Cooling Water Dissolved Gasses

[clay87]

We have a demineralized, low-oxygen cooling water system that has H2 as the cover gas in the surge tank. We have some air or O2 inleakage somewhere in the system that is raising the dissolved O2 in our cooling water (normally less than 3 ppb dissolved O2). In order to reduce the dissolved O2 levels, we began to sweep the tank with a slow ("feed and bleed") injection of N2 gas. The nitrogen is injected into the tank at less then 10 scfh and vents through the normal pressure relief valve at 5psig. By design, the relief valve can easily relieve 10 scfh at 5psig. After a day of injecting N2, we experienced an odd high-pressure event in the tank (up to at least 7 psig) along with an apparent reduction of flow in the cooling water system. I think we encountered a sudden restriction, such as relief valve plugging, in the relief line that led to the high pressure.

However, I am wondering about the dynamics of changing the cover gas from H2 to N2 and whether or not that could have played a role in this event. H2 is more soluble by volume and therefore more of that comes out than N2 goes in. Using rough numbers, the system water volume is about 2000 gallons, the tank total volume is about 100 ft3 with 25 ft3 of gas space, water temperature in the tank is around 140F. What would you expect happen as you slowly change the composition of the tank cover gas?

 

[bimr]

Once dissolved oxygen is removed from water, the concern focuses on how to maintain low dissolved oxygen levels in a storage tank. A
nitrogen blanket is typically employed to prevent atmospheric oxygen ingress into the tank. The nitrogen blanket does not to remove the oxygen from the water.

Perhaps some dissolved gas came out of solution and caused the piping to become air-bound for a short time.

Would not think that changing the gas would affect the pressure.

 

[Latexman]

Or, the tank vapor space with it's H2, O2, and N2 found an ignition source (static, a high temp, an arc somewhere) and underwent a deflagration. The N2 would have buffered/slowed down the flame front. Sounds like it wasn't close to a detonation. It may have been on the edge of fuel rich/flammable. The H2 or O2 was consumed; whichever was limiting. The pressure went up initially with the heat and then went down, possibly into a vacuum, onced cooled. The flow stopped until things equilibrated.

Possible?

Good luck,
Latexman

Need help writing a question or understanding a reply? forum1529

 

[Compositepro]

Most likely your tiny flow of purge gas was leaking through the tiniest crack at the seat of your relief valve and this crack was temporarily sealed by a few drops of water or debris. The pressure had to rise to move the valve and reopen it.
Relief valves are not precision back-pressure regulators. Most are called pop-off valves because they are designed to be fully opened or closed.

 

[clay87]

Thanks for the responses.

- the nitrogen is used to "sweep" out O2 as well as provide a blanket. This is to keep O2 levels from building up too much in the tank and therefore keeps the levels in the water low.

- I don't think it was ignition. The normal arrangement is 100% H2 tank gas, the high pressure event did not occur until we had been introducing N2 for a while.

- Compositepro, we have been thinking along the same lines as you. However, we rely on this relief valve all the time to relieve the tank of H2 gas that is expected to leak into the water (the application is electric generator cooling). We haven't had this high pressure event occur under those normal H2 in leakage rates (say .25 ft3/hr).

 

[rmw]

I'm voting for hysterisis in the relief valve. Check its cracking pressure and its re-setting pressure. Anything that caused it to reseat for any reason would then put it in a position that required it to 'crack' whereupon it would arrive at its relieving pressure. You need more of a tank vent than a RV for this application IMO.

Also remember that as you are 'relieving' H2, N2 or whatever from the tank, that air is diffusing back into the tank. People smarter than me will have to predict at what rate, but it will happen. Laws of physics, and all....

Can the O2 that you are finding be being pulled out of any of the containment metallurgy. Water that pure is 'hungry' and will pull ions out of whatever contains it (except glass I think) in order to equalize with its surroundings.

If this is a generator H2 system, you aren't leaking air into the H2 gas past the oil seals are you?

rmw

 

[bimr]

What type of process are you operating?

 

[rmw]

The post prior to my last post states that the service is an electrical generator cooling therefore my comments are based on the assumption that the cooling water is for the generator bars (coils), and that the generator is hydrogen cooled. I may have jumped to conclusions.

rmw

 

[clay87]

Correct rmw, the water cools stator and hydrogen cools rotor. Not sure the source of O2 but we don't suspect air is leaking through the seals for various reasons.

