HOW NEAR ARE FUEL-CELL MOTORCARS?
Reproduced from 'New Scientist' (Volume 24, No. 414 - 22nd October 1964) with kind permission of the authors and the publishers - The Certificated Engineer May 1965.
Traction is perhaps the most difficult application for fuel cells-particularly road traction where they have to compete with 70 years' development of the internal combustion engine. The massive investment in vehicles thus powered, and competition from other devices seeking a hold in the traction market superficially give cause to doubt whether fuel cells are worth considering at all. In time, the oil may become less plentiful and cheap, and rechargeable fuel cells, possibly using nuclear-synthesised fuel, will become mandatory; but this is not yet a pressing economic reason for developing such systems.
The fuel cell is simply a kind of battery; but one which instead of needing to be recharged, is continuously fed with fuel. This fuel the cell 'burns in electrochemical reactions, converting it directly into electricity. One of the earliest applications foreseen for the fuel cell was electric traction, and F. T. Bacon, who pioneered modern research in the field, demonstrated his hydrogen-burning 6-kW cell operating a fork-lift truck as long ago as 1959.
For traction purposes, fuel cells must compete for either or both of the following reasons: because they are cheaper, or because they have especially desirable properties. The results of past analyses for fuel-cell traction have not been very encouraging on either count. Technical development of fuel cells was insufficiently advanced and the estimates of future performance were not sufficiently accurate; there was too much generalisation-and also a tendency to copy the internal combustion vehicle, and there has not been coherent activity on all the problems of fuel-cell traction.
Our purpose here is not to reiterate the basic features of electric traction, nor to elaborate on the somewhat intangible advantages it can afford, such as silence and freedom from exhaust fumes. Rather we aim to show that several technical improvements, reported in isolation over the last 18 months, can greatly change one's picture of the economics of fuel-cell traction. What factors have changed is borne out by the current activity of many motor manufacturers in building experimental electric vehicles, though only a minority have as yet powered them with fuel cells.
There are three main aspects we must take into account. The first concerns the power-to-weight ratio of an electric power-plant. For some applications, a great weight is acceptable and even desirable, for instance in agricultural machines or for rail traction engines. But for road vehicles, especially motor cars, it seemed that to realise an acceptable performance with a load it was possible to have either a fuel cell or a motor-but, not both! Typical specific weights were, until recently, 100 Ib per kilowatt for the fuel cell and 15 Ib per kilowatt for the electric motor. Not surprisingly, then, the more urgent problem was to reduce the fuel cell's weight.
Volume is usually less of a problem in mall vehicles and for this reason, cells with the lightest materials of construction possible are preferable. Here is one reason why cells with a low operating temperature made largely of plastics may be chosen (Fig. 1). Cell powers of 10 kW per cu. ft. are attainable with specific weights of about 40 Ib per cu. ft.; in other words, we have a power-to-weight ratio of 4 Ib per kW. If a 25 per cent increase is allowed for auxiliaries, a powerplant weighing 5 Ib per kW is a realistic figure.
The figure is comparable with existing internal combustion engines, which lie in the range of 5-9 Ib per kW. (Experimental internal combustion engines, it is true, have weights as low as 1.3 Ib per kW; but some experimental fuel cells could provide power-plants weighing only 2 Ib per kW though there is little to be gained by going below 3 Ib per kW. The other factor in this connection is to note that, whereas large internal combustion engines have a better power-to-weight ratio than small ones, fuel cells show relatively little improvement as they are built upon a modular system and any gains relate only to the percentage of auxiliaries.
The second of the there aspects relates to the costs, both capital and operating. It is difficult to quote costs per kilowatt for fuel cells as they are being produced only in experimental quantities. Prices today range from £1 000 per kW downward. The market for fuel cells for special applications becomes significant at a figure of £100 per kW. And at £30 per kW fuel cells can compete with most portable powerplants. For comparison, mass-produced internal combustion engines are made for £3-5 per kW.
It has generally been assumed that fuel cell costs of £40 per kW are about the best that we can expect, but several manufacturers, developing simple construction techniques and using extremely cheap materials, have recently been able to cost fuel cells (without controls) at from £2-7 per kW, throwing completely new light on their utility.
In some ways, the development of fuel cells in competition with the internal combustion engine is rather a terrifying prospect but no doubt the challenge will be accepted. To this end major programmes are being undertaken throughout the world; in the U.S.A., notable programmes are those of the General Electric Company, Pratt and Whitney and Allis Chalmers, the latter having exhibited a variety of experimental fuel cell vehicles. ASEA in Sweden have an experimental vehicle, and Brown Boveri in Switzerland and Varta in Germany al 0 have big research efforts. In Britain, the Chloride Electric Company, Energy Conversion Limited and Shell Research Limited confirm the view that fuel cells have a big future-though their views on when a fuel cell motor car will become economically attractive still differ greatly.
