Heat transfer loss in laminar flow 4

[Ohlsson]

Hi everyone!

I have a specific problem and i hope this is the right forum.

I was wondering how to calculate the heat transfer capacity in a coil when the flow is:

1. laminar, Re under 2300
2. transient, 2300 < Re < 4000

Is there an easy method to determine the capacity when the flow is laminar if the capacity is know for turbulent flow?

Hope you understand what i mean and thanks in advance.

 

[25362]

If I understood your query, you are after heat transfer by forced convection. If so, there are graphs -in any book on heat transfer- from classical measurements by Sieder and Tate- expressing:

(Nu)(Pr)^-1/3([&mu;]/[&mu;]_w)^-0.14 = [&fnof;](Re,L/D)​

In which Re values range from 200 up to 10⁶, and cover quite well the transition zone.

Nu = Nusselt number, hD/k
Pr = Prandtl number, C_p[&mu;]/k
C_p = fluid's specific heat at constant pressure, J/(kg K)
L = length of pipe or duct, m
D = hydraulic diameter, m
[&mu;] = bulk viscosity, N s/m²
[&mu;]_w = viscosity at the surface temperature of the pipe or duct
k = thermal conductivity of the fluid, W/(m K)

For most cases ([&mu;]/[&mu;]_w)^-0.14 is practically 1.0.

Having estimated the heat transfer coefficient h, W/(m² K), from the above, the HTC for a helical coil with diameter D_c, would be: h_c [&asymp;] (1 + 3.5 D/D_c) h

Good luck.

 

[Ohlsson]

Thanks a lot for the reply 25362.

Is was helpful and really many thanks for pointing me towards Sieder and Tate, I found a lot of new good material with their findings.

Regards

 

[Montemayor]

Ohlsson:

I can attest, from personal experience, that process heat transfer using a spiral or helical coil is not easy. It is, however, fortunate that the Sieder-Tate relationships yield a conservative result.

I have designed and fabricated both helical and spiral coils in gas and liquid process service as well as condensers. I was (and still am) fascinated by the design very early in my career and had the opportunity to design and fabricate my own units when I managed and operated process plants overseas. Since I had the empowerment and I needed rapid and economic responses to my operating needs overseas, I took the initiative and employed this type of design where I could justify it. In the smaller sizes (10 to 75 ft² of heat transfer surface) and clean service, I found no better efficient or economic equals. I had great success employing this design and even developed mechanical fabricating techniques for specific applications. I applied this type of intercooler on many multi-stage reciprocating compressors with pressures up to 5,000 psig. Most of these units are still working today (30+ years later) to my knowledge.

I discovered very early on that my process design yielded over-sized units and it wasn’t until many years later when I had the luxury of using HTRI as a heat exchanger design tool that I discovered that HTRI had also recognized the same idiosyncrasies relating to the predictable sizing of coil units. They have done extensive research and work in this area and since I’m not at home at present, I’m unable to directly quote some of their findings. Suffice it to say that the heat transfer coefficient is definitely accentuated by what are described as “eddy” currents existing as the flow traces a coiled trajectory. This effect, in my experience is heightened even more so in the case of spiral flow as opposed to helical flow. A spiral is defined by me here as a pattern following the famous Archimedes Spiral curve. For many years (even before me) the Graham Manufacturing Company featured this advantageous effect in their “Graham Heliflow” heat exchangers. This model of exchanger’s name is a misnomer because their flow pattern is strictly in accordance with a spiral and not with a helix. HTRI documentation confirms that coiled flows are very desirable in heat transfer operation because of their eddy currents and the inherent improved film coefficients. Graham has never attested, to my knowledge, that findings confirm the advantages of their design. Since they control the majority of the market in this specialty item I would suspect that they have good knowledge of it and prefer to treat it as proprietary know-how. They have always made a very good, robust, and dependable product.

I would certainly follow 25362’s recommendations in carrying out your calculations for the sizing of this type of unit. However, do not be disappointed by the conservative results. In my opinion you need to have a lot of specific and proprietary expertise in this type of fluid flow and heat transfer before you can start getting very accurate in sizing. I found my units to be approximately 30-75% over-sized. However, this was no “big-deal”, since the total amount of area wasn’t that big. The compactness and the additional, future capacity paid off in the end.

