机电产品拆卸分析基础理论及回收评估方法的研究

【作者】 刘学平；

【导师】 刘光复；

【作者基本信息】 合肥工业大学 ， 机械制造及其自动化， 2000， 博士

【摘要】 本文对机电产品拆卸的信息建模、规划分析、评价方法等基础理论进行了深入、系统、全面的研究，探讨了基于神经网络的机电产品回收评估方法。主要研究内容如下： 1 对拆卸信息进行了分析及建模研究，建立了可拆卸方向范围、拆卸时间、拆卸能量的量化模型。构造了进行产品信息表达的网络图及零件可拆卸制约矩阵。 2 提出了零件间关联度的概念，作为零件间结合紧密程度的度量，利用层次分析法对零件间的关联度进行了定量分析。 3 依据关联度构造了零件间的关联矩阵。通过零件间的关联度矩阵，对构成产品的零件进行了子装配体的聚类，并通过空间几何约束矩阵的判别，及子装配体稳定性判别来判断结果的可行性。提出了拆卸路径的决策模型，实现了拆卸路径的自动生成。 4 探讨了基于人工神经网络的机电产品可回收性的评估模型和方法。通过把人工神经网络作为一个黑箱，从专家的判断样本集中提炼出专家的思想。实现用该网络代替专家进行评估判断。 5 设计了拆卸回收分析及评估系统，系统集成了多个模块的功能，包括产品的信息描述，网络模型的建立、拆卸分析、产品数据处理，拆卸回收评估等。更多还原

【Abstract】 In this dissertation, The base theory of disassembly and assessing method of recyclability is studied deeply and systematically. The main content are as followings: 1 The product disassembly information and its source are analyzed. The quantitative models of disassembly direction area, disassembly energy and disassembly time are built. The disassembly information is expressed by product network chart. The styles of parts correlative relations and mathematics judgement method are studied. Part disassembly restrict matrix is built, and By which, whether a part or an assembly can be disassembled is ascertained. 2 The concept of part relating degree is put forward, and it is measurement as combining tightness between two parts. The part relating degree is also analyzed quantitatively by using the analytic hierarchy process (AHP) in this dissertation. 3 According to the part relating degree, the relating matrix is built. The sub- assembly clustering can be realized by using the matrix. When some parts can be disassembled as a sub-assembly. the analysis difficulty of disassembly sequence decision-making can be simplified enormously, and the problem of disassembly sequence combination explode can be resolved efficiency. Disassembly analysis theory has been studied rather perfectly. Disassembly sequence decision-making model is built. The sub-assembly clustering can be realized by using the part-relating matrix. Based on the dimensional geometry, dimensional restrict matrix, sub-assembly stability, disassembly sequence can be auto built finally. 4 The recyclabitity assessment model based on Artificial Neural Network (ANN) has been established. The calculation results of some examples show that the method in this dissertation is effective. By using the ANN as a black-box, the result obtained from exports is used as sample set to train the ANN. The exports?idea (3) refined from the set can be employed as recycling assessment. 5 The software system of design for disassembly and assessment for recyclability is developed. The system contains many modulars, such as disassembly information express, product data processing. disassembly analysis, recyclability assessment.更多还原