Thermal State Analysis And Thermal Error Research Of Motorized Spindle In Turning Center

Due to it has the advantages of simple structure, high transmission efficiency, the moment of inertia and fast response, the high speed motorized spindle has been widely used in high speed machining equipment. As motorized spindle speed become higher, the spindle fever also increasingly serious, thermal error of the motorized spindle is also become bigger, the motorized spindle as the core component of high-speed high-precision equipment, thermal error on the precision of the equipment damage is very large. To achieve the purpose of control the thermal error of spindle, the temperature and thermal error of motorized spindle is need for further research, to explore the regularity of motorized spindle thermal errors, the motorized spindle for accurate thermal error compensation is of great significance. This article around the hot state characteristics of high speed motorized spindle and the error of prediction model based on the thorough analysis research. The main research contents are including:(1)Thermal analysis parameters is calculated based on the thermodynamics knowledge, analyzes the main forms of heat transfer and characteristics of the spindle system, and the experience in heat transfer calculation formula is given out, detailed calculation of heat production and heat transfer parameters of the spindle system. The steady-state analysis of motorized spindle are completed by using finite element method model, and analyzed the thermal state performance of motorized spindle also.(2) Thermal error of spindle experiments is finished, motorized spindle temperature of measuring point is measured precisely and its real-time thermal displacement, contrast analysis of the theoretical calculation and experimental data; On the base of the experiment data, based on three different temperature variable, electric spindle thermal error prediction model is established by using BP neural network modeling method, and the analysis of the network to the fitting of experimental data, residual error and the error percentage, etc. and the limitation of BP neural network is proposed.(3) Put forward the method that use genetic algorithm (GA) to optimize the BP neural network, and then apply this method to set up in front of the three kinds of BP neural network to optimize operation, the curves of fitness and error sum of squares curve of network are draw and analyzed, and analyzed the optimized network with the experimental data fitting, residual error and percentage error performance parameters, and so on.Through analysis show that the spindle temperature highest position is located at the motor rotor, spindle cooling device can effectively reduce the motor stator temperature, shaft bearing temperature rise is bigger and can take measures to reduce further. At the same time, the spindle thermal deformation are analyzed, the maximum in the axial direction spindle thermal deformation, therefore, should take measures to compensate the axial displacement, in order to reduce the spindle machining error.The error analysis of network modeling shows that the BP neural network has strong nonlinear fitting and generalization ability, model based on three different temperature variable can fit the experimental data very well, the BP neural network model based on relative temperature variables has minimum residual error, it is 3.35%;Based on three different temperature variable residual error percentage of GA optimized BP network compared with BP neural network improved substantially, and simply explains the stability of the network got improved after optimization, using genetic algorithm to optimize BP network has higher convergence speed and better stability and accuracy.