University of Ferrara run-to-failure vibration dataset of self-aligning double-row ball bearings

Abstract

Data from real systems is an important resource for research in machine diagnostics and prognostics. The demand for data has increased exponentially in recent years due to the growing interest in prognostics and the development of AI technologies for predictive maintenance. When used for fault detection and predictive maintenance, data must be able to provide information about the degradation phenomena that occur in machines. In addition, one goal of prognostics is to predict the remaining useful life (RUL), which requires a large amount of data to apply data-driven techniques or validate physics-based models. Bearings are subject to a wide range of loads and fatigue stresses, and their failure can be catastrophic for the entire machine or plant. The Department of Engineering of the University of Ferrara has carried out an extensive experimental campaign to record the evolution of vibration signals throughout the life of self-aligning double row rolling element bearings. Six accelerated run-to-failure tests were performed, while the acceleration signals were continuously recorded by a uniaxial accelerometer. A radial load was applied to the bearing housing and controlled by a load cell. The shaft speed was kept constant and controlled by an electric motor driven by an inverter. The data set provided contains acceleration signals in the radial direction for the entire duration of the tests and can be used for research or industrial purposes.

Keywords

• Prognostics
• Diagnostics
• Rotating machines
• Accelerated test
• Accelerometer
• Predictive maintenance