Sensorless compensation system for thermal deformations of ball screws in machine tools drives

Abstract The article presents constructional, technological and operational issues associated with the compensation of thermal deformations of ball screw drives. Further, it demonstrates the analysis of a new sensorless compensation method relying on coordinated computation of data fed directly from the drive and the control system in combination with the information pertaining to the operational history of the servo drive, retrieved with the use of an artificial neural networks (ANN)-based learning system. Preliminary ANN-based models, developed to simulate energy dissipation resulting from the friction in the screw-cap assembly and convection of heat are expounded upon, as are the processes of data selection and ANN learning. In conclusion, the article presents the results of simulation studies and preliminary experimental evidence confirming the applicability of the proposed method, efficiently compensating for the thermal elongation of the ball screw in machine tool drives.