[Gasification] heat exchanger turbulent flow

[Mark Ludlow]

Dear M. Loftin,

 

Thank you for your thoughtful response.  I realize that my attempt to
explain the functional difference in the heat exchange characteristics of
laminar v. turbulent flows was over-simplified.

 

Obviously, good design considers all of the factors you have mentioned and
seeks a balance between them all.

 

As far as access  to CFD codes, I would refer you to
http://www.openchannelfoundation.org/cosmic/ , the source of over 500 public
domain fluid dynamics codes, most from NASA.

 

As always, good design is a blend of technical discipline and art. CFD is
not a cure-all but a tool, just as the slide rule was 50 years ago. A
seasoned design professional as yourself can likely create a near-optimal
design based on intuition and experience. But the ubiquity of inexpensive
CFD codes makes good heat exchanger design a reality for the rest of the
technical community who lack your experience and wisdom.

 

Sincerely,

Mark

 

 

 

From: Comcast Mail Server [mailto:moloftin at comcast.net] 
Sent: Thursday, December 27, 2007 8:26 PM
To: mark at ludlow.com; Discussion of biomass pyrolysis and gasification
Subject: Re: [Gasification] heat exchanger turbulent flow

 

Mark and the List,
In my years of designing and rating heat transfer equipment (from a process
engineer's perspective), I have formulated the following concepts in regard
to this subject, which hopefully will be of some interest to the group and
this discussion.

The transfer of thermal energy between these two points can be practically
simplified into three concepts: 1) fluid turbulence of the hot stream, 2)
the thermal resistance of the barrier, and 3) fluid turbulence of the cold
stream.  

The second point is easy: a metallic barrier will produce higher thermal
energy transfer than teflon barrier. But, sometimes you have to sacrifice
heat transfer rate through the barrier for corrosion resistance.  That can
often be compensated for by increasing area (and $).

The other two points involve "turbulence" of either the cold or hot stream.
"Turbulence" here simply means the stirring action of the bulk moving fluid,
which mixes molecules of different temperature created by the thermal
gradient between the hot and cold fluids.  

Turbulence can be created by different ways.  There is the Reynolds number
"turbulence" which is merely a correlation of the rate of internal eddy
formations in a bulk fluid as a function of density, viscosity, and
velocity.  It is not a magical concept at all.  It is just an
empirically-derived property that can be useful for certain calculations.
The point here is that "Reynolds Number turbulence" is an easy way to create
turbulence within a fluid but is not necessarily the most "efficient" way to
produce "optimum" turbulence for heat transfer purposes.  Some appreciable
portion of the eddys formed by high bulk fluid velocities do not "scour" the
boundary layer and as such largely represent wasted energy for purposes of
heat transfer.  Most modern heat exchanger designers push the Reynolds
number well into the turbulent regime in hopes that "eddys everywhere" will
keep them from getting into trouble with their supervisor once the plant
starts up.  The price paid is a high exchanger pressure drop, which cost
money continuously.  I recently had the chance to work with one of the best
heat exchanger designers in the the world on this subject (works for a major
oil company) and he taught me much on this issue.  Experienced heat transfer
equipment designers know that everything is a trade-off between energy
costs, capital costs and fouling rate. 

Yes, CFD is a much more powerful tool for producing controlled boundary
layer eddys for optimizing design of heat transfer equipment.
Unfortunately, only 1 ppb of society can afford that option.  That is why
engineering schools teach the Reynolds number approach because it is
relatively simply to apply for entry-level engineering labor and that is
what most large corps want.  Plus, you generally do not get these complex
geometrical tube/fin shapes CFD analysis tends to yield, which are often
monstrously expensive to produce.

The bottom line here is that conventional Reynolds number based heat
transfer design is good enough to make money with.  The results will be
affordable and practical at the expense of some energy waste (energy is
cheap, right?).  It does not work in all situations nor is it theoretically
"optimum" for a number of reasons. 

Respectfully,

M. Loftin