In this first video, I managed to capture the performance of Owen’s Revolution running with a strong tailwind. The lake is quite shallow and, at this section, the wind is beginning to create 300mm waves at around 3m pitch. Extended segment skirts are not know to adapt much to the surface, but you can see they were doing a pretty passable job.

I was being shaken around so much I was surprised that I got any usable vision at all. Unfortunately my finger kept creeping over the lens.

This video is later in the day. We fitted four prototype segments around the front and took the craft out again, looking for those same conditions. Unfortunately the wind had died down a little and the waves weren’t as high. The ground contact point of these segments are much closer to the limit line but they appeared to behave quite normally. We are looking to find a skirt shape with just a bit more responsiveness (up and down float) but I think we can go further still.

1. reduce the spray coming over the bow
2. contain the skirt on the sides to allow loading on to the trailer .

The first design showed that it is possible to make a bag skirt to replace extended segments. While Peter’s preferred outcome was a bag skirt, we learned that there were features of the pure bag that were undesirable and a hybrid skirt would be the outcome. He decided that the skirt should have extended segments around the bow and and that the ground contact for

So we went back to the drawing board.

I started by looking at the various forms of skirts that we could put on. I used Solidworks sketches as the design tool because I could create a graphical representation of the forces in the skirt profile. That allowed me to change key dimensions of the profile and ensure that the skirt balance was maintained.

First I reviewed the idea of just a pure bag. As expected, it could be fitted but it’s obviously was never going to meet the requirement of fitting on the trailer. It just extends out too much.

To achieve the desired hard structure clearance the inside profile is very optimistic. In reality the skirt would crush down on the surface, lowering the hull clearance

The second type considered was a loop finger skirt. It seemed a logical next step. The finger is contained under the loop and the loop is a smaller radius than a bag skirt.

This is a balanced profile. Peter had tried a form of loop segment trying to sort his original skirt and wasn’t particularly keen to try it. The profile at the bow would have to be raked back to minimse plough-in which would alter the pitch trim the wrong way

The loop bag profile was the selected solution. The loop and bag extend the same distance outside the outer gunwale and the inside profile of the bag is not too extreme. This profile requires a third attachment between the inner and outer.

The dashed lines in the drawing are a representation of the tension forces in the skirt.

This is the front profile that was considered, trying to meet the objective of the same approximate ground contact point a but providing some protect from the generated spray.

Like the original skirt, we thought that at first the skirt would be just sewn together so that sections could be replaced if they were unsatisfactory. We didn’t proceed with this design. The transition sections around the corner would have been difficult to design.

I also investigated fitting a segment to the same loop as a profile for the bow, in order to minimise spray. It looked to be a feasible solution but we decided that simply using segments at the front and rear was the simplest known approach.

The ground contact point of the segment needed to be further out than usual to match the ground contact point of the bag.

This is the profile of the extended segment designed to match the loop bag skirt. Again the ground contact is further outboard that usual. We have a loot of experience with this form and although it would probably be a little more bouncy than normal PMX profiles, it should be stable.

The only issue remaining was to provide a pressure feed for the loop bag when lifting off from water. With the bag only fitted along the side of the craft, cushion pressure would still be restricted when floating. We had originally intended to have sort of fabric pipe running through corner segments near the bag but Peter came up whit the vastly more pragmatic solution. He attached some PVC drain pipe to the hull to hold the segments off the hull surfaces. At liftoff on water, sufficient air at pressure is pumped to the bag to inflate it and lift the hull. Once enough of the hull is clear of the water, the cushion is fed to the rear and craft fully lifts off.

The following two videos show the performance of the skirt. It meets all the design intent.

The craft is much stiffer in roll and Peter doesn’t use use body movement as he previously had with extended segments. He had to re-learn his flying technique to push the craft through corners more. He also found that the bags tend to resist rudder control at speed. When the thrust pitches the nose sown slightly the front of the bag tends to plough the water a little and create extra drag.

