Viscous Flow, Boundary Layers, and External Flow

Introduction

Fluid motion is governed not only by pressure and gravity but also by viscosity—the internal resistance that develops when adjacent fluid layers move at different velocities. Viscous effects are especially important near solid surfaces, where the fluid velocity changes from zero at a stationary wall to the external-flow velocity away from the wall. This thin region, called the boundary layer, controls frictional resistance, heat and mass transfer, and the onset of flow separation.

Construction engineers encounter these phenomena in wind loading on buildings and bridges, flow around vehicles and cranes, sediment and particle transport, pipeline discharge, ventilation, water structures, and the design of streamlined or bluff structural elements. A sound understanding of viscous and external flow is therefore essential for estimating forces, interpreting wind-tunnel results, and selecting appropriate design models.

Learning Objectives

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

1. Explain viscosity, shear stress, and Newtonian fluid behavior.
2. Distinguish laminar, transitional, and turbulent flow using Reynolds number.
3. Describe boundary-layer growth, velocity profiles, displacement thickness, and momentum thickness.
4. Explain adverse pressure gradients and boundary-layer separation.
5. Calculate drag, lift, skin-friction, and pressure forces in basic external-flow problems.
6. Apply these ideas to construction elements, vehicles, structures, and particles.
7. Recognize the limitations of simplified correlations and the importance of experimental or computational validation.

1. Viscosity and Shear Stress

1.1 Molecular origin of viscosity

In a moving fluid, neighboring layers may have different velocities. Molecular momentum exchange transfers momentum from faster layers to slower layers, producing a tangential force that opposes relative motion. This property is viscosity.

For a fluid between two parallel plates, let the lower plate be stationary and the upper plate move with velocity U. If the spacing is h and the velocity profile is approximately linear, the velocity gradient is U/h. Newton’s law of viscosity is

where τ is shear stress (Pa or N/m²), μ is dynamic viscosity (Pa·s), u is velocity parallel to the surface (m/s), and y is the coordinate normal to the surface (m). The relation applies to a Newtonian fluid and to a locally defined shear rate. The sign depends on the selected coordinate direction; the resisting stress acts opposite to the relative motion.

The kinematic viscosity is

where ν is kinematic viscosity (m²/s) and ρ is density (kg/m³). Water at ordinary atmospheric conditions has a much lower viscosity than air in absolute terms, although density must also be considered when comparing momentum diffusion.

1.2 Newtonian and non-Newtonian fluids

A Newtonian fluid has a linear relationship between shear stress and velocity gradient. Water, air, and many light oils are treated as Newtonian over ordinary engineering ranges. Concrete slurry, fresh cement paste, drilling mud, paint, and some polymer solutions may be non-Newtonian. Their apparent viscosity changes with shear rate, and a single constant μ is not sufficient to describe their behavior.

For a stationary wall, the no-slip condition states that the fluid velocity at the wall equals the wall velocity. Thus, for a stationary surface, u = 0 at y = 0. This condition is central to the formation of a boundary layer.

2. Laminar, Transitional, and Turbulent Flow

2.1 Reynolds number

The relative importance of inertial and viscous effects is represented by Reynolds number:

Here U is a characteristic velocity (m/s), L is a characteristic length (m), ρ is density, μ is dynamic viscosity, and ν is kinematic viscosity. Reynolds number is dimensionless. It is not a universal flow-state label without a specified geometry and disturbance environment.

Low Reynolds number indicates that viscous effects strongly influence the flow. At high Reynolds number, inertia dominates in much of the flow, although viscous effects remain decisive near walls and in wakes.

In a smooth circular pipe, fully developed flow is commonly classified approximately as laminar for Re < 2,300, transitional for about 2,300–4,000, and turbulent for Re > 4,000, where L is pipe diameter and U is mean velocity. These limits are empirical and can shift with inlet disturbances and roughness. External flow has no single universal transition value; a flat-plate transition often occurs near a local Reynolds number of several hundred thousand, depending strongly on surface roughness, turbulence, pressure gradient, and free-stream disturbances.

2.2 Laminar and turbulent motion

Laminar flow is relatively orderly, with limited transverse mixing. Momentum transfer occurs mainly through molecular action. Turbulent flow contains irregular fluctuations and eddies over a range of scales. Turbulence enhances momentum exchange, producing a fuller average velocity profile and a larger wall shear stress than a laminar flow at the same nominal conditions.

