基于无线传感器网络数据融合的滚动轴承故障诊断

Rolling Bearing Fault Diagnosis Based on Wireless Sensor Network Data Fusion

针对无线传感器网络(Wireless Sensor Network,WSN)在检测和诊断大型机械设备运行状态和机械故障中的应用,提出了基于无线传感器网络的滚动轴承故障诊断模型,结合主元分析和神经网络技术对无线传感器网络中的数据在网内进行融合处理,将原始的振动信号转换为故障类型,达到了减少通信量、节省网络能量消耗的目的.在实验中采用了工程上最常见的单列深沟球轴承,对传感器采集到的正常状态和故障状态数据进行了分析计算,在簇头节点进行特征级融合,提取出故障特征,得出相应的故障类型.实验结果证明了该模型和数据融合算法的可靠性和有效性.

It is a future trend to use wireless sensor network to detect operation state of large machinery and carry out fault diagnosis.A rolling bearing fault diagnosis model based on wireless sensor network was proposed,and principal component analysis and neural network technology were used to carry out wireless network data fusion in the network,with which original vibration signals were converted into fault type,in this way,correspondence and network energy consumption were reduced.In the experiment,the test subject was a deep groove ball bearing which is commonly used in engineering.Vibration data in normal and fault state collected by sensors were analyzed and calculated,data fusion of feature level was carried out on cluster nodes to extract fault feature in order to find out fault type.Experiment results prove that the model and the data fusion algorithm are reliable and effective.