Flywheel development has been directed primarily at two areas: vehicular propulsion and storage as part of an electrical system. Flywheels are particularly compatible with electric and electric hybrid vehicles, where they can serve as a primary storage medium or as a surge power source enabling enhanced vehicle acceleration and battery life.  They are also compatible with the regenerative braking systems common to these vehicles.  They have been investigated for vehicular purposes since the late 1960's, when Oerlikon operated several buses in South Africa utilizing flywheels as primary propulsion.

The technology advanced significantly with the development of carbon composite materials for the rotor.  Replacing steel, these materials provide more strength with less weight, and greatly reduce the risks previously associated with rotor disintegration. Various prototypes employing composite rotors are presently in use, including one in a BMW demonstration vehicle.  Another is to be installed in an electric vehicle presently under test at Hanscom Air Force Base in Massachusetts.

Flywheels are also under development or in prototype demonstration for several modes of use in electric systems.  Team member TSI provided a prototype, production-size flywheel battery for telecommunications backup power to Bellcore, the baby Bell operating company's research arm at Bell Labs.  TSI participated with the regional Bell Telephone Operating Companies developing the Generic Requirements for backup telecommunications power units to be purchased for beta-sites in the next few months and for widespread deployment next year.  Prices for production quantities up to one million annually have been quoted.

TSI's recent application to flywheels of solid-lubricated hybrid ceramic bearings and sliding surfaces greatly reduced flywheel cost in comparison to designs which use magnetic bearings.  This breakthrough evolved from TSI technology and products developed over three decades.  These bearings are used not only in TSI's own flywheels, but as primary or backup bearings in wheels made by United Technologies Corporation, including the units in the BMW and Air Force vehicles described above.

Much flywheel research has focused on use in vehicles and spacecraft, leading to an emphasis on minimizing the size and weight of the unit.  Thus, much work has been directed at high- speed (up to 90,000 rpm) units, which theoretically require magnetic bearings for longevity.  This work has not been fully successful to date.

The size/weight constraints of the typical PV system are far less severe; while transportation is certainly a consideration for many sites, present components (e.g. batteries) are heavy and fairly large.  Since the flywheel rotor may be heavier for such systems, rotational speed may be less, and such units are farther along the development curve than high-speed wheels.  The TSI wheel turns at 30,000 rpm, and has demonstrated longevity and reliabililty using ceramic bearings.  With respect to bearings, containment, and rotor design, the TSI product represents current state of the art in this type of flywheel, and will be used as a comparison in much of this report.

COMPATIBILITY WITH REMOTE SITES.  The present and, to a greater extent, the projected characteristics of flywheels make them exceptionally well-suited to the remote locations that are typical of PV power systems.  In part, this is because they share many of the characteristics of the photovoltaic module.  These characteristics are discussed below.  Flywheels are just beginning the prototype deployment stage in remote sites; assuming experience is favorable, it is probable that they will assume a major storage role in remote power.

The lead-acid battery, although widely used, is a very poor fit to many remote power systems, especially when compared to the PV component.  Batteries require significant service, must be replaced (2-7-year intervals, depending on various factors), are expensive to transport and install, generate explosive gas, are easily damaged by misuse or poor maintenance, and use hazardous materials.

LONGEVITY.  Performance of prototype units at TSI facilities and computer modeling indicate a life expectancy for TSI flywheels, in most applications, in excess of 20 years.  One TSI unit has been in operation since the 1950's without relubrication.  Present fatigue design criteria are for 100,000 charge-discharge cycles, which - at one cycle/day - equates to 274-year life.  The calculated L10 life (the period over which 10% of units would be expected to fail) is 90 years.

In comparison, the batteries of a PV system require replacement at least three times, and as frequently as six times - depending on severity of cycling and thermal stresses - over what PV designers have considered the nominal lifespan of a PV system: 20 years.

As experience with deployed PV systems has accumulated, major module manufacturers have gained confidence in the ability of their products to exceed this period.  Siemens recently extended the warranty period on their large modules to 25 years; Solarex has introduced a new series of large modules, the GSX series, with a 30-year warranty.

Furthermore, the failure mechanisms of deployed PV modules are, in general, either infancy failures or long-term gradual output degradation, which is not a true "failure," although it may eventually cause inability to support the load.  Once through infancy, modules are extremely likely to function effectively for at least 30 years.  As the PV industry has matured, the cause of infancy failures (thermal cycling of interconnects, material incompatibility, water migration, etc.) have been identified and dealt with, and the failure rate has been dramatically reduced.

It's appropriate, therefore, to consider a lifespan longer than 20 years for PV systems and their components.  Most of the comparisons in this report will be based on a 30-year system life.

LACK OF MAINTENANCE.  For all intents and purposes, the TSI flywheel is maintenance-free as a result of its hybrid ceramic bearings and solid lubrication system.  Alternative bearing systems are either very expensive (magnetic bearings) or incompatible with operation in a vacuum.  Conventional bearings require periodic lubrication, typically with volatile petroleum-based lubricants which contaminate the vacuum.