A useful decomposition is

where u is instantaneous velocity, Ū is time-averaged velocity, and u′ is the fluctuation. The product of density and the averaged product of velocity fluctuations creates the Reynolds shear stress, written in a simple two-dimensional form as −ρ overbar(u′v′). This additional momentum transport is why turbulent-flow analysis requires closure models or empirical information.

3. Boundary-Layer Development

Consider uniform flow of velocity U∞ approaching the leading edge of a smooth flat plate aligned with the flow. At the leading edge the boundary layer begins at nearly zero thickness. Moving downstream, viscous effects diffuse away from the plate and the boundary layer grows.

Outside the boundary layer, velocity is approximately U∞ and viscous gradients are small. Inside it, velocity increases from zero at the wall to nearly U∞. The boundary-layer edge is conventionally defined where u/U∞ reaches approximately 0.99, although this is a practical convention rather than a sharp physical boundary.

Figure: Growth of a boundary layer over a flat plate; its thickness increases downstream.

3.1 Flat-plate estimates

For a zero-pressure-gradient, smooth flat plate with uniform free-stream velocity, commonly used laminar estimates are

where δ is boundary-layer thickness, x is distance from the leading edge, Reₓ = U∞x/ν, and C_f,x is local skin-friction coefficient. These relations assume steady, incompressible, two-dimensional flow, a smooth plate, negligible pressure gradient, and laminar flow.

The average skin-friction coefficient from the leading edge to length L is

The wall shear force on one side of a plate of width b is

For turbulent flow over a smooth flat plate, one widely used engineering correlation, valid over a stated empirical range and subject to convention, is

The correction term is used when the boundary layer is assumed laminar near the leading edge and turbulent farther downstream. Correlations must not be extrapolated blindly, especially for rough surfaces or strong pressure gradients.

3.2 Integral thickness measures

Boundary-layer thickness alone does not fully describe its effect. Displacement thickness is

It represents the apparent outward displacement of external streamlines caused by the loss of flow near the wall. Momentum thickness is

It measures the momentum deficit in the boundary layer. The shape factor is

For a typical laminar flat-plate boundary layer, H is about 2.6; turbulent boundary layers commonly have lower values, often near 1.3–1.4, though the value depends on pressure gradient and surface condition. A rising shape factor generally signals a fuller loss of near-wall momentum and increasing susceptibility to separation.

4. Pressure Gradients and Separation

A favorable pressure gradient occurs when pressure decreases in the direction of flow. It tends to accelerate the external fluid and stabilize the boundary layer. An adverse pressure gradient occurs when pressure increases downstream. The low-momentum fluid near the wall may then be unable to move against the increasing pressure.

At the wall, separation is associated with vanishing wall shear stress:

Beyond separation, the near-wall flow may reverse, and a recirculating wake forms. Separation increases pressure drag, causes unsteady vortex shedding, vibration, noise, and fluctuating structural loads. Rounded leading regions and gradual turning can delay separation, whereas abrupt corners, sharp changes in geometry, and strong adverse pressure gradients promote it.

Turbulent boundary layers have greater near-wall momentum transport and often resist separation better than laminar layers, but they also produce greater skin friction. This trade-off explains why controlled transition or surface treatments can sometimes reduce total drag.

5. Drag and Lift in External Flow

5.1 Force coefficients

The drag force component parallel to the free stream is represented by

and the lift force perpendicular to the free stream by

where F_D and F_L are forces (N), A is a reference area (m²), and ½ρU∞² is dynamic pressure (Pa). The reference area must be stated: it may be frontal area for a building or vehicle, planform area for an airfoil, or projected particle area.

Drag has two principal components:

• Skin-friction drag, caused by tangential viscous shear on the surface.
• Pressure or form drag, caused by pressure differences between the front and rear of a body, usually associated with separation and the wake.

For streamlined bodies skin friction can be significant, while for bluff bodies such as cylinders, rectangular signs, scaffolding components, and building blocks, pressure drag usually dominates.

