Bearing Fault Diagnosis Based On Binary Harris Hawk Optimization And Extreme Learning Machine

Rolling bearings are one of the most crucial parts of rotating machinery. Finding bearing defects early on might help to avoid impacting the overall operation of the manufacturing system. Machine learning for bearing failure Identification has recently become a particularly attractive topic due to its methods, which do not require a large amount of training data, as well as the fact that the collection of vibration data is typically the initial point of inquiry. A variety of defected bearing datasets have been published and are available. "The Case Western Reserve University's Bearing Center" is the most extensively utilized public dataset. In this research, a new methodology has been suggested using binary Harris Hawk optimization and extreme learning machines for bearing fault identification. First, the feature extraction has been retrieved from the bearing vibration signals. Following that, a strong feature selection approach is presented and used to remove irrelevant and redundant features using binary Harris Hawk optimization. Finally, artificial neural networks and extreme learning machines are used separately as classifiers. The results demonstrate that the suggested approach of binary Harris Hawk optimization and extreme learning machines has achieved 98.8% bearing defect diagnostic accuracy. The findings reveal that this approach has the benefits of bearing defect diagnostic accuracy and stability.