You wrote "air is diffusing". I'm not sure what you mean by this...back through the relief valve??? Fwiw, the vent/relief line has a loop seal in it.

 

[rmw]

Do a search on this site on the word 'diffusion'. Quite a discussion some years ago that I was in the middle of. If your loop seal is exposed to atmospheric oxygen at the outlet end, then that water (assumption - maybe if it is water you need to use oil) absorbs atmospheric air and that air can be given up to the gas upstream of it which can then diffuse back through the RV into your tank.

Diffusion is a very good thing when it comes to the CO2 we breath out not surrounding us and suffocating us, but it is a troublesome thing when you are trying to keep air out of a vessel with flow to atmosphere.

I am a ME and diffusion is "black science" to me, but I learned about it dealing with Zirconium O2 analyzers many years ago and have found the effects in lots of places I would have never suspected since that time.

rmw

 

[clay87]

Thanks rmw, I see what you mean. I'd like to focus on my original question:

Is it possible that switching cover gas from H2 to N2 causes a weird transient/event that would lead to tank over pressure and/or flow reduction. Such as H2 "fizzing" out of the fluid and causing a lot of bubbles.

 

[bimr]

As the attached article states, the oxygen is entering the system with fresh DI water, from a surge tank open to the atmosphere, and possibly from the open loop seal. The oxygen is typically absorbed from the demineralized water by a reaction with the copper in the system, not from some "sweeping" action in the surge tank.

The nitrogen blanket is not causing any removal of oxygen from the water. Nitrogen is used to prevent the water from contacting and absorbing oxygen from the atmosphere.

At the low pressure that you are operating at as well as with the high temperature, there should not be significant concentrations of gases in the demineralized water. And you are not releasing enough pressure to cause the gases to fizz. You have stated that you have only 3 ppb of oxygen.

The blanketing gas that should be used is nitrogen, not hydrogen for obvious reasons. You should upgrade the gas bleed off device as a pressure relief valve is not appropriate for this application.

Evaluate the use of a membrane device on the fresh makeup water to remove the oxygen.

I agree with you that the relief valve probably hung up causing the observed pressure transient. The pressure transient had nothing to do with the nitrogen gas.

http://www.powermag.com/water/Forgotten-water-Stator-cooling-water-chemistry_235_p2.html

 

[25362]

Purity of N₂ gas?

 

[clay87]

Thanks for the responses. For the purposes of this thread, I am not interested in the source of O2. If it were easy to identify and resolve we would have done that by now. H2 is used as the normal tank cover gas because it occurs "naturally" due to the system design. I only mentioned the presence of dissolved O2 as context to why we are introducing an external gas to the tank to feed and bleed. We chose N2 for this purpose for obvious reasons. Injecting plant N2 into the tank was effective in controlling dissolved O2 levels presumably by sweeping out built up air/O2 from the tank (allowing the water to deareate).

It is my understanding that as we change the composition of gas in the tank (isobaric), some dissolved H2 will come out of solution. Likewise, some N2 will go into solution at a different rate. Based on the responses I've seen, any system effect from this should be minimal. Sound about right?

 

[bimr]

How does the backpressure of the hydrogen gas compare to the backpressure of the new nitrogen gas supply? If the relief valve hung up previously when operating on hydrogen, what would the resulting pressure spike have been?

Perhaps the set pressure on the nitrogen regulating valve is too high.

 

[clay87]

I started off with a more theoretical question. I'll add some more info on our actual event:

We orignally started of with the N2 injection regulator set at 6 psig providing 2.5 SCFH, which was able to improve O2 levels to about 6 ppb (tank pressure is 5 psig). We did this for over a day. Then, in order to drive the dissolved O2 levels even lower, we increased N2 pressure regulator setting to 7 psig and throttled our injection rate to 7.25 SCFH. Seven hours later, we noticed high tank pressure (7 psig) and low system flow rates (based on water temperatures). After isolating the N2 supply, the tank pressure returned to normal and system temperatures returned to normal. We have since replaced the relief valve and proved that it is capable of exhausting 7.25 SCFH at its cracking pressure (5 psig).