5.2 Lift and circulation

Lift may arise from body shape, angle of attack, circulation, or asymmetric separation. For a two-dimensional ideal flow around a cylinder with circulation Γ, the Kutta–Joukowski lift per unit span is

where L′ is lift per unit span (N/m) and Γ is circulation (m²/s). Real fluids possess viscosity and may separate, so this ideal relation is not by itself sufficient for predicting lift on a real building or vehicle. It is most useful as a conceptual and aerodynamic model under appropriate assumptions.

5.3 Bluff-body wakes and vortex shedding

Flow around a circular cylinder can separate alternately from opposite sides, generating a von Kármán vortex street. The shedding frequency is often correlated using the Strouhal number:

where f is shedding frequency (Hz), D is cylinder diameter (m), and U∞ is free-stream velocity (m/s). For many subcritical flows around a circular cylinder, St is approximately 0.2, but it varies with Reynolds number, roughness, turbulence, and confinement. If the shedding frequency approaches a structure’s natural frequency, vortex-induced vibration may occur.

6. Flow Around Construction Elements and Vehicles

Wind acting on a building is influenced by height, terrain roughness, neighboring buildings, atmospheric turbulence, corners, balconies, and openings. A rectangular building generally experiences separated flow at sharp edges, producing large pressure differences and fluctuating wake forces. Design therefore requires more than a single steady drag coefficient: local pressures, torsion, cross-wind response, and dynamic amplification may be important.

For bridges, cables, sign supports, towers, and temporary works, slender members may experience drag and vortex-induced vibration. Surface roughness, helical strakes, fairings, spacing, and altered cross-sections can modify separation and vortex formation. Such interventions must be checked for their effects on strength, fatigue, maintenance, and debris accumulation.

Vehicles benefit from rounded noses, controlled rear contraction, underbody management, and careful separation control. A smooth shape is not automatically low-drag: a long body may increase skin friction, while a poorly designed rear can create a large wake and pressure drag.

For particles in water or air, drag controls settling, suspension, and transport. At low particle Reynolds number, a spherical particle follows Stokes’ law:

where d is particle diameter, V_r is relative velocity, and the assumptions include a sphere, creeping flow (Re_p ≪ 1), and a Newtonian fluid. At larger particle Reynolds numbers, the drag coefficient is not constant and must be obtained from an appropriate correlation or experiment.

7. Worked Example: Laminar Plate Friction

A smooth plate is 2.0 m long and 1.0 m wide. Air flows over it at U∞ = 10 m/s. Take ρ = 1.20 kg/m³ and ν = 1.50 × 10⁻⁵ m²/s. Estimate the average laminar skin-friction drag, assuming the entire plate remains laminar.

First calculate Reynolds number:

The assumed laminar correlation gives

Dynamic pressure is

The wetted area of one side is A = bL = (1.0)(2.0) = 2.0 m². Therefore,

The result is approximately 0.14 N for one side. In reality, Re_L is high enough that transition would commonly occur before the plate end unless disturbances were exceptionally small; therefore this is an idealized calculation, not necessarily a design prediction.

8. Applied Case: Wind-Induced Loads on a Rectangular Site Office

Background

Consider a temporary rectangular site office measuring 12 m long, 6 m wide, and 3 m high. It is located in an exposed construction site where a design wind speed of 30 m/s is specified. For illustration, take air density as 1.225 kg/m³ and an assumed net pressure coefficient of 1.2 on the windward projected face. Actual design must follow the governing national wind standard and account for terrain, gusts, internal pressure, shielding, and structural dynamics.

Problem

Estimate the idealized horizontal wind force on the 12 m × 3 m windward face and identify fluid-mechanics issues that may make the estimate incomplete.

Analysis and approach

The dynamic pressure is

The projected area is

Using F = C_p qA,

This estimate represents a resultant force under the stated coefficient convention. Flow separates at the roof and side edges, creating a turbulent wake and suction pressures that may be more important for cladding than the net force. Open doors or vents can produce internal pressure, and neighboring temporary buildings can either shield or amplify local loads. Flexible components may also experience dynamic response.

Outcome and learning points

The calculation provides an initial order-of-magnitude force, but it is not a complete structural design. The design process should use the applicable code, appropriate external and internal pressure coefficients, gust or dynamic factors, load combinations, connection checks, and local cladding pressures. The key lesson is that a coefficient is meaningful only when its reference area, flow condition, geometry, and loading definition are known.

