基于神经网络切削毛刺专家系统的开发与其应用.pdfVIP

2012 年中国（国际）光整加工技术及表面工程学术会议论文集  基于神经网络切削毛刺专家系统的开发及其应用 王贵成 陈俊平 朱云明 周丹 （江苏大学机械工程学院，江苏镇江，212013 ） WANG Gui-cheng, CHEN Jun-ping, ZHU Yun-ming, Zhou Dan （College of Mechanical Engineering ,Jiangsu University，Zhenjiang 212013 ，China ） 摘要：金属切削毛刺是切削加工中产生的常见现象，它的形成与变化直接影响工件的加工质量和生产效率，已成为 影响和制约精密与超精密加工和自动化加工的关键技术之一。毛刺专家系统是毛刺形成机理研究与毛刺去除技术研 究的沟通平台，也是推进少无毛刺切削加工技术的重要技术渠道，其研究倍受国内外机械工程专家和学者们的高度 关注。本研究利用神经网络技术和专家知识系统技术，紧密结合金属切削毛刺实验研究，面向去除毛刺技术应用， 开发出具有网络功能的神经网络（NN ）金属切削毛刺专家系统。通过结合铣削实验中毛刺的形成与变化，利用本研 究开发的专家系统进行了毛刺尺寸的预测预报研究。结果表明，本系统的预测结果与实验结果有良好的吻合性，证 实了本系统的可行性与有效性。 关键词:专家系统; 金属切削毛刺;预报 Abstract: The formation of burr is a common phenomenon in metal cutting, its formation and variation directly affects processing quality and production efficiency of the workpiece. It is one of the key technologies which was applied in precision or super-precision manufacture and automatic processing. Burr expert system is not only a communication platform between burr formation mechanism and burr removal technology, but also an important channel that promote the development of burr-free cutting technology. It was highly concerned by mechanical engineering experts and scholars at home and abroad. This study develops neural network (NN) metal cutting burr expert system with network function, facing the application of burr removal technology, making use of neural network technology and expert knowledge system technology, and closely integrating experimental study of metal-cutting burr. Combined with milling, the study on prediction burr sizes using the system is carried out. The results show that the predicted results and the experimental results are in good agreement, which demonstrates the system is feasible and effective. K