[The Following is an excerpt from a letter to Allen Meece]

[Updated 4 September 2002]

I wrote earlier that the lift cells will not rub against the shroud during inflation. This statement was not wrong, but it turns out it will take a little intelligent design to make it so. The key to not scuffing the lift cells on the shroud lies in how the lift cells fold when not at full pressure.

The behavior of a partially inflated lift cell is similar to that of a plastic bag partially filled with water. The two experience similar forces. The perimeter of a vertical cross section of these vessels does not change regardless of the degree of inflation, because the fabric is pulled straight up and down and is not given an opportunity to fold horizontally. Vertical folds are another matter. Because the lift gas does not pull sideways once pressure equalizes, partially inflated lift cells do not have the horizontal tension necessary to pull vertical folds out of the balloon fabric.

Folds in the ballon fabric cause a reduction in the exposed surface area of the lift cell. As the cell expands during ascent, its exposed surface area expands, too. The additional surface area comes from pulling out these folds. In a smooth, radially symmetric lift cell, these folds form more or less randomly during deflation. Once formed, they tend to be quite stable, and it is necessary to apply some force (such as surface tension due to pressure) in order to alter them or pull them out. Also, while their tendancy is to form a random but roughly even distribution on uniform surfaces, if they are pulled into an uneven distribution they tend to stay that way. It is physically possible to pull all of the extra surface area on a partially inflated lift cell into a single fold, where it will stay until pulled out during inflation.

Fabric is pulled out of the fold at the open end. Unfolded fabric on the cell needn't move during this process. It is possible to arrange contact between two expanding lift cells while hudreds of meters of fabric are unfolded without letting one scuff the other. Consider three partially folded lift cells in contact with each other.

Where the lift cells are in contact with each other, equal air pressure tends to force their surfaces into a flat plane. To curve where they meet, one cell would have to be at a higher pressure than the other. Folded fabric can still exist at the point of contact, but it takes additional force to make a new fold where the cells contact each other because doing so must overcome both friction and air pressure. Thus, the contact surfaces are stable too.

If all of the cells expand at the same rate, their exposed surface areas increase at the same rate, too. If the folds in cells 1 & 2 where the cells meet are both the same size, they will both pull out at roughly the same rate during inflation. This means that their surfaces will not slide along each other during expansion. There will be no relative motion at any single point of contact. However, if the fold in cell 2 where it contacts cell 3 has to be pulled out against the unfolded (and unmoving) surface of cell 3, it will cause abrasion at the point of contact.

Unfolded surfaces don't move horizontally relative to each other because there is no expansion of surface area for either. If two flat, unfolded surfaces are in contact with each other during inflation, they will tend to stay that way unless something moves the entire cell.

So, we can keep balloon fabric from rubbing by pulling and folding it. It is possible to select the cut of the balloon fabric so that it will always fold along the same contours. Once folded, the fabric will remain that way as long as there is vertical tension, without need for straps or other measures to hold it in place. If we know where the lift cells will contact each other, we can place these folds where the fabric in them won't rub against anything when it's being pulled out. The same idea applies to keeping the lift cells in contact with the shroud.

This could allow us to use vertically oriented lift cell schemes where the cells remain in contact with each other during the entire ascent. Being able to transfer loads between cells would give the platform greater strength and endurance. (The two crude pictures above are drawn to scale. Note how the cell spacing differs between them. The cell axes can be allowed to move during inflation in this scenario, as long as the points of contact remain the same.)

This could provide us with the stability of the horizontal lift cell schemes seen in derigibles while still allowing us to use smaller vertical lift cells that a crew could replace at altitude. It would also require a shorter beam length for the platform, to keep the cells in contact when underinflated. Shorter struts mean less weight, and allow a smaller starting cross section.