基于桶消元的数控机床故障模糊贝叶斯网络推理方法

用模糊数描述节点的多种故障状态，用模糊子集描述根节点的故障概率，用贝叶斯网络的条件概率表描述节点间的不确定关系，构建数控机床故障的模糊贝叶斯网络；采用桶消元算法，对数控机床故障的模糊贝叶斯网络进行推理和计算，获得根节点的后验概率，为数控机床故障诊断与分析提供依据；最后，以数控机床刀架系统故障的贝叶斯网络推理过程为例，验证所提理论和方法的有效性。

Reasoning Method of Fuzzy Bayesian Network for CNC Machine Tool Fault Based on Bucket Elimination

Aiming at the fault information of CNC machine tools with incompleteness, inaccuracy and uncertainty, a reasoning method of fuzzy Bayesian network for CNC machine tool fault based on bucket elimination was proposed. The fuzzy numbers were used to describe the various fault states of the nodes, the fuzzy subsets were used to describe the failure probabilities of the root nodes, and the conditional probability tables (CPT) of the Bayesian network were used to describe the uncertainty relations between the nodes. The fuzzy Bayesian network for CNC machine tool fault was established. The bucket elimination algorithm was introduced to reason and compute the fuzzy Bayesian network for CNC machine tool fault, and the posterior probabilities of root nodes was obtained to provide the basis for diagnosing and analyzing CNC machine tool fault. The fault reasoning process of Bayesian network for CNC machine tool carrier system was taken as an example to verify the effectiveness of proposed theory and method.