Heat Transfer through Helical Coils - Concerns

[Iapyx]

A concept for the room:

We are working with a continuous flow process that, for efficiency purposes, requires very significant (delta T ~350F) heat exchange between two high pressure (1600-1700psig) streams. Each stream is a viscous fluid (85cP cold, ~60cP hot), is in a state of laminar flow and is at a low flow rate (15gpm in 2" sch160 CS pipe). Additionally, too low of velocities present problems with fouling and too high of a pressure drop presents operational problems.

After looking into several heat exchangers (shell/tube, plate, tube/tube, etc) it is apparent that in addition to the excessive pressure drops required to maintain acceptable velocities (>2ft/sec), the required quantity of HEXs to meet the necessary square footage is massive and unrealistic (this is true for exchanging the heat directly between the two fluids or through an intermediary heat transfer fluid, e.g. oil synthetics).

The proposed solution is a set of vessels with internal helical coils and agitation. The coils carry the streams that are transferring the heat and the vessel is filled with a heat transfer fluid, thus immersing the coil. The heat transfer fluid is agitated and constantly recirculates between the two tanks in order to transfer the heat. We are looking into static mixers for inside the coils, but have not reached any decisions on what would be acceptable. Calculations and some small scale lab testing indicate that the coil length is significantly shorter than the length required by any of the HEXs.

Our team has been immersed in the design and review of this section of our process for months now and feel confident about this potential solution. Our concern is that our self immersion might limit some level of objectivity and a failure to see the proverbial forest for the trees, so to speak. Can anyone see a fallacy in the premise of this design? Has anyone had experience with similar challenges or a similar design? Any comments/criticism is appreciated.

 

[moltenmetal]

Heat exchange between two viscous, high pressure streams at a low flowrate: I presume you're doing this for energy conservation reasons?!

At 15 gpm, is the energy you'll save via this cross exchange really worth the cost in capital and operating/maintenance of two high pressure agitated vessels with internal coils, and all the associated equipment, instruments, valves and controls? The energy to drive the agitators will not be trivial, nor will heat losses from the vessels themselves be negligible. Is the NET energy exchanged worth these costs? If not, it would be easier to heat one stream and cool the other with utilities.

If low velocities lead to fouling (i.e. for understood reasons, not just because someone thinks that 2 ft/s is a minimum velocity for any kind of exchanger), it would seem that you have few choices other than the agitated vessels to accomplish the cross exchange. If you go with utilities, you can put the viscous process fluid in tubes with internal static mixers and the utility in the shell.

If fouling is NOT a real problem, then you could go with either a tube in tube exchanger with static mixers in both the inner tube and the annulus (yes, such things can be made- talk to Chemineer Kenics or Sulzer etc.), or just a really long, laminar flow, tube in tube exchanger. In the case where the shellside is not controlling, the Nu = 0.023 Re^0.8 Pr ^0.33 correlation reduces to a minimum required LENGTH of tube and tube exchanger when flow becomes laminar- you'd have to check what happens when both shell and tubeside film coefficients are important and both are laminar. Heat exchange in laminar flow is still possible- it's just very inefficient in the use of exchanger area.

 

[davefitz]

It would seem that fouling would be the dominant issue with the coils as well. The heat transfer limits that would prevent fouling will need to be copletely understood, for both fluids.

While the agitation of the intermediate HT fluid helps as far as theoretical heat transfer, it does nothing to prevent the buildup of a laminar layer on the viscous fluid sides, which worsens the fouling propensity.

If the fouling can be prevented by limiting the temperature that occurs in the boundary layer on the viscous fluid side, then the use of a 2 phase intermediate HT fluid , with the 2 phase sat temp being less than the temp required for fouling, may be one design objective.

AS I recall, many tech papers on the science of fouling were published by E. Somerscales in the 1970-1980's.

Have you evalueated the use of plate & frame HX to overcome teh surface are issues?

 

[ione]

Take care to coils packing, as a too “dense” coil coupled to an agitated vessel could affect the mixing effectiveness (presence unmixed areas). Maintenance operations related to cleaning purposes could be a problem as well. Low velocity and laminar flow should also slash the heat transfer coefficient, anyway if you’ve performed lab tests and found it is acceptable, then no bother on that.

