Basics of flywheel UPSs

Today's electrical power systems are plagued by a variety of disturbances ranging from short-duration sags, swells, and transients to long-term interruptions.


Today's electrical power systems are plagued by a variety of disturbances ranging from short-duration sags, swells, and transients to long-term interruptions. These problems can be caused by motorized equipment within the plant starting and stopping, internal or external faults, or thunderstorm and lightning effects on the utility lines. In addition, power utilities in many areas of the country are operating at or near capacity, resulting in frequent "load shedding" (rolling blackouts) (Fig. 1).

Studies have shown that approximately 85% of all power events are voltage sags lasting less than 2 sec, with many under 1 sec. A study by Bell Labs found that 87% of downtime is caused by disturbances lasting no more than 0.5 sec.

Aggravating the problem is extensive use of electronic drives on large motors, preventing them from helping to maintain the bus voltage during disruptions. The motor is effectively decoupled from the system by drive electronics. Thus, industrial power systems today are much more vulnerable to power disturbances than they were in the past. Although electronic systems such as computers are highly susceptible, even electromagnetic contactors can be affected. A disturbance of only a few milliseconds duration can cause the shutdown of one or more individual machines, or an entire plant through a chain-reaction. The resulting cost to a company from downtime, product loss, and equipment damage can be hundreds of thousands of dollars.

Traditional UPS systems

To avoid these problems and ensure uninterrupted plant operation, more and more facilities are installing some form of uninterruptible power supply (UPS). Long-term needs are typically served by engine-generators. While they can provide power for an unlimited time as long as fuel is available, they require up to several minutes to start and synchronize, making them ineffective in handling the most prevalent short-duration disturbances.

For decades, electrochemical battery UPS systems have been used to handle short disruptions in power. These consist of large banks of rechargeable batteries, a rectifier/charger which provides dc charging current to the batteries when power is available, and inverter electronics that convert the dc voltage to ac to feed the critical load bus when needed. These systems are online and switch to battery power in a matter of milliseconds, and can provide power for up to several hours, which is known as ride-through. While battery systems can provide relatively large amounts of energy, they suffer from several disadvantages including large space requirements, high maintenance, environmental issues, and a limited charge/discharge cycle life.

So is there an option? It would appear that today the answer is yes.

New technology or old idea?

Since the earliest pottery wheels, people have known about the ability of a flywheel to store energy. This fact has been put to use in many machines, from the first steam engine to modern engines of all types, where flywheels provide smooth shaft rotation by eliminating engine pulsation. Sir Isaac Newton formalized this effect in his first law of motion, which states that a mass will tend to maintain its velocity unless acted upon by an external force.

This inertia principle has been used since the 1970s in rotary UPS systems, which consist of a conventional motor-generator with a flywheel installed on the shaft. In operation, the motor takes power from the supply bus, and the generator is always supplying power to the load. During brief interruptions, flywheel energy is used to maintain rotation of the generator and provide uninterrupted power. The rotational speed and generator frequency decay rapidly due to bearing and air friction and the loss of kinetic energy as it is converted to electrical energy. As a result, these systems can recover only a relatively small percentage of the available flywheel energy, and are limited to short duration ride-through. They can, however, supply large amounts of power, making them useful for handling short power events on large systems.

The new breed

Today there is a new generation of flywheel UPS systems, known by various names including kinetic battery, electromechanical battery (EMB), or flywheel energy storage system (FESS). They use high-speed flywheels rotating on extremely low-friction bearings in a near-perfect vacuum. They can store large amounts of energy and then deliver it within a few milliseconds when needed.

One drawback to flywheels is the energy loss associated with keeping the wheel spinning. Placing the units in a vacuum and using special bearings help, but the system will have continuous energy losses of approximately 1%. Depending on the kW rating and the user's energy cost, this expense should be taken into consideration. It may be greater than the cost of maintaining a charge in a conventional UPS battery system.

How they work

For the past decade or so, a number of industry consortiums, government agencies, and universities have been developing state-of-the-art flywheel technologies and systems. Two technologies have emerged from the laboratory and are commercially available today. One uses a steel flywheel, the other a composite flywheel.

Steel flywheels have limited energy storage capacities, due to their mass and structural considerations, which restrict them to rotational speeds under 10,000 rpm. Because of their lower speeds, they are considered by some to be safer, and they can use conventional bearings. Steel flywheels deliver power from several seconds to several minutes of discharge time, which is probably their ultimate limit.

Much recent activity is in composite flywheels (Fig. 2). These flywheels are constructed of various arrangements of carbon and glass fibers, and are considerably lighter than steel flywheels. Typical operation is at rotational speeds up to 50,000 rpm, although the capability exists for speeds close to 100,000 rpm. Discharge times from several minutes to several hours are now available at low power levels. The flywheel is only a portion of the total system cost, becoming more dominant in low-power, long discharge time applications. In systems that provide high power for short discharge times, less expensive flywheels can be used and the power electronics cost dominates.

Both types of flywheels are completely enclosed within a containment vessel (Fig. 3) that serves two purposes. First, it provides a vacuum chamber that eliminates air friction with the flywheel. Typical vacuum levels as low as 10

The second function of the containment vessel is to provide protection against flying debris in the event of a catastrophic failure of the flywheel. This protection is generally accomplished in combination with several additional barriers, since the containment vessel alone may not have sufficient strength. Although not required, the entire containment assembly could be buried underground, with the electronics above, both for safety reasons and to provide a reduced footprint.

Many of the systems today, especially those operating at lower rotational speeds, use low-friction ceramic ball bearings. For higher speed operation, either passive magnet or electromagnetic bearings provide lower friction, noncontact support. Various hybrid bearing designs are also being developed which combine the advantages of several of these technologies, with the objective of balancing performance, reliability, and cost.

The motor and generator in most flywheel systems are rotating field ac devices, with magnets attached to or embedded in the flywheel, and stationary coils surrounding it. In many cases, the motor and generator share common components, resulting in a very compact package. Although some motor-generator designs connect directly to the ac bus, the most common arrangement uses variable frequency drive circuitry either from the ac supply or from the dc bus, to power the motor, and rectifiers to produce dc voltage from the variable ac generator output. DC operating voltages allow the flywheel system to directly interface with, or replace, batteries.

When ac power is required, the rectified dc voltage from the generator is fed to an inverter, which produces constant, utility-grade ac, regardless of flywheel speed. The flywheel UPS uses the same inverter technology as a battery system. Some flywheel systems are being marketed without the inverter to compete directly with batteries. Traditionally, power inverters have used silicon control rectifiers (SCRs) as the switching devices, but they switch relatively slowly and require significant current and external commutation signals to operate. In newer designs, SCRs are being replaced by gate turn off-thyristers (GTOs), which are twice as fast and self-commutating, but still require significant operating current. The most promising switching device for immediate inverter designs is the insulated gate bipolar transistor (IGBT), which has much faster switching speeds and lower current consumption.

Sophisticated control systems monitor all aspects of operation of the UPS. Flywheel sensors supply information about its rotation, vibration, bearing temperature, and other parameters that provide an early warning of any pending failure.


Flywheel UPS systems can be used in several different configurations to meet the needs of a particular application. For a given energy storage capacity, there is a trade-off between power and discharge time. Both need to be adequate to do the job.