Problems with heat trasnfer of heat exchanger 1

[MedicineEng]

Dear All:
I am facing a big issue regarding one heat exchanger that I have installed in my plant. Let me just give you some insight about the system:
This is a monofluid system (secondary system) that has 3 different tanks (and respective HE and pumps) at different temperatures.
The one that is in negative temperatures is the one that is giving me a lot of problems due to the low thermal transfer.
One side has brine at -22/-24C and in the other has the fluid that will will cooled and pumped to Production areas.
The problem is as I already said the low temperature transfer across the HE. These were the actions that we already performed with reduced/minimum improvement:
-Cleaning of the HE in both sides;
-Checking of the brine modulating valve (it is operating fine with no restriction);
-Checking of pumps (Ok);
-Checking of brine pumps (Ok, since I don't have complaints about any other area fed by the pumps)

The brine system is able to withstand the thermal needs since the brine temperature is always in the range.
When we are in heavy production the temperature can only go down around 2C ( from -13 to -15C approx)that means a difference of at least 7/8C to the brine side.
Am I missing something?
Thanks for the help.

 

[25362]

Is it possible that the approach of the brine outlet temperature to -15 deg C on the process side is too narrow ? What correction factor for the LMTD are you obtaining ? What type of exchanger is being used ? Is the process-side prone to fouling ? What type of brine is being used ?

 

[MedicineEng]

25362:
Thanks a lot for your input.
Let me just give you some more information regarding this system:
This particular low temperature system has given us a lot of problems in the past. By that time, we were convinced that the problem were the chillers that were quite bad and inneficcient. Last year we replaced the chillers and installed some temperature probes in the system to monitor. The fact is that the brine temperature doesn't go up even in the top of the usage what leads me to conlcude that our chiller si able to cope with the demand.
Maybe I already had that problem before but it was hidden due to the low performance of the chillers. Now it came up.

You mention the "approach of the brine side is too narrow" let me see if I understand what you mean: You are refering tha thte diameter of hte pipe taht feeds the HE in the brine side is too small? Well it is according with the design and Idon't think that it is small.

As I said in my other post, we cleaned both sides of the HE (the procees side was not so dirty);

Sorry for my ignorance but what is "LMTD"?

The HE is a type of shell and tube where the tubes are squeezed and have an oval shape.
the brine is a mixture of methanol and water (around 50/50)
Thanks a lot for your help.

 

[ddot]

Hi MedicineEng

I might not be the best to comment on this, but here we go. First of all, LMTD is the acronym for the "log mean temperature difference," referring to the temperature difference between your two working fluids. Conceptually it refers the effective temperature difference

It is calculated using the equation:
LMTD = (A-B)/(ln(A/B)

where A = temperature of hot fluid in - temperature cold fluid out

B = temperature hot out - temperature cold in

If as you said you've cleaned the insides and outsides of your tubes and the transfer is not occurring as it used to occur, then I have two suggestions. It seems that if the system worked under the same conditions as originally installed and has recently been cleaned, you should check the flow rates of your two fluids. If the flow rates have changed, you could be trying to work the exchanger outside of its original design parameters.

On the other hand, if you've never had this system perform well, then it is quite possible that the design is inherently flawed. The approach that 25362 is referring to is the appraoch temperature difference of only 7 degrees C. I'm working on obtaining/designing a heat exchanger with similarly low approach temperatures. A lot of heat exchange area is required to perform low temperature difference heat transfer. Although I do not personally design heat exchangers, as I am coming to understand them, an inadequate design, in your case, would be lacking one or more of the following:

Sufficient heat transfer area (provided by the tubes) to meet your needs.
Sufficiently Turbulent flow (to allow for convection within the two fluids)
Improper flow type (parallel,cross-current, or countercurrent)

Of these things, you can only really alter the flow velocity (unless you have the ability to physically/intellignetly add more tubes!). If you increase the velocity, you probably will crease the turbulence, which would be good if there is a film effect preventing heat exchange. Otherwise, lowering the velocity might help to increase the time for the fluid to transfer heat. In any case, I hope that its a matter of changing the working conditions and not one of replacing your exchanger.