 

[Iapyx]

moltenmetal: Yes, the purpose of the heat exchange is for energy conservation and yes, it is worth it. Part of the reason for the coil in the vessel is that the vessel could be rated lower (approx. 150psig @ 700F with lower operating conditions) as the heat transfer fluid does not require the same pressures as the process fluid. And again, yes, the energy exchanged is still worth the cost of the mixers (in a large part because of the lower vessel design ratings.

Low velocities have proven that fouling occurs. The exact value is unknown, but it is known is that at 2ft/s no or negligible fouling has occurred (we are doing additional testing now to find that exact value, but, unfortunately, everything cannot stop while we wait for all data).

We did look into static mixing with a tube/tube exchanger (Sulzer and Chemineer two of many), but the issue then became pressure drop.

Davefitz: The reason we feel that fouling is less of a risk in the coil is that we can use the 2” sch160 pipe and obtain our proven velocity. We are testing the BTU transfer per square inch to see how that will effect fouling as well.

We have seen some studies on the helical coil that state the secondary flows in the coil create 2 “vortices symmetrical about a horizontal plane through the tube center.” (I only have a segment of this document and cannot fully cite) Purportedly this secondary flow disrupts the boundary layer and induces a certain amount of turbulent flow. There are some implications that some of this research is preliminary and limited in quantity.

We did look into plate/frame, but all manufacturing groups contacted did not want to deal with the pressure/temperature combination.

We partly need to see if there are any issues with the helical coil that are not intuitive, but also are looking to see if there are any other designs that are less than traditional means of heat transfer.

 

[davefitz]

There are other ways to induce turbulence within in the coil tubes,including:
a) extended internal fins, helical wound
b) Internal spiral ribbon turbulators
c) internal coil spring
d) rifled tubes.

 

[Compositepro]

The main problem I see is that for heat recovery counter-flow heat exchange is required. With the mixing vessel concept the heat transfer fluid will beat an intermediate temperature between the hot and cold streams. The hot stream will never get colder than this or the cold stream hotter. This will also reduce the delta T and therefore require more tubing.

A better and cheaper solution is to braze two tubes together length-wise and them coil them. Then run the flows counter-current. With viscous fluids the main resistance to heat transfer is the boundary layer in the fluid.

 

[moltenmetal]

OK, I was missing something because it never occurred to me that you would be agitating the LOW PRESSURE, presumably low viscosity, heat transfer fluid! You will still be tubeside limited, such that the shellside conditions don't matter much- the tubeside will entirely determine the required heat transfer area.

Is it the heat transfer fluid that is fouling and needs the 2 ft/s velocity? If so, consider a different heat transfer fluid...otherwise, shellside velocity won't matter much.

It's also scary to consider putting a 1700 psig tubeside coil inside a 150 psig MAWP shell, agitated or not. Tube rupture/failure will likely result in vessel rupture, regardless how big your relief valve is. You might get away with it in an open tank, but if you have to pressurize the shell side it's best to stick with the conventional rule of shellside MAWP = 2/3 of the tubeside- that's going to make your agitated vessel approach very expensive indeed.

Perhaps 2" inside 4" pipe and coil them both? It'd take a hell of a coiling machine since the 2" is sch160 and the shell will be at least sch80 if not heavier...

The point that has already been made about the uniform temperature of the heat transfer fluid in a mixed tank. Presume your 350 F dT renders that moot.

 

[gruntguru]

Tube-in-tube makes sense, with finning or helical turbulence inducers on both sides of the inner tube. The outer tube will require a lot of insulation - especially at the hot end.

 

[ione]

To expand a bit my previous post on the presence of unmixed regions. Empirical correlations usually adopted to evaluate heat transfer coefficients assume the temperature is the same all over the tank, but uneven mixing leads to temperature deviations from a presumed average value and this is likely to lead to an overrated heat transfer coefficients.

 

[TJOrlowski]

Tube-in-tube makes sense, with finning or helical turbulence inducers on both sides of the inner tube. The outer tube will require a lot of insulation - especially at the hot end.

Finning on the OD of the inner tube/pipe would make fabricating the coil nearly impossible. Turbulator wire would be much more feasible.

it is apparent that in addition to the excessive pressure drops required to maintain acceptable velocities (>2ft/sec)

When I read this, I had assumed that a double pipe heat exchanger had been evaluated. I'm unclear as to why a coaxial coil or double pipe heat exchanger would create pressure drop problems. If you're using two helical coils in a tank, and your velocities and pressure drops are where you want them, couldn't you just use a coaxial coil where the annular space between inner pipe OD and outer pipe ID was equal to the in^2 of the ID of 2" Sch.160 pipe at the cross-section?

