Research On Thermal Error Modeling And Compensation Method Of Dry-cut Hobbing Machine Based On Gradient Boosting Decision Tree Algorithm

Gears are widely used in various mechanical transmission systems,and their manufacturing accuracy directly affects the transmission accuracy of the system.The gear structure is complex and there are many precision items.It is easy to produce various errors during manufacturing,and each error comes from multiple error sources.The thermal error source occupies a large weight.In gear manufacturing,the hobbing process accounts for about 50%,so it is of great scientific and engineering significance to study the influence of the thermal error of the gear hobbing machine on the precision of the machining gear and its compensation method.The main contents of this paper are as follows:(1)The heat sources of a hobbing machine tool are analyzed,and the position deviation formulas between the hob spindle and the workpiece spindle caused by the thermal deformation of the hobbing machine tool are calculated theoretically.Based on the knowledge of multi-body theory and coordinate transformation,the geometric error model of the hobbing machine tool is theoretically analyzed and established,and the influence law of the position deviation of the machine tool along the X and Y directions on the tooth surface error is explored.(2)By comparing and analyzing the advantages and disadvantages of various sensors,the most suitable temperature and displacement sensors for experiments are selected according to the actual working conditions.According to the working principle and characteristics of the eddy current sensor,it is calibrated to ensure that the subsequent temperature-thermal error data acquisition is more effective.After analyzing the strategy of selecting temperature measuring points,temperature sensors are installed at several locations of the machine tool,and displacement sensors are used to obtain the variation of thermal error after each gear is machined.(3)Stepwise regression method is used to screen many temperature variables.Optimized temperature variables and thermal error data are used to model the temperature-thermal error.The modeling methods include stepwise regression method,BP neural network method,GA-BP neural network method and gradient lifting(GBDT)method.The results show that the fitting and forecasting effects of GBDT model are optimal,so the GBDT method is chosen to model.(4)The thermal error compensation module is developed and processed on the machine tool after compensation.Then the compensation effect is tested by measuring the span M value of the machined gear before and after compensation.The results show that the compensated M is worth improving,which shows that the compensation effect is remarkable and the compensation method is feasible.