Vacuum Deaerator Tower Troubleshooting

[meanyunny]

My plant has a vacuum deaerator tower (deaerates by reducing the seawater [partial] pressure in the tower using vacuum pump while seawater flows through packing, where entrained air will evolved from the water). The water discharge from the tower is connected to a HP centrifugal water pump. System has been running for around 3 years.

Lately we discovered that the discharge line of the tower (suction line of the water pump) shows vacuum condition after few minutes starting up the tower's vacuum pump; i.e. We opened up a sampling point at the pump suction line, initially before vacuum pump is running (while water pump is running), water continuously flows out from the sampling point. However few minutes after starting up the vacuum pump and as tower's pressure drops to ~0.05bar (normal operating pressure of tower), the water flow from the sampling point gradually decreases and eventually vacuum is observed (suction could be felt at the sampling point). While this was happening, however, all parameters remained and maintained at normal operating condition, i.e. the inlet water flowrate of the tower, tower pressure, the bottom water level of the tower, and the discharge of the water pump (flowrate and pressure).

We would like to know if this is possible or normal? Is it possible that the pressure of the water in the tower discharge line is so low that water would not be able to flow out from the sampling points (2 points - 2inches and 3/4inches; sampling outlet to atmospheric)? But yet in fact there is actually still continuous flow of water in the pipelines?
Previously, water has always been able to flow out from the sampling point, until recently after a servicing of the vacuum pump. we are unsure if these events are related.

 

[georgeverghese]

Yes, if tower top pressure is 0.05bar abs, then there should be no water flow out of the sampling point. The de aerated sea water injection pumps then boost this sub atmospheric stream up to injection pressure.

 

[meanyunny]

Hi georgeverghese,

Thanks a lot for your response. It seems like you have experience in similar system since you could tell it is a seawater injection system right away.

We are currently concern because previously water could flow out from the sampling points (which is also required for continuous sampling of the water for some parameter monitoring). We still could not understand why would this happen now suddenly, while the tower's pressure seems to be operating at similar sub-atm pressures as before (and also all other parameters).

Do you foresee any possible issue that may arise? or this is still considered perfectly normal as from your experience in similar systems.
Kindly advise.

Thanks.

 

[georgeverghese]

I've worked on the process design audit for a seawater injection system some years ago.
If you want to enable a sample flow from the pump suction side, you could try compensating for the sub atmospheric pressure by extending the sample line to some lower elevation.

Otherwise, sampling from the pump discharge side is also possible, but maybe the pressure is too high here.

 

[georgeverghese]

Going down to a lower elevation may not be possible in some cases if you dont have some 8-10metres to go down - if so, some kind of small flow sampling pump may be the only solution.

 

[itdepends]

Do you need a continuous flow for sampling? If yes- then you'll need to install a pump or run the line lower to get one.
If you just need to sample more regularly- construct, what we would call a sample bomb.

A small vessel or pipe section with small valves on each end that match your sample valve size.

Sample point must be on the bottom of your pipe to be sampled (or at least below horizontal)

Connect pipe section/sample bomb to the sample point
Open sample valve and top bomb valve, bottom bomb valve remains closed
Air in the bomb will flow up into the pipe and the water will flow down to fill the void. The sample bomb is now at the same pressure as the pipe
Close sample valve and top bomb valve
Open bottom bomb valve and collect sample
Close bottom bomb valve

Repeat the above if you want to ensure you have a representative sample (i.e. by flushing the bomb)

The top bomb valve is redundant- you could just have sample valve--bomb---bottom sample valve

As a chem eng/metallurgist the first part of any answer I give starts with "It Depends"

 

[georgeverghese]

For a fast loop sampling line, consider also a regulated low flow slipstream from the HP injection pump discharge back to pump suction. This line can be fitted with a max set PCV on a line parallel to the sensor, and a PSV on another parallel line for pressure protection also.

 

[LittleInch]

meanyunny.

