Rotor fault diagnosis for machinery fault simulator under varied loads

Machine fault diagnosis is a field of mechanical engineering concerned with finding faults arising in machines. In this paper, we use the Bayesian network (BN) classifiers and data mining technology to diagnose different kinds of rotor faults in machinery fault simulator (MFS) under varied loads. First of all, three kinds of popular BN classifiers are introduced as the diagnosis model for rotor fault, and the fault diagnosis modeling methods based on BN classifiers is established by data mining. Then, a MFS is introduced and applied to generate the vibration data of system with different rotor faults under varied loads, as dataset 1, dataset 2 and dataset 3. At last, the dataset 1 generated by MFS is used to demonstrate the rotor fault diagnosis process with BN classifiers. The same procedures are also implemented for dataset 2 and dataset 3 to show the difference of diagnosis results under varied loads.