Pressure resilient linear actuator for subsea applications

[RedNicStone]

TL;DR Summary: I'm designing a subsea linear actuator for an ROV.

The device should be able to withstand high external hydrostatic pressures (> 5MPa) for our application. I'm unsure what the best arrangement of seals for this is, so I'd appreciate any input on this and any other potential issues with my design.

I'm part of a student run society who focus on building ROV's. As part of that I am working on a linear actuator that can be used to to manipulate objects. We had previously inquired at two companies that build subsea linear actuators, but both times we got a quote that would have blown our budget even if we only got a single actuator.

The design I came up with uses a NEMA 17 linear stepper motor with an ACME screw and waterproofs the lead screw by using two carbon fiber tubes and a set of seals. The stepper motor itself is also waterproofed in a separate process that is not shown in the design.
Here is the design so far:

There are three parts (yellow) that are turned out of aluminium. They would be manufactured in either 6061 or 7075 (Is there any benefit of 7075 in this application?) then bead blasted and finally hard coat anodized.
The blue and orange tube objects are the carbon fibre tubes mentioned previously. These are made from roll wrapped plain wave carbon fibre, then sanded down to the desired diameter and finally painted with a thin layer of epoxy to restore the surface finish.
There are three rods (light blue and gray here) that prevent the rotation of the ACME screw. These are made from pultruded carbon fibre which has been ground to the desired diameter. The traveller glides along those rails using an oil free bushing with a surface layer of PTFE (is this material suited for operation on carbon fibre rods?).

The objects in turquoise are seals. Four of these are standard ISO O-rings which are not very interesting. The other ones are seals intended for hydraulic pistons. Here is a close up:

All of the seals are taken from the Trelleborg catalogue (which is easy for us to order). From left to right the sealing elements are 1. a Zurcon Scraper DA22, 2. & 3. a Turcon Stepseal 2K with O-Ring, 4. Zurcon Slydring.
The scraper has a pressure bypass to equalize the pressure across it. Im not sure this is needed.
I'm aware that placing the pressure relief ring at the end is probably not the best, but it makes the design quite a bit simpler.
This arangement is pretty much just taken from a hydraulic cylinder and reversed. Would this design work in subsea conditionds at all?

Currently the actuator would be filled with air. I am concidering adding a pressure relief and filling it with oil which I have seen in other designs, but Im not sure this is needed.

If the actuator ends up working well we are planning to document the build process and making the design open-source.

I know little to nothing about seals (I'm not even an engineering student, I study physics), so I am looking for some feedback on the design. I don't really know where exactly to get help for this sort of stuff, so I was quite happy when I came across this forum. I would very much appreciate any input on this :D.

 

[berkeman]

Welcome to PF.

What are the specs for this linear actuator? Max force, actuation distance, etc? Are there any loads on the actuator other than thrust?

My initial inclination would be to try to do this without seals, and use magnetic force through a wall structure to rotate the screw... Have you considered anything like that?

 

[Lnewqban]

Welcome!

Carbon fiber may not withstand the high pressure (725 psi) and cyclic changes (think recent submergible accident).

The material is not suitable for achieving the roundness and surface finishing that any seal manufacturer requires for proper and consisting seal effect on a sliding ram.

The nut, rather than the Acme screw, is desirable not to turn.

 

[RedNicStone]

Welcome to PF.

What are the specs for this linear actuator? Max force, actuation distance, etc? Are there any loads on the actuator other than thrust?

My initial inclination would be to try to do this without seals, and use magnetic force through a wall structure to rotate the screw... Have you considered anything like that?

I designed the linear actuator with a target force of 400N in mind. This is why I chose a lead screw with 2.54mm of lead per step. The actuation distance is limited by the stepper to 150mm.

The load of on the actuator would only be the thrust (we cant lift any object that is heavier than our thrust, so this is the limit). We have not decided on what motors we use so its hard to estimate it, but it likely will exceed 400N. That's why I chose to use an ACME screw since those are very hard to manipulate the other way around.

I have considered using magnetic coupling for the motor but this has some other challenges like finding an oil-less ACME screw and bearing.

Welcome!

Carbon fiber may not withstand the high pressure (725 psi) and cyclic changes (think recent submergible accident).

The material is not suitable for achieving the roundness and surface finishing that any seal manufacturer requires for proper and consisting seal effect on a sliding ram.

The nut, rather than the Acme screw, is desirable not to turn.

I ran some buckle calculations some time ago and estimated that these carbon fiber tubes should be suitable for at least 4MPa with a 50% leeway to the physical limit. This is just a rough calculation though and I have no idea how accurate that estimate may be.
I do have some concerns regarding roundness. We were also concidering using aluminum tubing but this could cause some problems down the line with weight. I feel this is something that we probably need to test (and probably sacrifice some prototypes for that).

Also why was this thread moved to the homework section? If I posted this in the wrong section im sorry about that.

 

[berkeman]

Also why was this thread moved to the homework section?

It's a schoolwork project, so it belongs in the schoolwork forums. That was mentioned in the PM I sent you when moving the thread.

Don't worry -- many of the same people help out in the schoolwork forums and the technical forums. You are showing lots of work and effort on your part, which is what we look for in schoolwork projects and problems.

 

[Baluncore]

Maybe I missed it, but the external pressure will push the actuator rod back into the housing, if the body is not filled with oil, and allowed to match the external pressure, plus lube the screw. That would require a significant bladder volume, since the volume change of the actuator body is significant.

The only alternative I see, would be to design it as a two ended device, where extension at one end is matched by a retraction at the other. Oil would still be needed, but the change in volume would be less. That would eliminate the majority of the seal problems by reducing the differential pressure.

 

[Baluncore]

There is a simple, low-cost solution.
Replace the inner CF tube with a square stainless steel tube that can become filled with external water. The acme threaded rod, also stainless steel, is driven by the motor through a couple of lip seals on the shaft. The motor is filled with kerosene that has pressure equalised to the outside through a small flexible diaphragm. There is little change in the motor volume, just compression of the liquid and liquid thermal expansion. Kerosene is used in underwater electrical generators as it prevents entry of water, and cools the wires without ventilation.

The acme nut is attached to the square section tube, so does not rotate as the tube passes through a square end-guide. That eliminates, and removes, the side forces from the internal guide rods. Those rods would have bent and fatigued anyway, due to the triangle of forces near half-travel.

The external CF tube could be replaced with a square stainless steel tube to act as a guide for the inner square tube. Both tubes then have a few holes drilled to allow water to come and go from either end. The acme nut is lubricated by water.