PDA

View Full Version : NH3 separator desing







Rio69
15-12-2006, 12:33 PM
Hi,
I am designing high stage NH3 separator. On ref. system there are booster compressors. Since discharge pipe from the boosters is going into the separator I have following question.
What is recommended velocity through holes on the discharge pipe where ammonia gas is going out into liquid ammonia in the separator?

Velocity through discharge pipe is cca 23m/s, but I believe that velocity through the holes should be much lower, otherwise can cause troubles for high stage compressors?


Thanks

US Iceman
15-12-2006, 03:35 PM
Well Rio69. That is a good question for your first post here.

Here is my thought. If the gas velocity is 23 m/sec through the discharge pipe into the intercooler any additional holes you provide in this pipe below the liquid surface will help to reduce the overall gas velocity.

You certainly do not want to cut too many holes as this may weaken the structural integrity of the pipe.

I am not sure if there is any "magic" number for gas velocity. At least I have not seen or heard of one.

As a best guess I might suggest holes in a vertical pattern (under the liquid level) at 90 degree intervals around the pipe perimeter.

One thing that may be helpful is a supression plate right at the liquid level. This is a perforated plate that helps to prevent the high velocity gas jets from raising liquid up into the gas stream.

I'm sure some of the other ammonia guys will have some comments for you also.

Josip
03-01-2007, 05:57 PM
Hi, Rio69 :)


Hi,
I am designing high stage NH3 separator. On ref. system there are booster compressors. Since discharge pipe from the boosters is going into the separator I have following question.
What is recommended velocity through holes on the discharge pipe where ammonia gas is going out into liquid ammonia in the separator?

Velocity through discharge pipe is cca 23m/s, but I believe that velocity through the holes should be much lower, otherwise can cause troubles for high stage compressors?
Thanks


Your question is very nice, but still today there is not any uniform answer with


I am not sure if there is any "magic" number for gas velocity. At least I have not seen or heard of one. exact values i.e. velocities.

As I remember for horizontal separator speed of gas and droplets must not pass 0,4-0,7 m/s at -10C evaporating temperature up to 1 m/s for -40C evaporative temperature.

For vertical separator all those values are different but I do not know about....

According to this, it seems you have to make a lot of holes;) in your submerged pipe/s to reduce outlet gas speed under this values. I think it is better to make more holes then needed ;) or even to split that pipe in two branches under liquid....

It is very important to make separator with good diameter (to have gas area above booster pipe to separate droplets from gas before entering into suction pipe down to compressor) and with good length to give a time to droplets (from system suction pipe) to separate from gas.

Of course all of this depend on evaporating pressure, system capacity... and most important is Mr. Budget.

Hope this is of some help

Best regards, Josip :)

US Iceman
03-01-2007, 07:02 PM
Hi Josip,

I think the numbers you are using for velocity are based on the actual separation velocities for the liquid droplets of a certain particle diameter.

Intercoolers are interesting vessels. With the lo-stage discharge gas flow into the liquid pool this creates a lot of disturbance and I think this contributes to liquid carryover.

Most of the older people I have talked to made a general recommendation of having an extra 2-3 feet (0.61 to 0.91 meters) of vessel height to accommodate liquid splashing inside the intercooler.

All of the hot gas flowing into this vessel is like installing a high pressure compressed air hose into a bucket of water. The water goes all over the place when the air is turned on!

Personally, I think the gas entrance velocity should be low, but as Rio69 asked; what is the velocity that should be used?

The risk of too many holes will make the pipe look like Swiss cheese, and I'm not sure of the mechanical integrity of the pipe if this is done.

Almost all of the intercoolers I have seen use single discharge pipe down to the bottom of the intercooler. The pipe opening is the only outlet for gas.

The inherent problem with this is when the gas bubbles rise, they take some liquid with them when they break the liquid surface. Sort of like dropping a rock in water... you get a splash.

If the vessel diameter is not large enough then the liquid flows out with the vapor. This is why I think the old-timers said to increase the vessel shell length. The extra disengagement length had a better chance to separate liquid.

This is another one of those interesting facts on ammonia system design, that not too many people question, but do not know where it came from either.:D

Josip
03-01-2007, 09:02 PM
Hi, US Iceman :)


The risk of too many holes will make the pipe look like Swiss cheese, and I'm not sure of the mechanical integrity of the pipe if this is done.

Almost all of the intercoolers I have seen use single discharge pipe down to the bottom of the intercooler. The pipe opening is the only outlet for gas.


When I was working in construction department we made a holes on the bottom of that pipe within 120 degree and cross section of all holes were much bigger then cross section of pipe itself just to reduce high speed (as Rio69 said 23 m/s) down. Pipe was like a letter L with longer part laying horizontally (I_____ ) and almost long as vessel. Unfortunately I do not remember about any "rule" how many holes we made.

We weld additional pipe supports and some of those vessels are in use still today:eek: .

Not in use thanks to our good job, but to luck of money for replace;)



The inherent problem with this is when the gas bubbles rise, they take some liquid with them when they break the liquid surface. Sort of like dropping a rock in water... you get a splash.

Above that booster pipe we install a pipe snake to subcool feeding refrigerant for low pressure vessel to make additional barrier to buoyancy bubbles.


If the vessel diameter is not large enough then the liquid flows out with the vapor. This is why I think the old-timers said to increase the vessel shell length. The extra disengagement length had a better chance to separate liquid.

There were some ratio between length and diameter of separator L/D but I forget what was that (maybe 4 or 5) of course for horizontal separators.

I think there are some ASHRAE recommendations about separator design.

But, again we have a customer at the end of that chain;) - asking: "please, can you reduce the price?"

Best regards, Josip :)

US Iceman
03-01-2007, 09:33 PM
When I was working in construction department we made a holes on the bottom of that pipe within 120 degree and cross section of all holes were much bigger then cross section of pipe itself just to reduce high speed (as Rio69 said 23 m/s) down. Pipe was like a letter L with longer part laying horizontally (I_____ ) and almost long as vessel. Unfortunately I do not remember about any "rule" how many holes we made.


I have done the same thing in horizontal intercoolers used in a skid mounted package. Exactly as you describe and it does work well. The start-up people said the superheat to the high stage compressors was quite reasonable.

I suspect we both did the same thing for the same reasons. Unfortunately, I do not remember what "rule" I followed either.:D

I just know it was a lot of small holes. In my case we provided a weld cap on the end of the pipe, so the gas had to flow through all of the holes and into the liquid. My thought was without the weld cap, the gas would take the path of least resistance.

The L/D ratio I remember is around 3-5:1 for best economy, but every increase helped the separation too.:)

It is always about price. Now if we can only figure out how to give away refrigeration systems and still make a profit. We would have happy customers.:rolleyes: