A Technique Of Fault Diagnosis On Gear Reducer In Slag Roller Mill Based On Wavelet Analysis

Gear box is common power transmission components in all kinds of machinery and its study on fault diagnosis has been achieved great development.If inner race of rolling bearing was pitted seriously and the pit was enough big to jam two balls,the drive shaft had a lot of torque,resulting in being twisted off.Slight fault occurred in gear,bearings etc.components were found early,which could avoid follow-up big loss.This is very significant in engineering.The vibration signal was always unsteady.Wavelet transform,empirical mode decomposition etc.time-frequency domain analysis methods are often used to analyze and deal with.Because of wavelet transform having good time-frequency characteristic and localization capability,this thesis uses it to study the fault diagnosis of gear box.And result was tested by fact.The main results are as following:(1)According to analyzing gear box fault's mechanism and characteristic and the correlation of gear box and time-frequency characteristic,a technique of fault diagnosis on gear box was put forward based on wavelet analysis and frequency domain parameters regarding as fault characteristic.(2)Analyzing the frequency characteristic of wavelet filter,a wavelet function is chosen due to its best characteristic.The wavelet function was verified by simulation signal.Through analyzing the influence of sampling frequency,frequency resolution and wavelet packet decomposition on decomposition result,layer of decomposition and order of node frequency are determined.The error rate of decomposition was discussed by wavelet packet decomposition in eight layers simulation experiment.(3)Extracting the signal component whose frequency was behind bearing's passing frequency,gear's meshing frequency and shaft's rotating frequency through wavelet packet decomposition in eight layers,which could be as fault characteristic.Aiming at gear fault characteristic,it is necessary to do zoom spectrum analysis for confirming its fault condition.A way of extracting fault characteristic was put forward,which only displayed amplitude where every characteristic frequency in frequency,their double and triple.Thus it is effective to observe and judge the vibration.(4)The vibration signal under real and producing form three work conditions(normal,gear's pitting corrosion and feedstock moisture)were used to verify the way of extracting.When gear pit corrosion,the amplitude of meshing frequency were much bigger than normal;when feedstock was wet,it was found that the amplitude of roller mill's main frequency was bigger than normal and it has nothing to do with the fault of gear reducer.