| 英文摘要 |
Under the circular economic initiative, countries are proposing plastic-limitation policies, and international brands are pledging to gradually replace new plastics with recycled materials by 2030. The Environmental Protection Administration Executive Yuan, R.O.C proposed a policy that focuses on "Reversing the Course of Waste toward a Circular Future." In February 2023, a proposal was drafted for "Internet Shopping Packaging Restrictions and Implementation," aimed at increasing the use of over 25% recycled materials in plastic packaging. Global recycling and reuse rates for waste, excluding bottles and some electronic plastics, remain low. To increase the reuse rates of plastic packaging, this project aims to plan a recycling chain for electronics plastic packaging waste and help develop a "Smart Automated Sorting and Circular Recycling Resource Platform." The project will use three core technologies: smart image recognition, smart material sorting, and plastic packaging regeneration.
This project was started on March 1, 2023 and is to be concluded on December 15, 2023, with a completion rate of 100%. The project team successfully conducted a comprehensive review and analysis of domestic and international techniques for image recognition, material sorting, and the regeneration cycle of electronic plastic packaging. The key achievements of this project are as follows:
(1) Two expert advisory committee meetings have been conducted, namely the "Smart Automated Sorting and Intelligent Recycling Technology Platform for Waste Plastics-Recycling Cycle Technology" and a discussion meeting with recycling experts. These meetings provided insights into the needs of the domestic waste-recycling industry and progress achieved in the classification and recycling of electronic plastic packaging. Subsequently, a recycling system was devised that can map out routes for the processing of regenerated materials and present the current state of the technological blueprint for the recycling of electronic plastic packaging.
(2) We have completed the development of an intelligent recognition technology, designed and established an optical module for image recognition, and constructed a database with 10,227 images. As such, we could obtain clear images of recyclables that featured different packaging types. By developing a type recognition algorithm using deep learning models, materials with different forms, patterns, and other variations were discerned. Such packaging can be recognizably categorized into five types, namely trays, foam, bubble wrap, boxes, and bags, and with different colors. The final verification results exceeded the target of 90%, with an accuracy rate of 96.4% at the speed of 16 items/s. In addition, a survey of the materials used in electronic plastic packaging was conducted, totaling 315 entries. We also managed the development of suitable material-classification technology for electronic plastic packaging waste and established a database containing 6,100 infrared reflection spectroscopy data entries. By using the spectral angle mapper (SAM) in the rapid spectral-recognition technique with spectral characteristics, we could classify the packaging materials into PE, PET, PP, PS, cellulose, and PVC , achieving a speed of 18 items/s and maintaining an average accuracy of over 97.5%.
(3) To develop innovative regeneration technologies for electronic packaging materials, we evaluated the devolatilization of flat packaging of a single plastic material (PP, PS, PET), performed pretreatment of foamed materials (EPP, EPO, EPS) for volume reduction, and used separation and purification technology to assess multilayer films. Furthermore, by using a melt-granulation process and testing fundamental physical properties, along with the use of structural-reinforcement techniques, we determined the optimal treatment methods for packaging. This involves the drafting of innovative, sorting, and regeneration-application-process planning techniques to assist in evaluating the regeneration before the electronic packaging material waste enters the recycling chain. Moreover, we assessed the feasibility of the integrated techniques for classifying and reusing packaging materials.
|