Mixing of flashing water with cold water 2

[GeorgeGrande]

Case description:
Saturated brine (9 bara) is flashed (5 bara) over a valve. The two phase fluid is than mixed with cold brine (<100oC) discharged from a heat exchanger.
What we see (or hear) on field are heavy and noisy bumps that seem close to take the pipe out of the concrete.
1. Are there any flow elements that can help with that kind of mixing?
2. Are there any rules of thumb for the upstream minimum length for the two phase before it mixed with the cold fluid?
3. Any reference/paper deals with that kind of mixing?
4. Any good field experience?
Thanks

 

[MortenA]

It sounds more like cavitation to me - but are you surfe its not from the valve?

Best regards

Morten

 

[rustbuster]

Sound like hydrualic shock (water hammer). Perhaps an injection quill would help?

 

[GeorgeGrande]

MortenA
It will be not responsible to say that I am sure that it is not from the valve. The all line was taking the bumps and the valve is very close to the mixing point. It is very hard to identify if the bumps are from the valve or from the mixer. For the mixer, and that goes for RustBuster as well, we use common mixer with the cold fluid line is inserted in the main line tangentially.
My questions to you still stands.

 

[ione]

Is it possible to temporarily exclude the cold water injection just to evaluate whether the problem resides in the pressure reducing unit and not in the mixing phase?

 

[StewPad]

Sounds like you create a slug flow - I would move the mixing point before the valve (reduce the amount of flashing) but if you can't do that (pressure issues) I had a good experience with spraying cold fluid in a 2 phase separator vessel (similar to direct condensing)
Tell me if you need a draft.

 

[MortenA]

George - my point about cavitation was that IF it is cavitation then you cant avoid it by "mixing" - ithe the vapour pockets collapsing due to the cooling when mixing with cold water. This cannot be avoided whit smart gadgets. Mixing in a vessel as mentioned may do the trick.

But to test my other theory about the noise comming from the valve you would only have to turn of the cold water for a while - if possible. Then the noise should move "downstream" the cavitation coming later due to slower cooling.

Best regrads

Morten

 

[GeorgeGrande]

Thank you all for your suggestions.
The mixing should be done inline so vessel is not applicable.
I will try to arrange a test to check IONE and MORTENA suggestions.
For the sake of good practice, let’s consider the following options:
1. Flashing the saturated fluid on a valve and enlarging the pipe so the flow will become annular (assuming enough straight run). Than injecting the cold fluid.
2. Same as one but not enlarging the pipe to allow dispersed flow for better mixing.
3. Injecting the cold fluid just after the flash valve to avoid two phase flow regime development.
After deciding on the best location what is the best way:
1. Type of mixer.
2. Inject the cold fluid from bottom/side/top of the line.
Further information that I dropped is that the cold and saturated fluids flow is similar (by weight) and lines are 12”.

 

[Latexman]

I agree with rustbuster, sounds like classical "water hammer". The sat'd steam comes into direct contact with cold brine. There is very little/almost no resistance to heat transfer between steam and liquid water, so the steam collapses violently.

Good luck,
Latexman

 

[ione]

Re-reading more carefully the whole thread, I’ve realized I had disregarded this statement:

close to take the pipe out of the concrete

I have to admit that the phenomenon theorized by rustbuster and Latexman is very likely to happen (kudos to them). It is defined as condensation induced waterhammer. Condensation promoted by cold brine and flashing brine mixing is associated with a fast volume reduction. Consequent depressurization leads to voids formation, which are rapidly replenished by the liquid phase. The resulting collision of liquid phase provokes surge.