摘要
加工质量实时预测是工件多工序加工质量控制的关键。航空制造领域关键零部件的异形空间大尺寸、材料难加工与小批量加工等特性,导致加工样本数据不足与加工误差监测困难。针对上述问题,提出一种基于赋值型误差传递网络的多工序加工质量预测建模方法。通过将质量特征引入多工序误差传递网络来描述加工过程中节点间的影响关系,形成赋值型的误差传递网络。并以关键质量特征节点为基础,采用基于粒子群算法优化的支持矢量回归机方法,构建单工序质量预测模型。在此基础上,基于赋值型误差传递网络的拓扑结构,合并单工序加工质量预测模型,以构建多工序加工质量预测模型。最后,开发了一个面向多工序加工质量预测的软件平台并以起落架零件的加工为例验证上述模型,结果表明该方法能够有效地预测加工误差,并从多工序的角度为异形零件的加工过程控制提供依据。
It is the key issue to predict the machining quality in real time for machining quality control in multistage machining processes（MMPs）. For aircraft manufacturing, the characteristics of special and large space size, hard machining materials, and small batch processing always lead to insufficient sample data and difficult monitoring of machining error. Considering the above issue, a quality prediction method is proposed based on assigned error propagation network（AEPN） in MMPs. Quality features（QFs） are introduced into a machining error propagation network（MEPN） for describing the influence relation between each node in machining process, and an AEPN is constructed too. Based on key QF nodes, a single process predict model（SPPM） is established by employing the support vector regression machine（SVRM）, which is optimized by the particle swarm optimization（PSO） algorithm. Based on this, the SPPM is merged based on the topology structure of the AEPN, and a multi-processes predict model（MPPM） is further constructed, A software platform for machining quality prediction in MMPs is developed, and a landing gear part is used to verify the applicability of the above method. The result shows that these methods can effectively predict machining error and provide foundation for the machining process control of special parts from the perspective of MMPs.
出处
《机械工程学报》
EI
CAS
CSCD
北大核心
2013年第6期160-170,共11页
Journal of Mechanical Engineering
基金
国家重点基础研究发展计划(973计划
2011CB706805)
国家自然科学基金(50975223)
湖南省科技厅工业支撑计划(S2011G20123359)资助项目
关键词
多工序
误差传递
支持矢量回归
质量预测
Multistage machining processes Error propagation Support vector regression Quality prediction
