Testing the Wind Turbine Journal Bearing Flex-Pin Gearbox
ONYX InSight and NREL have been testing the Next Generation Drivetrain (NGD) on the 2.5MW dynamometer at the National Wind Test Center near Boulder, Colorado. ONYX InSight designed, instrumented and assembled the planetary single stage gearbox.
The testing started in the autumn of 2015 with carefully applied load (torque) steps, followed by progressively aggressive testing sequences. The test culminated in spring 2016 with many weeks of continuous start-stop cycles and stationary dithering, all in an effort to establish the threshold of abuse that planet journal bearings can endure for wind turbine applications.
Let’s see some test data!
The challenge with a 4 planet gearbox is equally distributing the torque loading amongst all planets. The planet shafts are cantilevered to allow them to flex and self-adjust to variations in torque, alignment, deflections and dynamic loading common to wind turbine gearboxes. The plots below show the bending behavior of the planet flex pins as they optimize the load share across the four planet gears orbiting around the ring gear. At very low torque the load share is uneven, moving to high torque the load share is exceptional. This was the design intent, so smiles all around!
The cantilevered planet flex pins were instrumented with strain gauges, along with 32 ring gear gauges. The sinusoidal plot is the variation of bending forces on each planet pin through a full rotation of the carrier, for planets 1-4. The results of the self-equalizing planets are shown in the bar chart in terms of relative load carried. The two cases are low load (21 kNm) and high (rated) load (417kNm). At rated load the planet load share improves to an exceptional result for a 4 planet system.
Another novelty of the Next Generation Drivetrain project was using NREL’s Controllable Grid Interface system to subject grid faults on the drivetrain. These are the much feared and rarely measured, ‘shock’ loads that originate in the electric utility grid causing the generator to introduce torque spikes on the gearbox high speed shaft. They are often talked about as a potential source of gearbox failure. Perched on the edge of our seats in the dynamometer control room as we commenced these tests we were surprised to see grid fault torque responses greater than 2 times the nominal torque. The figure below shows the nasty shock loading the gearbox was subjected to over and over…..
The rated torque of the gearbox is 417 kNm, to be conservative in our testing we did the grid fault testing at half that value. The orange line starts at 200 kNm then the brief grid event causes a torque spike of 475 kNm or 2.25x. Consider what that result would have been if tested at rated torque and that wind turbines may experience these grid events.
The journals performed flawlessly throughout these tests, but the true assessment was during disassembly inspection. These teardown results will be featured in our next blog on the NGD gearbox.
So what’s the benefit of developing and proving out this technology?
Benefit 1: Using journal bearings = 100’s of less moving parts in the planet stage = higher reliability and lower cost
Benefit 2: 4 planets + good load share = a lighter wind turbine gearbox for the same input torque = lower tower top mass
Here’s the dither test in action. Our drivetrain engineer Jesse Graeter and program lead, Jon Keller from NREL look on as the actuator dithers the input shaft back and forth.
Keep coming back to see when Part 3 of this story is published!