TE Connectivity

AI-based Real-time Molding Ejection Process Video Analysis

Molding processes use heat and pressure to fill a mold with certain materials. The mold can be opened and closed by controlling the two parts of a mold, Mold Base A and Base B. When the molding starts, the mold will be closed, and the materials will be injected. Once the injection is complete, the mold will be opened to eject the molding parts from the cavity. In this project, students will aim to monitor the ejection process through computer vision technologies. One objective this student team will work toward is to track the part ejection process to determine the optimal duration for opening the two bases, thus reducing the cycle time. The second objective is to develop a real-time monitoring functionality that can inform operators when parts are not ejected from cavity successfully. This student team will work to detect accuracy of parts during molding ejection and computational speed. This student team will work to i) train the model for detecting molding parts of different part numbers; ii) mold base opening cycle time optimization; iii) monitor part ejection failure. Several deliverables this student team will work to achieve are: i) an AI model that can detect molding parts of different part numbers in real-time (can be multiple models, to be determined) ii) corresponding training pipeline iii) a pipeline for molding ejection time optimization and related data analysis iv) a software that embeds the AI model and connects to the camera for real-time monitoring of part ejection to allow cycle time recording and alarm operators when ejection failure happens