Abstract: For the actual demand of high-speed Numerical Control(NC)equipment with complicated nonlinear features, a new fault diagnosis method was proposed based on improved evidence theory and nonlinear featurs. The fault features of high-speed NC equipment with multiple sensors were analyzed, and a group of feature vectors which reflected the system's nonlinear dynamic feature was extracted. The multiple fault diagnosis model based on evidence theory was presented and each evidence's mass functions were obtained according to the similarity among different modes. To solve the composition problem of multiple evidence in the situation of evidence conflicting, a dynamic parameter conflicting evidence combination method was proposed based on average trust degree. Experimental result indicated that this method had high recognition rate when the evidences were conflicting, and had practical value for diagnosis algorithm of nonlinear high-speed numerical control equipment.

Key words: nonlinear system, high-speed numerical control equipment, fault diagnosis, evidence theory

摘要： 针对高速数控装备的复杂非线性特性，基于改进证据理论和非线性特征提出了一种新的故障诊断方法。分析了高速运行状态下数控装备的故障特性，提取出一组反映系统非线性特性变化的故障特征向量。给出了基于证据理论的多类型故障识别模型，并利用模式之间的相似度获取各个证据的mass函数。为解决冲突情况下的多个证据合成问题，提出一种基于平均信任度的动态参数冲突证据合成方法。仿真实验结果表明，在证据存在冲突的情况下，该方法识别率高，适合于具有非线性特性的高速装备故障诊断。

关键词: 非线性系统, 高速数控装备, 故障诊断, 证据理论