The running costs of fuel cells have, too, compared unfavourably with internal combustion engines. Given the previously high cost of fuel cells, it was thought desirable to operate the fuel cell near to its maximum output at about 40 per cent thermal efficiency. Moreover, expensive fuels were considered. While it is not yet possible to transport hydrogen as such economically, many cells are still basically hydrogen-burning cells of the kind originally developed by Bacon. So recourse to reforming or partial oxidation of methanol (methyl alcohol) or hydrocarbons is undertaken.
The trend now is away from reformers and toward the direct oxidation of carbonaceous fuels. With methanol, it is relatively easy, but direct low-temperature hydrocarbon cells are much more experimental and expensive; indeed the answers to some of the technical problems may never be found. And so methanol emerges as the preferred fuel for the moment. Naturally, the costs per mile are as variable as for conventional transport, but the consensus is that overall savings of from 10-20 per cent of the cost of motoring would result. Hydrocarbons used directly could increase the figure to 30 per cent, eventually.
The third aspect concerns the unique properties of fuel-cell vehicles; these need not impinge on the user but are important about the overall economics. Here are some of them.
A vehicle is generally expected to tart instantly and thus there is some incentive to use cells which give at least some power when cold and which can 'bootstrap' themselves up to their normal operating temperature. For vehicles that are in use throughout the day, it is possible to use high-temperature cells without the cost of keeping them hot during the idle time becoming important. But even this load is likely to shorten the fuel cell's life, for it means that it is always in an 'active' state, and hence is always exposed to corrosive attack. However, lives of 2 000 hours for medium-temperature cells are now about average; and for low-temperature cells, a life over 10 000 hours (say, one year means that, in intermittent use, they should last the life of a vehicle.
Another factor relates to the speed control of the motor. Even in quite small vehicles, the current will be of the order of hundreds of amperes and the switchgear will still be expensive. Control based on the use of silicon-controlled rectifiers has been tried but the operating efficiency of conventional traction motors is adversely affected when running on pulses of current. New motor designs use semiconductors in place of the commutators combining control with the commutator's switching function.
The cost of semiconductor control has been high but could soon come down to a few shillings a kilowatt, given the demand. Nevertheless, there is room for the economy. There are several methods available but perhaps the most promising is the idea of fixing the fuel cell's output at a given current by feeding back a small pulsating current, typically 0.1 per cent of the output. The mechanism here is not yet well understood. It may also be possible to operate (the fuel cell in a pulsed mode suitable for a direct Connection to some of the motors we shall describe later.
Great strides have also been made in simplifying the controls for the fuel cell itself, brought about mainly by the choice of temperatures and pressures much nearer normal. The fuel flow is regulated quite simply by the power demand. Very often it is not economic to use the last few per cent of the fuel, which might be used for other purposes such as preheating the cold, fresh gases by direct combustion.
In general, the dynamics of each system have to be determined by experiment and calculation, and the stability regimes plotted. These restrictions reflect on the cost and economics of operation and as we implied earlier, the controls may cost about' 20 per cent at the few-kilowatt level and absorb about 3 per cent of the gross electrical output. This is a bi, improvement, especially when we take into account the reduction in cell costs.
Let us turn now to the mechanical parts of an electric car. Electric traction is more complicated than conventional internal combustion traction because two separate stages of conversion are necessary, chemical to electrical, then electrical to mechanical. In the internal combustion engine, we have chemical to thermal to mechanical conversions taking place in the same chamber. Even if the fuel cell weighed nothing, the conventional electric motor would give us a poorer power-to-weight ratio than the internal combustion engine vehicle.
We must, therefore, look to new and, particularly, to lighter ways of building a motor. The torque of a motor is proportional to its physical size: however, the horsepower depends on the torque multiplied by the speed, so a very small motor can produce large horsepower if it is run fast enough. Conventional traction motors run up to speeds of 3 000 rev/min, at a power-to-weight ratio of 10-20 Ib per h.p., but a high-speed motor rotating at 100 000 revs/min can have a specific weight of only 0.4 Ib per h.p. This specific weight, combined with an advanced fuel cell weighing some 5 lb per kilowatt, makes a traction system competitive with internal combustion engines.
The problems of running motors at such high speeds have been greatly over-stated; gas turbines run at 60 000 revs/min despite the high temperatures and corrosive conditions.
First, let us take the electrical problems. Owing to the high speed, the 'back emf' set up will be high, too. The windings must, therefore, consist of a very small number of turns, in most cases only one. Such windings can, therefore, consist of specially-shaped conductors containing cooling ducts, and can be manufactured from anodised aluminium, dispensing with all other insulation. What we now have is an electric engine which bears little relation to conventional electrical engineering. But the high speed complicates the problem of feeding current into the
rotating conductor so we must turn to an electronic system, using semiconductors.
