Fast identification of machine tool spindle system temperature rise based on multi-model fusion and improved D-S evidence theory

Abstract

Thermal equilibrium test is the key means to obtain the thermal characteristics of machine tools. In order to shorten the test period and reduce the research and development cost, a novel fast temperature rise identification method for machine tool spindle systems is proposed. The existing prediction identification methods ignore the limitation of the single prediction model, leading to large error fluctuations in different environments. In this study, various intelligent prediction models are combined with the improved D-S evidence theory to improve the accuracy and robustness of the prediction. Firstly, based on the virtual prediction, the evidence identification framework is established through the multiple evaluations of the data information in the evidence segment. Then, the weight allocation of each basic prediction model is carried out by the evidence combination theory. In this process, the evidence identification framework is reconstructed according to the improved strategy to avoid the high conflict problem in classical evidence theory. Finally, the fusion prediction of multiple models can be realized. The VM-850L machining center was selected as the research object for the thermal equilibrium test to evaluate the proposed method. The results show that the proposed multi-model fusion prediction method can accurately predict the temperature rise of selected points in a short time. Moreover, the prediction accuracy is significantly improved compared with the traditional single model. The proposed method has good universality and is expected to be popularized and applied more widely.