CAJ | 학술논문

针对机床热误差补偿技术中的关键温度测点选取问题，提出了一种温度测点优化新方法.首先采用简单相关分析，剔除掉与热误差明显不相关的测点.然后对初步筛选出的测点进行模糊聚类分析，以消除温度变量间的复共线性问题，同时进行灰色综合关联度分析，判断各测点与热误差间的紧密程度.根据分析结果，建立多个不同测点的热误差模型，对模型进行基于统计学理论的分析，确定出关键温度变量，将温度测点由20个减少至4个.根据优化结果，重新建立多元线性回归模型.误差预测结果表明，主轴 Z 向最大热误差从17.903μm 减小到1.850μm.
침대궤상열오차보상기술중적관건온도측점선취문제，제출료일충온도측점우화신방법.수선채용간단상관분석，척제도여열오차명현불상관적측점.연후대초보사선출적측점진행모호취류분석，이소제온도변량간적복공선성문제，동시진행회색종합관련도분석，판단각측점여열오차간적긴밀정도.근거분석결과，건립다개불동측점적열오차모형，대모형진행기우통계학이론적분석，학정출관건온도변량，장온도측점유20개감소지4개.근거우화결과，중신건립다원선성회귀모형.오차예측결과표명，주축 Z 향최대열오차종17.903μm 감소도1.850μm.
A new method is proposed to optimize the temperature measuring points of thermal error compensation technology for machine tools. A simple correlation analysis is used to weed out measuring points which are unrelated to thermal error apparently.Then,the selected temperature measuring points are classified by fuzzy clustering analysis to eliminate the multicollinearity problem between temperature variables.Simultaneously,the grey synthetic degree of association is used to determine the correlation between the measuring points and thermal error.Several thermal error models are established based on the results.The models are analyzed based on principles of statistics to identify key measuring points.The number of measuring points is reduced from 20 to 4.Then,a new multiple linear regression model is built based on these key points.The measuring results show that the Z-axis thermal error is reduced from 17.903 μm to 1.850 μm.