-TJ Orlowski

 

[ione]

If you're using two helical coils in a tank

I thought the OP were using just one helical coil immersed in an agitated vessel

 

[Iapyx]

ione: Are you familiar with any research studies on stratification of heat throughout a tank (or, conversely, the uniformity of heat throughout a tank)? We have done lab tests - and are doing more - but we are trying to pay close attention to the challenges deriving from scale.

gruntguru: The challenge we ran into repeatedly with tube/tube exchangers is that over the length required at the tube size required (to maintain velocity) the pressure drop becomes significant. We have not thrown this out, but part of our fluid composition runs a high risk of fouling. Even through increased turbulence (i.e. static mixing) the pressure drop issue exists.

moltenmental: It is the process fluid - not the heat transfer fluid - in the coil that runs the risk of fouling. We are agitating the low pressure simply to get the highest efficiency out of that material's heat transfer potential and are now attempting to resolve the issue within the coil and laminar flow fluid. You mention (essentially) coiling a tube/tube exchanger, do you think this will improve heat transfer on both sides versus the transfer of a straight tube/tube exchanger.

 

[Iapyx]

ione: The concept we are focused on - and described originally - is a single helical coil immersed in an agitated, recirculating heat transfer fluid.

TJ: It was explained to us that because of the large surface area required by a double pipe HEX the number of elbows and tight passages amounts to a large drop by the end of the exchangers. We had not reviewed a coiled tube/tube exchanger.

 

[TJOrlowski]

Ok I guess I misunderstood the application.

requires very significant (delta T ~350F) heat exchange between two high pressure (1600-1700psig) streams.

The proposed solution is a set of vessels with internal helical coils and agitation. The coils carry the streams that are transferring the heat and the vessel is filled with a heat transfer fluid, thus immersing the coil.

Reading this, I thought Coil A in the tank had hot fluid, Coil B had cold fluid, and the goal was to transfer heat between A&B. Thus, both coils are immersed in a single agitated tank containing heat transfer fluid.

What's really happenining is you have two tanks, each with a single coil. Tank A has a coil with hot fluid, which heats the the heat transfer fluid, and tank B has a coil with cold fluid, which cools the heat transfer fluid. The heat transfer fluid from A is then piped to B, and vice versa? Am I understanding that correctly?

-TJ Orlowski

 

[Iapyx]

TJ: Yes, that is correct. The heat transfer fluid is being employed to take heat from the coil inside of tank B and then put that heat into the coil in tank A.

 

[TJOrlowski]

Have you considered putting two helical coils inside one tank, and transfer the heat through radiation?

Nest the coils inside one another, arbitrary numbers:

36" OD large coil

28-7/8" OD small coil

Then you could either put an agitator inside the small coil, or a shroud inside the small coil to help redirect radiation away from the natural centerline of both coils.

If it worked you could eliminate an entire tank, and much of the peripherals to pump the heat transfer fluid back and forth.

-TJ Orlowski

 

[moltenmetal]

Take TJO's idea of nested coils, but replace the "tank" with a sand mould which you then fill with molten aluminum, and perhaps you've got something. At least then there's no expensive vessel(s) and agitators...presuming you're operating well below the melting point of aluminum that is.

Will a coiled tube exchanger give you better performance than a straight tube if both are in laminar flow and hence entirely tube-side controlled? A coil does increase axial dispersion even in turbulent flow so there is some hope- but the effect is nothing when compared against the performance of static mixers inside the tube. Static mixers do not need to induce turbulence or develop huge pressure drop in order to accomplish the mixing you require to obtain better heat transfer. Static mixers are also really good at reducing stagnation and hence the tendency to foul at the wall. You're probably not fairly comparing a coil against a straight tube anyway- you're comparing a coil against a series of straight tubes inside separate shells, with a series of jumpers which connect the tubes in series. The jumpers will tend to introduce mixing to some degree as well- but again, not much.

 

[MJCronin]

You guys should consider a spiral.

There are pressure limitations.....but a spiral will handle a viscous to viscous application.

Call Alfa....

http://www.seterm.com/pdf/39spiral.pdf

-MJC

 

[TJOrlowski]

If they can design and build one to withstand the pressures, a spiral would work similar to the way a tube-in-tube helical coil would, but I shutter to think about the maintenance/upkeep/repair/replacement cost of one of those.

-TJ Orlowski