【作者】 刘伟；

【导师】 任成祖； 王太勇；

【作者基本信息】 天津大学 ， 机械制造及其自动化， 2010， 博士

【摘要】 工艺设计是机械制造业的重要基础工作,是连接产品设计与产品制造的桥梁,是一个经验性很强且随制造环境变化而变化的决策过程。工艺设计的质量和效率直接影响企业制造资源的配置与优化、产品质量与成本、生产组织效率等。当前,随着制造业的飞速发展,产品更新换代频繁,多品种、小批量的生产模式已占主导地位。因为传统的工艺设计存在着一致性差、效率低、难以保证数据的准确性等缺陷,又过于依赖工艺设计人员个人的经验和水平;而早期的CAPP专家系统又片面追求工艺决策的自动化,知识获取困难、推理方式单一,所以均难以满足现代制造业的发展需要。本文针对上述问题,结合人工智能特别是计算智能技术的研究成果,对智能CAPP系统中的知识表示、工艺路线和切削参数的决策等问题进行了研究与探索,研究内容和成果主要体现在以下几方面:1.在对传统的CAPP专家系统的不足进行分析的基础上,构建了基于知识的智能CAPP系统的总体框架,分析了它的组成及各个组成部分之间的信息传递,对系统各部分功能的实现方法和相关技术进行了多角度的研究。2.结合零件特征建模技术,综合面向对象的表示方法、框架表示法等多种知识表示方法的特点,提出一种基于零件特征的知识表示方法。将零件特征信息划分成特征元、加工元和切削元三个层次,然后采用XML语言对零件特征信息进行了结构化描述,从而构建了零件信息知识库。3.针对智能CAPP系统中的工艺路线的决策问题,提出了基于改进蚁群算法的优化方法。在将被加工零件划分为若干特征元和加工元的基础上,根据加工元的属性,用加权海明距离表示它们之间的相似程度;在蚁群算法中设置前趋表,用来表示加工元之间的约束条件,从而使得算法在对解空间进行搜索的时候,既要遵循基本蚁群算法中的禁忌准则,又要受到前趋表的限制。这样不仅保证了计算结果的有效性,而且提高了算法运行的效率。算法运行的结果是一条最优或近优的工艺路线。4.针对智能CAPP系统中的切削用量决策问题,将Pareto最优解的概念和遗传算法结合起来,提出一种基于先寻优后决策求解模式的优化算法。以切削效率和刀具耐用度为目标建立多目标优化模型;采用置零法处理不符合约束条件的个体;通过竞争的方法构造进化过程中所产生的Pareto最优解,并将其保存在非劣解集中直接保留到下一代,从而保证了算法的搜索方向;基于小生境技术,建立一种排挤机制,抑制个体的近亲繁殖,以提高种群的多样性;使用混合交叉算子和步长变异算子进行基因重组。算法运行的结果是一组沿Pareto前沿面均匀分布的优化解。在对以上基本理论进行研究的基础上,本文以天大精益公司的ERP系统为开发背景,采用基于Browser/Server模式的三层架构,进行了基于计算智能技术的CAPP系统的开发。更多还原

【Abstract】 The process design is an important basic work in mechanical manufacturing industry and the bridge which connects the product design and the product manufacturing and an empirical decision-making process which varies with the manufacturing environment. The configuration and optimization of enterprise manufacturing resources and the product quality and cost and the production efficiency are influenced directly by the quality and efficiency of process design. At present, products are upgraded frequently and the enterprises have to produce a large variety of products in little batch sizes with the fast development of manufacturing industry. There are a lot of shortcomings of traditional process design such as poor consistency, inefficiency and it is difficult to ensure correct data. Early CAPP expert systems put one-sided emphasis on automation decisions and the knowledge acquisition is very difficult and the reasoning method is simple. It can be seen that traditional process design and early CAPP expert systems are unable to meet the development needs of modern manufacturing industry. This paper presents related research and exploration on such issues based on research results of artificial intelligence especially computational intelligence technology. The main research content and achievements are as follows.1. The overall frame of an intelligent CAPP system based on knowledge is set up on the analysis of shortcomings of traditional CAPP expert systems. The system composition and the information transmission are analyzed and the implementation and related technologies of the function of its sections are studied from multiple perspectives.2. According to the characteristics of representation method of knowledge frame and object-oriented, an approach of knowledge representation based on part feature is introduced on the basis of the feature-based modeling technology. The part feature information is divided into three layers, namely, feature cell, operation cell, cutting cell. Then the part feature information is described structurally by using XML language and the knowledge base of part feature information is built in this way.3. Based on the improved ant colony algorithm, an algorithm is put forward for the process route optimization to solve the decision-making problems of the process route in a intelligent CAPP system. First each part was divided into some feature cells, and the operation cells were generated by machining chains of feature cells. The similarity of every two operation cells was represented by the weighted Hamming distance between them on the basis of their properties. According to the constraints of the operation cells,the preference sequences of the operation cells can be determined, and the fore cells of each operation cell can be obtained and were stored in the fore tables. The validity of calculations will be ensured and the efficiency will be promoted. A process route was generated and optimized under the control of the tabu criterion and constraints.4. Based on the Pareto genetic algorithm, an algorithm based on a solving mode of optimization before decision is proposed for the cutting parameters selection and optimization to solve the decision-making problems of the cutting parameters in a intelligent CAPP system. First, a muti-objective model was built by analysis of restraint with cutting speed and feed as optimization variables and cutting efficiency and the tool life as optimization objectives. Second, the selection operator was improved. The fitness of individuals which fail to meet the constraint requirements equal to zero.In order to ensure the search direction, a non-inferior set was set up to save Pareto optimal solutions which were generated by competition during evolutionary processes. The crowing mechanism based on niche technology was established to keep population diversity. And then, genes were recombined by means of mixed crossover operator and step-size mutation operator and an optimal set which distributed uniformly along the Pareto front was obtained after a few times iteration.According to the theoretical studies above, this paper which takes the ERP system of TDNC corporation as the background introduces the development of a CAPP system based on computational intelligence technology by using the 3-tiered architecture based on the Browser/Server model.更多还原