The 225 Wedge was the craft that Owen Ellis and I developed back in 1984. It ias a plan/kit craft based around the Robin EC44 twin cylinder 2 stroke engine that was around at the time.

Part of our thinking was that the builder would be best person to maintain a hovercraft – they would literally know it inside out.

This is a video of an inspection of Paul Mott’s craft. He is one of the early builders of the Wedge.

Our idea worked out. Paul has run and maintained his 225 Wedge for 34 years now – it’s still one of the fastest craft around.

I think the pilot (with very few hours) was trying to do a spin but got caught by the weight shift of their passenger who was videoing the camera craft.

The leading side goes down, and a section of the rear part of the hull contacts the water. The craft has simply rolled.

The skirts don’t collapse or roll under as is commonly promoted. They just crush but continue to due their job.

On contact, water is bulldozed forward in the direction of flight. It spurts out between the segments. I observed these water spurts hundreds of times many years ago when we were trying to understand the boundaries of planing hull design.

Fortunately the craft is trimmed nose up so only the rear of the hull contacts the water. It would have been more dramatic if the craft was trimmed level and the whole of the side contacted the water at once.

Normally, we all tend to lean into a spin and avoid creating this kind of situation. Experience pilots find it hard not to lean in – doing a spin with flat roll trim is harder than you think.

We figure that sideways contact is such a rare occurance, and is almost wholly a consequence of pilot skill, that pilot training and warning is as good a risk control strategy as complex hull configurations.

Video courtesy of Owen Ellis

This article examines the design limits of two attachment point segment skirt systems. By two point I mean that the skirt is only attached to a craft at an outer point and again at an inner point. There are no other intermediate attachments to the hull and the skirt on one side of the craft isn’t connected to the skirt on the other side.

In general , two point attachment skirts include bag skirts, straight segment skirts, extended segment skirts, loop segment and bag finger skirts. Jupe and C skirts are not included.

The analysis will look at straight segment skirts then apply those results to extended segment skirts. The result is applicable to all the skirt types mentioned and they will be the subject of future articles.

This article is not about the roll stiffness of a hovercraft. Roll stiffness is the ability of a craft to right itself when one side is deflected down. That response is a product of all the elements of the lift system – fan, ducting, compartmentalisation and skirt deflection.

This is an example of a straight segment. This form of segment is usually used as a finger on a bag and finger skirt and only had a brief usage in the late 60s as a whole skirt system.

An example of use

Terms

Outer face – that part of the skirt that is exposed to the atmosphere

Effective outer face – an area defined by a line from the outer attachment point to the ground contact point by the segment width

Assumptions

Pressure exerts a force perpendicular to a surface

Pressure on a surface results in a force perpendicular to the surface. Anywhere in space, pressure acts in all directions, but at a surface it appears as a force at right angles to the surface.

Another way of saying this is that pressure can’t make a surface slide sideways.

Distributed pressure force acts through a single point

The pressure on the outer face of the segment acts along the whole surface.

This net pressure force can be modeled as a single force acting through the centre of the area. i.e. through the middle of the outer face.

No compressive loads in the fabric

It’s fabric. It only resists tension. Compress it from the edges and it wrinkles and collapses.

External shape is not important

This is a free body analysis. The only important geometries are the lines from the inner attach, outer attach and the ground contact point. And to simplify even further, we’ll flatten the outer face.

Design principle

The external forces on the skirt are always balanced. No skirt fabric has compressive forces.

The analysis model

Based on the above assumptions, this is the profile for the flat straight segment as fitted to a hull.

Angle S is the profile angle of the outer face relative to ground at the ground contact point. As a starting point for this analysis, it is about 60°. For regular extended segments, this angle is around 45°. This straight segment would be a make-able segment but compared with the extended segment, the craft would have a much smaller hard structure clearance.

