AUTOMATED MEASUREMENT AND COMPENSATION OF THERMALLY-INDUCED ERROR MAPS IN MACHINE-TOOLS

In this paper, a direct method of machine fool calibration is adopted to model and predict thermally induced errors in machine tools. This m ethod uses a laser ball bar (LBB) as the calibration instrument and is implemented on a two-axis computerized numerical control turning cent er (CNC). Rather than individually measuring the parametric errors to build the error model of the machine, the total positioning errors at the cuffing tool and spindle thermal drifts are rapidly measured using the LBB within the same experimental setup. Unlike conventional appro aches, the spindle thermal drifts are derived from the true spindle po sition and orientation measured by the LBB. A neural network is used t o build a machine model in an incremental fashion by correlating the m easured errors with temperature gradients of the various heat sources during a regular thermal duty cycle. The machine model developed by th e neural network is further tested using random thermal duty cycles. T he performance of the system is also evaluated through cutting tests u nder various thermal conditions. A substantial improvement in the over all accuracy was obtained. (C) Elsevier Science Inc., 1996