Polymath A (mostly) technical weblog for Archivale.com

December 2, 2009

Hovercraft Technology Brief

Filed under: Ground Effect — piolenc @ 6:53 pm

Hovercraft Technology Brief

Technology Brief—Amphibious Hovercraft

Author: F. Marc de Piolenc, Consultant (piolenc@archivale.com)

What is a Hovercraft?

A hovercraft is a surface-skimming aircraft sustained by a cushion of air below it that is maintained at a pressure slightly higher than atmospheric—gauge pressures run from 15 pounds per square foot for small machines to 140 or so for the largest built so far. The cushion of air is typically confined by a fabric “skirt” around the periphery of the craft. Because the skirt does allow air to escape from the cushion (a small air gap between the skirt and the ground is an essential feature, reducing resistance to the craft’s motion to a minimum), it must be maintained by a high-volume, low pressure blower or “lift fan.” Propulsion is usually provided by air propellers,
open or shrouded, though some vehicles use internal fans exhausting through nozzles or skirt vents.

The typical, skirted hovercraft described above is called an amphibious hovercraft in the title to distinguish it from the less common sidewall hovercraft or surface-effect ship (SES), in which the cushion is confined at the sides by rigid sidewalls placed catamaran-fashion, and at bow and stern by flexible seals. Such craft can only operate over water, and are often propelled by conventional marine propellers or water jets.

Capabilities and Limitations

Skirted hovercraft can make a smooth transition from water to beach or boat ramp. They can travel over any reasonably level terrain, including ground so soft that no wheeled or tracked vehicle can pass. Their very low ground pressure (compare hovercraft cushion pressures of typically 15 pounds per square foot to typical wheeled-vehicle tire pressures of 30 pounds per square inch) makes them a natural choice for providing safe and convenient access to environmentally sensitive areas. Their impact on such sensitive terrain as tundra is in fact less than that of a human footfall. With draft measured in inches, no propeller in the water and a flexible skirt, hovercraft can pass over fishnets without damaging them, and over logjams, reeds and the like without sustaining damage. They are excellent water-rescue vehicles, as the soft skirt can safely over-run a swimmer without injuring him, and there is no propeller in the water.

Although sensitive to sloping terrain on land, hovercraft have successfully negotiated rapids on such challenging rivers as the Kali Gandaki, in both the upstream and downstream directions.


In more mundane applications like water taxi and passenger and vehicle ferry service, hovercraft are unique in allowing passengers to disembark dryshod from a perfectly immobile platform onto solid ground. Because of their reduced friction drag, they can economically achieve higher speeds than conventional surface craft. Amphibious capability allows them to operate from beaches and sandbanks where port facilities are unavailable, silted-up, overcrowded or inaccessible.

Having extolled their virtues, it is only fair to mention their limitations.
On land, hovercraft have only very limited ability to handle slopes—even crowned highways sometimes give trouble. Deep, coarse gravel beds, which allow the cushion air to leak away faster than the lift fan can renew it, will halt a hovercraft, as will very deep, loosely matted vegetation. Because the hovercraft has no ground traction, steering it requires special skills not applicable to speedboats, automobiles or ATVs. The first generation of small hovercraft were often powered by two-stroke engines, giving hovercraft a reputation for being inherently noisy which has made them unwelcome in some areas despite their excellent safety record; this prejudice is being slowly overcome by the current generation of mostly 4-stroke powered machines. Reducing propeller tip speeds has also lowered noise.

Recreational Use and Amateur Construction

The versatility and relatively low cost of the amphibious hovercraft has attracted interest from recreational users. In Europe and in North America there are thousands of owner/builders who have assembled their vehicles from commercially available plans or kits. This provides a pool of trained pilots from which the relatively small number of commercial hovercraft pilots can easily be drawn.

