Fault Diagnosis of Rolling Bearing Based on Improved Mathematical Morphology

Aiming at weak fault features extraction of rolling bearing, an improved mathematical morphology method based on feature amplitude energy index (FAEI) and autocorrelation energy ratio (AER) was proposed. Firstly, seven morphological operators were used to process the signal, and the optimal operator is selected based on FAEI. Then, the length parameters of flat structure element were adaptively determined using AER criterion. The bearing fault experiments show that AER criterion can select the optimal length of structural elements, and combine the morphological operator to perform optimal morphological filter for fault signals. The results show that the proposed method can extract impact pulse signals in a noisy backgroud, and accomplish weak features extraction of bearing fault. Compared with the morphological method based on kurtosis, the improved method makes the characteristic frequency amplitude increase by 100% in fault detection.