摘要
基于切削试验探讨颤振征兆早期识别的模糊信息融合方法。试验中在同一个测量区内使用了功率传感器和加速度传感器,利用模糊数学与Dempster-Shafer证据论相结合的方法对两种传感器信息进行了分析融合,对切削状态进行了描述。试验证实:利用证据理论与模糊推理相结合的信息融合方法进行颤振征兆的早期模糊识别得出的目标切削状态的隶属度介于单一征兆隶属度之间,对过于敏感的传感器,会考虑其他传感器的信息予以修正,降低其评价隶属度;对于不敏感传感器,会考虑其他传感器的信息予以补偿。这种方法弥补了最大隶属度原则的缺陷,即在模糊推理中,系统对某一状态的隶属程度实际上是由模式特征集中贡献最大的那个特征决定的,而没有用到其他特征提供的信息,这说明证据理论与模糊推理相结合的信息融合方法在进行颤振征兆早期识别时具有更高的可靠度。
The fuzzy information combination method of chatter symptom early recognition is studied. The power sensor and acceleration sensor are engaged in the same time to measure the information of chatter. The Dempster-Shafer rule of information combination and fuzzy theory are applied in the data processed, the description of cutting state is given. The test shows that fuzzy information combination method is better than fuzzy logic method on chatter detection during the cutting process. Fuzzy information combination method uses more information, and makes up for the lacuna of rule of maximum membership degree which is determine by the biggest single character in the pattern's character set. The veracity of chatter symptom early recognition is improved by using information combination in fuzzy reasoning.
出处
《机械工程学报》
EI
CAS
CSCD
北大核心
2004年第2期108-111,共4页
Journal of Mechanical Engineering
基金
吉林省自然科学基金(20020623)
西安交通大学机械制造系统国家重点实验室开放基金资助项目