Legal and Regulatory Status

Although technically aircraft, hovercraft are usually regulated as marine vehicles. In the USA for example, recreational use is subject to the same Coast Guard regulations and inspections as recreational boating. Commercial hovercraft are classified like ships by the marine classification societies, some of which have developed expertise in this particular class of vehicle. This has the advantage that obtaining and maintaining regulatory approval for a hovercraft is cheaper than for a conventional aircraft of the same capacity. There is one apparent disadvantage by comparison with aircraft certification, which is that in marine practice no “type certificate” or “production certificate” exists—each commercial ship or boat is classified as an individual entity. This practice does, however, facilitate incremental improvement in hovercraft design without incurring the huge paperwork burden involved in recertifying a modified commercial aircraft.


Size Range and Applications

The light hovercraft market resembles the light aircraft market in that
small-hovercraft manufacturers have not done well, the few that remain operating essentially as cottage industries. As noted, the market for plans and kits is thriving, and many light hovercraft are built and successfully operated by their owners.

At the opposite end of the hovercraft spectrum, large commercial hovercraft operating over short sea routes in the developed world have been gradually withdrawn from service, the victims of rising fuel and labor costs. These machines were designed in the Sixties, when the only practical powerplants for them were open-cycle gas turbines—a circumstance which barred them from competition for short ferry routes in the Third World. In developed countries, where the ready availability of conventional port facilities allows many alternatives, they have lost out to fast catamaran ferries. The will to upgrade their powerplants for economical service in the archipelagic Third World seems to be lacking, though the technical possibility  exists. This writer is convinced that future applications of very large hovercraft will center on less developed countries with limited port  facilities. Machines to serve this market will be powered by diesel or CoDaG (combined diesel and gas turbine) plants, and will take  advantage of relatively low labor costs for skirt maintenance.

In the middle of the hovercraft spectrum is a lively market that centers on military and rescue applications, served by a small but stable group of manufacturers based in the industrialized world. Military missions include landing small commando or reconnaissance units on beaches or riverbanks, river and swamp patrol, fisheries surveillance and antipiracy duties. In rescue work, hovercraft have excelled in areas where rescues have to be carried out over combinations of ice, snow, mud and water. Another application served by this group is oil-platform service, especially in cold climates where access is over ice during much of the year. Unfortunately, the cost of manufacturing medium-payload hovercraft in industrialized countries is limiting their application in the Third World, where they could be put to good use if they were affordable.

In short, opportunities exist for manufacturers based in low-labor-cost areas to create and serve local markets. With further effort, they could penetrate the market in industrialized countries, undercutting the prices of manufacturers based there. If Russia persists in taxing its own exports of machinery, it may even be possible to build large hovercraft profitably in the Third World.


Configurations of Existing Machines

In general, hovercraft are categorized by their propulsion and lift arrangements, and by the types of skirt systems that they use. Means used for control and stabilization are a further discriminator. We will tackle them in that order.

Propulsion and Lift Arrangements

These are distinguished by whether a single engine is used for both propulsion and lift, or each function is assumed by a separate engine. When only one engine is used, it may drive a single fan that provides both propulsion and lift, or two separate fans. Thus we have:

Single engine, single fan: this is typical of racing hovercraft and first-generation sports machines. These typically use ducted fans for propulsion. There is a “splitter” plate in the duct just downstream of the fan, which diverts about 1/3 of the duct flow to feed the cushion. The obvious disadvantage of these machines is that they cannot hover on-cushion.

Single engine, dual fan: this configuration is adopted by the very popular SEVTEC series of amateur-built hovercraft (though some builders have converted their machines to two engine, two fan configuration). These have somewhat better control than the single engine, single fan machines, but still require some skill to park. The SEVTEC machines have ports in the front skirt that allow braking and rapid dumping of skirt air, partly compensating for the thrust of the propeller at low speeds and allowing the machine to be brought
to a halt with a fair degree of control.

Two engines, two fans: This is the configuration adopted by another long-established hovercraft plans publisher, Universal  Hovercraft. The advantage is that lift and propulsion are completely independent. Unfortunately, there are now two engines to maintain, and the machine is immobilized if even one engine fails, so there is no additional security in having two powerplants on board.

