Recycling Strategy and Recyclability Assessment Model Based on the Artificial Neural Network

作     者：LIU Zhi-feng, LIU Xue-Ping, WANG Shu-wang, LIU Guang-fu (College of Mechanical & Auto Engineering, Hefei University of Techno logy, Hefei 230009, China) 

出 版 物：《厦门大学学报（自然科学版）》 (Journal of Xiamen University：Natural Science)

年 卷 期：2002年第41卷第S1期

页      面：153-154页

核心收录：

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 08[工学] 0835[工学-软件工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：ProjectsupportedbytheNationalNaturalScienceFoundationofChina (5 993 5 1 2 0 )andNaturalScienceFoundationoftheAnhuiProvince (0 0 0 4 75 0 9) 

主　　题：recycling strategy product recycling artificial neural network assessment model design for recycling 

摘      要：Now, a rapidly growing concern for the environmental protection and resource utilization has stimulated many new activities in the in dustrialized world for coping with urgent environmental problems created by the steadily increasing consumption of industrial products. Increasingly stringent r egulations and widely expressed public concern for the environment highlight the importance of disposing solid waste generated from industrial and consumable pr oducts. How to efficiently recycle and tackle this problem has been a very impo rtant issue over the world. Designing products for recyclability is driven by environmental and economic goals. To obtain good recyclability, two measures can be adopted. One is better recycling strategy and technology; the other is design for recycling (DFR). The recycling strategies of products generally inclu de: reuse, service, remanufacturing, recycling of production scraps during the p roduct usage, recycle (separation first) and disposal. Recyclability assessment is a very important content in DFR. This paper first discusses the content of D FR and strategies and types related to products recyclability, and points out th at easy or difficult recyclability depends on the design phase. Then method and procedure of recyclability assessment based on ANN is explored in detail. The pr ocess consists of selection of the ANN input and output parameters, control of t he sample quality and construction and training of the neural network. At la st, the case study shows this method is simple and operative.