【摘要】 针对传统故障诊断方法未考虑到参数数据可信度以及历史故障诊断信息导致的误报率高的问题,文中利用故障诊断信息提出了经验迭代算法,确定了每种诊断参数权重数值,在诊断之前对数据进行可信度加权使其中的野值剔除或者减轻其权重,将各个参数的数据融合得出概率值并判断发动机是否出现故障。通过对添加可信度计算的故障诊断与不添加可信度的诊断进行对比,结果表明,故障数据在没有异常的情况下两种方法基本持平;在出现异常的情况下,文中方法较之传统方法,准确率明显提高。更多还原

【Abstract】 Traditional fault diagnosis methods have the limitations of low reliability of parameter data and high false alarm rate caused by historical fault diagnosis information.The paper presents an empirical iteration algorithm based on fault diagnosis information with the weight values of each diagnostic parameter determined.Before diagnosis,the reliability of the data is weighted to eliminate outliers or reduce their weight.The probability values are obtained by fusing the data of each parameter,on the basis of which the engine failure is judged.The comparison between the fault diagnosis methods with and without adding credibility shows that both methods are basically the same in the absence of abnormal fault data;when fault data are abnormal,the accuracy of the proposed method is significantly higher than that of the traditional methods.更多还原