According to the disadvantages about the error information opaque and incomplete in the thermal error modeling,a new OE-CM prediction model was proposed herein based on combining the self characteristic of GM(1,n) and LS-SVM algorithm. A online modeling experiment was designed on the three-axis CNC machine tool,and the results show that OE-CM model has the characteristics of high precision forecasting and good robustness.This model is superior to GM(1,n)
model and SVM model,which are recommended to be applied to predict and compensate the spindle thermal errors under different working environments.Firstly,according to the experimental data of the spindle thermal errors,this model used GM(1,n) algorithm to establish the prediction model and linear superposition respectively.Then,the forecasting effectiveness measure was used to adjust the weighting coefficients,and finally the optimal prediction effect was obtained.In order to verify the validity of this model,thermal error online compensation experiments were made on the spindle,and the z-direction maximum error is reduced from 23.8μm to 8μm,which can improve the precision of the machine tool,and have preliminary engineering popularizing prospect.
