Modelling and real-time compensation of cutting-force-induced error on a numerical control twin-spindle lathe

Cutting-force-induced errors are one of the major sources of error in numerical control (NC) machine tools. The error compensation technique is an effective way to improve the manufacturing accuracy of NC machine tools. Effective compensation relies on an accurate error model that can predict the errors exactly during the machining process. In the present paper a robust and accurate cutting-force-induced error model is built using a back-propagation (BP) neural network and a genetic algorithm (GA) for an NC twin-spindle lathe. The GA—BP neural network modelling technique not only enhances the prediction accuracy of the model but also reduces the training time of the BP neural network. A real-time compensation system of the cutting-force-induced error on the lathe is developed based on the cutting-force-induced error model. The errors were reduced by about 38 per cent after real-time compensation in a machining experiment.