Research On Photoelastic Fringes Identification Of Transparent Injection-molded Parts

Residual stress is one of important indexes of evaluating the quality of injection molded parts, especially for optical products. Residual stress involves many factors, such as polymer resins, molding parameters, equipments, and so on. And it is the focus of injection molding research fields. Birefringence phenomena of transparent injection molded parts is usually used to measure the residual stress by photoelastic experiment which is simple and need not destroy parts. Currently, interference patterns of photo-elastic experiment are usually identified by naked eye. But it is difficult to identify birefringence fringes which are dense. As for automatic identification by computer, the birefringence fringes are treated as digital information and the little difference can be recognized. Therefore, compared with naked eye, the computer has more advantages in birefringence fringes treatments.In this thesis, photoelastic experiments of transparent plastic parts are studied carefully and then find the experiments results characters, its fringes with different density, opened and closed fringe patterns, and interference of molding and instruments. After reviewed the corresponding literatures, methods for identification of stress fringes with monochromatic light source are provided. Based on fringes-identification arithmetic, the computer program is developed using Microsoft Visual C++, which has the following functions, 1) input different image file type, e.g. bmp, jpg, etc., 2) identify interference fringe pattern and eliminate interference, 3) postprocess for identified lines of interference fringes.In the context of chapter one, comments on fringe identification and its arithmetic are made. Then, the fringe features of transparent polymer parts are discussed and the optic-stress equations of photoelastic experiments are induced, which help to understand the reasons for selected identified methods. The procedure of fringe automated recognizing is finished in chapter three, preprocessing the picture from photoelastic experiments, drawing the skeleton-line of birefringence fringes by Fringe Direction Method, postprocessing the skeleton-line by refining, distinguishing bright and dark fringe, reducing dashed line, and eliminating interference and error. Flow charts for the program are followed. Finally, the examples are given to show the theory and numerical methods and program are valid and reliable.In conclusion, the jobs have done in this thesis. On one hand, image processing system is built, including a set of image libraries with plug-in interfaces to support multi-format image files, matrix process function modules and class libraries with C++ templates to fulfill image arithmetic efficiently. On the other hand, a set of numerical methods are provided to identify the fringes of transparent injection-molded parts, including deducing equations and solution steps of fringes, single point processing method and small-area eliminating interference.