I hope this experience helps you out.

 

[Ohlsson]

I will give You a more detailed view of my problem.

I work as a technical salesman for a manufacturer of HVAC equipment and our main product (which i am "boss" over) is fancoil units for indirect systems. For those unfamiliar with that term, indirect means that water alternatively water/glycol mixture is used instead of refrigerant.

Also I work in Sweden where we have a thing best translated as "district cooling" where cold water is being distributed from a large central to a large area of users.

For economical reasons the company distributing the chilled water want to have a high delta-t in their system (meaning smaller pipes, pumps and so on).

Now we approach the problem:

Standard, massproduced fancoils are designed for a delta-t of approx. 5-7 degrees and in our "district cooling" systems we have 10 or 11 degrees delta-t.

This means that ordinary coils designed for say 5K dt will experience laminar flow at these low water flows (Re well below 2000 and even below 1200 in many cases).

My company however has embraced the problem long ago and we have made special coils for these applications, theses coils have fewer circuits meaning higher water speeds in the tubing.

My job is to convince Swedens consulting engineers to use our special coils in these applications and so far i´ve been successful.

I have been reading up on my thermodynamics and fluid mechanics, but so far i haven´t been able to find a good way to calculate what i most need.

Imagine the following:

We have a fancoil and we know its cooling capacity exactly when we have turbulent flow (or Re approx 4000 or higher).

Imaginge the same coil with the same working conditions but with the flow being fully laminar (Re <2300) or transient ( 2300 < Re < 4000).

How much does these kind of products lose capacity-wise due to laminar /transient flow instead of turbulent. (Disregarding the fact that we lose cooling capacity due to higher delta-t, lower flow and so on) Just the difference in cooling /heating capacity due to the state of flow.


At last i apologize for any errors in grammar /spelling as english isn´t my native language and sorry for the long post.

Best Regards and hopefully thanks for any help.

/
Ohlsson

 

[BigInch]

If you are talking about reducing the size of the pipes, won't you have increased velocity and hence more turbulence inside, providing of course you keep the same flowrate?

I think you'll lose a lot of heat transfer capability in laminar flow as the boundry layers in the pipe are thicker and the hot fluid will tend to flow down the centerline of the piping, leaving the cooler fluid "sticking" to the walls.

 

[moltenmetal]

Ohlsson:

For a long, straight tube heat exchanger with fully developed laminar flow in the tube, you approach a constant Nusselt number. A peculiarity of this result is that at a given flowrate and a near-constant shellside temperature, the diameter of the tube basically no longer matters: the required length is the same for a tube of 3/8" diameter as for one of 1/2" diameter etc. so long as the flow remains laminar. To imagine conceptually what is happening, the fluid near the walls is being cooled, but the fluid in the centre of the tube is passing through essentially uncooled.

Unlike in the turbulent region, whatever frictional pressure drop you're encountering in the laminar region is of basically no benefit to you since it generates no net radial mixing (i.e. no mixing which is of use to you to move heat from the fluid in the centre of the pipe to the wall where the shellside can take it away). For viscous fluids, one must introduce radial mixing in some other way such as via the use of static mixing elements etc.

As Montemayor said, a coil is more complex than a straight tube, and fortunately for the better. Though the flow may remain laminar, the coiling does result in some radial mixing, which reduces the required length somewhat versus a straight tube.

What happens in the transitional region is very geometry dependant. Film coefficients calculated for fully developed laminar and for fully developed turbulent flow will be significantly different from one another, so the geometry of your coil will determine the transition between laminar and turbulent flow and what the resulting exit temperature will be. You only can say for certain that in transitional conditions, the performance of the exchanger will be somewhere between these two extremes.

If your coil is made from integrally finned or spiral corrugated tubing, your situation becomes even more complex.

Sounds like some testing work is in order for you! Correlations can only take you so far.

 

[BigInch]

MoltenMetal.

Very good point. I must admit that I was thinking straight lengths of pipe. I think its worth a star.