基于粗糙集与支持向量机的加工质量预测  

Processing Quality Prediction Based on Rough Set Diagnosis and Support Vector Machine

作　　者：王新海[1] 孙培禄[1] 郑锐 WANG Xin-Hai;SUN Pei-Lu;ZHENG Rui(Department of Mechatronics Engineering,Yuncheng University,Yuncheng Shanxi044000,China;Xi'an Xiangxun Technology Co.,Ltd.,Xi'an Shaanxi710000,China)

机构地区：[1]运城学院机电工程系,山西运城044000 [2]西安翔迅科技有限责任公司,陕西西安710000

出　　处：《机电产品开发与创新》2020年第6期15-19,共5页Development & Innovation of Machinery & Electrical Products

基　　金：山西省高校科技创新项目(2019L0854);运城学院科研项目(CY-2018021);运城学院博士科研启动项目(YQ-2017009)。

摘　　要：针对航空零件质量预测中质量关联因素复杂和样本数据不足的问题,提出了一种基于粗糙集诊断与支持向量机相结合的加工质量智能预测方法。针对产品质量数据不断更新的情况,采用基于模糊粗糙集的增量属性约简方法计算各前道工序质量特征及本工序加工要素对被预测质量特征的影响程度,得到关键条件属性,并根据这些关键条件属性进行样本数据筛选;然后经过灰狼算法优化的支持向量机对样本数据进行训练,得到最终的质量预测模型。最后,以飞机起落架零件的加工质量预测为例讨论了该方法的可行性和可靠性。Aiming at the problem of the complex association among processing factors and insufficient sample data in the small batch process for the aviation parts’quality prediction,a processing quality intelligent prediction method is proposed based on rough sets diagnosis and support vector machine.Using rough set theory,the i mpact of factors on the predicted characteristics’quality can be calculated,which include the quality characteristics of each front-step process and the processing elements.Thus we can derive the critical condition attributes in prediction,and sample data is filtered according to these attributes.Then we used support vector machines to train the sample data and got the final quality prediction model.Finally,we gave an example of aircraft landing gear parts machining quality forecasts to illustrate the feasibility and advantages of this method.

关 键 词：质量预测 粗糙集诊断 支持向量机 

分 类 号：V264.1[航空宇航科学与技术—航空宇航制造工程]