As a side note, if you do decide to replace your exchanger, consider plate heat exchangers as they are sometimes recommended over shell heat exchangers for lower approach temperatures in the temperature range you're working in. This is contingent on sufficiently low working pressures, however.

 

[25362]

LMTD is the logarithmic mean temperature-difference in the expression

Q = UA LMTD​

Where Q is the heat duty, U is the overall heat transfer coefficient, A is the heat exchange surface.

It is calculated by the following equation: ([Δ]t₁-[Δ]t₂) [÷] ln([Δ]t₁[÷][Δ]t₂)

where [Δ]t₁ and [Δ]t₂ are the terminal temperature differences.

In a regular multipass heat exchanger part of the flows are cocurrent and part countercurrent. As a result the LMTD is corrected by a factor F, which = 1 for pure countercurrent flow. The more the flow is cocurrent the lower F, and the efficiency of the exchanger comes down.

There are formulas and computer programs to calculate F, but graphs are generally used. Any book on heat transfer bring these graphs based on the terminal temperatures.

By approach I mean the temperature-difference between both
streams.

 

[panduru]

@MedicineEng:

I feel you must seriously consider referring to a qualified design engineer who will assess the actual working parameters vis-a-vis the design of the equipment.

If lucky, you may only have to regulate the flow pattern to achieve the desired results. Otherwise, you may have to replace the exchanger with a suitably designed one.

I am not sure if familiarizing yourself with the theory of heat transfer will be of much help in this instance (if you are keen, you may gain some valuable insight from Kern*).

(*) Kern, Donald Q., Process Heat Transfer, McGraw-Hill Book Company, New York, 1950

K Panduru

 

[25362]

I fully agree with panduru. Let me add that the LMTD can be replaced by the arithmetic average [&frac12;](?t₁+?t₂) with an error of less than 1%, as long as 1<?t₁÷?t₂<1.4.

 

[StoneCold]

Medicineng
I also have a "brine" system running methanol. We are running between 75 and 85 vol/vol% methanol at about -40C.
We run into problems on dead legs. Long or short. There seems to be a phase separation there and the water will freeze out of the system. Not a huge issue except it happens on standby pumps.

I would add that you should measure the inlet temperatures at the exchanger and the outlet temperatures at the exchanger. This should tell you something about your exchanger set up. If your "brine" is coming out cold then you probably don't have enough area. (I am assuming that you are pushing enough flow through the exchanger to get turbulent flow.) If your brine is comming out hot then you probably need more flow.

Another thing you could try is to reduce the temperature of the brine. Our system runs at -40C but we generally never bring our processes lower than -20C. -30C is achievable in our systems but it is slow. A comparable -33C (7C approach) would be very difficult. Not impossible but slow enough the we would resort to addional cooling from Liquid nitrogen injection. (Time is money).

I hope this helps some.

Regards
StoneCold

 

[25362]

When the chiller was cleaned on both sides, what kind of fouling has been found and removed ?

One should hope the methanol-water mixture or the process fluid have no dissolved gases which are released in the exchanger impairing heat transfer.

 

[MedicineEng]

Guys:
Thanks a lot for your ideas. Just to reply to some doubts here raised:
-When we cleaned the HE in the brine part we found rust and in the process part we found also some rust but not significative;
-Lowering the temperature of the brine is not an option. We are already running the chiller in its lowest point;

From the design documentation for this HE I got the following information:

Circuit A (process side)
Flow: 114400 Kg/h;
Temperature (in/out): -10/-20C

Circuit B (brine side)
Flow: 118400 Kh/h
Temperature (in/out):-25/-19.79C

Thermal load of the HE: 587888 W

So, by design, I should be able to get 10C difference in teh process side, but what I get is maximum 6C and this only if I slow down the production process. If I maintain the same speed, then I only have like 3C across the HE and my process side remains about -13/-15C.

