Control suggestion for a hyperbaric chamber

[mcutin]

Hello, guys.

I was designated to think about a solution to control the pressure inside a hyperbaric chamber we have here at the workshop that is used to perform pressure tests on equipment designed for subsea operation.

The point is today the technicians have to manually control the water ingress inside the hyperbaric chamber to control the pressure inside it until it reaches the test pressure. Pressurized water is provided by a pneumatic pump that is mounted in the hydraulic circuit outside the chamber. There is a pressure transmitter installed in the chamber inlet to monitor its pressure. We start to pressurize the chamber after completely filling it with water.

Recently we were informed of a requirement that has not been considered since then. We need to reach the test pressure in 30 minutes +/-1 minute. Therefore, it is quite difficult to meet this requirement manually controlling the pressure inside the chamber.

Do you have any suggestion on how I can implement this control? We have a small PLC available that can be used as part of the hardware needed for the solution. However, I don't know if it would be best to control the air pressure in the pneumatic pump pilot (and this way controlling the flow rate in its outlet), or if a control valve should be used to control the water flow based on the pressure transmitter readings.

Any guess? I appreciate any help.

Thank you,

Marcelo

 

[thebard3]

Does your 30 minute requirement include fill time? If not it shouldn't be a problem to achieve pressure in just a few minutes.

Brad Waybright

It's all okay as long as it's okay.

 

[mcutin]

Hi, Brad.

No, it does not include fill time. My doubt is more related to what I should control to meet the requirement.

The system itself is capable of achieving the test pressure in less time, indeed. The main difficulty is making it gets from 0 psi to, let's say, 3,900 psi in 30 minutes. We have already lost some tests because the technicians were not able to manually keep this rise time and the test pressure was reached only after 40-45 minutes (or, on the other hand, on 20-25 minutes).

I believe a PI control would be appropriate for this process. However, I still do not know what I should control.

Thanks once again,

Marcelo

 

[thebard3]

If this is in North America, I assume this is an ASME code pressure vessel. I believe your description of the control element is that a piston pump is used introduce a relatively small amount of water to the vessel after it is filled and the majority of air displaced. You are correct that PID control will accomplish the task, but the control element must be sized accordingly. Also, assuming the piston pump process is used, the less air that is left in the filled vessel, the shorter the pressurization time. I realize this is outside the parameters of your question, but when working with pressure vessels, particularly such high pressures, it's important to ensure that proper and standard safeguards are employed to prevent over pressurization, opening under pressure, and the like. If this equipment is already in place, then that's a good thing. If not, you should thoroughly research the requirements. You may find that you need more equipment and devices than the 'small PLC' that you refer to can support.

Brad Waybright

It's all okay as long as it's okay.

 

[mcutin]

You are absolutely right.

This hyperbaric chamber has overpressure protection devices already installed (relief valves). Although we are not in North America, indeed it is an ASME code pressure vessel plus our local regulations.

 

[thebard3]

I have just re-read your initial inquiry-

I don't know if it would be best to control the air pressure in the pneumatic pump pilot (and this way controlling the flow rate in its outlet), or if a control valve should be used to control the water flow based on the pressure transmitter readings.

Actually, if the system works as I think it does (HP piston pump forces water into vessel to pressurize it), then you really don't even need a PID control unless there is some kind of loss in the system (e.g. a leak, etc.). You should be able to simply run the pump until the desired pressure is achieved, and then stop it. If you need to modulate the pump output, then I believe it would be best to use a control valve on the pump air supply to control the speed of the pump, but from the description of your process, I think you probabaly just want it to run at full speed until the pressure is reached. BTW, as I mentioned before, the amount of air remaining in the vessel when you begin pressurizing it will play a direct role in the time needed to reach pressure. You may only need to work on flushing as much air as possible from the vessel before your pressurization sequence begins.

Brad Waybright

It's all okay as long as it's okay.

 

[itsmoked]

Here's what I'd do, and no a PID or PI is pointless unless it's a steady-state process where you must continually do something to maintain the pressure.

If you have a system that fixes the amount of air then this would be trivial. For instance, a dribble port. You fill the tank from the bottom until the water overflows out the port then close the port. This easily traps the same amount of air every cycle. If that's your system, you use the PLC to run the pump at a specific rate completely dead-reckoned. A few trials would nail the pumping rate needed.


If you have no way to provide the repeated air amount then you add to the dead-reckoned method. A monitoring subroutine watches the first minute of pumping at a dead-reconed rate and at the end of the first minute assesses the rate at which pressure is building then sets the dead-reckoned speed constant for the next 29 minutes.

I would use what's called an I to P to control the constant air normally provided to the pump. This would allow full speed control of the pump.
I/Ps take a 4-20mA signal and convert it to 0-Xpsi. The more air pressure out the back the faster your pump will cycle. You could also use a dumber pneumatic valve that will work the same way.

Sounds like a lot of fun.



Keith Cress
kcress -

http://www.flaminsystems.com

 

[waross]

I like it Keith.
I would also consider a PID controller with a ramp and soak function.
Set the terminal pressure and set the ramp time to 30 Minutes.
You could use an accumulator rather than leaving air in the vessel.
The PID with ramp and soak would also piggyback onto your trapped air scheme.
Here is a Ramp and Soak process controller from Omega for under $200.Link

Bill
--------------------
"Why not the best?"
Jimmy Carter