Fault Diagnosis Of Rolling Bearing Based On Acoustic Signal Measurement And Blind Source Separation

Machinery has played a key role in both production of the society and people’s lives,and most of its breakdowns are caused by abnormal operation of its machinery parts.As one of the important part of the components,the rolling bearing can directly affect the function of the whole machinery.While rolling bearing is quite vulnerable with a rather high possibility to breakdown,and according to statistics,about 30% of rotating machinery failure is caused by the rolling bearing,which means it is necessary to monitor the rolling bearing’s state and carry on fault diagnosis.Because the vibration signal generally has higher signal-to-noise ratio and is sensitive to the early fault of the bearing,most methods are based on the bearing vibration signal to diagnose,but the vibration signal acquisition needs to install the vibration sensor on the surface of the equipment.In the case,the contact measurement way is susceptible to the characteristics of equipment or work environment restrictions.Compared with vibration signal,the gathering method of sound signal has advantages of convince,non-contact measuring and non-disturbance,etc,especially in some occasions not suitable for contact measurement,it can give full play to advantages.Moreover,the sound signal during the operation of the bearing radiation can also reflect the state of the bearing.This paper aims to classify and diagnose the state of rolling bearing by measuring and analyzing the sound signal during its running in the method of blind source separation.Here follows the main work:(1)Introduce the structure of rolling bearing and explain the causes of breakdown,analyze the inner connection between vibration and sound signal,indicate the feasibility of using sound signal to diagnose,and state the trait of noise-low signal-noise ratio and easily to be affected by background noise and reverberation.All these analysis can give a direction for further research on fault diagnosis by sound signal.(2)On the basis of studying the existing blind source separation algorithm,aconvolution blind source separation(blind deconvolution)algorithm is improved and applied to fault diagnosis of rolling bearing.(3)For signal processing,in accordance with the low signal-noise ratio,propose a method for fault diagnosis of rolling bearing by wavelet transform and blind source separation after the study of current diagnose method.The wavelet transform is for improving the signal-noise ratio,blind source separation is for classifying acoustic signals with different characteristic to enhance the capacity of distinguishing causes.It has been proven that the method can effectively reduce the impact of noise,and the diagnosis is accurate.(4)For signal measurement,pressure gradient is select due to the strong anti-noise ability to the background noise,after measuring the pressure gradient,analyze the signal for fault diagnosis of rolling bearing based on blind source separation,and the effectivity of the method is also proved by experiment.