Work Required for Heat Transfer Enhancement 1

[BretCahill]

This concerns the mechanical energy required to get a certain heat duty and/or effectiveness while factors such as heat exchanger volume are set.

One strategy might be to use long thin tubes and have a lot of surface area and the only work required will be moving the fluids through the tubes.

Another strategy might use larger tubes, less surface area and less pressure drop -- less work to pump the fluids -- but would accomplish the same heat transfer by independently forcing convection thereby requiring some additional work that is independent of the fluid flow pressure drop work.

Is there any number, parameter, equation, graphs etc. already in use that is used to compare, say, the work required to pump the fluids through the tubes to the work required for any added forced convection to the heat transfer?

In other words, is there a quick way to determine if creating turbulence requires more work that what it's worth?


Bret Cahill

 

[Montemayor]

Bret:

Can you express your concern in simpler and specific english terms? Your first paragraph is very difficult to interpret. What does "factors such as heat exchanger volume" really mean? The volume inside a heat exchanger doesn't come into play in any process heat transfer equations that I've ever seen. Do you mean heat transfer area?

There is no "quick" way to optimize the mechanical configuration of a heat exchanger - if that's what you are after. If you are concerned about creating turbulence to increase the heat transfer film coefficients, then calculate the required heat exchanger that way and compare it with a heat exchanger calculated with laminar flow. And which side of the heat exchanger are you concerned with - the shell side or the tube side? Is your application one of sensible heat or latent heat ? Or a combination of both?

Hardly any heat exchanger user that I have ever known or worked for has any interest in designing heat exchangers. The Duponts, Huntsmans, BP's, Dows, BASF, Bayers, etc, etc, of the world all leave this work (and any optimization) to the real heat exchanger experts: the designer/fabricators that compete with each other in the open marketplace.

Is this an industrial project or an academic question? Unless you've already accumulated manhours designing and fabricating heat exchangers, you're not going to come up with credible results or answers.

Sorry, but this is one subject that I've had some experience in -designing and building exchangers - and the answer to your concerns is always the same - leave it to the competitive experts. There are some general, accepted rules of thumb - like for example, longer tubes make for less costly exchangers (per ft2 of surface). But this entails other tradeoffs. If you are really concerned about a specific application, then consult with an expert heat exchanger designer and fabricator. If you have a real project, they will work with you in optimizing the ultimate exchanger to do the job - and they'll warrant the performance.

Art Montemayor
Spring, TX

 

[BretCahill]

I may have fortuitously stumbled across a situation that might be replicated in other heat exchangers to enhance transfer. It now seems so obvious it's hard to believe I can't find anything similar in the literature. In fact, it's hard to believe it took _me_ 5 years to see it even though it was staring me right in the face, dropping hints as subtle as a daisy cutter bomb.

But at the very minimum, some engineer somewhere BETTER have some empirical or other correlations showing that forced convection works here and not there or I'm going to call an acquaintance on the National Academy of Engineering and demand a blue ribbon commission investigate this omission.


Bret Cahill

 

[rmw]

Look up and learn about "reynolds number" for starters. I think that is your starting place.

rmw

 

[tickle]

What type of HE's?

As an ex S&T HE designer. The general rule of thumb was to win the job. To do so was to design the cheapest HE possible. Which in turn required the smallest diameter, and usually required 3/4" tubes.

One way of increasing heat transfer is to use turbulence promoters (HiTran) or static mixers that increased heat transfer coefficients on the tubeside and also tube side pressure drops. However the smaller tube length usually made the pressure drop similar. Similarly on the shell side there are finned tubes. Shell side coefficients can also be increased using twisted tubes.

You are obviously excited by your discovery, however like Montemayor I too found it a bit difficult to understand what you are asking. Can I suggest that you stand back and determine what you want to know and ask it straight.

 

[IRstuff]

If the other side is air, wouldn't that dominate the HE performance?

TTFN

 

[BretCahill]

It's not all that big a deal to assume a flow rate and get some Nusselt number(s), boundary layer thicknesses, temperature differences, conductivities, surface areas etc. to determine a Q for ONE situation -- especially if you don't need two or 3 decimal place accuracy.

