废旧零部件再制造工艺路线模型及其机器学习研究

发布时间：2018-12-14 02:23

【摘要】：再制造是将废旧产品制造成“如新品一样好”的再制造产品的再循环过程，对缓解资源短缺和资源浪费的矛盾，保护生态环境，具有重要的现实意义。论文结合国家自然科学基金“废旧机电产品再制造系统不确定性工艺规划与生产调度问题研究”（项目编号：51105395），对再制造工艺规划中的废旧零部件分类、再制造工艺路线等问题展开研究。首先，对国内外再制造、再制造工艺规划的发展现状和发展趋势进行综述与分析；分析废旧零部件的不确定性，建立一种包含再制造工艺规划任务层、再制造工艺规划关键技术层、再制造工艺规划支撑层等三个层次的再制造工艺规划体系结构，并进一步分析不确定性对再制造工艺规划的影响。其次，结合废旧零部件常见失效类型分析，建立一种综合考虑废旧损坏形式、损坏程度等特征的废旧零部件分类指标体系；基于此，建立一种基于模糊聚类的废旧零部件归类方法，并把该方法应用到废旧机床主轴分类中。再次，针对再制造工艺路线的不确定性问题，引入GERT图方法，建立一种基于GERT图的再制造工艺路线模型，该模型可描述各废旧零部件的所有可能工艺路线，并可建立工艺的概率分支与各工序加工时间分布函数；通过信号流图以及矩母函数对模型进行求解，能得到废旧零部件到达某一节点的概率与期望时间；结合废旧机床主轴再制造工艺，对所建立模型进行验证分析。然后，针对基于GERT图的废旧零部件工艺路线模型中工艺概率不确定性问题，建立一种基于贝叶斯网的GERT图再制造工艺概率参数机器学习模型，并将该方法应用到废旧机床主轴再制造工艺概率参数学习实践中。最后，，结合重庆机床集团再制造实践，采用Arena仿真软件对基于GERT图的废旧机床主轴再制造工艺路线模型进行仿真验证；同时，采用Matlab的贝叶斯工具对基于贝叶斯网的GERT图再制造工艺概率参数学习模型进行计算。
[Abstract]:Remanufacturing is a recycling process in which waste and used products are manufactured as "as good as new products". It is of great practical significance to alleviate the contradiction between the shortage of resources and the waste of resources and to protect the ecological environment. Combined with the National Natural Science Foundation of China "Research on uncertain process Planning and production scheduling of waste Mechanical and Electrical products remanufacturing system" (item number: 51105395), the paper classifies waste parts in remanufacturing process planning. Research on remanufacturing process route and so on. Firstly, the development status and trend of remanufacturing and remanufacturing process planning at home and abroad are summarized and analyzed. The uncertainty of waste parts is analyzed, and a system structure of remanufacturing process planning is established, which includes the task layer of remanufacturing process planning, the key technology layer of remanufacturing process planning, and the support layer of remanufacturing process planning. The influence of uncertainty on remanufacturing process planning is further analyzed. Secondly, based on the analysis of common failure types of waste parts, a classification index system of waste parts is established, which considers the characteristics of damage form and damage degree. Based on this, a method of classifying waste parts based on fuzzy clustering is established, and the method is applied to the classification of waste machine tool spindle. Thirdly, aiming at the uncertainty of remanufacturing process route, GERT diagram method is introduced to establish a remanufacturing process route model based on GERT diagram, which can describe all possible process routes of waste parts and components. The probability branch of the process and the distribution function of the processing time of each process can be established. By solving the model by signal flow diagram and moment generating function, the probability and expected time of waste parts reaching a certain node can be obtained, and the model is verified and analyzed by combining with the remanufacturing process of the spindle of waste machine tool. Then, aiming at the uncertainty of process probability in the process route model of waste parts based on GERT diagram, a machine learning model of GERT remanufacturing process probability parameters based on Bayesian network is established. The method is applied to the learning practice of the process probability parameters of the spindle remanufacturing of the waste machine tool. Finally, combined with the remanufacturing practice of Chongqing Machine tool Group, Arena simulation software is used to verify the remanufacturing process route model of waste machine tool spindle based on GERT diagram. At the same time, the Bayesian tool of Matlab is used to calculate the probability parameter learning model of GERT remanufacturing process based on Bayesian network.
【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP181;TH16

【参考文献】