Forces on the segment

Force due to cushion pressure

There is an even pressure (cushion pressure) over the whole outer face. The net force is

\begin{aligned}
F_c&=P_c \times A &\\
\end{aligned}

where

    [katex]F_c[/katex]   is the force on the outer face
    [katex]P_c[/katex]   is the cushion pressure
    [katex]A[/katex]   is the area of the outer face

\begin{aligned}
A&=l \times w &\\
\end{aligned}

where

    [katex]l[/katex]   is the distance from the outer attachment point to the ground contact point
    [katex]w[/katex]   is the width of the segment

The force [katex]F_c[/katex] acts in the middle of the line from the outer attachment point to the ground contact point.

Force at the inner attachment

The force at the inner attachment counters the pressure force and acts in a line from the inner attachment to the middle of the effective outer face.

In normal skirts the inner attachment force has a component that directly counters the cushion pressure force in size and direction and another component at right angles that tends to pull the skirt downwards. It is the second component that keeps the skirt under the craft and allows it to lift off.

Force at the outer attachment

The force at the outer attachment counters the downwards component of the inner attachment force. It’s a tension force.

This diagram shows each of the three external forces on a straight segment.

Force balance

The three external forces on the segment must be balanced. That means that a force in any one direction has to have forces in other directions to counter it. An unbalanced force means that the segment will move (relative to the hull). That’s something we don’t want to happen.

The three forces can be visualised as all acting through a point in the middle of the outer face. This point is a bit like the fulcrum of a balance.

The external forces on the segment at the mid point of the outer face, point M, look like this:

Any two of the three forces combine to exactly counter the third force. In engineering, we often use a force vector circuit to see that the forces are balanced.

Two of the forces have been shifted to create a circuit. The forces all balance. The skirt won’t move under changing cushion pressure.

Increasing the length of the segment

Next, we lengthen the outer face of the segment.

In practise, this would create geometries that can’t be made, so we’ll only examine the forces on the outer face like the previous two diagrams.

As we increase the length of the outer face, the size of the forces become larger but they remain in balance. The angle between the outer face and the direction of the force at the inner attachment point, angle A, becomes larger.

Eventually, when we extend the length enough, angle A becomes 90° and the force on the outer attachment point become zero. The only forces on the skirt are the equivalent force of the cushion pressure and the force at the inner attachment point.

If we extend the length of the outer face further, the force at the outer attachment has to become a pushing force to keep the forces in the segment in balance.

However, since the skirt is fabric and can’t withstand a compression force. It will just wrinkle and collapse.

This is a limit to the amount that the outer face can be lengthened.

Other skirt angles

Now to examine what happens when we change the angle of the outer face of the segment.

The same outcome occurs – the length of the outer face can be extended until the limit is reached and there is no force on the outer attachment on the outer attachment point.

Compared width previous example, the main difference is the lower hard structure clearance, the hover height. The forces are balance and at their limit but just at a reduced hover height.

If we repeat the analysis for many different angles, we find that each one has a different hard structure clearance.

Tracing the ground contact point G for different outer face angles results in a limit line.

The limit line

The limit line is a circular arc, centred on the inner attachment point, with a radius of the distance between the inner and outer attachment points. The line is the design limit for the ground contact point of a two point attached skirt.

This means that any segment that has a ground contact point outside the limit line is likely to collapse and won’t contain the air cushion. All straight segments will fall within the limit line.

The main implication of the line is that it sets a maximum theoretical hard structure clearance for a 2 point attachment system.

The maximum height is from the inner attachment point to the ground is the distance between the inner and outer attachment points.

Limits of extended segment geometry

In the previous analysis, we disregarded the physical practicality of making skirt shapes that verge on the limits of the geometry. Making a straight segment large enough to approach the limit line is almost impossible. The segment would perform poorly well before that limit. Straight segments also have the disadvantage that they reduce the cushion area and generally allow for less hard structure clearance compared with other skirt systems.

Straight skirts quickly evolved into extended segments, commonly used on many small hovercraft today.

Extended segments can be analysed in the same manner as straight segments.

If we make an imaginary cut in an extended segment from the outer attachment point to the ground contact, we create a straight segment and the extended part

The two elements can be separated for the purposes of the analysis. The element on the left is like a straight segment. If we imagine that there was a flat outer face across the cut, then we have the straight segment per the previous analysis.

