【摘要】 为了提高复杂系统故障诊断的效率,简化诊断模型的结构,提出了基于模糊综合评判的故障诊断方法,构建了基于多分类器融合的诊断模型,并应用于数控机床滚动轴承的故障诊断。首先,多个分类器分别对高维故障特征进行信息融合,作出初步诊断;然后,采用模糊综合评判的方法,对多个分类器的初步诊断结果进行决策融合,建立了滚动轴承的单级模糊诊断模型,构造了以分类器输出信息熵为核心的评价函数,探讨了熵值权重分配方法,开展了数控机床滚动轴承磨损故障诊断的相关实验。实验结果表明:在存在诊断冲突的样本中,70%的样本经过模糊综合评判融合后可以正确识别故障类型,与单一分类器方法相比,故障识别率得到了提高,达到97.06%,证明了基于多分类器融合与模糊综合评判的诊断方法与模型的有效性。更多还原

【Abstract】 In order to improve the fault diagnosis efficiency of complex system and simplify the structure of diagnosis model,a fuzzy comprehensive evaluation method of fault diagnosis was proposed.The diagnosis model based on multiple classifiers fusion was established and applied to the fault diagnosis of rolling bearing in numerical control machine.Firstly,different classifiers that fuse the high dimensional fault characteristics were used to obtain the preliminary diagnosis of samples.Then the preliminary diagnosis results were fused by the method of fuzzy comprehensive evaluation to achieve the final diagnosis conclusion.A singlestage fuzzy diagnosis model was established accordingly.The evaluation function was constructed based on the information entropy.The method of determining weights was discussed with entropy coefficient.Experiments were carried out on the rolling bearing of numerical control machine.The results show that 70% of the samples having diagnostic conflicts can be identified correctly by the fuzzy comprehensive evaluation and the fault identification rate has been improved up to 97.06%as compared with the single classifier methodreaches which,which proves the effectiveness of the proposed method and model.更多还原