Defect size estimation in rolling element bearings using vibration time waveform

In this paper a new approach to determine the size of defects in rolling element bearings is proposed. This approach is based on the statistical characteristics of the vibration signals generated by rolling element bearings. Although some traditional statistical parameters such as r.m.s. and kurtosis have some diagnostic capabilities, these parameters are not suitable for quantifying the defect size. The proposed new model was validated experimentally for both inner and outer race defects and it is concluded that the newly proposed model can determine the defect length of rolling bearings.