Study On Assessment Strategies And Measuring Methods For Double-line Parallelism

With the Industry 4.0 era approaching,domestic manufacturing industries urgently need to break through bottlenecks in traditional manufacturing,shift to upmarket equipment manufacturing technology and realize precision,ultra-precision,efficient and intelligent manufacturing.High-precision,de-artificialized online detection of geometry error of parts is increasing demanded in the process of machining,detection and use of equipment,but there are no unified and normalized measurement and assessment methods for line-to-line parallelism tolerance yet.At the same time,sometimes the reference is undefined at design time,but requirement for parallelism error between two lines is proposed.To address this issue,in this paper,an assessment strategy for parallelism errors and an evaluation method for accuracy assessment are proposed for mutual benchmarking double-line elements and a measuring system is built to realize efficient precision detection of parallelism errors of the subj ect in its state of rest or of uniform motion.Researches on assessment theories and methods for parallelism are carried out based on the new generation of standard system of Geometrical Product Specification(GPS);the development of a visual detection system for the projection line of the blade edges of high-speed Variable Slit(VS)is accomplished,according to a MATLAB image processing algorithm for software design;finally,parallelism errors of the blade edge of Variable Slit is measured,assessment examples are given and assessment theories for parallelism are validated.Research methods used in this paper and the main contents of the research are as follows:(1)The definition of the new generation of GPS for geometrical tolerances is analyzed,two-and three-dimensional theoretical models for assessment of benchmarking fitting and parallelism errors are created using the least squares method and the minimum zone method respectively.Exaction elements to be detected are simulated with MATLAB for emulation experiments and performances in the fitting and assessment processes.The experiment results show that the minimum zone method based on Particle Swarm Optimization(PSO)simultaneously has advantages like precision in terms of single-element fitting,accuracy and consistency in aspects of assessment of mutual benchmarking elements.(2)A measuring system is built for image acquisition,a lighting mode combining transmission-type lighting and reflection-type lighting is used,an array CMOS Image sensor with a single pixel of 5.5?m and a limited lens resolution of 4.70?m and an optical amplification of 1.252 is selected and a X64 Xcelera-CL PX4 Full image acquisition card is used to coordinate computer software to accomplish image collection.The pixel equivalent is 4.391?m,calibrated by Zhang's camera calibration method.(3)The image to be detected is de-noised using adaptive filters and enhanced by gray-level transformation and a dual-peak adaptive threshold method is put forward to cut images,ineffective cutting units are filled with morphological processing methods to remove unrelated information,and finally,the information of location and shape of projection lines of the blade edge in the image to be detected is exacted using the Canny operator.The mapping and reducing of each point on the edge from the image coordinate system to the space-based coordinate system for real objects are completed.The user interface of software is designed to improve experience and detection efficiency.(4)Parallelism errors of projection lines of the blade edge are measured,calibration of pixel equivalent and distortion correction are accomplished using Open CV.According to the results of measurement,the static parallelism error limit is 0.032 mm,the difference of parallelism is 0.016mm;the dynamic parallelism error limit is 0.035 mm and the difference of parallelism is 0.017mm;the difference of parallelism in the minimum zone method is always less than that in the least squares method.The analysis of error sources and the modeling of the unc ertainty for the measuring system are completed,The expanded uncertainty of parallelism measurement is 0.009 mm,k=3;and the expanded uncertainty of the difference of parallelism is 0.012 mm,k=3.In short,the method of parallelism assessment proposed in th is paper resolves the issue related to non-uniqueness of assessment results of mutual benchmarking line elements to be detected,an evaluation thinking of accuracy in assessment methods with benchmarking straightness errors as the basis of evaluation is pu t forward;measurement methods are applicable to parallelism detection of double lines at rest and in motion in the plane,which have a significance of reference on the promotion of normalization of directional tolerance measurement and assessment and the establishment of unified standards.