Convection Fundamentals and Boundary Layers

Introduction

Convection is heat transfer between a surface and a moving fluid. It combines molecular conduction within the fluid adjacent to a surface with bulk fluid motion. Because fluid motion strongly affects the temperature and velocity distributions near a wall, convection cannot be understood adequately without the concept of a boundary layer.

Convection is central to the design of heat exchangers, turbine blades, electronic cooling systems, vehicle bodies, solar collectors, boilers, condensers, and building services equipment. The engineer must determine whether motion is imposed by a fan, pump, or vehicle—or generated by buoyancy—and must select appropriate dimensionless correlations for the resulting flow.

Learning Objectives

By the end of this chapter, you should be able to:

1. Explain momentum and thermal boundary layers.
2. Distinguish forced, natural, and mixed convection.
3. Define and interpret Reynolds, Prandtl, Grashof, Rayleigh, and Nusselt numbers.
4. Use dimensional analysis and similarity principles to organize convection problems.
5. Apply standard external-flow boundary-layer correlations.
6. Evaluate assumptions, limitations, and engineering consequences of convection models.

1. Physical Basis of Convection

At a solid surface, the no-slip condition requires the fluid velocity to equal the surface velocity. For a stationary wall, fluid velocity is zero at the wall and increases away from it. If the wall temperature differs from the fluid temperature, the fluid temperature also changes rapidly near the wall.

The momentum boundary layer is the region in which viscous effects cause the fluid velocity to change from its wall value to approximately the free-stream value. The thermal boundary layer is the region in which temperature changes from the wall temperature to approximately the free-stream temperature.

For an external flow over a flat plate, the boundary layer begins at the leading edge and grows downstream. Outside it, velocity and temperature gradients are usually small. Within it, gradients are large, and therefore viscous shear and heat transfer are concentrated there.

Figure: Growth of momentum and thermal boundary layers over a flat plate.

A conventional boundary-layer thickness is the distance at which velocity reaches 99% of the free-stream velocity:

where u is local velocity, U∞ is free-stream velocity, and δ is momentum boundary-layer thickness. The thermal thickness is often defined using a similar 99% temperature criterion, although other definitions are used.

2. Momentum Boundary Layer

For steady, two-dimensional, incompressible flow over a flat plate, the boundary-layer approximation reduces the Navier–Stokes equations to:

Here u and v are velocity components in the x and y directions, ν is kinematic viscosity in m²/s, and U∞ is the external velocity. The equations assume steady flow, constant properties, negligible body-force effects in the streamwise momentum equation, and a thin boundary layer in which gradients normal to the wall dominate gradients along it.

For a zero-pressure-gradient flat plate, U∞ is constant and dU∞/dx = 0. The local wall shear stress is:

where τₛ is wall shear stress in Pa and μ is dynamic viscosity in Pa·s. The local skin-friction coefficient is:

where ρ is density in kg/m³. For a laminar flat-plate boundary layer:

and the average coefficient from the leading edge to x is:

where:

These results apply to a smooth flat plate, zero pressure gradient, steady incompressible laminar flow, and approximately constant properties.

For a turbulent boundary layer, one commonly used smooth-plate correlation is:

for sufficiently high Reynolds number and away from transition. Correlations differ according to the transition model, surface roughness, pressure gradient, and Reynolds-number range; consequently, their stated validity range must always be checked.

3. Thermal Boundary Layer

The thermal boundary layer develops because fluid at the wall is heated or cooled by conduction, while fluid farther away is transported by bulk motion. For constant properties and no internal heat generation, the energy equation in a boundary layer is:

where T is temperature in K and α is thermal diffusivity in m²/s, defined by:

Here k is thermal conductivity in W/(m·K), ρ is density, and cₚ is specific heat capacity in J/(kg·K).

At a wall with temperature Tₛ and free-stream temperature T∞, the local convection coefficient is defined by Newton’s law of cooling:

where q″ is surface heat flux in W/m² and hₓ is the local convection coefficient in W/(m²·K). At the wall, heat is transferred by conduction into the fluid, so:

The local Nusselt number is:

The Nusselt number compares convection at a surface with conduction across a characteristic fluid length. For a laminar isothermal flat plate:

and the average value over length L is:

These expressions are appropriate approximately for 0.6 ≤ Pr ≤ 50,000, constant wall temperature, zero pressure gradient, and laminar flow.

4. Dimensionless Groups and Similarity

Dimensional analysis reduces many variables to a smaller set of dimensionless groups. The most important groups in convection are summarized below.