Integrated or integral arrangements: Several attempts have been made to use a “buried” axial- or radial-flow fan or fans feeding a plenum that is common to both the lift and the propulsion system. In these craft, propulsion air escapes through a controllable nozzle at the rear of the craft, while a continuous stream of air is drawn into the cushion. Several amateur-built craft have had this arrangement, as had the CushionCraft series built in Britain. More recent attempts to revive this configuration include the Australian Hovershuttle and Don Sainty’s Condor 2000. For a number of reasons which will be discussed in greater detail in connection with control, this is a very desirable configuration. Suffice it to say here that vehicles with this type of machinery have a lower profile than any other type and can also be made very quiet because all propulsive and cushion flow is internal. There are difficulties, of which the present author has made an extensive study, discussed later in connection with the proposed work.


Skirt Systems

While the heroic era of hovercraft development, extending from the Fifties through the mid-Seventies, saw many skirt systems tried, those listed below are the only ones that survive.The most popular skirt system in light hovercraft is the simple “ bag” (“loop” in British parlance) skirt, and variations on it. It is the second lightest and most economical (after the simple cone skirt) and also second in the amount of labor required for its assembly. Unlike the cone skirt, it can be fitted to any planform.

The “open” bag, a variation on the bag skirt originating with this writer, has the advantage of not requiring internal rigging; that is, it is attached to the body of the hovercraft only at its top edge. It is not as versatile as the plain bag skirt, being applicable only to highly curved sections of the hovercraft’s planform. It is mentioned because it may offer advantages when applied to the bow section of an otherwise conventional bag skirt.


The plain “finger” (“segment” in British parlance) skirt is common in racing hovercraft and in derivative single engine, single fan, splitter-equipped hovercraft. It is reputed to have better sealing  characteristics than a plain bag skirt, especially at speed over rough ground. It requires much more material and many more fasteners and assembly operations for a given perimeter than a bag skirt. Maintenance is also expensive, as the individual fingers are easily torn loose. The number of different skirt parts is high, because  different types of fingers must be provided for forward, side and aft portions of the skirt planform.

The bag and finger skirts are often found in combination, especially in British hovercraft and British-influenced designs. In this system, fingers are attached to the bottom of a bag skirt and fed through holes in the bag.The actual ground cushion seal is thus provided by the finger component, while the bag portion provides additional cushion depth. This arrangement has all the maintenance difficulties of the pure finger type.

The “curtain” skirt appears to be an invention of SEVTEC principal Barry Palmer; to the best of this writer’s knowledge it is not found on any other designer’s craft. It is included here because of the ubiquity of Palmer’s designs. This skirt type, used in the bow section and transverse partition of the Fan-Tastic and SEVTEC machines, has no internal rigging. Unlike every other skirt system known to this writer, it relies on the resistance of an inflated membrane in shear to retain its shape. While very economical of material and easy to build, it has the disadvantage of “raising its skirts” like a can-can dancer when it is momentarily depressurized – say, in running over rapids. It is used only at the bow of Palmer’s designs, the rest of the skirt being a conventional bag, and probably could not be used alone unless struts were provided to take lower-corner tensions.

Cone skirts are best known through the work of French engineer Jean Bertin, though a few North American designers have used them as well. These are simply truncated cones tapering downward at a half-angle of about eight degrees. Mechanically they are the simplest and easiest to build of all skirts. Their disadvantages are that they lend themselves only to circular planforms and have “footprint” areas smaller than the planform areas that they occupy. As circular planforms are not the most desirable, hovercraft equipped with cone skirts tend to use multiple skirts, each with its own feeder duct. To recover some of the lost area between cone skirts and that lost to taper, these skirt systems are sometimes surrounded by another, peripheral skirt, which might itself be a cone or connected conic segments, or a bag skirt. These practical arrangements nullify the potential economy of the cone skirt. Another problem with cones is their tendency to scoop water or catch on projecting obstacles. Multi-cone skirts do have a high degree of stability, both static and dynamic. Unfortunately they are difficult to inspect and require complex skirt feeder systems.

