Research On Extraction And Evaluation Of Spherical Contour In Sphericity Error Measurement

With the development of national defense industry and aerospace technology,the precision of spherical inertial devices is required to be higher.Faced with the urgent need of high precision sphericity measurement,this paper focuses on the spherical contour form feature extraction and sphericity error evaluation methods in sphericity error measurement.The problem of precise spherical contour reconstruction based on two rotational degrees of freedom for centrifugal sampling system and the rapid and accurate evaluation of sphericity errors of minimum zone sphere(MZS),maximum inscribed sphere(MIS)and minimum circumscribed sphere(MCS)criterion are analyzed theoretically,modeled,simulated and experimentally studied.This dissertation intends to address:In order to extract spherical contour features accurately,mathematical modeling of the centroid measurement system and sampling strategy is carried out for spherical contour;a Gauss zero phase shift digital filtering method for unequal interval sampling data is designed,and the weight function of the Gauss filter is corrected based on the fitting angle;for compensating the measurement errors caused by the initial value difference of the sensor,the eccentricity of the position of the measurement piece and the reading error of the rotation angle in the sampling process,a spherical contour matching fusion method based on three-dimensional orthogonal datum is proposed.Under the condition of spherical contour matching,the center alignment between the measured contour and the main datum contour is completed;the objective function is constructed according to the measured difference between the measured contour and the main datum contour,the sub-datum contour A and the sub-datum contour B,and the rotation angle is corrected by optimizing the minimum corresponding angle of the objective function.On this basis,the size matching of the measured profile is realized according to the mean value theorem.The simulation and experimental results show that the proposed contour matching and data fusion method can effectively restore the real features of the measured contour.In order to evaluate sphericity error accurately and fast,a sphericity error evaluation method based on heuristic search and feature point model is designed.According to the macro-convergence,symmetry and multi-feature point characteristics of sphericity distribution,the improved strategies of sphericity error evaluation algorithm for rough positioning of spherical center search,optimizationof search path and correction of evaluation results are studied.The test results of data sets show that the proposed method can achieve a maximum compression ratio of 1/512 of the spherical center search range;the multi-directional adaptive search search algorithm owns both the convergence speed and global optimization performance;the spherical solution with Nano-precision can be obtained by 50~75iterations,and the solution time is less than 0.025 s,which is only 1/10 of the traditional algorithm.The feature points model correction method can build the feature point model based on MIS,MCS and MZS criteria and solve the ideal spherical center according to the current optimal solution location when iteration reaches the specified scale.Combining with heuristic search method,the solution efficiency can be improved by 26~61%.On the basis of the above work,the experimental system for measuring the center spherical profile is built.The center alignment and the stability experiments of the lifting device are carried out.The validity of the proposed method for spherical contour surface feature extraction and spherical error evaluation is verified by relevant experiments.