There are two basic switching systems available. One involves taking a positional signal from the rotor and switching the stator windings in the correct sequence, and the second involves setting up a rotating magnetic field that is independent of the rotor in its phase but related to it in speed. They could be described as basically 'synchronous' and 'induction' motor techniques. Silicon-controlled rectifiers (SCRs) provide a convenient method for switching on the power to a particular winding, although special methods are required to switch it off again.
The induction motor has advantages here in mechanical simplicity, but requires more sophisticated control; for instance, a separate semiconductor device is necessary to switch off. However, a multi-wheeled vehicle can be driven off one pulse generator, with the advantage that while the wheels are not locked together they must rotate at the same speed within a few per cent. Any wheel with a tendency to spin automatically reduces the power.
All controllers need a pulse generator to provide the 'trigger' pulses at the correct times, giving a very flexible control system, with accelerating and braking controlled by electrical inputs not only from the accelerator pedal but from transducers too. These transducers give signals proportional to wheel acceleration and vehicle acceleration, which combine to give an anti-skid capability. Temperature signals from the motor will allow high overloads to be used for short periods to provide high ace 1 ration. Other inputs such as rate of turn, speed limits and proximity devices can be added very cheaply to the basic electronics.
The solid-state type of controller has this advantage over other controllers: it is cheap and easy to add refinements for safety and performance. The cost of SCRs, at present quite unrealistic, will come down very sharply because material costs are only a few shillings; made in numbers which the motor car market implies, it should cost no more than ten shillings for a 70-amp device.
For a motor to run at 100 000 revs/min the torque requirements are extremely modest-a few ft-Ib for a car traction motor. The main problem is gearing this torque to provide a larger torque at a lower speed. A total ratio of approximately 100: 1 is required. Also, bearings are required to support the rotor and provide reaction for the primary reduction drive. This requirement may, however, be avoided by using a balanced drive system, when the rotor bearings have no work to do other than to support the rotor.
Bearings can be eliminated by allowing the planetary members of the reduction drive to support the rotor, in which case a separate reduction drive will be needed at each end. Owing to the low driving torque at the high-speed end the primary reduction gears need have no teeth, hardened steel rolling surfaces being substituted. A gearbox only 1.25 inches in diameter will handle 12 h.p. at 100 000 revs/min.
Before the design of an electric vehicle can begin we must define our requirements. Consider first the family saloon. It requires good acceleration and high cruising speed for motorway use, comfortable seating for at least four adults, and considerable luggage capacity. These demands imply a rating of 70-90 brake horsepower for internal combustion engines but electric vehicles with their lower transmission losses and high overload capacity could achieve similar performance on about 40 brake horsepower. This power can be divided among two or four motors driving the appropriate wheels. Power supplies for a range of vehicles could be devised on a modular basis using a 12 h.p. motor and a 10 kW fuel cell.
For such a vehicle the unsprung weight must be kept to a minimum for good road holding so that a motor inside the wheel hub is not feasible. A much better method is to employ a toothed transmission belt inside the suspension link, providing both drive and universal action. No brake drums or discs are required as the feed-back' or 'regenerative' braking we mentioned earlier can be inherent in the system. As an emergency brake to comply with the law, the second means of brake operation could be a simple and very light 'one-shot' device. And the motor will provide magnetic lock-up for parking.
The conditions are rather different for the town carriage or 'runabout' because of the limited maximum speed required and the small duty cycle of the motor. Maximum power is required only for acceleration immediately followed by the deceleration and regenerative braking. Also, the efficiency of the runabout is unimportant about the saloon, so the fuel cell can be run near maximum power to economise in weight and space. It seems probable that the vehicle with a single driven front wheel will emerge as the most satisfactory electric drive as it can turn in its length. But it is practical to produce a short car on this principle only if we adopt an electric drive. Single rocking pedal control with kick down for exceeding present speed limits would eliminate all other engine instruments and controls except the steering wheel.
Finally, a look at the economics: assuming production of 200 000 fuel cell vehicles a year, a selling price of as little as £20 for the motor and control equipment might be achieved.
To summarise, then, our case for believing the fuel cell car is at hand we contend first that power-toweight ratios are now available for fuel cells which compare favourably with those for internal combustion engines. Secondly, projected costs for the fuel cell itself now leave a realistic margin for the electric motor, making the system as a whole competitive as a prime mover.
A third point is that a fuel cell's running costs could show a 10-20 per cent saving, given a sufficiently long life. Again, control systems for the fuel cell are close approaching the twin ideals of very high efficiency and negligible cost. And the last point is that compact, efficient and potentially cheap electric drives are now available.
These five points add up, in our view, to the probability that fuel-cell traction for private vehicles will be a reality very soon.