Pressure build up in a pipe system containing solvent

[EnginE3R]

Ok, having a bit of trouble here. there seems to be a gremlin in a system at work that is causing a pressure build up but we cannot find the cause.

in essence, the system is run off a huge vertical storage vessel, the solvent is pumped out via a pump at approximately 40psig. This then feeds two dispensing guns, 4 large mixers, 5 washers and a dispensing room with three outlets. The pump is only activated when a dispensing point is operational, and stops as soon as operations cease. The issue at hand is that there is a large pressure build up, intermittently, of up 150psig. Enough so that it can blow through a dispensing guns trigger operation.

Im looking for any help or ideas at what could cause such a build up within this kind of system and to this magnitude. Any ideas are welcome and happy to explain further details if you need more information.

thanks

Sean

 

[Tmoose]

What is the pump discharge pressure, and the height of the storage vessel relative to the dispensing points when full?

What event/sensor or ?? causes the pump to run?

Are there flow check valves installed, and where are they.

I'm guessing it a positive displacement pump.

Are certain dispensing points more likely to cause the pressure build up?

Regardless, I'd be thinking about stuff like accumulators, pressure relief valves, and circuits to bleed/recirculate fluid from the pump discharge back to the main storage vessel.

6 down, 14 to go.

 

[MacGyverS2000]

Is the feed pressure that high? If so, sounds like a leak at the regulator, allowing for the pressure to build up to the main tank pressure.

If the main pressure isn't supposed to be that high, possibly something is allowing the compressor to run unattended, forcing the system pressure to rise.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[LittleInch]

You really haven't given us a lot to go on here. Initially sounds like thermal expansion or vapour creation but need system details, pump details, any valves, relief valves, sequence of operations. How does the pump start and stop? What is the flow min and max?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[moltenmetal]

Sounds like you have a positive displacement pump which is dead-heading against (hopefully) its relief valve before there is a chance to get the discharge open or the pump stopped. But to help you we need a lot more information about the system, otherwise we're merely guessing.

 

[EnginE3R]

I shall try to attain all the information you're asking for, although you will have to excuse if some of this is new to me as I've only just started at my position here last week and been thrown a little in the deep end.

In regards to the tank, it is its self approximately 5 meters high and operates at atmospheric pressure. the exit to the tank is on the same level at the pump and all the dispensing points other than the 4 mixers which are at about 3 meters above ground.

The pump it automatically operated by sensors and switches, depending on the operation tool. Although this has already been tested and the pump shuts off immediately when operations cease.

I don't believe there are any flow check valves but would have to double check that one.

the pumps are centrifugal pumps that operate at approximately 3 bar and pump at a rate of 12000L/h.

I'll try to get any other information tomorrow (home time for me) and I appreciate the advice guys.

Current theory is possibly water hammer effect that gets "stuck" behind the only NTV which is directly after the pump but its a hard one to prove.

 

[LittleInch]

Check all sources of heat rise. you will have an NRV somewhere in the system otherwise the pressure would just go backwards through the pump.

If this issue only arrives when you've got one dispenser running then it might be surge, but only a few degrees C will make a massive difference in pressure.

It could be something like a stretch of pipe is exposed to sunlight at certain times of the day and not others.

You'll need to try and get as amuch as you can from the pople who see it and try and find out what is going on when these pressure spikes appear.

Are they fast rise, slow rise?
How often does it happen?
Can it be replicated?
Try and formulate a set of trials to work out if its surge or temperature.


Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[moon161]

Why not water hammer? Do the spikes occur when flow suddenly stops? Could some capacitance in the system absorb this?

 

[BrianPetersen]

If you have a valve arrangement such that you have created a closed vessel (e.g. between a pump outlet check valve and closed valves at each of the subsystems that your pipe feeds) and the circumstances are such that the temperature is rising somewhere ... you are getting thermal expansion with nowhere for the fluid to escape.

 

[Jboggs]

If the valve at the dispensing tool is closing even a millisecond before the pump stops, you could have an immediate spike in pressure. You're dealing with an incompressible fluid here, right? I would redesign the system so that the pump has fully stopped before the dispensing valve closes. Or let all the dispensing points feed from a common pressurized header, with an accumulator to provide some flexibility. Then the pump would only come on when the pressure in the header drops below a certain level.

I would also strongly recommend installing pressure relief valves throughout the system to prevent potential human injury.

 

[EnginE3R]

All your advice guys is very helpful thus far. I shall be looking into each suggestion to see if it affects the system, although the company recently has a large data storage swap over so getting certain information is difficult, the joys of modern technology.

 

[EnginE3R]

something else I should probably mention that just came to mind. The pressure build up seems intermittent, and build up seems to be slow, and not instantaneous after the line closes. Although this is hard to tell as the hand dispensing points where the problem is most visible isn't used that often, mostly the line is dispensed into closed mixers. While this would suggest some form of Thermal expansion, the pipe goes along the same pipe bridge, and through the same areas as about 8 other pipes, yet only this line has a pressure problem.

Again I know this isn't much to go on, but I'm new in this role and just looking for any spitball ideas that I can look at or test.