Air curtains were the original means of maintaining a cushion in the earliest modern hovercraft. Instead of a fabric skirt, these machines used a thin peripheral jet of high-velocity air, directed downward and inward, to confine the cushion and maintain the necessary pressure discontinuity between the cushion and the atmosphere. With the advent of fabric skirts these systems were mostly abandoned because of their very high power requirements. They are still worth mentioning because they can cross terrain such as coarse gravel beds where a fabric skirt is useless.

Control and Stabilization Systems

It is a truism of hovercraft operation that hovercraft don’t go where they are pointed; they go where they are pushed. They resemble spacecraft in tending to continue in a given direction regardless of heading, and in requiring the application of thrust for any change in course. Thus the question of control systems design is a particularly important one, and is inseparable from propulsion design.

The most common means of control in small and medium hovercraft is a cascade of rudders immersed in the slipstream of a conventional open or ducted propeller. In some machines, a horizontal surface is added to provide dynamic pitch trim, and still more rarely roll control is provided. A smaller number of machines is provided with reverse thrust arrangements. These range from the simple though bulky “buckets” of some Neoteric craft to the complex, but much more compact reverser cascades of the CanAir machines.

Smaller still in number are machines provided with the means of generating side thrust and turning moments independent of forward thrust. Means used include “puff ports” fed from the skirt plenum; more satisfactory ones use air diverted from a propulsive fan.

The most sophisticated light-hovercraft controls developed so far are those of the River Rover series designed by T.J.R. (Tim) Longley and those embodied in the CanAir machines by Ron Fishlock.

The River Rovers used two ducted fans, each with a horizontal control surface in the exit tract acting as an “elevon.” Differential deflection provided roll control, primarily to increase drag on the inside of a turn and enhance skirt venting to the outside. At larger deflections the elevon on the inside duct would partly close it off, spoiling thrust on that side and producing differential thrust. Conventional rudders were included as well, as they were found to be superior to the elevons in a crosswind.

Fishlock’s designs use a single, larger duct with a system of cascades in the sidewalls. The duct can be completely closed off aft and the sidewall cascades opened individually to pivot the machine in place or together for reverse thrust.

Design Considerations

The design of light hovercraft resembles the design of light aircraft in that it is possible, with minimal skill and training, to build a craft that works—gets up on cushion and advances under its own power with some degree of control. But as is the case in aircraft design, there is a gap between the efficiency of just barely flyable machines and the best achievable performance. What is more, much design work is being done by persons with minimal engineering background. This results in “copycat” designs which converge to a few accepted configurations, neglecting the many other possibilities that exist. Thus there remain, despite the mature state of the technology, many opportunities for innovation and improvement.


Lift System: As with any powered-lift aircraft, it is important to reduce the power required for lift to an absolute minimum. How one goes about this in hovercraft, however, differs markedly from the practice for other forms of aircraft. In VTOL aircraft and helicopters, minimizing lift power means achieving the highest possible mass flow through the lift system; in hovercraft we try to achieve the opposite. In a hovercraft, the power required by the lift system increases with mass flow. So, too, does the momentum drag of the lift airflow, which increases the power required for propulsion. Thus the power penalty for excess mass flow increases with the hovercraft’s speed, unlike drag due to lift in conventional aircraft which drops rapidly as forward speed increases.

Increasing the efficiency of the lift system means minimizing mass flow and reducing losses in the fan, ductwork and plenum. Of the two, the first has priority because it affects both lift power and through momentum drag) the cruise power. Despite this, reducing mass flow has had relatively little attention because of the apparent difficulty of achieving a favorable result. There is in fact essentially only one way of reducing lift system mass flow, all other things remaining equal (as they usually must), and that is to reduce the effective leakage area. This in turn means reducing skirt perimeter and/or decreasing the effective nozzle coefficient of the skirt-to-ground gap.

Reducing skirt perimeter for a given area forces the designer to adopt planforms that are circular or nearly so, to take advantage of the circle’s minimum ratio of perimeter to area. While very desirable from a construction standpoint, a circular planform is usually not optimum for operation, so the designer has very little latitude here.