In these definitions, U is characteristic velocity, L characteristic length, g gravitational acceleration, β volumetric thermal-expansion coefficient in K⁻¹, ΔT characteristic temperature difference, k_s solid conductivity, and L_c a solid characteristic length.

The Reynolds number determines whether a flow is likely to be laminar or turbulent, but the transition value is geometry-dependent. For a smooth flat plate, transition often begins near Reₓ ≈ 5 × 10⁵, although free-stream turbulence, roughness, pressure gradients, and disturbances may shift it substantially.

The Prandtl number indicates the relative thickness of momentum and thermal layers. When Pr > 1, thermal diffusion is relatively slow and the thermal layer tends to be thinner than the velocity layer. When Pr < 1, thermal diffusion is rapid and the thermal layer tends to be thicker. For air, Pr is approximately 0.7; for liquid metals it is much less than 1; for oils it may be much greater than 1.

The Grashof number plays a role in natural convection analogous to that of Reynolds number in forced convection. The Rayleigh number is normally the principal correlation parameter for natural convection.

Similarity requires that geometrically similar systems have matching relevant dimensionless groups. Thus, a model experiment can represent a prototype only when important groups—such as Re, Pr, and sometimes Gr or Ra—are adequately matched.

5. Forced, Natural, and Mixed Convection

Forced convection

In forced convection, motion is produced by an external device or imposed relative motion. Examples include air flowing over a vehicle, water pumped through a heat exchanger, and coolant passing through an electronics channel. Correlations are usually expressed in terms of Re and Pr:

where C, m, and n depend on geometry and flow regime.

Natural convection

Natural convection arises from buoyancy caused by density differences. Under the Boussinesq approximation, density variation is neglected except in the buoyancy term:

This approximation is suitable when temperature differences are small enough that property variations are limited. Natural-convection correlations are commonly written as:

with constants determined by geometry, orientation, and Rayleigh-number range.

Mixed convection

Mixed convection occurs when forced motion and buoyancy are both important. A useful indicator is the Richardson number:

If Ri ≪ 1, forced convection generally dominates; if Ri ≫ 1, natural convection generally dominates; and if Ri is of order unity, mixed convection must be considered. These are engineering guidelines rather than universal boundaries.

Buoyancy can assist the forced flow, oppose it, or act transversely to it. Opposing buoyancy may cause separation, flow reversal, or enhanced unsteadiness, so a simple forced-convection correlation may be unreliable.

6. Boundary-Layer Transition and Separation

A laminar boundary layer has relatively orderly motion and lower skin friction, but it is vulnerable to disturbances. A turbulent boundary layer has strong mixing, higher wall shear, and generally greater heat transfer.

Transition is affected by:

• Reynolds number;
• free-stream turbulence;
• wall roughness;
• acoustic and mechanical disturbances;
• pressure gradients; and
• surface curvature.

When external velocity decreases in the flow direction, pressure increases. This adverse pressure gradient reduces near-wall momentum and can cause the boundary layer to separate. Separation occurs when wall shear approaches zero and then changes sign:

Separated flow produces wakes, pressure drag, recirculation, and often unsteady heat-transfer distributions. Flat-plate correlations should not be applied blindly to strongly curved or separated flows.

7. A Systematic Method for Convection Problems

A reliable solution procedure is:

1. Define the geometry and characteristic length. For a plate, L may be plate length; for a tube, diameter D is commonly used.
2. Identify the driving mechanism. Decide whether motion is forced, natural, or mixed.
3. Select fluid properties. Evaluate μ, ρ, k, cₚ, and β at a suitable film or bulk temperature unless the correlation specifies otherwise.
4. Calculate dimensionless groups. Determine Re and Pr for forced flow, or Gr and Ra for natural flow.
5. Determine the flow regime. Check laminar, transitional, turbulent, and developing-flow conditions.
6. Select a correlation within its validity range. Confirm geometry, boundary condition, property range, and characteristic length.
7. Calculate h and heat rate. Use h = Nu k/L and q̇ = hA(Tₛ − T∞) for a uniform surface coefficient.
8. Check units and physical plausibility. Large velocity, high conductivity, or strong mixing generally increases h, although the precise dependence is correlation-specific.

Worked Example: Laminar Air Flow over a Heated Plate

Air at 25 °C flows at 5 m/s over a smooth isothermal plate 0.50 m long and 0.30 m wide. The plate is maintained at 75 °C. Use approximate air properties at the film temperature 50 °C: k = 0.028 W/(m·K), ν = 1.80 × 10⁻⁵ m²/s, and Pr = 0.70. Estimate the average heat-transfer coefficient and heat-transfer rate, assuming the boundary layer remains laminar.

