Abstract: According to the complex natures of fault diagnosis for CNC machine tool,the fault diagnosis based on graph theory was researched.The multi-fault propagating model was proposed to represent these complex natures by characterizing the relationships of fault propagation,and this model would be processed by the matrix algorithm and the hierarchy algorithm.The priority of source faults was sorted by the risk-based fault localization algorithm.The effectiveness and correctness of proposed method was proved by a case of ram feed system of CNC boring machine tool FB260.

Key words: numerical control machine tool, fault diagnosis, graph theory, multi- fault propagating model

摘要： 针对数控机床故障诊断问题的复杂性特点,研究了基于图论算法的数控机床故障诊断的方法。对故障传播关系的表征建立了多故障传播模型,并对其进行矩阵化和层次化处理,再利用基于全局风险影响度的故障原因定位算法对多故障进行排序,确定故障原因的优先级。以数控镗床FB260的滑枕进给子系统的故障诊断为例,论证了该方法的正确性及有效性。

关键词: 数控机床, 故障诊断, 图论, 多故障传播模型