论文部分内容阅读
提出在航空发动机多个部位安装多个振动传感器组成传感器网络.采用多传感器信息融合技术进行早期振动故障的诊断方法,并引入Dezert-Smarandache理论(DSmT)来处理由早期微弱故障本身所导致的各个传感器信息相互冲突的问题.在构建的早期微弱故障诊断系统框架中,采用基于本征模态函数(IMF)的信息熵特征提取方法提取各路振动数据的特征,采用反向传播(BP)神经网络完成对故障属性的判断并生成各种故障模式的基本置信分配,最后根据DSmT融合规则得到最终的诊断结果.算例表明采用该方法可以有效地解决早期微弱故障条件下的高冲突信息融合问题,故障诊断结果准确可靠.
Aiming at the problems that early vibration faults are diagnosed by multi-sensor information fusion technology and Dezert-Smarandache theory (DSmT) is introduced to deal with the problems caused by early weak fault itself Sensor information conflict with each other.In the framework of early weak fault diagnosis system, the information entropy feature extraction method based on intrinsic mode function (IMF) was used to extract the characteristics of vibration data of each road, and the back propagation (BP) nerve The network completes the judgment of the fault attributes and generates the basic confidence assignment of various failure modes, and finally obtains the final diagnosis result according to the DSmT fusion rules.The example shows that this method can effectively solve the problem of high conflict information fusion under the condition of early weak fault , Fault diagnosis results are accurate and reliable.