The film temperature is:

First calculate Reynolds number at the plate end:

This is below the commonly used flat-plate transition estimate of approximately 5 × 10⁵, so a laminar correlation is appropriate under the stated assumptions.

The average Nusselt number is:

Since Re_L^(1/2) ≈ 373 and Pr^(1/3) ≈ 0.888,

Therefore:

The plate area is A = 0.50 × 0.30 = 0.15 m², and the temperature difference is 50 K. Thus:

The estimated heat transfer is therefore approximately 92 W. Radiation, leading-edge effects, property variation, and possible external disturbances have been neglected.

Applied Case: Cooling an Electronic Enclosure

Background

Consider an illustrative design case in which a sealed electronic enclosure dissipates 120 W. The enclosure has an external vertical surface area of 1.2 m² and is exposed to still room air at 25 °C. The average enclosure surface temperature must remain below 55 °C.

Problem

Determine whether natural convection alone can remove the heat, recognizing that the exact result requires a vertical-plate natural-convection correlation and careful property evaluation.

Analysis and approach

The temperature difference is 30 K. The characteristic length is the vertical height, assumed here to be 0.60 m. Air properties should be evaluated at the film temperature, 40 °C. The Rayleigh number is calculated from:

For air near 40 °C, representative values are β ≈ 1/313 K⁻¹, ν ≈ 1.70 × 10⁻⁵ m²/s, α ≈ 2.40 × 10⁻⁵ m²/s. Substitution gives:

This lies in a range where natural convection is vigorous but may approach transitional behavior depending on surface condition and geometry. A standard vertical-plate correlation should be selected for the precise range. If an estimated average coefficient of 5 W/(m²·K) is obtained from an appropriate correlation, the heat removal is:

The predicted capacity exceeds 120 W, giving a nominal margin. However, the design must also consider enclosure orientation, obstruction of rising air, radiation, internal hot spots, allowable component temperatures, and uncertainty in h. If the enclosure is installed horizontally or in a confined cabinet, the natural-convection coefficient may be much lower. Adding fins, vents, or a fan may be necessary.

Learning points

1. Orientation is a governing variable in natural convection.
2. A calculated average coefficient does not guarantee safe local component temperatures.
3. Correlation validity and installation conditions are as important as arithmetic.
4. Radiation may contribute significantly and should not be omitted in a final thermal design.

Limitations and Best Practices

Convection correlations are empirical or semi-empirical representations of complex flow. Their limitations include idealized geometry, uncertain transition behavior, variable properties, surface roughness, nonuniform heat flux, entrance effects, and three-dimensional flow.

Best practices include:

• use properties at the prescribed reference temperature;
• distinguish local from average coefficients;
• use hydraulic diameter for noncircular internal passages only when the correlation supports it;
• avoid applying flat-plate formulas to separated, curved, or strongly developing flows;
• include radiation when surface temperatures are substantially different from surroundings;
• report assumptions and uncertainty, not only a numerical value;
• validate critical designs experimentally or with a suitably verified computational model.

Computational fluid dynamics can resolve velocity and temperature fields, but it does not remove the need for correct boundary conditions, turbulence models, mesh-quality checks, and validation. AI-assisted design tools may help search geometries or correlations, but their outputs require engineering verification and traceable sources.

Chapter Summary

Convection transfers heat between a surface and a moving fluid. The momentum boundary layer describes the region influenced by viscosity, while the thermal boundary layer describes the region influenced by temperature gradients. Reynolds number characterizes inertial-to-viscous effects; Prandtl number compares momentum and thermal diffusivities; and Nusselt number expresses the enhancement of heat transfer by convection relative to conduction. Natural convection is governed primarily by buoyancy through Grashof and Rayleigh numbers, whereas forced convection is governed primarily by Reynolds and Prandtl numbers. The Richardson number helps identify mixed convection. Correct characteristic lengths, property values, flow regimes, and correlation limits are essential to reliable engineering calculations.

Key Terms

Boundary layer: Thin region adjacent to a surface where velocity or temperature changes rapidly.

Momentum boundary layer: Region affected by viscous momentum diffusion.

Thermal boundary layer: Region containing the principal fluid temperature gradient.

No-slip condition: Fluid velocity equals surface velocity at a solid wall.

Convection coefficient: Proportionality constant h relating heat flux to surface-to-fluid temperature difference.

Reynolds number: Ratio of inertial to viscous effects.

Prandtl number: Ratio of momentum diffusivity to thermal diffusivity.

Nusselt number: Ratio of convective to conductive heat transfer across a characteristic length.

Grashof number: Ratio measuring buoyancy relative to viscous effects.

