High Pressure Valve Opening - Collision Mechanics

[redowling]

Hello all - this isn't a homework question; it's a real problem I'm struggling to solve. Apparently I'm being a bit of a mechanics 'tard, so any help would be appreciated.

I have a high pressure reservoir (~60MPa) sealed off from a low pressure area by means of a metal plate. The plate has a hole in of 40mm diameter, kept just to one side so the metal plate provides an initial seal. I need this hole to be fully open to the high pressure reservoir in 2.5ms. The plate is likely to be clad with PTFE to reduce the frictional forces (mu = 0.2, PTFE to Steel). All of the normal force will be provided by the pressure acting on the plate.

In order to open the valve in this time, I am investigating striking the plate (m=0.37kg) with a larger mass which will achieve the required opening speed.

I'm struggling with taking into account the large frictional forces on the plate to work out the mass & velocity I need from the larger mass in order to achieve the right opening speed. I don't think it's realistic to model as a perfectly inelastic collision where the masses stick together (because the physical design of the system probably will need to stop the larger mass once it has struck the smaller).

This seems like it should be really simple mechanics question - but I can't seem to get my head around it.

 

[K^2]

Why not use a butterfly valve? These are ideal for fast open/close operations under high pressure.

 

[Q_Goest]

Hi redowling, welcome... I'm having a hard time envisioning what you want to do. Is the impact going to slide the disk off the hole or lift it? Generally, metal to metal impacts have a low coefficient of restitution and I think you'd be conservative assuming the two stay together after impact.

You might also consider using the pressure to assist in opening, similar to the way relief valves work. The areas exposed to the pressure change once the valve starts lifting, accelerating the valve open.

 

[redowling]

K^2 - A butterfly valve doesn't allow us to shape the pressure rise, which is an important improvement that we need later on. I'm hoping that altering the hole profile will allow us to shape how the pressure increases in the low pressure section.

Q_Goest - the impact will strike the plate, sliding it along such that the hole in the plate is exposed to provide a channel for the gas to flow into the low pressure area. Essentially I need to design a valve which will allow me to shape the pressure-time curve downstream over a total event duration of approximately 5ms.

 

[Q_Goest]

... the impact will strike the plate, sliding it along such that the hole in the plate is exposed to provide a channel for the gas to flow into the low pressure area. Essentially I need to design a valve which will allow me to shape the pressure-time curve downstream over a total event duration of approximately 5ms.

Sounds exactly like what Worcester ball valves did to create a control valve. A ball valve is very similar to what you have in the sense that the ball is like your sliding plate that uncovers the hole from one side to the other, changing the flow area as it does. But what they did was to create a specially designed area so that as the ball uncovers the area, the flow area changes non-linearly as a function of rotation angle.

Take a look at page 6 of http://www.flowserve.com/files/Files/Literature/Products/Flowcontrol/WorcesterControls/WCE0C44.pdf" . They show various styles of openings that are inserted into the ball valve to get different flow characteristics as the ball rotates and uncovers more of the area.

You can do the same thing with a sliding plug. Just have a hole that isn't simply round. The hole shape can be contoured to give you the flow characteristics you desire as a function of time.

One concern however about the sealing material. At 60 MPa, you're well above the compressive strength of most (not all) filled Teflons. To seal, you need to have a contact stress above the differential pressure, so the contact stress is necessarily going to be above the compressive strength of most Teflons. The material will likely deform, conforming to the shape of the disk you're using to seal the hole. That isn't to say it won't work, it just may not work well. Performance might be erratic. If you don't need to use Teflon, I have some suggestions on alternative materials that would work much better, just provide the fluids being sealed and temperature range you're operating in.

 

[redowling]

Thanks, that's a useful bit of information.

So in terms of actuation speed, is there anything that could feasibly open a ball valve like that in the time required? I haven't been able to find anything that will open the valve in around 2.5ms!

 

[Q_Goest]

I don't think you're going to find an off-the-shelf solution. It's not just the speed, it's also the high pressure and large opening (40 mm) you want. There are ball valves that handle the pressure but they're smaller than 40 mm opening. You might contact a few manufacturers and see if they'll make you something, or just make something yourself.

But since you don't seem to be stuck on the idea of a sliding plate... If I were to do this, I'd design a balanced globe valve in order to reduce pressure loads. For actuation I'd try what you're doing; using an impact type of arrangement to actuate the valve stem. If I had to contour the flow, I'd do that by shaping the valve plug. To get a high speed pressure trace, you'll need to use a fairly sophistocated pressure transducer such as what http://www.kistler.com/pressure-sensor_en" makes. I've used them before for ballistics testing, and the pressure trace could see pressure waves bouncing from one end of a 60,000 psi vessel to the other, the vessel being less than a foot long.