Taking Seriously Negative Pressure Designs

I need to take a moment to register one of the biggest shifts in thinking I've had the the flow dividers up until this point. In almost every case, I have taken the same approach to the problem that I'll call the "pressure walk" problem.

This problem involves first setting a boundary condition for pressure, establishing that you know pressure at some given point.

• Points where we know the pressure is equal to (or roughly equal to) ambient
• All points fully outside of the tube
• The open ends of the tube
• Basically the entirely centerline inside of the tube (the axis of rotation)
• Points where the pressure is higher than ambient
• The habitat surface, as the air above it is centrifuged in the "down" direction
• All points between flow dividers, as these need a positive pressure

Of course, I'll make a diagram here.

It is important to conceptualize that the connection point, at the opening at the end, has no real pressure difference from ambient except for what's necessary for the air bleed system. Also, positive pressure is needed to maintain shape, so extra positive pressure may be needed for this.

What is Different About Negative Pressure

The one key change we will make for negative pressure designs is that the dividers will be wide open. I imagine a porosity of 10% or more. So they are flow dividers, in that they present large surfaces to disrupt large flow patterns, but they don't hold air in. Because of this, the pressure walk goes straight through the dividers. Start from ambient and go to lower radii, and pressure decreases (for the same reason it increases going the other direction.

The challenge that you've created for yourself is now that you have a highly negative pressure region, and you somehow have to seal that over a rotating seal. We have pressure over a rotating self in the prior design, but this is different. The other design set no minimum of the pressure difference.

Here, I tried to give some illustration of the ingress over the rotating seal, due to those pressure differences. How much is the pressure difference?

I made a table and got numbers, but they are basically the same as the pressure of the stages in the old design. This is driving by the rotation rate and radius. Again, the difference is that this pressure is exposed to a rotating seal. There is a minor difference, because of accounting for the walk through the stages before entering the rigid part, but it's mostly this effect.

Recapping the above points, but for negative pressure designs:

• Points where we know the pressure is equal to (or roughly equal to) ambient
• All points fully outside of the tube
• The open ends of the tube
• Basically the entirely centerline inside of the tube (the axis of rotation)
• Points where the pressure is higher than ambient
• The habitat surface, as the air above it is centrifuged in the "down" direction
• Points where the pressure is lower than ambient
• Everywhere between the flow dividers

This is an odd juxtaposition of positive and negative pressures. In fact, the inner-most stage which will drive the design winds up being negative pressure by about the same amount that the habitat is positive.

This is very seriously and legitimately weird. It's so new to me that I'm still actually not sure if there's a better way to design it so that the pressure across the rotating seal can be further reduced. Having connection points at larger radii would probably help, but the innermost stage is still the design driver.

Also, there are very interesting structural implications of this design. The flow dividers wouldn't just be blocking flow, but would also have to contain members that could push back against the air pressure on the rotating end plates. Because nothing else is moving at the required speed, so it has to be integrated into the stages.

I am extremely unsure which approach is better. The pressure over a rotating seal is an argument against this. But it's also a very attractive feature to have porous flow dividers. I learn towards the negative pressure design being worse, but it's also entirely plausible that bizarre flow mechanics lead to discovery of new effects where the porous design allows it to perform dramatically better. So for a good while into the future, I expect these to be 2 valid and competing design paths and one shouldn't be dismissed for the other.

Speaking more practically, I can see the negative pressure designs working better for small scale experiments.