关键词: 数字孪生, 几何模型, 测量方法, 误差模型, 智能制造

Abstract: In order to obtain high-precision measurement data and build an accurate digital twin geometric model,which is different from the previous non-contact methods such as three-dimensional scanning,a method of obtaining measurement data by contact was proposed. Taking the measurement of bearing ring as an example,the measurement method of bearing ring size detection line was improved,and the outer diameter error models based on sine approximation and Convolutional Neural Networks (CNN) were established respectively,and the optimal parameters of CNN outer diameter error model were selected by orthogonal test,and then the optimal outer diameter error model was selected for the outer diameter measurement of bearing ring big end and small end through comparative test. The experiment results show that after the improvement,the measurement errors of the outer diameter of the big end and the small end are reduced by 63.20 % and 97.98 % respectively. Finally,the calculation method of single point outer diameter value was proposed and verified,which solves the problem that the measured data is difficult to be applied to the construction of digital twin geometric model.

Key words: digital twin, geometric model, measurement method, error model, intelligent manufacturing