[Gasification] heat exchanger turbulent flow

[Mark Ludlow]

Hi Toby.

To begin with, the sciences of heat exchange as well as fluid dynamics are
largely empirical. Both make use of lots of dimensionless-numbers, some of
which have been derived from first-principles, but many more that came from
observing or trying to visualize, in some form or another, physical
phenomena occurring. All of this pre-dated the development of CFD
(computational fluid dynamics) which today has influenced the design of
everything from heat exchangers to aircraft to sailboats.

Much of what you "intuit" is correct. For instance, a turbulent flow has
more energy in it and the energy must come from somewhere, often a pump,
which has to work harder. There is more intermolecular interaction in the
fluid stream and this also equates to dissipative energy losses. Pressure
drops are higher.

Returning for a moment to your reference to a "turbulator": I think we both
agree that if there's a fixed mass of gas entering a flow conduit, (and
there is conservation of mass), that the same mass of gas, per unit of time,
will appear at the other end of the conduit. But, as we are causing the gas
to move in a swirling motion and the shortest distance between two points is
a straight line (in Euclidian geometry!), the gas velocity must increase.

So is this swirling flow stream laminar or turbulent? It could be either.
Here's where the Reynolds number come into play. The Reynolds number is,
simplistically, a ratio between inertial and viscous forces. If the Reynolds
number is small, the amount of inertial force is small, compared to the
viscous force. So the flow streams tend to be layered or laminar or "stuck
together". As the velocity increases the influence of the inertial force
dominates so that, for instance, a tiny projection in the path of the flow
stream can be the germ of a subsequent eddy, (which we are all familiar with
by staring at water flowing down a creek and past/around a rock).

The Reynolds number itself is usually referenced, as I recall, to flow over
a flat plate. So at some velocity--depending on the physical properties of
the fluid such as viscosity and density--the flow across the flat plate will
become turbulent, or consist of numerous eddy currents rather than a
neatly-layered stack of stream lines. My bet is that your Turbulator induces
turbulent flow.

What does this mean? First one has to conceive of the notion that in a
rapidly moving turbulent stream there are actually molecules that are
traveling against the direction of bulk transport. There are random
fluctuations in the direction and velocity of the fluid and the flow is
inherently unsteady. The flow is, from a practical point of view, nearly
impossible to visually describe with classical tools such as the
Navier-Stokes equations and our present-day knowledge is largely the result
of the numerical methods of CFD and more recently to methods pioneered by
Mandelbrot.

>From a practitioner's point of view--thankfully!--intimate knowledge of the
behavior of individual molecules, is of little practical interest because
design solutions are almost always concerned with the time-mean or averaged
properties. It's enough to know that turbulent flows provide increase heat
and mass transfer rates although not necessarily at lower expenses of
energy. In one respect, it is this loss of energy from the flow stream that
is the goal of heat exchanger design.

In terms of the boundary layer, turbulent flows can be mathematically shown
to increase shear stress which results in greater mixing and therefore
better heat, mass and momentum transfer. Small eddies are constantly forming
and being re-absorbed and they are efficient in normalizing the physical
properties of the flow stream and thus increasing the efficiency of heat
transfer.

Toby, bear in mind that a boundary layer is stagnant. Anything that you can
do to help it exchange heat/mass with the primary flow stream is beneficial.
The idea is to force as much interaction with the core of the flow stream
and the boundary layer as possible. This is best accomplished with a
fully-developed turbulent flow and is the goal of heat exchanger design
where fluids are not of such high viscosity that turbulent flows are
impractical to achieve.

Sorry for the long-winded reply! 

Mark




Mark and Andrew,
   
  I'm working on a cyclone that is inside of a downdraft gasifier. It's not
the point of this discussion as much as what I have had as a general
understanding of heat exchange.  The hot gases are to be cooled while the
fuel on the other side of the ss is heated.  As what I'm suggesting is a two
stage method, the gas is very hot (high delta T), being part synthesis gas.
It's not far from the DTU gasification system.  I hear that they are now
moving to a similar method.
   
  I worked for several years at Bryan Steam Company with liquid to liquid
heat exchangers.  That's been many years ago now.  My boss/mentor was Tom
Milton who was the ASME Section 7 Chairman.  We had set formulas as heat
exchangers inside of the boilers were "standardized".  When the flow reached
turbulent level, it was like moving a thick fluid or having a
restriction...less flow rather than more.  
   
  I remain perplexed.  In the turbulator example, the flow path is not
straight but spiral.  A circular movement is made by the fins.  The shortest
distance through the tube is straight and as the spiraling increases, so
does the length that any molecule in contact with the surface must go.  Yes
the velocity increases as well. The flow path is thus longer but not
necessarily the residency time.  But this flow is not turbulent.  In fact
the terminology as a "turbulator" seems incorrect.  It causes a flow path
that wipes the surface much better by reducing the boundary layer.  That's
not turbulent flow.  Are we talking about two different things?  Turbulence
in aircraft creates parasitic drag.  Is that the type of turbulence I
believe you are suggesting is more efficient?
   
  Mark, in your description you describe the reduction of turbulent flow on
the surface of the heat exchanger.  Moving that flow out of the center and
along the surface does increase the velocity, wiping the surface better.
But as that flow reaches the point that it becomes turbulent, eddys form
that slow the surface movement and when that flow is increased more, they
move further out from the surface, creating more parasitic drag and I
believe a larger, thicker boundary layer.  
   
  The optimum would seem to be a high flow rate that is just under the point
that turbulence develop.
   
   
  Best Regards,
   
  Toby Seiler

       
---------------------------------
Never miss a thing.   Make Yahoo your homepage.
_______________________________________________
Gasification mailing list
Gasification at listserv.repp.org
http://listserv.repp.org/mailman/listinfo/gasification_listserv.repp.org
http://gasifiers.bioenergylists.org
http://info.bioenergylists.org