Pre-compensation of contour errors for five-axis machine tools through constructing a model reference adaptive control

Abstract

Existing contour error pre-compensation methods are sensitive to modeling errors during the prediction of contour errors. This article proposes an adaptive contour error pre-compensation method for five-axis machine tools. A model reference adaptive control (MRAC) for a feed drive system is constructed. Through transforming the control signals from the external proportional–proportional–integral (PPI) controller, MRAC makes the response of the real plant match that of the nominal model. Both feedback and feedforward laws are combined in the transformation. The transformation coefficients are updated in real time by the parameter adaption law. The model matching conditions are theoretically derived. Through model matching, the tool tip and tool orientation contour errors can be precisely predicted by the nominal model regardless of the modeling errors. Through utilizing the position Jacobian and the orientation Jacobian, the compensation signals are directly solved in the workpiece coordinate system (WCS) by the model predictive control (MPC) algorithm. Simulations and experiments show that the proposed method can obviously reduce the prediction errors and thus improve the contouring performances.