对数控机床在主轴空转和实切状态下的热误差特性进行了比对分析。利用模糊聚类和F统计量确定了最佳的分类及分类阈值,根据温度与热误差之间的灰色关联度确定出温度敏感点,进而建立补偿模型。对实验结果的分析表明,温度敏感点在两种状态下是动态变化的,不同状态下的补偿模型并不通用;实际生产中的热误差补偿应优选实切状态下的热误差模型。
Thermal errors of CNC machines were different under actual cutting and spindle idling. The methods of fuzzy clustering and F statistics were used to classify temperature variables,as well as to confirm the best threshold. Then the temperature sensitive points were selected according to grey correlation among temperature variables and thermal errors. At last,a model among temperature sen- sitive points and thermal errors was built, The results show that temperature sensitive points will be changed under two situations, the model under different situations can not be used mutually-interchan- ging. The model under actual cutting should he used prior during actual production.