Using Fourier Analysis and Theoretical Equation to Predict Bearing Raceway Waviness

Large size bearing’s raceway waviness is a most common NVH issue that presenting in manufactured bearing. Unlike small or medium size bearing with its raceway been honed after grinding. Typically, for large size bearing’s raceway are not honed due to its size limitation, unless there are special requirements from customers. Hence, it will inevitably generate great number of waves on the raceway circumference, which lead to a significant source excitation to the system. Although great amount of experiments substantiated that bearing waviness will not affect bearing life. However, annoyance noise is the main reason that customer reject bearings for quality concern. Hence, at the early stage of identifying raceway waviness issue at customer site could prevent from economic loss. Generally, the energy level caused by large size bearing waviness could be distinctive. Testing engineer will capture time domain vibration data by using radial accelerometer cells mounted on the housing for bearing health assessment. Then the vibration frequency spectrum can be converted by using Fast Fourier Transformation method. But, the preliminary step of using FFT is that people should know exact wave numbers on raceway circumference in order to trace back waviness excitation frequency. In this paper, a theoretical equation is developed, combining with common Fourier Analysis method. pinpoint raceway wave numbers in a fast and direct way become reality. A case example of CRB been used in a 5.5MW wind turbine generator and its NVH data will be analyzed for waviness equation prediction. Finally, the problematic bearing is also checked by waviness analyzer to confirm exact wave numbers on raceway circumference.