分形法在发动机隐含故障诊断中的应用

在对发动机的机械故障诊断中 ,由于其结构复杂常使其故障特征不明显 ,为此提出用分形法对某些隐含故障进行诊断。介绍了分形法诊断故障的基本原理 ,并利用奇异值降噪技术和分形维数法对隐含故障的特征进行了提取 ,实现了对发动机轴承隐含故障的诊断。诊断结果表明 ,发动机轴承正常状态下的分形维数为 4.4,出现明显故障时轴承的分形维数为 5 .3,出现隐含故障时轴承的分形维数为 5 .37。由此可以看出 ,用分形维数可以诊断隐含故障 ,说明分形诊断法能从本质上反映故障的结构特征。

Diagnosis on cryptic fault of engine with fractal method

The fault attributes are not obvious for some equipment such as engine because of the complex structures, so the normal methods are not effective for diagnosing some cryptic faults. The noise reduction technique and the fractal dimension were used to abstract the characteristics of fault. The basic principles of fractal method applied in fault diagnosis were described. This method was successfully used to diagnose the fault of bearing in engine. The result implies that the fractal of good bearing is 4.4, and the fractal of faulty bearing is 5.3, and fractal of bearing with hidden faults is 5.37. It is illustrated that this method is very effective for describing the fault features and diagnosing the cryptic fault of engine.