LIttle attention seems to have been paid to reducing the orifice coefficient of the skirt leakage gap, but there would seem to be room for improvement here. While the bag skirt is popular because of its ease of construction, it forms what amounts to a half-venturi nozzle with the ground, giving a very high orifice coefficient and a flow very close to the theoretical maximum for the nominal leakage area. The finger skirt appears to have the advantage here, as its gap is closer to being a sharp-edged orifice, at least on one side. This writer suspects, however, that leakage between fingers more than compensates for this apparent advantage in actual operation.

It would be very desirable to find some way to modify the bag skirt so as to retain its advantages while reducing its leakage rate. This writer has considered one possible means, namely the use of an “open” bag over the bow portion of the hovercraft planform. Such a bag ends at the ground contact line, forming a more sharp-edged orifice than the conventional skirt. It is not universally applicable, and it does impose some constraints of its own. As a practical matter it would probably be usable only at the bow, and the bow would likely have to be semicircular in planform to keep the patterning of the skirt reasonably simple, though in theory any convex-curved section of the planform could be accommodated.

Another possibility that is worth investigating is modifying the ground-contact area of an otherwise conventional bag skirt to approximate a labyrinth seal. As originally conceived, labyrinth seals required both sealing surfaces to have protruding and interlocking “fingers,” but subsequent research has shown that labyrinth seals can be effective even if they are one-sided, one of the sealing surfaces remaining smooth. This opens up the possibility of using strips on the bottom of the bag skirt whose spacing and depth would allow them to act as a labyrinth seal, with a considerable reduction in skirt leakage. So far as this writer can determine, this has never been tried.

Improving the efficiency of the fan and ductwork will be discussed in connection with stability and control.

Lift/Propulsion Integration: Hovecraft use a substantial part of their total installed power for lift. In craft with completely separate lift and propulsion systems, considerable excess power must be provided for lift, both to ensure reasonable life for the lift engine and to allow the fan to be overdriven when extra lift air is needed. This is especially important for machines too large to man-haul off a deep gravel bed or slashpile. There is obvious merit in a scheme that allows considerable excess power to be used by the lift system in such an emergency. On the other hand, efficiency at cruise requires that lift power be held to the minimum required by conditions, and that the powerplant furnishing lift power be running
in the most efficient part of its operating map. It is obviously impossible to fulfill both conditions with a separate lift engine.

The single engine, two fan system is somewhat better in that a single,
larger engine is usually more efficient than two smaller ones—not to mention cheaper—and it is at least theoretically possible, by changing the gearing ratio between the propulsion and lift drives, to overspeed the lift fan and transfer more power to it than it would otherwise receive. The writer is unaware of any machine actually embodying this feature, however, though one at least is in the planning stages. The mechanical complications involved are not trivial, especially as ratio changes must be made under load.


From an operational standpoint, the ideal system would be a single fan (or set of fans operating in parallel) feeding a single plenum from which both lift air and propulsive airflow would be drawn. In this scheme, the engine would operate at constant speed and load transfer would occur pneumatically. That is, as the propulsive nozzle opened, the load on the fan would increase and the constant-speed governor would open the throttle to maintain engine speed. When the nozzle closed, the engine would be throttled back automatically to prevent overspeeding. The advantage of this scheme is that, so far as the fan and engine are concerned, it doesn’t matter whether the additional load occurs because additional lift air is needed or because the vehicle is accelerating; the air simply goes where it is needed, and the engine and fan maintain pressure in the plenum. Such a system would be very easy to control, especially if the fan or fans were centrifugal-flow rather than axial-flow. The pilot would simply open or close the propulsive nozzles to get the thrust and control forces that he needed, while the engine governor would ensure that the plenum pressure remained within narrow limits.

