Research On The On-machine Measurement Uncertainty Evaluation And Experiment Of Complex Surface

On-machine measurement technology based on the principle of spatially modulated optical ranging has been widely used in the fields of reverse engineering and in-situ detection in precision manufacturing due to its advantages such as non-contact,high efficiency,and convenient in-situ measurement.The on-machine measurement system is used to obtain the 3D pose and morphology information of the parts,which is the data support for realizing the optimization of the surface redesign and aided manufacturing.However,the actual measurement data of the on-machine measurement system has dispersion and uncertainty,which requires quantitative evaluation of its dispersion.Therefore,the measurement uncertainty is introduced into the on-machine measurement system to evaluate the measurement uncertainty.The structure of the on-machine measurement system is complicated,and the environmental disturbance factors are multiple,there are dozens of sources of uncertainty,and the transmission coupling effect is unknown,so it is difficult to evaluate.In addition,because there is measurement uncertainty in the on-machine measurement,the processing surface reconstruction / surface splicing based on the on-machine measurement data also inherits the uncertainty,which affects the processing accuracy.Therefore,it is necessary to consider the measurement uncertainty effects to the surface construction and splicing.In order to study the influencing factors of on-machine measurement uncertainty,the on-machine measurement architecture is analyzed firstly,and the spatial coordinate transformation model and the measuring point coordinate extraction model are established.Based on the on-machine measurement architecture,the traceability of on-machine measurement uncertainty influence factor is performed from many aspects aiming to clarify the influencing factors and influence patterns.Due to the characteristics of high coupling effects and difficult to obtain transmission laws among influencing factors of on-machine measurement uncertainty,a measurement uncertainty analysis method combining black-box model and quantitative characteristic analysis is proposed,the repeatability measurement component,reproducibility measurement component,temperature influence component,and data round-off component are used to simplify the influencing factors to achieve influence factor analysis.Aiming at the problem of uncertainty measurement evaluation of on-machine measurement system,a dynamic measurement uncertainty evaluation framework for on-machine measurement is designed.Based on the characteristics of on-machine measurement,two types of sample set planning methods,cubic lattice sample set and orthogonal continuous sample set,are proposed.These kinds of sample sets can effectively reflect the entire domain of the measurable space or the specified space.To achieve effective sampling,a space-filled synchronous motion sampling method is proposed.A method of fitting sampling data according to the principle of maximum likelihood estimation or nonparametric statistical method is proposed,and the best fit distribution and probability density function of the uncertainty components are obtained.The mathematical model of measurement uncertainty evaluation for on-machine measurement is established.The transfer synthesis of uncertainty based on adaptive Monte Carlo method are accomplished.Using the constructed on-machine measurement experimental system,a on-machine measurement uncertainty evaluation experiment was carried out to verify the feasibility of the method.Aiming at the actual measurement of the inner wall's external contour of the rocket nozzle,the measurement method of cyclic reciprocating copying scanning is designed and optimized,and the on-machine measurement efficiency is doubled.The measurement uncertainty of the rocket nozzle slot-milling machine on-machine measurement system is evaluated in multiple dimensions.The rocket nozzle surface reconstruction/surface splicing considering the uncertainty of multidimensional measurement is studied,and an internal and external envelope method based on multi-dimensional uncertainty is proposed for surface fitting,and the selecting principle of envelope data points is proposed,Cubic spline interpolation envelope fitting and NURBS envelope fitting are realized.Completed the algorithm design of cyclic reciprocating copying scanning on-machine measurement and the interactive algorithm design of the main and auxiliary side surface splicing,completed the software writing based on VC ++,and realized the engineering application.