9. Measurement, Computation, and Design Practice

Wind tunnels and water channels can measure mean forces, fluctuating forces, pressure distributions, and vortex frequencies. Similarity requires attention to Reynolds number, geometric scale, turbulence intensity, surface roughness, and blockage. Exact simultaneous matching of all dimensionless parameters is often impossible, so the engineer must identify which effects dominate.

Computational fluid dynamics can resolve or model external flow. Reynolds-averaged Navier–Stokes methods are economical for engineering averages but depend on turbulence models. Large-eddy simulation resolves more turbulent structures at higher computational cost. Direct numerical simulation is generally restricted to relatively low Reynolds numbers and research applications. Mesh quality, boundary conditions, wall treatment, convergence, and validation against measurements are essential.

AI and data-driven methods can assist with surrogate models, image-based flow diagnostics, anomaly detection, and rapid prediction of force coefficients. They do not remove the need for dimensional analysis, physical validation, uncertainty assessment, or professional judgment. A model trained outside its range of Reynolds number, geometry, or roughness may produce confidently incorrect results.

10. Limitations, Safety, and Professional Responsibility

External-flow correlations are empirical and geometry-dependent. Important limitations include three-dimensionality, surface roughness, atmospheric turbulence, unsteady separation, compressibility at high speed, free-surface effects, blockage, scale effects, and uncertain material or fluid properties. Engineers should document assumptions and report uncertainty rather than presenting a coefficient as universally valid.

For construction, fluid-force predictions affect public safety, temporary works, cranes, façade panels, roof systems, bridges, and environmental protection. Designs should comply with applicable building, wind, hydraulic, and occupational-safety regulations. Ethical practice requires transparent assumptions, traceable calculations, independent checking for critical work, and conservative treatment of uncertain or fluctuating loads.

Best-Practice Checklist

• Define the characteristic length, velocity, area, and fluid properties.
• Calculate Reynolds number before selecting a correlation.
• Distinguish local from average coefficients.
• State whether drag includes skin friction, pressure drag, or both.
• Check for separation, wake unsteadiness, and possible resonance.
• Account for roughness, turbulence, neighboring bodies, and blockage.
• Validate important predictions with code guidance, testing, or credible computational analysis.
• Record assumptions, units, reference areas, and uncertainty.

Chapter Summary

Viscosity produces shear stress whenever velocity varies across a fluid. Newtonian fluids obey a linear stress–velocity-gradient relation, while non-Newtonian fluids require more specialized constitutive models. Reynolds number compares inertial and viscous effects and helps identify laminar, transitional, or turbulent regimes, although transition depends on geometry and disturbances.

A boundary layer develops because of the no-slip condition at a wall. Its thickness and integral measures grow downstream, producing skin-friction drag. Adverse pressure gradients reduce near-wall momentum and may cause separation, recirculation, pressure drag, fluctuating forces, and vortex shedding. Drag and lift coefficients relate fluid forces to dynamic pressure and a stated reference area.

These principles apply to buildings, bridges, vehicles, temporary works, structural members, and particles. Engineering correlations are useful only within their assumptions. Reliable design combines dimensional analysis, appropriate empirical or computational models, experimental evidence where necessary, code requirements, and professional judgment.

Key Terms

Viscosity: Internal resistance to relative fluid motion.

Dynamic viscosity: Proportionality constant between shear stress and velocity gradient for a Newtonian fluid.

Kinematic viscosity: Dynamic viscosity divided by density.

Shear stress: Tangential force per unit area.

Newtonian fluid: Fluid with a linear shear-stress/velocity-gradient relationship.

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

Reynolds number: Dimensionless ratio of inertial to viscous effects.

Laminar flow: Relatively orderly flow with limited transverse mixing.

Turbulent flow: Flow containing irregular velocity fluctuations and eddies.

Boundary layer: Near-wall region strongly affected by viscosity.

Displacement thickness: Equivalent outward displacement caused by boundary-layer flow deficit.

Momentum thickness: Measure of momentum loss within a boundary layer.

Shape factor: Ratio of displacement thickness to momentum thickness.

Pressure gradient: Rate of pressure change in the flow direction.

Flow separation: Detachment of near-wall flow from a surface.

Skin-friction drag: Drag caused by wall shear stress.