In contrast, lead-acid batteries require inspection, watering, and terminal cleaning between two and twelve times a year, depending on cycling and thermal climate.

INSENSITIVE TO DEEP CYCLING.  Even the best deep-cycle batteries suffer from radically shortened lifetimes if they are cycled beyond their specified depth-of-discharge limit, or are cycled frequently. This forces system owners and designers to trade off reliability against cost, either by oversizing a battery to ensure continuous power to the load or by including a low-voltage disconnect circuit which protects the battery but sacrifices the load.

Complete discharge is catastrophic to a battery, but flywheels are unaffected by it, both structurally and in terms of longevity, regardless of discharge frequency.  This characteristic frees the system designer from the costly sizing/disconnect consideration above.

SURGE CAPABILITY.  Ability to deliver demand surges can be both a positive and a negative characteristic of a flywheel, depending on the design of the flywheel, the duration of the surge, and the criticality of the load's power requirement.  Although PV battery banks generally have the ability to meet demand surges, if the surge is sufficiently large it will shorten the life of the battery bank.  In systems which anticipate such surges, adding a fast- response flywheel or a capacitor to the power system may improve overall power quality and extend battery life.

TOLERANCE OF AMBIENT TEMPERATURE EXTREMES.  With solid lubricant, TSI flywheels are unaffected by any terrestrial ambient temperature, in terms of efficiency, longevity, and storage capacity.  In contrast, batteries are compromised by and can be destroyed by temperature extremes.

Battery capacity and life is typically optimized for, and rated at, 25 C operating temperature, and any significant variation from that temperature compromises operating life or storage capacity. Manufacturers' literature indicates that a battery's rated life is halved if it's operated continuously at 35 C, and that batteries fail very quickly if operated above 50 C.  Battery capacity is approximately halved (from rating) at 0 C.

A discharged battery may be destroyed by subfreezing temperatures. At 80% depth-of-discharge, battery electrolyte freezes, with permanent damage to plates and case, at -10 C; a fully discharged battery freezes only slightly below 0 C.

LACK OF ENVIRONMENTAL IMPACT.  Batteries use poisonous materials - among them lead, sulfuric acid, and cadmium - and thus present an environmental threat in manufacture, use and disposal at end-of- life.  The materials used in flywheels present no environmental threat.

Safety concerns about containment of flywheel components in case of catastrophic failure have been addressed by several projects and studies, the most recent being a DARPA project which included extensive analysis in conjunction with NREL.

The containment system of the TSI flywheel is based on multiple burst tests (in all of which the rotor was completely contained) and advanced computerized testing.  The high kinetic energy of a bursting flywheel is expressed almost entirely in the form of high circumferential rubbing speeds of fragments centrifugally loading the containment housing.  TSI's unique containment design, tested on the DARPA project, is patent pending, and is currently undergoing additional proof testing.

In planned deployment of the TSI product as a battery for communications repeaters, the unit will be buried below ground, This technique is applicable to many PV applications, and provides redundant protection in case of flywheel failure.  Although sometimes employed with chemical battery systems, primarily to dampen temperature swings, it is more practical for flywheel systems since they require no periodic inspection or servicing.

A spreadsheet was used to analyze the costs over time of various system configurations.  This analysis provided the cost comparisons which are referenced in other sections of this report.

We analyzed three comparable systems using the lowest projected costs for a nominal 1 kW PV system.  The primary differences are in the first costs of the storage systems and in their placement and maintenance.  It was assumed that the power electronics for all of the systems would need replacement on a ten-year cycle.  The chemical batteries would need to be replaced on a seven-year cycle, and the flywheel batteries would be maintained at five-year intervals with an annual inspection.  Chemical batteries were also assumed to be maintained once per year, A thirty-year system life was assumed, corresponding with the nominal design life of the PV modules.

A net present value (NPV) method was used to compute the lifecycle costs of the three systems.  In this method, the time value of money is accounted for by discounting future cash flows at a fixed discount rate.  This rate is the owner's after-tax cost of capital, and will vary with the type of entity owning the equipment, the tax status of the owner, the credit worthiness of the owner, the risk preferences of the owner, and the capital market situation in the country where installed.  A typical rate at the present time for U.S. corporations is 10%, so this rate was used in the analysis. Higher rates reduce the value of future costs and therefore make systems with lower initial costs but higher operating costs more favorable.  Lower rates have the opposite effect - making future cost streams relatively more significant.

The net present values of the flywheel/EC system and the lead-acid battery system with a 10% discount rate are close enough to be comparable, given the uncertainty over the future costs of the systems.  The analysis clearly shows that the new systems have the potential to compete effectively with chemical storage in PV applications and, as discussed in the Market Analysis Section, to extend the range of PV applications.