Back to the programme printer.gif Print




Delegates are invited to meet and discuss with the poster presenters in this topic directly after the session 'Innovative concepts for drive train components' taking place on Thursday, 13 March 2014 at 11:15-12:45. The meet-the-authors will take place in the poster area.

Melanie Michon University of Sheffield, United Kingdom
Co-authors:
Melanie Michon (1) F P Robert Holehouse (1) Bo Wang (1) Kais Atallah (1) Gary Johnstone (2) Liang Xu (2)
(1) University of Sheffield, Sheffield, United Kingdom (2) Romax Technology, Nottimgham, United Kingdom

Printer friendly version: printer.gif Print

Presenter's biography

Biographies are supplied directly by presenters at EWEA 2014 and are published here unedited

Melanie Michon received the BEng degree, in 1999 and the MSc degree, in 2004, both in Electrical Engineering from the Hogeschool Limburg, and Eindhoven University of Technology, the Netherlands, respectively. In 2008, she was awarded the PhD degree in Electrical Engineering from the University of Sheffield, United Kingdom.From 1999 to 2004 she worked as a Research Assistant at Eindhoven University of Technology, and since 2004, she has been a Research Associate at the University of Sheffield. Her research interests include electromagnetic analysis of permanent magnet devices, advanced drive-train concepts and electro-mechanical interactions in wind turbine drive-trains.

Abstract

The effects of unbalanced magnetic pull in wind turbine drivetrains

Introduction

In rotating electrical machines, in addition to circumferential shear stresses which produce torque, radial attractive stresses between rotor and stator also exist. Unbalanced magnetic pull (UMP) occurs when the rotor axis is displaced from the stator axis, i.e. eccentricity exists, which affects the generator and associated mechanical components [1-3]. This effect is significant in large wind power generators where rotor eccentricity may be unavoidable due to large rotor mass, manufacturing tolerance, bearing clearance and system deflections. The paper presents the effects of UMP on different drivetrain design parameters, and recommendations on how to consider UMP during the drivetrain design phase.

Approach

A 2D analytical method is used to predict the unbalanced magnetic pull in permanent magnet generators and the method is validated with 2D/3D finite element analysis and a small experimental test rig. Fig. 1 shows the variation of the unbalanced magnetic force with a uniform rotor displacement, where it can be seen that there is good agreement between the analytical and finite element predictions and measurements. This method is subsequently used to predict UMP due to rotor eccentricity for large medium speed permanent magnet generators, and the results are combined with mechanical models of the generator and drivetrain.
The resulting electro-mechanical analysis provides insight in the significant effect of UMP on key drivetrain design parameters and therefore leads to improved bearing life prediction, improved understanding of the interactions between the gearbox and the generator and improved prediction of natural frequencies, all resulting in reducing the risk of drivetrain failure.
RomaxWIND provides a suitable platform for this electro-mechanical analysis, considering its well-established drivetrain mechanical analysis and the straightforward integration of combined generator and gearbox models. Its mechanical analysis provides accurate calculation of shaft and bearing deflections and resulting bearing life based on Beam Theory and stiffness analysis. Furthermore, dynamic analysis provides drivetrain natural frequencies and corresponding critical speeds.


Fig. 1 Variation of unbalanced magnetic force with eccentricity

It can be seen in Fig. 1 that the UMP varies linearly with rotor eccentricity, therefore can be modelled as a linear stiffness:

FUMP = -KUMP*d (1)
with FUMP the unbalanced magnetic force, KUMP the UMP stiffness and d the rotor displacement. However, as the UMP acts in the same direction as its originating force it should be modelled as a negative stiffness. RomaxWIND allows straightforward integration of UMP as a negative stiffness into its drivetrain models.


Main body of abstract


Fig. 2 Typical generator-bearing configuration

Fig. 2 shows the shaft and bearing arrangement for a typical medium speed wind turbine permanent magnet generator, consisting of a shaft supported by two deep groove ball bearings, on which the generator rotor is mounted. In this configuration gravity acting on the rotor will result in a deflection of the shaft and bearings, resulting in rotor eccentricity and, hence, an unbalanced magnetic pull. This UMP acts in the same direction as the gravitational force, resulting in equilibrium between the mechanical stiffness of the system and the total force on the shaft, where the increased load on the bearings will result in a significant reduction of the bearing life. The total force on the shaft is:

Ftot = FUMP + Fgrav (2)
with FUMP according to (1) and Fgrav the gravitational force on rotor and shaft. Assuming the UMP acts on the center of the shaft the bearing force is Fbearing = Ftot/2.


Fig. 3 RomaxWIND model of medium speed generator shaft assembly including UMP

Fig. 3 shows a RomaxWIND model of shaft and bearings for a typical medium speed wind turbine generator. The UMP is modelled as a set of linear negative stiffnesses along the rotor shaft and the total negative stiffness is calculated using the 2D analytical method, Fig. 1, resulting in typical values of 107-108N/m for medium speed permanent magnet generators.


