Adaptive Compensation of Spatial Contour Error based on Neural Networks in multi-axis machining

The spatial contour error which exists in multi-axis machine tools is believed relative with the track-lag error existing in the single axis, the set value of feed-rate and the curvature of the contour. A spatial error model is taken for presenting a vector method to calculate the contour error in three-axis machine tool and an adaptive cross coupled controller which is based on neural network is presented. The simulation results of cross coupled control in three-axis, indicate that the contour precision can be improved effectively which exists in the machine tool, and the errors influenced by the change of curvature and the set value of feed-rate can be obviously reduced.