Correlations for Heat Transfer Coefficients

Each flow geometry requires different correlations be used to obtain heat transfer coefficients. Initially, we will look at correlations for fluids flowing in conduits.

Most correlations will take the "Nusselt form":

Nusselt form

The correlations that follow are limited to conduit flow without phase change. Different geometries, boiling, and condensation will be covered in later lectures. Frictional heating (viscous dissipation) is not included in these correlations. This should not be a problem, since this phenomena is typically neglected except for highly viscous flows or gases at high mach numbers.

Unless otherwise specified, fluid properties should be evaluated at the "bulk average" temperature -- the arithmetic mean of the inlet and outlet temperatures:

Bulk Temperature

Choosing a Correlation

When choosing a correlation, begin by asking:

  1. What is the geometry? (Flow through a pipe, around an object, over a plane, etc.)
  2. Is there a phase change?
  3. What is the flow regime? (Check the Reynolds number to decide on laminar, transition, or turbulent flow.)
  4. If the flow is laminar, is natural convection important? (The Grashof number will be used for this.)

Turbulent Conduit Flow -- No Phase Change

The historic equation for use in turbulent conduit flow is the Dittus-Boelter Correlation (MSH Eq. 12.32):

Dittus-Boelter
The exponent on the Prandtl number depends on the service -- 0.4 is used for heating and 0.3 for cooling. Different values are needed because of the variation of viscosity with temperature.

Heating and cooling effect the velocity profile of a flowing fluid differently because of the temperature dependence of viscosity. Heating usually makes the fluid near the wall less viscous, so the flow profile becomes more "plug-like." Cooling has the opposite effect, increasing the viscosity near the wall and impeding heat transfer. The effect is most pronounced for viscous flows with large wall -- bulk temperature differences.

Instead of using different exponents for heating and cooling, a direct correction for viscosity can be used. This takes the form of the ratio of the viscosity at the bulk fluid temperature to the viscosity at the wall temperature. The ratio is then raised to the 0.14 power.

Viscosity Correction
When this is added, the result is the Seider-Tate Correlation (MSH Eq. 12.33), the correlation recommended for use in this class:
Seider-Tate Correlation

Seider-Tate applies to "normal" fluids in turbulent flow in long, straight pipes, so:
Seider-Tate Limitations

Multiplicative correction factors are available to adjust for the entrance/exit consequences of short tubes:

Short Tube Correction
and for pipe curvature
Coil Correction
If the conduit does not have a circular cross-section, the inside diameter should everywhere be replaced by the equivalent diameter
Equivalent Diameter

Transitional Conduit Flow -- No Phase Change

Levenspiel (1998) recommends the following correlation for transition flow. The entrance effect correction may be omitted for "long" conduits.

Transition Correlation

Laminar Conduit Flow -- No Phase Change

Many of the laminar flow correlations are set up in terms of the Graetz Number. McCabe et al. define this as

MSH Graetz Number
but many other works do not limit evaluation to cylindrical geometries, and use
Graetz Number
which differs from the MSH value by a factor of Pi/4. Consequently, you must be very careful to use the form that matches the correlation you are using. Those that follow are based on the standard form, NOT the MSH form.

Two correlations are provided for laminar flow, depending on the magnitude of the Graetz number. For Gz<100

Laminar Flow Gz<100
which approaches a limiting value of 3.66 in long pipes with uniform wall temperature (see MSH Fig, p. 344). For Gz>100, use
Laminar Flow Gz>100
Both of these may need to be corrected when natural convection is significant.

Laminar Flow and Free Convection

Heat usually causes the density of a fluid to change. Less dense fluid tends to rise, while the more dense fluid falls. The result is circulation -- "natural" or "free" convection. This movement raises h values in slow moving fluids near surfaces, but is rarely significant in turbulent flow. Thus, it is necessary to check and compensate for free convection only in laminar flow problems.

The Grashof Number is used to assess the impact of natural convection (MSH Eq. 12.66):

Grashof Number
It makes use of the coefficient of volume expansion:
Coefficient of Volume Expansion
The fluid properties used to calculate the Grashof number should be evaluated at the film temperature, the arithmetic mean between the bulk and wall temperatures. This will require determining an additional set of property values.

The Grashof Number provides a measure of the significance of natural convection. When the Grashof Number is greater than 1000, heat transfer coefficients should be corrected to reflect the increase due to free circulation. Multiplicative correction factors are available to apply to the Nusselt Number or the heat transfer coefficient (do NOT use both). These are:

Free Convection Correction

References:

  1. Brodkey, R.S. and H.C. Hershey, Transport Phenomena: A Unified Approach, McGraw-Hill, 1988, pp. 510-15.
  2. Levenspiel, O., Engineering Flow and Heat Exchange, Revised Edition, Plenum Press, 1998, pp. 173-78, 182-84.
  3. McCabe, W.L., J.C. Smith, and P. Harriott, Unit Operations of Chemical Engineering (5th Edition), McGraw-Hill, 1993, pp. 330-31, 333-34, 336-37, 339-41, 344-48, 353-55, 362-68.
  4. McCabe, W.L., J.C. Smith, and P. Harriott, Unit Operations of Chemical Engineering (6th Edition), McGraw-Hill, 2001, pp. 336-37, 339-40, 342-49, 365-71.

R.M. Price
Original: 12/8/99
Modified: 1/2/2001, 1/4/2002, 2/4/2003

Copyright 1999, 2001, 2002, 2003 by R.M. Price -- All Rights Reserved