Your question which included this "Previously, water has always been able to flow out from the sampling point, until recently after a servicing of the vacuum pump" shows to me that the tower previously was not operating at your design level of 0.05bara. Note the bara bit, i.e.it is operating below normal atmospheric pressure of 1bara.

You don't say how much water is above your sample point ( i.e. water level in the sump) to the sample point. If this is less than approx. 10m, then the air pressure in the atmosphere is more than the pressure of the fluid so air will flow in.

I'm struggling to understand why you don't just connect the sampling to the outlet of the pump?

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

 

[meanyunny]

Dear All,

Thank you all for your responses. Greatly appreciated.

FYI to all, sampling is meant to be continuous, where it is connected to an analyzer/meter which is also tied into the IPS system.
Currently not looking into design change or adding in pipelines, but great thanks to the ideas provided.

To George, I believe my colleague Francis recognized you and he sends his regards

To LittleInch,
Tower pressures previously from historical trending of past year do show that it Is operating at required operating conditions, i.e. sub-atmospheric of around 0.05bara. As mentioned in the original post, all parameters maintained almost the same from previous operating conditions, i.e. the inlet water flowrate of the tower, tower pressures, the bottom water level of the tower, and the discharge of the water pump (flowrate and pressure). Therefore it is alarming that this is happening only now but not previously.

Tower water level is around 4.5meters, and sampling point is around another 5m from bottom of tower. We now do understand that yes there is a possibility that the condition observed is due to the lack of head and the point's pressure is sub-atm. However, as mentioned repeatedly, we are still alarmed as previously this has not been the case and we have always been sampling from the point continuously.

We are indeed looking into rerouting the sampling point to the pump discharge line if all fails. However this involves design change and some time. Thanks a lot for your suggestion nevertheless.

 

[EmmanuelTop]

Maybe there is an obstacle to flow in the sampling loop and the water inside the tower doesn't have enough head to overcome the barrier. Since this is not the problem from day No.1, obviously you have done something differently in the way the system is operated or there are mechanical issues in the sampling loop. Maybe there are other options as well, but without sketch and drawings it is hard to troubleshoot.



Dejan IVANOVIC
Process Engineer, MSChE

 

[LittleInch]

The issue might be a very small change in vacuum pressure on the tower. If there was any sort of positive head at the sample point, even 10cm, you would get flow out of your sample point. If this has changed to a negative head of 10cm, you will get air flow in. A change of 20cm is 0.02bar. How accurate is your monitoring of the pressure and can you see this sort of change from before and now??

The same applies to your level control in the base of the tower. If this is now only a little bit less than it was before, then the same applies. At 9.5m head with a top pressure of 0.05bara, you are right on the limit of being at 1.0 bara at the sample point. Similar a change of only 20cm in the water level in the tower will make the difference between water flow out and air flow in. Again how accurate is your measurement of this and has it changed from before to the amount noted (20cm)??

If you can squeeze your level control up a little bit or watch the level to see what it is doing then you might be able to solve the issue there.

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

 

[sshep]

Dear Meanyunny,

If this is a new problem, consider that you have an air leak at the pump seal or in flanges on the suction line. The elevation required to keep the suction at positive pressure assumes clear liquid. The instant air gets pulled in, there is no longer sufficient clear liquid head above the leak, it will not correct itself. This can happen with a level bobble, or start-up filling of the line.

This is just a suggestion as a common troubleshooting finding for such systems.

best wishes,
Sean

 

[georgeverghese]

Check if the level control setting suggestion will help avoid this design change. Also check that the pump shaft seals cater for subatmospheric suction - else air may get pulled in through the seals.

 

[LittleInch]

Any update?

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

 

[LittleInch]

Any update?


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

 

[don1980]

Experience has taught me the value and validity of Occam's Razor.

Perhaps the velocity in your pipe is higher than it was before, with the result being that you have a vacuum at the sample port. This is how an eductor works.