Screw Grinder Online Error Compensation System Design And Research

This paper based on the background of Jiangsu Province Scholar's plan projects "the Intelligent grinding methods and key technologies of precision CNC thread grinder". Aiming at the problem of thread grinder error on-line detection and error compensation technology, an error compensation method which is based on time series prediction and self-adaptive feedback control is put forward. Solving the problem that the machining precision of domestic ball screw grinding is a little low. Improve ball screw grinding's quality as well as its precision level.After analysis the types and causes of ball screw grinding error, Think ball screw grinding machine errors have complex of factors, there is no need to measures and compensation separately, a "black box" theory is put forward, that all the influence of grinding the processing as the poles, and the error as the "black box" which don't know the internal structure, only compensate for the final error. Moreover, build a systemic overall design, setting up experiment systems, and introducing its functions as well as system flow.Error series forecast is the core of this paper. By analyzing the characteristic of the ball screw grinder error, from the practical and the unity principle, model is established based trend stationary series and the slashing of non-stationary series which to forecast errors are belonged to the two characteristics weighted least square method to estimate the modeling parameter. Deriving the model order number identification theory which orders number from low to high and judges whether it's suitable or not. Set a forecast method that timely amended the model's output. As the requirements that based on feedback compensate, building a Least Square Root Auto-adjust Control model.Setting up a prediction simulation and Real-time compensation control software system, to verify the methods mentioned. Prediction simulation adopted abundant processing data in all cases, so the conclusion has universal significance. The simulation has analysis the influence that different weighting coefficient and different sampling points brought to the prediction results, the best value range is gotten, it provides a basis for real system to choose parameter. The compensation control experiments verified that the predictive compensation control method which in this paper is effective, errors of prediction accuracy above 80%, average compensation rate was also above 80%, after compensated the error decreases more than 64%, and proved that it can improve the machining precision greatly.