摘要
针对机电设备故障诊断中存在的知识冗余和不确定性,从原始数据出发,利用决策表约简算法进行属性和属性值的约简,建立了故障诊断的规则库。给出了基于粗糙集的故障诊断和知识获取模型的一般结构。通过对旋转机械典型故障的分析,建立了决策表,通过对决策表的约简,减少了数据库中数据的数量,解决了故障诊断中知识获取的瓶颈问题。提出了决策表的属性值约简的一种简化算法。实验证明该方法是可行的。
In order to extract simple and effective decision rules for fault diagnosis from the available original fault data containing inconsistent and redundant information, the attributes and attribute value are reduced through the proposed decision table reduction algorithms. A knowledge acquisition model based on rough sets is proposed in detail. The decision table is built by analyzing typical faults of rotating machine. This method solves the bottleneck of knowledge acquisition. Diagnostic results show the validity of the approach.
出处
《北京机械工业学院学报》
2006年第2期5-8,共4页
Journal of Beijing Institute of Machinery
基金
国家自然科学基金资助项目(50375017)
北京市自然科学基金资助项目(3062008
3042006)
关键词
故障诊断
粗糙集
旋转机械
决策表
fault diagnosis
rough sets
rotating machine
dicision table