If we add a flat face to the element on the right then we simply have a type of balloon – the pressure in the balloon is cushion pressure. The shape is distorted from a normal balloon shape by the other skirts around it and the tailoring of the outer shape. The pressure in the balloon is balanced by the tension forces in the skirt fabric and the imaginary flat face of the cut. In this model, the balloon doesn’t contribute to the forces at the attachment point.

Calling this type of skirt an extended segment is very appropriate – it simply extends the shape of the straight segment.

The limit of the ground contact point still applies. Extended segments allow the geometry of the skirt to approach the limit.

The limit line also explains why extended segments should be designed to be inboard of the outer attachment. If the ground contact point is designed to be outside the limit line there will be a point that the segment simply won’t retain the air cushion and the craft won’t lift off the ground.

Conclusions

However the limit line is applicable to extended segments (and all other 2 point attachment skirt systems).

Extended segment skirts have portions of their geometry outside the limit line.

The rules-of-thumb for the geometry of skirts have a sound basis in physics.

The interesting consideration is when the craft rolls or deflects sufficiently to move the ground contact outside the limit line.

The limit line is mathematical outcome of an analysis based on simplifying assumptions. Rather than being a precise line, it is more a region. As the geometry of the segment approaches the limit, the segment will bounce and deflect more. Skirts that are well past the line will simply not contain the air cushion and the craft will not lift off.

An example of an extreme segment

In the early 1970s, some oil storage tanks in England were moved by attaching a segmented skirt and floating the tank to it’s new location.

This is the only media I could find about the project, but I remember seeing a video of installing the segments. This project could only work because the distance between the upper and lower attachment points was large enough to allow the segment ground contact point to be lower than the underside of the tank.

The segment was fully contained within the limit line and would certainly be at a safe value at the ground contact point.

Non-dimensional skirt characteristics.

Two attachment point skirts can be characterised by a non-dimensional number:

\begin{aligned}
\frac {l_{GC}} {l_a}
\end{aligned}

where

    [katex]l_{GC}[/katex]   is the distance from the inner attachment point to the effective ground contact point
    [katex]l_a[/katex]   is the distance between the inner and outer attachment points

The number represents the proximity of the ground contact point to the limit line and is commonly about 0.75 to 0.85 for well designed segment skirts.

The “flat” segment was investigating how little tailoring could applied to a segment design and yet it would still do the job.

We had been using compound segments for many years. Compound have different materials for different parts of the segment, are cheaper to make and mostly reusable. The basic concept of the segment was a given – a square wear panel of durable material and side and top panels of “carrier” material – material there to hold the wear panel in place.

In regular segments, the side and top sections of the segment are joined together, usually on curved seams. The tailored seams give the segment the curved shape.

The aims of these trials was to learn just how little the top and sides could be joined but still have a functioning segment.

There were two slightly different types – one made of all rectangular panels and the second with a wider outer panel.

They were not successful. Although they sealed in a fashion, they were too easily displaced and leaked. (This video is pretty poor and the commentary almost inaudible – but you might get something out of it)

Next test – the outer panels were pinned to the web panels at about the middle

The actual change is shown and then the next step

The results of lightly joining the outer panels together.

Their main drawback in long term operation is the wear at ground contact (true for all skirt systems) leading to sudden catastrophic failure. This lead me to also wonder if a bag skirt could be fitted with a dedicated wear panel similar to the one that we had in our segments. And rather than just being an afterthought, could the wear panel be designed to alter the characteristics of bag skirts?

Initial investigation showed that the forces in a dual bag skirt could be balanced. So time to try it out.

This is the profile diagram of the dual bag. It’s designed as a full flow bag so there’s no pressure differential between the bag and the cushion. The inner face of the bag is straight, as is the inner edge of the wear bag.

The broken lines represent the forces in the material. The skirt was designed by establishing the force balance and adjusting key parameters to achieve the design shape.

