An AT Load-Stabilizing Flywheel

An affordable, load-stabilizing flywheel will be essential to any off-grid system to partially replace batteries and to prevent over-heating of critical components from constant surging and potential brown-outs. Surges and brown-outs significantly reduce the lifetime of appliances energized by any off-grid, micro-power plant—very typically under-sized due to cost. So, just buy a bigger inverter, you might think. Buy the biggest, most expensive inverter you can afford, and it still won’t be big enough. Have you seen those tiny capacitors?

Using a flywheel on a small generator for load-stabilization is not new. The typical, off-grid diesel unit comes with horizontal flywheels. Such heavy, horizontal flywheels, however, cannot be retrofitted to non-diesel generators.


I was excited a year ago to discover the potential of DIY vertical-axis flywheels that could be added to an existing off-grid system. I’m sharing this idea because I’m betting other off-gridders will appreciate the cost-benefit of retro-fitting a vertical axis flywheel to their system as soon as they experience an inverter fire (cheap inverters have no heat shield behind them) or a brown-out ruins their refrigerator or computer. You’ll either buy a load-leveling flywheel, a new inverter, or a string of new appliances.

Buy the flywheel!

Better yet, make one yourself like I did. Here’s a sketch and photos of the prototype. Our design goal was to build a small, inexpensive (costs less than a refrigerator) AT flywheel that would produce 1000 watts of instant power at 1500 rpms for a few minutes—to handle a surge.


We started with a flywheel made of steel cable wound on a large tractor front wheel rim and then covered the cable with fiberglass and Bondo to make it smooth. We balanced it on a truck tire balancer, even though the unit will actually balance itself!

To drive the flywheel, we are using a 2hp, high torque, 1150 rpm motor. To relieve some of the start-up stress on this drive motor, we added a heavy duty, NORAM, three leg, centrifugal clutch which connects the drive motor to the generator motor. Once up to speed, the drive motor shuts off and the generator motor keeps spinning because of the flywheel inertia. The generator motor is a treadmill D/C motor that will produce 30 amps and 90 volts at 1750 rpms—that’s about 1000 useful watts D/C at our speed. This charge is available to the DC side of the inverters.

The bearings on this machine are over-kill. There are two aircraft engine, tapered ball bearings that can each handle an axial load of 1 ton and spin at over 10,000 rpms. The entire drive assembly is mounted on a tubular frame and welded to a double gimbal. It is this gimbal set-up that absorbs all the vibrations from multiple sources. We tried a single gimbal, but the shaking was uncontrolled. With a double gimbal, the entire apparatus settled down. The entire machine is quiet and stable at our current design speeds.

bearing .jpeg

We have yet to run controlled tests on exactly what it is outputting at each interval of time and rpms. We have no plans to use magnetic bearings or build a steel vacuum chamber used in expensive commercial units now on the market. My AT design criteria insist that what I use here in the USA must be reproducible in the rural workshops of East Africa. I think the unit in this picture meets that criteria.

vert axis flywhl.jpeg

I must conclude by saying that even this small flywheel is dangerous. The cable “cage” on the lower part of the 5 ft. high, tubular steel frame is intended to capture a loose flywheel. Let’s face it, if that thing comes loose at 1200 rpms, the cage only gives us the illusion of control. To ensure real safety, this type of flywheel must be buried inside a 500 gal. concrete septic tank that is carefully water-proofed. It must be buried in case any of the bearings or a weld should fail. A loose flywheel will destroy the machinery inside the concrete tank but not the expensive equipment and batteries in the power shed. Imagine what damage a 600 lb. flywheel could do if it got loose? You’d find it in the next county unless it was buried in a concrete tank too.

Since this small, prototype unit works so well with a tiny 120lb. flywheel at 1150 rpms, we are, in fact, hoping to scale one up to see if a 500-600 lb. flywheel spinning at 2-4000 rpms will give us enough “juice” to replace my battery bank for any length of time. Stay tuned for that truly scary project.

I’ll document how a flywheel is connected to the rest of the system and how it “sees” the surge demand on the inverter--after we install the thing here and take it for a spin when the new power shed is up and my monster methane generator is operational.


No tags yet.