Data-Driven Roller Bearing Diagnosis Using Degree of Randomness and Laplace Test

In this paper, we present a new diagnosis and prognosis method using the degree of randomness (DoR) measure and Laplace test procedure. The abnormal events are detected based on changes of randomness of vibration signals. The trend of randomness is resulted from faulty components such as roller bearings. We aim at the early detection of semi-failure events through the use of Laplace test statistic which measures the rate changes of abnormal event occurrence. Algorithms are data-driven and capable of making fault detections at its early stages. They have also been integrated into a realtime diagnosis system.