Support vector machine and neural network united system for NC machine tool thermal error modeling

In order to realize modeling and predicting for the thermal error of numerical control (NC) machine tool, a new united prediction model is introduced. The united prediction model combines the advantages of support vector machine (SVM) and neural network (NN) theory to show the excellent capability. The prediction precision of the hybrid prediction model for machine tool thermal errors is the highest among three kinds of models. The testing results show that the precision of the united prediction model is 0.5μm. The mean absolute percentage error (MAPE) of prediction model is 1.95%, outperforms any one of the two single prediction methods. Therefore, united predictive model can highly improve machine tool´s processing precision. Using the predicted thermal error model, the thermal deformation can be compensated.