摘要
机械加工中毛刺的形成与存在影响、制约着精密与超精密加工和自动化加工技术的发展。基于神经网络理论,构建并开发了金属切削毛刺专家系统(Ujs),通过对神经网络的设计,并作为推理机制,以数据库为训练样本,建立了毛刺形态、尺寸与切削条件之间的映射模型,应用于金属切削毛刺的控制及预报。结合车削实验,比较了系统预报结果与实验结果,取得了较好的吻合。神经网络专家系统的建立为解决毛刺预报与控制提供了一条有效的途径。
The development of precision and surperprecision and automation technology was influenced by the formation and existence of burr in machining processes. The UJS-expert system for metal cutting burrs was developed based on NN theory. Mapping mode between burr shape and size and cutting conditions was gained through, designed neural network and database acted as deduction mechanism and training sample separately. The expert system was applied to predict and control the burrs in metal cutting. Prediction results are in good agreement with experimental ones in turning test. The development of expert system provides an efficiency approach for burr controlling and prediction in metal cutting.
出处
《中国机械工程》
EI
CAS
CSCD
北大核心
2005年第12期1039-1041,1064,共4页
China Mechanical Engineering
基金
国家自然科学基金资助项目(50275066)
关键词
金属切削
毛刺
预报
专家系统
神经网络
metal cutting
burrs
prediction
expert system
neural network(NN)
