Since 2008, our ONYX InSight team has been involved in the NREL GRC (Gearbox Reliability Collaborative) with modelling, analysis, design and testing activities.
“Gearbox 3 (GB3)” is a redesign of the gearbox previously tested GB2 with modifications to improve reliability and to support specific research goals.
InSight’s staff and NREL tested the GB3 gearbox on the 2.5MW dynamometer during the Fall of 2016. Throughout this Winter we have been assessing the data to determine how this redesigned gearbox performed. The graphic shows the design improvements over the original OEM gearbox.
Improving gearbox reliability is important for the wind industry as it reduces the levelized cost of electricity (LCOE). NREL has now completed testing of the redesigned GRC Gearbox 3 and it has greatly improved performance. Expertise gathered from this R&D project is now being applied by InSight to support owners with improving their gearbox life.
The test article at NREL is set up to allow a replication of the actual field operation of a wind turbine. See the figure below, where the entire drivetrain (main bearing, shaft, gearbox, high-speed coupling, generator and power electronics) are mounted on a turbine main frame supplied years ago from a wind farm here in northern Colorado. The dynamometer provides torque and speed input while the rotor loads from the wind are introduced by large hydraulic actuators. The very capable dyno allowed complicated testing, for example, one of test sequences takes a field recorded time series during 20m/s winds and replays those dynamic loads back into the drivetrain.
The drivetrain is mounted as in the nacelle and tested with loads that are reproduced just like in a real turbine
One of the innovative instrumentation activities on this gearbox was to place strain gauges within the planet bearings. There were 10 gauges bonded to the bearing inner race in custom machined grooves. During the dyno testing, we subjected the gearbox to 100% torque loading and then extreme bending moments, applied by the hydraulic actuators onto the main shaft. Due to the instrumentation, we were able to measure the load change and variation in the load zone on the gearbox’s planet bearings. The load zone measurements correlated closely with the RomaxWIND analysis conducted in the design phase of the project and will be available soon in an NREL report.
Strain gauges were installed on the planet bearings in specific locations for capturing the magnitude and shape of the load zone.
Planet bearing reliability – L10 life
The significant increase in planet bearing L10 life as a result of larger bearings can be seen between GB2 (pre-redesign) and GB3 (post-redesign). RomaxWIND was used to calculate the predicted planetary fatigue life for a predetermined drivetrain load spectrum. An increase in bearing life and a levelling of bearing damage can be seen between the upwind (UW) and downwind (DW) planet bearings. This is due to the stiffening of the gear assembly from the preloaded taper roller bearings (TRB) and due to a reduction in the wind-up of the carrier assembly. Note that GB2 used cylindrical roller bearings in this location.
To reduce failures of planet bearings it is important to achieve equal load sharing between the upwind (UW) and downwind (DW) bearings that support each planet. It can be seen that GB3 has vastly improved load sharing over the previous gearbox design (GB2). A major change in between the gearboxes was to upgrade from cylindrical roller bearings to taper roller bearings.
Often in the wind industry, bearings are stated as being designed (selected) for a 20-year life, by an L10 life calculation, which outputs a theoretical probability of less than 10% of the bearings failing during the design life. One must keep in mind that this is for each bearing, not a group of many bearings. Summing many individual bearing failure probabilities together obviously gives a lower system life. Not surprisingly wind turbine gearboxes generally require service to replace one or more of the bearings earlier than this 20 years. For the gearboxes, GB2 and GB3, it is important we compare the “bearing system life” of the planetary stage. The predicted L10 life for the eight planetary stage bearings (six planet TRB’s and 2 carrier TRB’s), shows an increase in the bearing system life from 3.4 to 11.9 years (3.5x increase) over the GB2 design. This system life is often skewed by the lowest life component, which in the case of GB2 is the poor performance of the UW bearings, reducing this design’s system life considerably. The carrier bearings far outperform the planet bearings for L10 life and thus have little influence on the system life calculation for both GB2 and GB3. The lesson here, from a bearing design for life perspective, is that one or two underperforming bearings in a gearbox can reduce the system life considerably.
The lifetime of a gearbox depends not on the design life rating of each bearing but rather on the system life of all bearings within it. It can be seen that the original gearbox design had a low combined L10 life for the planetary stage “system” that is greatly improved in the new design.
This GRC project is a collaboration of public and private partners supported by the US Department of Energy. As such, all the test data is available to the public. Feel free to contact us for links to more information on this project.
Need help with your gearbox? ONYX InSight provides the root cause analysis, engineering upgrades, instrumentation and supply chain support needed to improve your gearbox life. Get in touch.