrew bearings and the cutting tool-workpiece interface. Rather than attempting to eliminate errors, this study provides a methodology for predicting errors and compensating for them in real-time, thus improving the accuracy of machined workpieces. A general mathematical model is developed, determining the total error at the cutting tool tip contributed by the errors of each machine elements and their thermally-induced variations. Homogeneous coordinate transformations for each element of the machine tool are employed. In order to predict each error component, a methodology for the machine tool error calibration is determined. A flexible and modular software compensation system is developed based on the models created in the methodology. The compensation system predicts the errors of the machine tool using a combination of data taken from various sensors on the machine tool and established error relationships. Finally, the methodology is implemented on a two-axis turning center. The predictions for geometric and thermally-induced errors for this machine are generated using least squares curve fitting techniques on the error data. A single-board microprocessor-based system translates errors into servo counts, which are injected into the machine tool controller in real-time. To demonstrate the validity of the model and methodology, cutting tests were performed under transient thermal conditions on a computerized numerical control (CNC) turning center. Accuracy enhancements of up to 20 times were obtained.

Degree

Ph.D.

Subject Area

Industrial engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS