The effect of external dynamic loads on the lifetime of rolling element bearings: accurate measurement of the bearing behaviour

William Jacobs, René Boonen, Paul Sas, David MoensJournal of Physics Conference Series 2012 · DOI

Abstract

Accurate prediction of the lifetime of rolling element bearings is a crucial step towards a reliable design of many rotating machines. Recent research emphasizes an important influence of external dynamic loads on the lifetime of bearings. However, most lifetime calculations of bearings are based on the classical ISO 281 standard, neglecting this influence. For bearings subjected to highly varying loads, this leads to inaccurate estimations of the lifetime, and therefore excessive safety factors during the design and unexpected failures during operation. This paper presents a novel test rig, developed to analyse the behaviour of rolling element bearings subjected to highly varying loads. Since bearings are very precise machine components, their motion can only be measured in an accurately controlled environment. The test rig is optimised to perform accurate measurements of the bearing behaviour, utilizing a modular structure and a fully controlled multi-axial static and dynamic load system.

Introduction

The industry standard for calculating bearing life, ISO 281, primarily considers static or slowly varying loads. However, in modern applications like wind turbines, bearings are subjected to highly dynamic, multi-axial forces. When these dynamic loads occur, the thin film of lubricant separating the metal surfaces can break down, leading to metal-to-metal contact and "fretting" wear. Because current models don't account for this well, engineers often over-design components or face unexpected failures. The challenge lies in measuring a bearing's microscopic behavior under heavy vibration, as noise and thermal expansion usually drown out the data.

Experimental Methodology

Researchers at KU Leuven have developed a custom "mechatronic" test environment to bridge the gap between theoretical models and industrial reality. The rig is designed to subject rolling element bearings to complex, multi-axial static and dynamic loads while achieving nanometer-scale measurement precision.

Figure 1: (a) Schematic of the loading configuration and (b) 3D CAD model of the experimental setup showing the main components.

Spindle Design and Thermal Stability

To ensure the shaft itself does not introduce measurement errors, the rig uses a rigid spindle supported by tapered roller bearings in an "O" (back-to-back) arrangement. This specific geometry ensures that as the shaft heats up and expands, the bearing preload remains nearly constant, providing a thermally neutral environment for long-duration testing.

Figure 2: Cross-sectional view of the main spindle assembly highlighting the 'O' arrangement of the support bearings and the main shaft.

Hybrid Loading and Decoupling

The system decouples static and dynamic forces to ensure measurement integrity. Air springs provide static loads up to 10 kN, while electrodynamic shakers introduce broadband vibrations (25–500 Hz). Because air springs have very low lateral stiffness, the shaker's vibrations are not absorbed by the static load system. Furthermore, by ensuring the Center of Gravity (COG) of the test housing coincides with the center of the bearing, parasitic tilting moments are eliminated.

Modular Mounting and Precision Sensing

The rig features a modular mounting system using a collet chuck and intermediate sleeves, allowing it to accommodate different bearing types and sizes (10–52 mm diameter) without losing alignment precision. For measurement, capacitive proximity probes are used to detect bearing displacement. To reach an accuracy of 8 nm, the team characterized the shaft's inherent runout using a reference air-bearing spindle and subtracted this "signature" error during post-processing.

Key Results and Discussion

The test rig successfully demonstrates the ability to apply up to 10 kN of static force and dynamic loads up to 500 Hz. A 9.1 kW servo motor allows precise speed control from 0 to 3000 rpm. Finite Element Method (FEM) simulations confirmed that the rig’s structural resonances (first mode at 695 Hz) are safely above the operating frequency of the shakers, preventing interference. The achieved measurement resolution of 8 nm allows for the detection of microscopic lubricant film breakdowns, providing a window into the fundamental causes of bearing failure under dynamic conditions.

Conclusion

This instrumentation paper details a significant advancement in bearing testing technology. While the primary limitation is that it describes the instrument rather than long-term wear results, the rig provides the necessary precision to generate empirical data for updating ISO 281 standards. Future work will focus on using this platform to develop more accurate lifetime models that account for the complex vibrations found in modern industrial applications.

Summary updated

Abstract

Accurate prediction of the lifetime of rolling element bearings is a crucial step towards a reliable design of many rotating machines. Recent research emphasizes an important influence of external dynamic loads on the lifetime of bearings. However, most lifetime calculations of bearings are based on the classical ISO 281 standard, neglecting this influence. For bearings subjected to highly varying loads, this leads to inaccurate estimations of the lifetime, and therefore excessive safety factors during the design and unexpected failures during operation. This paper presents a novel test rig, developed to analyse the behaviour of rolling element bearings subjected to highly varying loads. Since bearings are very precise machine components, their motion can only be measured in an accurately controlled environment. Otherwise, noise from other components and external influences such as temperature variations will dominate the measurements. The test rig is optimised to perform accurate measurements of the bearing behaviour. Also, the test bearing is fitted in a modular structure, which guarantees precise mounting and allows testing different types and sizes of bearings. Finally, a fully controlled multi-axial static and dynamic load is imposed on the bearing, while its behaviour is monitored with capacitive proximity probes.
BibTeX
@article{jacobs2012effect,
  author = {William Jacobs and René Boonen and Paul Sas and David Moens},
  title = {{The effect of external dynamic loads on the lifetime of rolling element bearings: accurate measurement of the bearing behaviour}},
  journal = {Journal of Physics Conference Series},
  year = {2012},
  doi = {10.1088/1742-6596/364/1/012037},
}
Figure