【作者】 詹晓娟；

【导师】 王培东；

【作者基本信息】 哈尔滨理工大学 ， 计算机应用技术， 2007， 硕士

【摘要】 铣削加工作为机械制造中一种常用的切削加工工艺,被广泛应用于汽车、航空及模具制造业中机械零件的粗、精加工。目前大多数企业在常规铣削加工中,往往采用经验数据或是参考切削用量手册来选择铣削加工参数。如何提高刀具切削加工的生产效率,降低加工成本,一直是机械加工领域专家们不断探索和致力研究的重大课题。因而,在铣削加工工艺设计中,需要采用快速、合理地确定铣削加工切削参数的新方法。建立铣削数据库,为机械制造业提供合理或优化的刀具数据,是提高切削加工效率和经济效益的最有效措施之一。神经网络中应用最为广泛的是BP算法,但有收敛速度慢、易于陷入局部极小的缺点;而蚁群算法是一种新型的模拟进化算法,有正反馈、分布式计算、全局收敛、启发式学习等特点。在分析了蚁群算法和人工神经网络技术基本特点的基础上,本文提出了一种有效的蚁群算法和神经网络相结合的方法,即借助蚁群算法优化神经网络,从而实现切削参数的合理选择并使切削参数的选择具有一定的智力水平。该方法有利于提高系统的运行速度和运算效率并可避免BP算法的缺陷。本文还使用该方法设计并实现了基于蚁群神经网络的铣削数据库系统,其特点是:系统从已有的实验数据中获得学习样本,并且通过用户向系统提供的刀具、材料、加工质量等参数,能够快捷且比较合理地制订出铣削加工的切削速度、每齿进给量、切削深度的大小。而且伴随着加工信息的反馈,系统能够通过自学习,不断完善数据库,提高自身决策能力,从而使决策结果更趋合理。实验表明,本文的方法能够快速合理的制定切削参数,应用本系统大大缩短了刀具选择的时间,减少了工人劳动,同时降低了加工成本,不但克服了刀具选择过程中人为因素的影响,还有效控制了加工质量。更多还原

【Abstract】 Milling is broadly used for the manufacture of profiled components in aerospace, automotive and mould/die industries. Cutting parameters in Milling are determined usually based on either experience or reference handbooks. How to improve machining efficiency and lower the cost is one of the important tasks explored and studied for a long time by machinists. New methods are needed to rapidly and accurately decide the parameters of a milling operation in designing the machining technology, among which building milling tools’ database and providing optimized data is one of the effective methods to improve machining efficiency and lower the cutting cost.The back–propagation (BP) algorithm is the most widely used variation in neural networks. However, it has some shortcomings, such as slow convergent speed and easy convergence to the local minimum points. And the ant colony system is a novel simulated evolutionary algorithm. It has positive feedback, distributed computation, global convergence, and uses a constructive greedy heurism. The paper first analyzes the characteristics of both the ant colony system and artificial neural net (ANN), and realizes a new ant colony system which is based on ANN for choosing cutter parameter. The method has certain level of intelligence. It can improve the computing efficiency of the system and avoid a lot of algorithmic defects of BP.The paper also implements a milling database system based on the algorithm. It characterizes: ability to obtain study samples from experimental data; able to rapidly and reasonably decide cutting speed, feed of a cutting-tool’s tooth, cutting depth with the condition of cutting tool, material of work piece, and the required quality; strong self-study ability from operation feedback to refine the database, improve judgment and make it more reasonable.Experiments show that the method proposed in this paper can rapidly and accurately decide cutting parameters of milling,save much time in choosing tools, cut down on labors, thus lower the processing cost. Besides, it overcomes the influence of anthropic factor during tool choice, and brings manufacturing quality under finer control.更多还原