Fault diagnosis of roller bearing based on EMD and auto-power spectrum method

The vibration signal of roller bearing is the non-stationary signal. And the conventional Fourier transform can't identify the local feature of the vibration signal, which has a few limitations. Therefore, a method for the fault diagnosis of roller bearing based on empirical mode decomposition (EMD) and auto-power spectrum was presented. The vibration signal of roller bearing was decomposed into intrinsic mode functions (IMFs) by the EMD method. Then the IMFs were analyzed with auto-power spectrum, which may include the fault frequency of roller bearing. And then, the fault type of roller bearing was identified. Through the verification of actual measured vibration data, this method is very effective for the fault diagnosis of roller bearing.