The author has given the above scheme a great deal of study and has learned from calculation that, in small machines, it cannot be implemented exactly as given above. The reason is that the ideal propulsion plenum pressure is higher than the pressure needed to feed the skirt. This holds true until the vehicle becomes large enough to have a sufficiently high skirt pressure to give a good match between propulsion and lift requirements.

In smaller machines, therefore, two fans and two plenums are required, but the constant-speed control system can be maintained and pneumatic load transfer is still possible. Supposing that the vehicle encounters a deep, coarse gravel bed during a river cruise and needs extra cushion air flow to get off. This is accomplished by closing the propulsion nozzle down to minimum flow to minimize the propulsion power, then overriding the speed control to overspeed both the lift fan and propulsive fan shafts. As both the output volume of the fan and the power absorbed vary as the cube of the speed, a small speed change can result in a considerable power transfer, all without changing gearing ratios. A further advantage of this system is that the normal lift fan speed and power can be chosen for best efficiency, knowing that a considerable momentary increase in output is possible. In the usual case, a considerable margin is included in dimensioning the fan to provide for errors in calculation and operational vicissitudes.

Several hovercraft, past and current, have approached this configuration. So far as is known, however, none has implemented it completely. Perhaps the closest designs to realizing a complete implementation thus far have been the Hovershuttle and the Condor 2000. The Hovershuttle used a single, axial flow fan to feed both the lift system and the propulsion nozzle. A serious problem with this machine was caused by the use of an axial-flow fan. These become unstable at low flow rates, where the slope of the delivery curve (pressure vs. rate of flow) becomes positive; that is, decreasing
delivery volume results in decreasing pressure, which in turn further reduces delivery. Because of this, the Hovershuttle had forward-facing ports that opened when the machine needed to hover or progress at low speed, allowing a flow rate high enough to ensure stability to be maintained while canceling the thrust of the propulsive nozzle. This had the unfortunate effect of requiring a high power setting even when the machine was hovering on-cushion. The Condor 2000 solves the problem by separating the lift and propulsion systems and by driving both hydraulically, the centrifugal lift fans at constant speed and the axial propulsion fan at varying speeds.


Implementing a true constant-speed system requires that the propulsive fan be sufficiently stable at low flow rates to allow load control by only opening and closing the propulsion nozzle. Centrifugal-flow fans already exist that come close to meeting this criterion, while every attempt to accomplish this with axial-flow fans has apparently failed. It is not clear why workers persist in using the axial-flow fan for this purpose. Achieving a completely stable flow-modulated fan output at constant speed, right down to zero flow, seems possible with a hybrid or intermediate type of fan—called mixed-flow. This type is essentially a centrifugal flow fan with the blading extended upstream so that the incoming air first meets the blading traveling axially rather than radially. From what literature the author has been able to consult on this relatively unexplored type, it is possible to achieve a negative slope to the delivery curve—i.e. stable operation—down to zero flow. Another benefit of this type of fan is that it allows higher specific speeds than pure centrifugal types, and lower specific diameters. This in turn means that mixed-flow fans should be more compact than centrifugal flow fans of equal delivery and efficiency.

Proposed Work

The objectives of the proposed work are:

  • To develop scalable internal flow lift, propulsion and control systems
    for hovercraft incorporating constant shaft speeds, with overspeed capability for “dash” propulsion or emergency augmentation of lift airflow.

  • To develop suitable fans for feeding such a system, incorporating whatever refinements are needed to allow stable propulsive fan operation down to zero flow.

  • To develop a range of hovercraft embodying the above advances, first for development and later for customer missions.

The above objectives are to be achieved as follows:

  • First, by rig testing of fans, lift systems, propulsion systems and controls.

  • Next, by the construction of a small hovercraft for vehicle-based development work, and for demonstrations and VIP transport once successful.

  • Thereafter, by design of a range of hovercraft suited to defined mission requirements submitted by customers or derived from market study.

Relevant books:

Elsley & Devereux: Hovercraft Design and Construction

Trillo: Marine Hovercraft Technology

Mantle: Air Cushion Craft Development

Relevant category on Archivale.com: hovercraft

Powered by WordPress