Pressure drag: Drag caused by pressure differences, commonly across a separated wake.

Lift: Force component perpendicular to the free-stream direction.

Strouhal number: Dimensionless vortex-shedding frequency.

Dynamic pressure: Quantity ½ρU² used in force-coefficient definitions.

Review Questions

Short Answer Questions

1. State Newton’s law of viscosity and define its variables.
2. What physical effects are compared by Reynolds number?
3. Why does a boundary layer form on a stationary solid surface?
4. Distinguish skin-friction drag from pressure drag.
5. What condition at the wall is associated with boundary-layer separation?

Long Answer Questions

1. Explain the development of a laminar boundary layer over a flat plate and discuss the assumptions behind common thickness and skin-friction correlations.
2. Compare laminar and turbulent boundary layers with respect to momentum transfer, wall shear, velocity profile, and separation resistance.
3. Explain how adverse pressure gradients produce separation and why separation is important in building and bridge design.
4. Discuss the meaning of drag and lift coefficients, including the importance of reference area and flow conditions.
5. Describe how wind-tunnel testing and computational fluid dynamics can support external-flow design, including their principal limitations.

Analytical / Application Questions

1. Air at ρ = 1.20 kg/m³ and ν = 1.50 × 10⁻⁵ m²/s flows at 15 m/s over a 3 m smooth plate. Calculate Re_L and determine whether the fully laminar flat-plate correlation is physically plausible without transition control.
2. A cylindrical construction cable has diameter 0.08 m and is exposed to wind at 25 m/s. Explain which dimensionless quantities and structural response mechanisms should be checked before estimating drag and vortex-induced vibration.
3. A façade panel is tested in a small wind tunnel, but the model Reynolds number is much lower than the full-scale value. Analyze which flow features may not scale correctly and propose methods for improving confidence in the design result.

Multiple Choice Questions with Answers

Question 1

Which equation defines Reynolds number using dynamic viscosity?

A. Re = μUL/ρ

B. Re = ρUL/μ

C. Re = ρμ/(UL)

D. Re = U/(ρμL)

Correct Answer: B. Re = ρUL/μ

Question 2

For a Newtonian fluid, shear stress is proportional to which quantity?

A. Pressure only

B. Fluid density only

C. Velocity gradient normal to the flow

D. Free-stream area

Correct Answer: C. Velocity gradient normal to the flow

Question 3

At a stationary no-slip wall, the fluid velocity is:

A. Equal to the free-stream velocity

B. Zero relative to the wall

C. Always turbulent

D. Equal to the pressure

Correct Answer: B. Zero relative to the wall

Question 4

Which condition most directly promotes boundary-layer separation?

A. A favorable pressure gradient

B. A uniform zero velocity gradient everywhere

C. An adverse pressure gradient

D. Reduced fluid density only

Correct Answer: C. An adverse pressure gradient

Question 5

The drag coefficient is defined using which reference quantity?

A. ρU

B. ½ρU²A

C. μU/L

D. ρνA

Correct Answer: B. ½ρU²A

Question 6

For a smooth laminar flat plate, the local skin-friction coefficient varies approximately as:

A. Re_x^(1/2)

B. 1/Re_x

C. 1/√Re_x

D. Re_x²

Correct Answer: C. 1/√Re_x

Question 7

Which drag component is usually dominant for a sharp-edged bluff body?

A. Pressure or form drag

B. Only molecular diffusion drag

C. Lift-induced drag in every case

D. Zero drag because of separation

Correct Answer: A. Pressure or form drag

Question 8

If a cylinder has diameter 0.10 m, free-stream speed 20 m/s, and shedding frequency 40 Hz, its Strouhal number is:

A. 0.02

B. 0.20

C. 2.0

D. 40

Correct Answer: B. 0.20

Question 9

For creeping flow around a small sphere, Stokes’ drag law is proportional to:

A. ρV_r²d²

B. μdV_r

C. μ/(dV_r)

D. ρg/d

Correct Answer: B. μdV_r

Question 10

Why must the reference area be stated when reporting a force coefficient?

A. Because coefficients have units of metres

B. Because the numerical coefficient depends on the selected area convention

C. Because viscosity is otherwise zero

D. Because Reynolds number cannot be calculated

Correct Answer: B. Because the numerical coefficient depends on the selected area convention