plate exhanger reboiler level control

[mecmanmike]

Hi,

We have recently tried to commission a small portable process where we flash EG/Water mixtures in a small brazed plate exchanger. Firstly, forgive me if my chemical engineering vocabulary is lacking but I am a lowly mechanical engineer. The heat side of the exchanger uses a closed loop high boiling point glycol based heat transfer fluid that circulates through one side of the heat exchanger at 150 C (300F) (inlet) at flows up to 20 USGPM. The process side of the exchanger is designed to run at just less than 1 USPGM. On startup we flood the process side of the exchanger and slowly bring up our temperature on the heat transfer fluid side. Everything appears to work well for a while until we get an unexplained drop in our discharge temperature (unable to maintain 280 F) on the process side. We're thinking that as we don't have a proper level control for the fluid in the exchanger and the exchanger may be slightly oversized, that we are trying to inject fluid into a dry hot exchanger which does not give us the proper heat transfer.

Can anyone give any guidance as to what we should be doing for a plate exchanger level controller given process conditions up to 150 C (300 F) and 25 in Hg? Are my assumptions regarding level correct for a plate heat exchanger?

Thanks,

Mike

 

[katmar]

Mike, it would be unusual to have level control in this type of reboiler. If you had a kettle type reboiler I would expect a level controller, but with a shell & tube or plate exchanger the reboiler is normally run flooded. The circulation is normally by "thermosyphon effect", which is by having one leg at a lower density because of the bubbles - the way a coffee percolator works. But it sounds like you are pumping the fluid into the reboiler?

It is possible to get the effect you describe in an oversized exchanger. This is called "vapor blanketing". Normally there is much more liquid than vapor circulating and you get small bubbles of vapor forming on the heat transfer surface and then carried away by the liquid. You get very high heat transfer coefficients with this mechanism, but once the heat transfer surface is coverered with vapor the heat transfer reduces rapidly and it takes a while to flush the vapor out. This can lead to surging in a thermosyphon reboiler. Have you tried reducing the temperature of the circulating glycol?

The problem could also be a high pressure drop in the return piping from the reboiler to the column or flash vessel. If you are using thermosyphon circulation, this piping must be sized for more than just the vapor generated. There could easily be 10x (by mass) as much liquid as vapor flowing through this piping.

Harvey

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[reena1957]

Dear mecmanmike,
Pl see the following threads for some ideas:
124-66970,124-95533,124-100423 and the latest 124-172116
Best wishes

 

[mecmanmike]

Hi,

Perhaps I should point out that this is not a reboiler. It is a feed pre-heat that has been sized to vaporize all of the feed (i.e. we do not re-circulate anything back to this exchanger). We've been told by the supplier that no one has ever tried to use a welded plate exchanger for this type of application. The problem I see is the way the exchanger is plumbed there is no way to guarantee that we are maintaining any kind of liquid level in the exchanger and if we don't have our feed flow rate exactly right, we either flood the exchanger to the point where we don't get the vapor fraction and temperature we are looking for or we lose all of our fluid and it the exchanger seems to vapor lock. We've been able to operate the process for a few hours until we hit either extreme and lose our temperature setpoint to the first separation drum.

Again any multi-phase fluid experts please throw in your two cents.

Thanks a million for the responses.

Mike

 

[katmar]

Mike, you are confusing me. In your last post you said "sized to vaporize all of the feed", and then later in the post you said "we don't get the vapor fraction and temperature we are looking for". Unless the vapor fraction you are looking for is zero, these are contradictory statements.

I cannot see how it would be possible to put level control on this unit. The only way I can see it working is to have flow control on the process feed. You will have to start at low process feed flow, and gradually increase it until you get to where you want to be. Hopefully your supplier sized it on the basis of the pressure drop assuming all the feed was vaporized. If not, you could well get a "vapor lock".

Harvey

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[reena1957]

Dear Mike, Why dont you put a control/throttle valve downstream of the exchanger and reduce the degree of vaporization. Then after the control valve the pressure drops and everything vaporizes by itslef not affecting the exchanger, vapor locking, etc.
Best wishes

 

[JOM]

Mike,

Most plate PHE applications operate with a flooded unit on both sides - process and heating medium. Level control inside the unit is not an issue.

Do you actually need a vapor space? If so, you're sort of operating it as a plate evaporator.

J.

J.