 

[StoneCold]

So if you have all this information then what is the assumed U value and what was the design area for the heat exchanger. Did you check the design area against the heat exchanger drawing or model numbers? Are the design exchanger and the actual exchanger both running counter current?

Regards
Stonecold

 

[25362]

Please note there is a temperature crossing in the original design. Can you detail the type of exchanger ?

 

[MedicineEng]

The heat type is a platular heat exchanger made by Barriquand model IIXS20+21/2+20X1750X430.

25632:
You point out a very interesting issue that I never noticed. There is a temperature crossing. Can this be possible?

 

[25362]

I've heard about, but not seen yet, a platular HE (combining plates and tubes). I assume it should be able to provide a narrower temperature approach than a conventional shell-and-tube unit, even with a degree of temperature crossing.

I suggest you consult the maker on this issue, especially on whether there could be a possibility of one of the fluids by-passing or short-circuiting part of the unit.

As a tutorial I suggest:

http://www.cheresources.com/heat_transfer_basics.shtml

 

[speco]

MedicineEng,

First, it looks to me like the application you have is nearly ideal for a plate and frame heat exchanger. This is especially true since it has a temperature cross-over, both fluids have relatively low viscosities and fouling potential, and the flows are of the same order of magnitude on both sides. I presume that the particular one you are using is single-pass on both the process side and the service side. You can confirm this by just looking at the front plate. If all four nozzle are on the front plate, then it is single-pass on both sides. This is the most common configuration for a PHE.

I suspect that the problem you are having is the result of one or two things:

1. The actual flow rate of the process side has increased far beyond the design, or conversely, the flow rate of the service (alcohol/water mixture) is reduced.

2. The heat exchanger was undesized in the first place. This is not terribly uncommon in the PHE world. Most PHE manufacturers refuse to include fouling factors in their designs, opting instead to use a "safety factor", which is usually in the 10-15% range or so. However, sometimes they will shave their ratings in order to be competitive. If you are able to confirm that your flow rates on both sides of the exchanger are close to the original design parameters, I would go back to the manufacturer and demand some additional plates to make up for the loss of performance. Typically the frames are designed with room for extra plates. They may tell you that it's out of warranty, in which case you would be better off to buy a replacement from an competitor. The price they usually charge for additional plates is usually so high that a complete new exchanger is a better deal.

Regards,

Speco (

http://www.stoneprocess.com)

 

[chicopee]

From what I read into your problem, it seems that your system is OK under normal production condition. Then under heavy production requirement, the working fluid does not reach that -13/-14 C range which means either that flow of the working fluid has increase or inlet temp. of the working fluid is much higher that during normal production.
It seems that this low temp HX is underdesigned for peak production.
Put some temp and pressure guages on the inlet and outlet of the working fluid lines and check pressure and temp drops for comparison w/ system specs.

 

[MedicineEng]

Thank you all for your valuable comments. We made some more tests and now we are lloking in other direction. We realized that our thermal monofluid has a small ammount of water (still to understand where it came from).
We are suspecting that this small ammount of water, that it is unmiscible in the thermal fluid, when it reaches the HE it freezes inside, creating a layer of ice that prevents the thermal exchange.
We are now trying to arrange a way to remove this water from the system and confirm is this was the cause.

 

[Yorkman]

Is there a way to run the solution through a cold trap (using dry ice and alcohol) and seperate it from the solution, leaving it in the cold trap as the thermal fluid runs through. Another approach would be to distill it, by boiling it off from the thermal fluid, depending on the boiling point of the thermal fluid. One last approach would involve dessicant dryers like we use in refigeration systems, these come in replaceable core driers and are installed in line with a bypass so the system can stay on line while the drier cores are change. Good luck

I'm not a real engineer, but I play one on T.V.
A.J. Gest, York Int./JCI

 

[MedicineEng]

Yorkman:
Thanks for your tips. We are considering for now to distill to see if we can remove some water.
During our shutdown, we were already speaking on getting some dessicant to see if we could trap the water.

 

[abeltio]

make sure the HE is properly vented. Air pockets will reduce thermal effectiveness dramatically.

saludos.
a.