The POD Project - Structure

Parts

In architecture, a structure is an assembly of parts that transmit forces. Forces are compressive (squeezing an object), tensile (stretching an object, torsional (twisting an object), bending, and rotational (pushing an object around a point). The structure must endure loads, or external fources, imposed it by natural sources like wind and gravity. A part or connection between parts may bend or twist a little, but it cannot break. So the task here is to design the POD in such a way that it will survive loads without any parts or connections breaking.

The structure of a POD is made up of three elements: beams, nodes, and cables. Beams transmit compression forces along their length. Nodes transfer part (but not all) of this force from beam to beam. What the nodes cannot handle is taken up by cables which are connected to the nodes. While nodes may dampen rotational forces to some extent, they really should be treated as idealized hinges. We will rely on cables to handle the rotational forces.

Note that panels do not in any way contribute to the structure. This is an important design point: panels merely cover space, but do not transmit loads. With this decision, we open up the possibility of many flexible but weak materials for covering the POD.

The diagram below shows how a downward force is transmitted through a cross-section of a POD. The cabling at the second tier is crucial for keeping the structure from collapsing outward.

PICTURE HERE

From above, the second- and third-tier cabling looks like this:

PICTURE HERE

This cabling provides tension, without which beams will pop out of nodes. They also counter the rotational force around nodes along axes parallel to the ground that would cause the top slice to cave in and the bottom slice to pancake.

To this cabling we add some X-shaped patterns. This will prevent the structure from squashing when exposed to lateral forces like wind:

PICTURE HERE

While the triangular faces should be self-stabilizing, squares are vulnerable to two types of rotational collapse. First, the beams may rotate around the node along an axis perpendicular to the face, much as a house of cards would collapse. In a POD this would result in the top slice spiraling around until it falls directly onto the bottom slice. We counter this by adding an X-shaped pattern inside the face, connecting opposite nodes:

PICTURE HERE

The second instability is where the face can fold around the diagonal. This is not as serious as the previous case, because the POD stucture will tend to resist it, but it can cause still distortions and strain in the panels that make up the walls. So to counter this, we must add cabling along the second and third tiers in rectangular patterns:

PICTURE HERE

The second tier differs from the third in an important way: it must also hold up a floor. This is a challenging problem I have not yet solved, though I have a few ideas. In one scenario, there are rigid beams extending radially from the center, perhaps connected to a vertical central pole. Floor tiles shaped like isosceles triangles would sit on top of these like slices of an octagonal pizza. In another scenario, there is a web of cabling with a spiral tension-adjusting strand that would hold up smaller tiles.

Rejected Design

When I first thought about the shell structure, I played around with the idea of passing a cable through the beams (i.e. the beam is a pipe). The idea was to link up all the beams around a face with a tight cable. Then other faces would be "chained" on in similar fashion. All my attempts to model this failed because I could never get the cable tight enough, and the result was a sloppy mess. Furthermore, there was no good way to transfer forces between beams. The edges would hit in random ways and could cause damage. Connecting cross-cables would be difficult too. This experience is proof to me that modelling is essential in testing out every idea before it becomes a very expensive lesson later.

Beams

The requirements for beams are:

• Light enough for one person to carry.

• Easily connected to nodes, preferrably without additional hardware.

• Strong enough to handle the compressive and bending loads.

• Durable for all kinds of weather.

• Minimal environmental impact.

The shape I settled on is a tube. It resists bending while using less material. Length is around six feet and width is anywhere from 3 to 6 inches, depending on the material and load requirements. Wall thickness is an important consideration too. In the future I will have calculations here for determining strength and ideal dimensions.

The first material I considered is aluminum. For a metal, it is strong, light, and not too expensive. It resists corrosion and has a nice whitish metallic look. It was R. Fuller's choice for the Dimaxion House. A big unknown for me at the moment is the environmental impact. I suspect that the energy required to produce aluminum is high compared to materials like wood and plastic. So I need to research this.

I recently started working with bamboo. Bamboo has some very nice properties. Compared to other kinds of wood, it is very strong for its weight and grows extremely fast. Its shape is ideal for POD beams, being long, cylindrical and naturally tubular, so it requires very little cutting and shaping. Its fast growth means that it can substitute for tree species that are at risk from over-logging. The species I am looking at first is Guadua Angustifolia, commonly called guadua. It grows over 20 feet with diameters over 6 inches.

Bamboo has some problems. Its shape is somewhat unpredictible, bending, tapering, wobbling along the length. This might not be a problem as long as both ends fit into nodes snugly and it is thick enough for the loads. More serious is its tendency to crack. This will require special protection on the ends, such as wrapping with wire. Being an organic material, bamboo is susceptible to rotting if it gets wet and dirty. However, there are chemical treatments that make it resistant to rot and insect damage.

Some architects have done fantastic things with bamboo. There is at least one company dealing in bamboo contruction hardware (Bambootech) that looks very interesting. It affixes metal connection points to the tips that can lock into joints with pins. I will have to see if I can get samples of this.

Nodes

Nodes are the most challenging part of the structure to design. They are connection points for both beams and cables. Since I have not seen anything on the market that matches my specific requirements, I have to design them myself from off-the-shelf parts. Here are the requirements:

• Must connect easily to beams without laborious nut-tightening.

• Fit snugly into beams, so they won't pop out when force is not applied.

• Each node must allow for at least 4 cables to be connected.

• Must dampen rotational movement of beams around any axis perpendicular to the beams.

• Must transmit compressive forces from beams.

• Has to be able to take high tensile force from cables.

• Must survive constant grating and flexing from vibrations and changing loads in the POD.

• No welding, drilling, or machining of parts.

I played around with many designs before finding one that seemed to satisfy all these requirements. It uses metal rods bent in various angles, fused with tape or wire into a four-pronged shape. An eye bolt with large washers is screwed through the middle as a connection point for cables. I wrapped this in foam padding to get it to fit snugly in the pipes I was using for beams.

It works well for a 1/4-scale POD, but I don't know how it will work for a full scale version. If it proves too weak, I will consider dropping the last requirement and taking up welding or machining. Also, the Bambootech option may prove workable and I can outsource node construction.