Bearing Fault Detection For Brushless DC Motor Based On Stator Current

Brushless DC motor has been used widely in the industry for its advantages of simply control,high power density and etc.So its safe operation state is getting more and more concern.Bearing fault is the most common type of all faults.When the bearing fault happens,it will probably cause many disastrous consequences such as the rotor blocked,motor sweep and so on.So it is of great significance to extract the bearing fault information at the early stage.This paper studies the characteristics of the bearing fault shown in the stator current,and focuses on the method to extract the fault information.This paper establishes mathematical model of the bearing fault of the brushless DC motor by the impulse function,and analyzes the characteristic frequency of the bearing fault in the stator current.In the analysis,it is shown that the phase current only conducts for 2/3 cycle in every electrical cycle,and the phase current wave is an irregular square wave.This means that the phase current of brushless DC motor contains many harmonics,which can cause the bearing fault information buried,and weaken the bearing fault detection effect.To avoid this defect,this paper proposes two methods to detect the bearing fault with the maximum value of three phase currents and bus current.The maximum value of the three phase currents and the bus current are all continuous quantities,and compared with phase current,they contain less harmonics.It represents that the fault information in them can get less interference than phase current.So the maximum of the three currents and the bus current are better analysis objects.Bearing fault signal of the BLDCM is very weak in the current,and it is a nonstationary signal.This paper extract the signal with WPT(wavelet package transform).WPT can divide the signal into different WPT nodes,and the fault signal can be extracted from the specific node when the frequency characteristic of the fault signal is obtained.This paper decomposed the currents in healthy and fault conditions respectively into different WPT nodes,and restructure the appropriate node to calculate the root mean square(RMS).The bearing fault can be detected by contrasting the RMS.This paper builds the experimental platform with a BLDCM,fault bearings and the load,and the motor is controlled by DSP,and corresponding software algorithm is designed.The experimental results show the merit and effectiveness of the proposed method.