As always, I appreciate all your suggestions and advice.

 

[Tmoose]

Does one pump feed all 8 pipes from an 8 legged manifold?
In order for just one line to develop high pressure, it suggests there is a check valve somewhere along each pipe, or at least THAT pipe.

If 100% positive prevention of back flow is not needed, a strategically place leak/bleed would prevent pressure build up.

19 down, just 1 to go.

 

[EnginE3R]

Okay, here is a stripped back version of the line. For obvious reasons I cannot include any other information. This is just to give you an idea of the layout of the system. So, you have the two tanks which store the same fluid (used one at a time). From these the solvent is pumped through the two centrifuge pumps (also only running one at a time) to the areas needed. The line (approximately 2 inch steel pipe, with either 3 or 5 mm thickness) goes into a single large building (approximately 75-125 meters of piping in this system from one end to the other) which feeds the four stirrers, 5 washes, two gun dispensers and a dispensing room.

From what I can tell, which is why were having so much issue with it, is that there shouldn't be any way, other than the pumps, that pressure is input into the system. All the washers and stirrers aren't pressurized and the fluid is poured straight in, and all dispensing points are essentially open ended. All solenoids are run on a slow down procedure, which means when they detect that whatever they're filling is reaching the desired weight, they slowly close the valves. The pipe goes over the same bridge as several other pipes, so if there were thermal expansion in this pipe, it would also be in all the others. While there are no relief points, the centrifugal pumps should not be able to pump any more that 3 bar and about 3L/s. These have also been tested and they turn off as soon as all operational points cease using the fluid. The storage tanks are not pressurized either, so the max pressure at the bottom of the tanks when full is no more than 0.5 Bar.

At this point there is no silly suggestions, anything outside the box is welcome. Anything I can test or research is a positive step forward.

Thank you for all the suggestions so far, they have all be most appreciated

 

[LittleInch]

Your pipe might be in the same pipe bridge, but you need to check out the vapour pressure versus temperature for this liquid - "solvent" could be virtually anything.

Exposed pipe in the sun can hit 60-70C and you might find that your liquid is boiling, but all depends on what it's physical properties are.

What are those vertical tube like things just downstream of the pumps? and on the lines in the solvent dispensing area? Why is only one connected in a through flow system?

you are only showing one PI upstream of the non return valves. Where are you measuring this 150 psi?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[EnginE3R]

The Vertical Tube type things were just pipe identification codes, just removed the codes for this particular drawing. All the lines carry similar solvents across the bridge, but I could request the R&D department to run some tests on the fluid to see if that is the issue.

As for the point at which the pressure is displayed, the Gauge is located at the dispensing gun in the production room. The Gauge hasnt been added to the drawing yet as it was only added fairly recently. But thank you for pointing that out, something I can do fairly quickly today.

 

[LittleInch]

Hmmm,

Only thing I can suggest more is for YOU to physically check the entire run of pipes from tank through to each dispensing point using your drawing just to make sure no one has added a tapping point or connected something to any of your blanked off valves. don't trust anyone else with this.

Also figure out how much of your pipework in volume is exposed to any form of solar heating. It might only be 20% of the entire volume, but could easily raise pressure by 7 bar if you get s ay a 20C rise in temperature. Solvents and hydrocarbons in general are notorious for thermal expansion / pressure rise issues.

Find out what it's volumetric temperature expansion co-efficient is. Some typicals for solvents are 0.0015 m3/m3/C. Water by example is 0.0002, so 8 times worse than water.

Try and get a history of these pressure rises to see when they occur and what was happening before and during it.

otherwise just fit a pressure relief line back to the tank or fit a pressure accumulator d/s the pumps NRV set at some suitable pressure which doesn't impact on the D/S equipment.

It could be your system is just new and the non return valves are seating better than any other system where there are few more leak paths.

Keep us informed.


Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[chicopee]

You have not said anything about ambient conditions, solar exposure, nearby heat sources and when (day or night) these high pressures occur. As alluded above, a heat source could be the culprit.

 

[EnginE3R]

Well, we had an "interesting" incident the other day. On of the large machines has some solvent dispensed into it. after which the pressure at the gauge reads an increase of just over 1 bar a minute. within 15 minutes it reaches 20bar. As the pressure is building, I go to check on the pump. As I expected, its off. At the point, when the pressure is 20bar, the dispensing gun is used to relieve pressure briefly, immediately after the build up ceases. We had a second incident using the same machine in the afternoon, as soon as it was done pressure built up (yet the pump was off, and its a centrifugal pump with a maximum output of 10 bar I might add). This time the gun was was used to relieve pressure at 10bar. Yet like last time the pressure stopped building afterwards. Now the machine that "triggered" the events is just an open top mixer, it has no pressure at all and the solvent is simply poured into it.

At this point I have given up trying to find the cause because this system simply shouldn't physically be able to build up 20bar in 15 minutes then just stop. So I've moved onto designing a modification to the system to allow it to be used safely.

 

[zeusfaber]

It really is time we knew what this solvent is.

Is the solvent stored somewhere that's much cooler than the rest of the factory?

A.