Innovative design process to optimise a spherical roller bearing for wind turbine rotors

Introduction

As wind turbines have evolved, their sub-systems and components have faced demands for greater performance within tighter economic constraints. The rotor mainshaft bearing is no exception.

This paper describes a novel engineering methodology, “to do old things in a new way”. Well established rolling bearing design methods were enhanced with modern calculation tools and combined in a process for design optimisation.

It illustrates the case of a spherical roller bearing; internal geometry was tailored to the particular nature of operation in wind turbines, providing a superior combination of mass and performance.


Approach

Spherical roller bearings were introduced by SKF in 1919. Recent milestones were the SKF Explorer performance class (1999) and more recent upgrades achieved by improvements in bearing geometry, materials and manufacturing methods.

Maximisation is not necessarily optimisation, however. The design team contrasted the typical wind turbine mainshaft against conditions found among diverse industrial applications;

Maximum speed up to only about 20 rpm (compared with otherwise up to 600 rpm).
Bearing temperatures starting from -40’C up to 80’C (lower than other machines)
High and variable loads, with relatively high axial component.

Such differences justify a dedicated bearing design.


Main body of abstract

The process involves 4 modules;
i. Generate a coherent geometry
ii. Analyse/Simulate with application inputs
iii. Apply unique SKF design criteria to qualify candidates
iv. Optimise with parameters reflecting user value

GEOMETRY GENERATOR
This is the core module, generating a geometry within given boundary dimensions. It is based on the main roller parameters; contact angle, length, diameter, pitch and osculation. Together with limitations from the required load cases, as well as clearance, lubrication and temperature inputs, this provides many feasible geometries. All need evaluation in application context.

ANALYSIS & SIMULATION
The SKF BEAST is a multibody simulation tool, a "virtual test rig", applicable to operational phenomena which are dynamic.
It enables study of the dynamic behaviour of all bearing components, including the cage, e.g. the forces and motions of the cage, skidding, skew and tilt behaviour of rolling elements.
SKF SimPro is an application engineering software suite. It considers all important factors of bearing application, including the effects of shaft elasticity and system deformations. Its versatile tool set includes clearance requirements, contact pressures & stress distribution, lubrication and contamination conditions, rating lives and advanced modules like flexibility, friction, etc.

DESIGN APPROVAL
Well established design rules must apply, to secure robustness. Analysis results are screened for fatigue loading conditions of the bearing components, their kinematics, static load limits, risk for fretting, C-value, etc.

OPTIMISATION
The process generates thousands of feasible designs! Visualization of the pertinent parameters (for example bearing mass and life) on a scatter diagram reveals a boundary line, the Pareto Front, representing optimum executions from which the engineer may flexibly select according to the user case.


Conclusion

With sometimes contradictory optimisation targets, there is no single solution. Selection can be guided by the priorities assigned to characteristic parameters for the intended application, which in this case were the load capacities, mass and life.

The end result was the new SKF spherical roller bearing for wind turbine mainshafts. Laboratory work was performed in verification and several pilots are currently underway in turbines of 2-4MW.



Learning objectives
Process-driven design tools can be applied to rolling bearings in wind turbines.

The Pareto Front is an effective means of guiding optimisation.

Process iteration in design stage is feasible for taking a multi-objective generic algorithm (MOGA) optimisation approach.

The MOGA approach can be enriched with new tools and qualitative values of a user in future, supporting evolution of a turbines.


N.B. The presentation during EWEA will provide an update on results from operation in the field.