Comprehensive Errors Prediction and Compensation System for CNC Turning Based on the Neural Network and Visual Basic

Abstract

In order to improve the machining accuracy for CNC turning, an error prediction and compensation system (EPCS) is developed based on the neural network (NN) and the visual basic (VB). The comprehensive errors of the machined shafts are simulated using the NN. The weights and bias of the trained NN are extracted to develop the EPCS through the VB. A correction coefficient and an equivalent diameter formula are proposed to achieve the error compensation under various working conditions. Using the EPCS, the maximum error of the machined shafts is reduced from 12.6 amu;m to 1.2 amu;m. The comprehensive errors are compensated by 90.5% compared with no compensation.