But to save everyone time, to quickly and clearly flag an optimum design as well as to to evaluate a lot of other situations, wouldn't it make sense to have some normalized or dimensionless quantity, something like Wfc/Q (work of forced convection divided by heat rate).

There must be something holding back more aggressive forced convection tactics to save money on materials and reduce the size of heat exchangers. It's hard to believe the one time cost of a motor-impeller or pump-nozzle is the problem. The only conclusion is that it is the constant overhead of energy costs.

Somewhere someone has studied this issue and has some useful parameter.


Bret Cahill

 

[25362]

To BretCahill, thinking aloud: have you considered cases in which, increasing the HTC on one side of an exchanger wouldn't sensibly affect its performance, since the resistance offered by the other side is determining the overall heat transfer coefficient ?

Then there are cases in which the minimum linear velocity of the streams is dictated by fouling effects ? Higher velocities are then used to dislodge any potentially sticking or coking deposits, without considering the pumping costs, since it would otherwise become a go/no-go situation.

Then, if I may comment on the basic philosophy behind your argument: I think it is the plant overall optimization that should be considered. A heat exchanger is just but one component of the whole bunch.

Consider, please, beside pumping work as suggested by you, these other factors: yields of products, ease of recovery of main products, investment, operability, maintenance, overall plant energy consumption.

I could think of several cases, in which the energy-optimization suggested by you around a particular heat exchanger (which I think is the leitmotiv of your post) may even interfere with -and degrade- the overall plant optimization.

Kindly comment.

 

[EdStainless]

The other side of the problem Bret is that in most cases heat exchangers are 'over sized'. People and fudge factors to heat loads and flows and in the end this assures that the heat exchanger is operating far from optimum conditions.

I have worked with a number of power plants where raising the flow rate (at the cost of electricity) saved enough fuel (at the cost of coal) to make it pay off.

= = = = = = = = = = = = = = = = = = = =
Corrosion never sleeps, but it can be managed.

http://www.trenttube.com/Trent/tech_form.htm

 

[25362]

To BretCahill

To be entirely fair, an optimizing equation as the one you are after exists, and is sometimes used for comparison purposes on S&T heat exchangers. But exclusively on the tube-side fluids on turbulent flow, when it is known that it is there where the dominant heat transfer resistance resides.

The cost-optimized heat transfer coefficient is

h=(0.023/D^0.143)[×][1.6K_sg_c/(0.023K_p)]^0.2857[×](k^0.6C_p^0.4[ρ]^0.571/[μ]^0.457)

where:

D = tube inside diameter
K_s = annualized cost per unit of surface area
K_p = annualized cost per unit of tubeside pumping power
k = thermal conductivity
C_p = heat capacity
[ρ] = density
[μ] = viscosity

Based on this equation the relative heat-transfer performance parameters of various heat transfer fluids can be compared as follows:

h₁/h₂ = (k₁/k₂)^0.6(C_p1/C_p2)^0.4([ρ]₁/[ρ]₂)^0.571([μ]₁/[μ]₂)^0.457

For more details, see, please, an article titled Get Fluent About Heat Transfer Fluids by Green, Larsen and Pauls (Monsanto) in the Chemical Engineering issue of February 1989, under the general heading of Engineering Feature.

 

[BretCahill]

For awhile I was afraid I was going to get a dimensionless number named after me!

I can rest easy now.


Bret

 

[speco]

Brett,

Here's an idea to consider. I work mostly with air-cooled exchangers. Some customers apply a cost per consumed horsepower on the fans as a way of accounting for the long-term cost of their operation. I have seen numbers anywhere up to about $2500 per horsepower.

Since there is always a trade-off between the capital cost and the operating cost, this gives the designer a way to evaluate various options for his design. Heat exchanger problems always have an infinite number of solutions, but some are much better than others.

You could also use a similar means to evaluate cost of pumping liquids.

Just a thought.

Regards,

speco

http://www.stoneprocess.com