Rayleigh number: Product of Grashof and Prandtl numbers.

Forced convection: Convection driven by an external flow-producing mechanism.

Natural convection: Convection caused by buoyancy arising from density differences.

Mixed convection: Simultaneous forced and buoyancy-driven convection.

Skin-friction coefficient: Dimensionless wall shear stress.

Boundary-layer separation: Detachment of near-wall flow under an adverse pressure gradient.

Film temperature: Average of surface and free-stream temperatures, commonly used for properties.

Similarity: Matching of relevant dimensionless groups between systems.

Boussinesq approximation: Treatment retaining density variation primarily in the buoyancy term.

Review Questions

Short Answer Questions

1. What physical effects define the momentum and thermal boundary layers?
2. Define Reynolds, Prandtl, and Nusselt numbers.
3. Why is the Prandtl number important in comparing boundary-layer thicknesses?
4. Distinguish forced convection from natural convection.
5. What does a Richardson number much greater than one suggest?

Long Answer Questions

1. Derive the physical meaning of the boundary-layer approximation and state its principal assumptions.
2. Explain how laminar-to-turbulent transition affects skin friction and heat transfer.
3. Discuss the role of dimensional analysis in developing convection correlations.
4. Explain buoyancy, the Boussinesq approximation, and the use of Rayleigh number in natural convection.
5. Describe a complete engineering procedure for selecting and applying a convection correlation.

Analytical / Application Questions

1. Air at 8 m/s flows over a 1 m plate with ν = 1.6 × 10⁻⁵ m²/s. Calculate Re at the trailing edge and assess whether a laminar flat-plate correlation is plausible under quiet free-stream conditions.
2. Explain how increasing fluid viscosity, velocity, and characteristic length affects forced-convection behavior using Reynolds number.
3. An electronic enclosure operates in still air but is installed horizontally rather than vertically. Identify the changes in physical mechanism, correlation selection, and design risk that should be evaluated.

Multiple Choice Questions with Answers

Question 1

Which condition defines the conventional 99% momentum boundary-layer thickness?

A. u = 0.01U∞

B. u = 0.50U∞

C. u = 0.99U∞

D. u = U∞ + 0.99U∞

Correct Answer: C. u = 0.99U∞

Question 2

The Reynolds number primarily compares which effects?

A. Buoyancy and conduction

B. Inertia and viscosity

C. Radiation and convection

D. Pressure and temperature

Correct Answer: B. Inertia and viscosity

Question 3

For a fluid with Prandtl number greater than one, the thermal boundary layer is generally:

A. Thicker than the momentum boundary layer

B. Thinner than the momentum boundary layer

C. Always turbulent

D. Absent

Correct Answer: B. Thinner than the momentum boundary layer

Question 4

Which dimensionless group is most directly associated with natural-convection buoyancy and thermal diffusion?

A. Rayleigh number

B. Reynolds number

C. Biot number

D. Mach number

Correct Answer: A. Rayleigh number

Question 5

If Richardson number is much less than one, which mode usually dominates?

A. Natural convection

B. Forced convection

C. Radiation

D. Conduction in the solid

Correct Answer: B. Forced convection

Question 6

For laminar flow over an isothermal flat plate, the average Nusselt number varies approximately as:

A. Re_L^(−1/2)Pr^(1/3)

B. Re_L^(1/2)Pr^(1/3)

C. Re_LPr^(−1/3)

D. Re_L²Pr

Correct Answer: B. Re_L^(1/2)Pr^(1/3)

Question 7

A plate has U∞ = 5 m/s, L = 0.50 m, and ν = 1.80 × 10⁻⁵ m²/s. What is Re_L approximately?

A. 1.39 × 10³

B. 1.39 × 10⁴

C. 1.39 × 10⁵

D. 1.39 × 10⁶

Correct Answer: C. 1.39 × 10⁵

Question 8

At boundary-layer separation, the wall shear stress is typically:

A. Maximum and positive

B. Zero before changing sign

C. Equal to free-stream pressure

D. Independent of velocity

Correct Answer: B. Zero before changing sign

Question 9

Which equation defines the local convective heat flux for a surface?

A. q″ = hₓ(Tₛ − T∞)

B. q″ = hₓ/(Tₛ − T∞)

C. q″ = kA(Tₛ − T∞)

D. q″ = ρcₚU∞

Correct Answer: A. q″ = hₓ(Tₛ − T∞)

Question 10

The Grashof number measures the relative importance of:

A. Buoyancy and viscous effects

B. Convection and radiation

C. Momentum and thermal diffusivity

D. Surface area and volume

Correct Answer: A. Buoyancy and viscous effects