The ground contact point is inboard of the outer gunwale, simply to bridge the distance to the ground contact line of the standard extended segments, which are even further inboard.

The wear bag was modeled as being circular, even on the interior face. It wouldn’t be because the air is accelerating to escape and the local static pressure is falling but I simply didn’t know enough about the mechanics at this point to make a better prediction. Hopefully, the difference compared with the actual shape will be small.

This is a proof of concept design. Would it even work at all?

I hadn’t really thought through sewing the bag and end cap. There was no allowance for a seam edge on the end cap panels, so I just sewed it anyway. The bag was intentionally too long; I figured I could trim it if needed. It ended p being an odd sausage shape.

The first tests were just engine on at fast idle.

I needed to check that the bag was inflating correctly and more importantly sealing to the adjacent segments. They were.

The wear bag at the bottom was inflating and looked to be in the right location and shape. It was open at either end – mainly to allow rapid draining of the main bag but also to create an air exhaust path. If the ends got sealed up or restricted, then the pressure would rise which seemed like a good thing.

Seems like we had a go code for flight test.

Peter Venn has been using hovercraft for more than 40 years. From weed control, mining support to extensive fun rides, he’s done most things.

His current workhorse was custom built for him. Essentially a widened, lengthened Viper, it has a separate lift engine at the front and a Subaru EA81 driving a large thrust fan.

Although based on a extended segment hull, the craft has no ducting to feed air to each segment. The intention was to fit a loop segment skirt, but that didn’t happen. Instead Peter fitted standard extended segments. While the craft would cruise and perform well, it was very difficult to take off from a floating start. Since there was no pressure distribution, Peter developed a wild rocking technique to get the hull to initially lift from the water.

He was never really happy with the string of changes and developments he tried to improve the skirt performance.

So he was very interested when I told him about the trials I had been doing with bag skirts on that hull profile.

Needless to say, we decided to do it.

The Design

We decided to go head first and do a skirt based on my 2017 trials – a dual bag.

The craft is a one off so we spent a lot of time getting the measurements for the hull. Then the skirt was 3D CAD modeled.

The obvious issue with fitting a bag skirt to a segment skirt hull is that the fixing points are the same profile all around the craft. The practical wisdom is that bag skirts need to be constrained differently at the front to minimise the pitch down motions. Sometimes this is done by tailoring – adding in panels and joins to make the skirt fit in a particular way.

Since this was quite experimental, we decided to stick with the same profile all around and just sew the panels together. If the shape of the front of the skirt was poor or behaved badly, then the skirt could be picked apart and a new front made.

The sections of the little wear bag at the ground contact point were not going to be joined to one another. At the intersection of two panels there was to be no join between the wear panels. The idea was that this was where the inflation air and any water would escape.

Build and test

Peter made the main bag from PVC material and sewed the joins. He decided to not add the lower bag before actually running the skirt.

Without the wear panel skirt, the ground contact point was much further outboard than intended – in fact getting very close to the stability limits for a bag skirt. However the craft hovered and ran.

The first issue that appeared was the lack of hoverheight compared with the segment skirt. Peter has a steep ramp to leave the river at the bottom of his property and the front of the hull was impacting the ramp.

Second was the spray over the front. After many years of very little spray and staying dry, it was annoying but manageable

The kicker was that it was difficult to load on to his trailer. The bag skirt extends out much further than the extended segments and that extra width is much lower. The trailer had simply not been designed for that width.

The skirt is bounces a lot which is consistent with the ground contact being so far out. Although Peter didn’t do a lot of operation with this skirt, he didn’t have a plough-in as you might expect. It looks like the propensity of the skirt to deflect upwards meant that it didn’t tend to get dragged under the craft, leading to a plough-in. Then again, it was probably Peter’s decades of driving experience.

The drawbacks of the skirt, especially the trailer, meant that we didn’t persist any further.

Back to the drawing board.

A video from 2011 when we did a trip up the Shoalhaven River which was in full flow. No camera stabilisation but you can get the feel of unstoppable.