基于数据驱动的叶片加工质量评估方法研究  

Research on the machining quality evaluation method of blade based on data-driven

作　　者：吴国新[1] 潘涛 刘秀丽[1] 徐小力[1] WU Guoxin;PAN Tao;LIU Xiuli;XU Xiaoli(Key Labratory of Modern Measurement and Control Technology,Ministry of Education,Beijing Information Science&Technology University,Beijing 100192,CHN)

机构地区：[1]北京信息科技大学现代测控技术教育部重点实验室,北京100192

出　　处：《制造技术与机床》2021年第12期105-109,共5页Manufacturing Technology & Machine Tool

基　　金：北京学者资助(2015-25);北京市重点实验室开放项目(KF20202223202);北京市教委科研计划项目(KM202011232001)。

摘　　要：通过数据分析方法能够分析出数据之间潜在的联系或规律。首先采用多元线性回归的方法,选取了某一叶盘上的27个叶片的测量数据,通过三坐标测量机采用截面测量的方法,测量了17组参数的数据,并根据理论值得出误差,其次分别以其中1个参数的误差值作为因变量,其余16个参数的误差值作为自变量,通过MATLAB软件分别得出多元线性回归方程的系数,最后根据模型检验,选择了最优模型,即以TE THK参数的误差值作为因变量,其他参数作为自变量。结果表明,参数TE THK的误差值与CV CONT MAX、CC CONT MAX的误差值之间存在联系且呈正相关性,与LE THK(CA)、WARP、SFT F/A、SFT CC/CV之间线性关系最小。Data analysis methods can analyze the potential connections or laws between data.In this paper,the method of multiple linear regression is used to select the measurement data of 27 blades on a certain leaf disc first.The data of 17 sets of parameters are measured by the three-coordinate measuring machine using the method of cross-section measurement,and the errors are calculated according to the theoretical values.And then,taking the error value of one parameter as the dependent variable and the error value of the remaining 16 parameters as the independent variable,the coefficients of the multiple linear regression equation were obtained through MATLAB software.Finally,according to the model test,the optimal model was selected,that is,TE THK the error value of the parameter is used as the dependent variable,and the other parameters are used as the independent variable.It is found that there is a strong linear relationship between the error value of the parameter TE THK and the error value of CV CONT MAX and CC CONT MAX,and a positive correlation;TE THK and LE THK(CA),WARP,SFT F/A,SFT CC/CV has the smallest linear relationship.

关 键 词：多元线性回归 叶片参数 误差值 

分 类 号：TB921[一般工业技术—计量学]