磨齿机电主轴热特性及热误差建模

针对磨齿机在磨削加工时，电主轴存在热致误差等问题，提出基于模糊神经网络（FNN）建立电主轴热误差模型的方法.分析电主轴内部的热生成和热传递机理，得到内部的传热规律.通过计算热载荷和边界条件，利用有限元分析（FEA）软件对电主轴系统的温度场和热变形进行数值模拟，得到电主轴系统中温升和热变形最大的部位.通过电主轴热误差实验获得温度和热变形数据，分别训练模糊神经网络和BP神经网络，建立温度场和热变形之间的热误差模型，对主轴热误差进行预测.结果显示：在电主轴径向热误差预测模型中，模糊神经网络模型和BP模型的建模精度分别为96.74%和89.77%.这表明模糊神经网络模型建立的热误差模型，在拟合和预测精度上优于BP神经网络模型.

Thermal characteristics and thermal error modeling analysis for motorized spindle of gear grinding machine tool

A new modeling method based on fuzzy neural network (FNN) was proposed to solve the problem of the thermal error caused by the motorized spindle in the grinding machining process. The internal heat generating and transfer mechanism of the spindle were analyzed to reveal the heat transfer law. The temperature field and the thermal deformation of the spindle system were numerically simulated by finite element analysis (FEA) software with the given thermal load and boundary condition. The maximum risen temperature and the largest thermal deformation were obtained. Fuzzy neural network model and BP neural network model were trained respectively, through acquiring temperature and thermal deformation values of the spindle in the thermal error experiment. The thermal error model between temperature field and its thermal deformation was established to predict the thermal error of the spindle. Results show that the modeling accuracy of the fuzzy neural network model and the BP model are 96.74% and 89.77% respectively in the prediction model of the radial thermal error. The thermal error model of FNN model is superior to that of BP neural network for the fitting and forecasting accuracy.