Fig. 4 Effect of UMP on bearing force

Fig. 4 shows the increase in bearing force under influence of UMP for varying shaft diameters, where it is apparent that UMP causes a significant increase in bearing force. Indeed, for a nominal shaft diameter the bearing force is 1.85 times the force without UMP.


Fig. 5 Effect of UMP on bearing life

Fig. 5 shows the result of this increased force on the bearing life for the deep groove ball bearings, using the ISO 281 standard, including bearing clearance CN, and at rated torque. It is apparent from Fig. 5 that the bearing life is reduced significantly due to the UMP. Indeed, for a nominal shaft diameter the bearing life is reduced to only 15% of the rated life without UMP. From this analysis it is shown that UMP analysis needs to be included in the generator design stage to accurately predict bearing life and create a robust and reliable generator.

In order to study the effects of UMP on wind turbine drivetrains and assess their robustness the UMP model is implemented into three medium speed drivetrain topologies:
• Compact semi-integrated medium speed drivetrain – A two stage gearbox where the generator is located inside the gearbox. The generator rotor shaft is supported by two bearings at the downwind end of the shaft.
• Highly-integrated medium speed drivetrain – A single stage gearbox, where the generator rotor shares the same shaft with the sun gear. There is no generator bearing.
• Semi-integrated medium speed drivetrain – A two stage gearbox, where the generator is hanging on the back of the gearbox. The generator rotor shaft is supported by two bearings on both ends.


Fig. 6 RomaxWIND models for medium speed drivetrain topologies

RomaxWIND models have been built for these three topologies, Fig. 6, where the generator UMP has been included in the model. Manufacturing tolerance and bearing clearance are also included. Fig. 7 shows a comparison of the variation in airgap length, where it can be seen that for all topologies the combination of UMP, manufacturing tolerances and bearing clearance significantly reduce the airgap length. The effect is most pronounced for the highly-integrated medium speed topology, whereas the compact semi-integrated topology is least sensitive.


Fig. 7 Effect of UMP on airgap length for different drivetrain lay-outs

The generator dynamic behaviour is also key to a good drivetrain design. Hence, RomaxWIND is used for dynamic analysis of the generator natural frequencies. Fig. 8 shows the effect of UMP on the first natural frequency for increasing shaft diameter, where it can be seen that the UMP significantly reduces the first natural frequency, which is 15% lower for nominal shaft diameter. Hence, UMP should be accounted for in dynamic analysis of generator and gearbox.


Fig. 8 Effect of UMP on first natural frequency


Conclusion

The paper shows that UMP has a significant effect on bearing performance and drivetrain dynamic behaviour, and, hence, is a key parameter in the drivetrain design stage. A good understanding of how to account for UMP in this design stage is key to improving predicted bearing life and avoiding unwanted electro-mechanical interactions through improved prediction of natural frequencies. Furthermore, sensitivity to UMP is a dominant factor in selecting a suitable drivetrain topology, alongside sensitivity to manufacturing tolerances and bearing clearances.

The paper presents an analysis of the electro-mechanical effects of UMP in wind turbine drivetrains. A 2D analytical method is used to predict UMP in large medium speed permanent magnet generators. This method is implemented in a combined electro-mechanical analysis of the generator rotor shaft in isolation. It is shown that UMP increases the forces on the generator bearings significantly, and, hence, reduces the bearing life considerably. The developed model is subsequently used to compare the effects of UMP for three different drivetrain topologies, showing that the compact semi-integrated topology is least sensitive to the effect of UMP. Furthermore, the generator dynamic behaviour is studied under UMP showing that the first natural frequency is reduced significantly.

It is shown that RomaxWIND provides a suitable platform for this electro-mechanical analysis. UMP is easily integrated into well-established mechanical analysis by means of a negative stiffness. The software provides accurate prediction of bearing life and dynamic analysis of system natural frequencies, which is key in its ability to perform detailed comparisons of different, conventional and novel, drivetrain topologies.



Learning objectives
The paper shows that UMP is a key factor in the generator and drivetrain design phase in order to reduce risk of failure. UMP strongly affects bearing life and the system natural frequencies, and a methodology and suitable software platform are introduced to include UMP in drivetrain design and analysis in order to improve performance and reduce risk of failure.


References
[1] Dorrell, D.G.; Min-Fu Hsieh; YouGuang Guo; , "Unbalanced Magnet Pull in Large Brushless Rare-Earth Permanent Magnet Motors With Rotor Eccentricity," Magnetics, IEEE Transactions on , vol.45, no.10, pp.4586-4589, Oct. 2009
[2] Kim, U.; Lieu, D.K.; , "Magnetic field calculation in permanent magnet motors with rotor eccentricity: without slotting effect," Magnetics, IEEE Transactions on , vol.34, no.4, pp.2243-2252, Jul 1998
[3] Kim, U.; Lieu, D.K.; , "Magnetic field calculation in permanent magnet motors with rotor eccentricity: with slotting effect considered," Magnetics, IEEE Transactions on , vol.34, no.4, pp.2253-2266, Jul 1998