Heat Transfer by Convection

What Is Convective Heat Transfer?

Convection occurs whenever a fluid—liquid or gas—moves over a surface that is at a different temperature. The fluid carries thermal energy with it as it flows, greatly enhancing the rate of heat transfer compared with conduction alone. At the molecular level, convection still relies on conduction within a thin layer of fluid adjacent to the surface (the boundary layer), but the bulk motion of the fluid continuously sweeps away the heated (or cooled) fluid and replaces it with fresh fluid at the freestream temperature, maintaining a steep temperature gradient and a high rate of energy exchange.

Convective heat transfer is characterised by the convective heat transfer coefficient $h$, which encapsulates the combined effects of fluid properties, flow velocity, surface geometry, and turbulence. A higher $h$ means more effective heat exchange—which is exactly why a fan-cooled heat sink outperforms a passive one by an order of magnitude.

Newton's Law of Cooling: The Fundamental Equation

The rate of convective heat transfer from a surface is expressed by Newton's law of cooling:

$${\displaystyle \dot{Q}_{\text{conv}} = h \, A_s \left(T_s - T_{\infty}\right)}$$

Where:

• $\dot{Q}_{\text{conv}}$ is the convective heat transfer rate $\left[\mathrm{W}\right]$.
• $h$ is the convective heat transfer coefficient $\left[\mathrm{W \, m^{-2} \, K^{-1}}\right]$.
• $A_s$ is the surface area exposed to the fluid $\left[\mathrm{m^2}\right]$.
• $T_s$ is the surface temperature and $T_{\infty}$ is the bulk fluid temperature $\left[\mathrm{K \; or \; °C}\right]$.

Deceptively simple in appearance, this equation hides enormous complexity within $h$. Determining $h$ accurately for a given situation is the central challenge of convective heat transfer analysis—and the reason why empirical correlations, dimensional analysis, and numerical simulation play such important roles.

Natural Convection vs. Forced Convection

Natural (Free) Convection

When fluid motion arises solely from buoyancy forces—density differences caused by temperature gradients—we speak of natural or free convection. A classic example is the plume of warm air rising above a hot radiator: as air near the surface heats up, it becomes less dense and rises, drawing cooler air in from below. Natural convection heat transfer coefficients are relatively low, typically in the range of 2–25 W/(m²·K) for gases and 50–1000 W/(m²·K) for liquids, because the fluid velocities involved are modest.

The dimensionless Grashof number quantifies the ratio of buoyancy forces to viscous forces and determines whether the natural convective flow will be laminar or turbulent:

$${\displaystyle \mathrm{Gr}_L = \frac{g \, \beta \left(T_s - T_{\infty}\right) L^3}{\nu^2}}$$

where $g$ is gravitational acceleration, $\beta$ is the volumetric thermal expansion coefficient, $L$ is a characteristic length, and $\nu$ is the kinematic viscosity. The Grashof number plays a role in natural convection analogous to the Reynolds number in forced convection.

Forced Convection

When the fluid motion is driven by an external mechanism—a pump, fan, or wind—the resulting heat transfer is called forced convection. Because external forcing generates much higher fluid velocities, forced convection yields significantly higher heat transfer coefficients: 25–250 W/(m²·K) for air and up to 100,000 W/(m²·K) for water or other liquids under turbulent flow conditions. Virtually all active cooling systems—from laptop fans to automotive radiator circuits—rely on forced convection.

The Boundary Layer Concept

Ludwig Prandtl introduced the boundary layer concept in 1904, transforming fluid mechanics and heat transfer theory. When fluid flows over a surface, a thin region develops near the surface where the velocity transitions from zero (no-slip condition at the wall) to the freestream value. This is the velocity boundary layer. Analogously, a thermal boundary layer forms where the temperature transitions from the surface value to the freestream temperature.

The relative thickness of these two boundary layers determines the nature of the convective process. Their ratio is characterised by the Prandtl number:

$${\displaystyle \mathrm{Pr} = \frac{\nu}{\alpha} = \frac{\mu \, c_p}{k}}$$

where $\nu$ is kinematic viscosity, $\alpha$ is thermal diffusivity, $\mu$ is dynamic viscosity, $c_p$ is specific heat, and $k$ is thermal conductivity. For air, $\mathrm{Pr} \approx 0.71$, meaning the thermal boundary layer is slightly thicker than the velocity boundary layer. For water, $\mathrm{Pr} \approx 7$ at room temperature, and for engine oils, $\mathrm{Pr}$ can exceed 1000—indicating a very thin thermal boundary layer compared with the velocity boundary layer.

Key Dimensionless Numbers

Dimensional analysis is the workhorse of convective heat transfer. By expressing the governing equations in dimensionless form, engineers can correlate experimental data into universally applicable relationships. The most important dimensionless groups are:

Reynolds Number

$${\displaystyle \mathrm{Re}_L = \frac{\rho \, u \, L}{\mu} = \frac{u \, L}{\nu}}$$

The ratio of inertial to viscous forces. Below a critical Reynolds number (about 5 × 10⁵ for flow over a flat plate), the flow remains laminar; above it, the flow transitions to turbulence, dramatically enhancing mixing and heat transfer.

Nusselt Number

$${\displaystyle \mathrm{Nu}_L = \frac{h \, L}{k_f}}$$

The ratio of convective to conductive heat transfer across the boundary layer. A Nusselt number of 1 would mean convection offers no advantage over pure conduction. In practice, Nusselt numbers range from around 10 for gentle laminar flows to several hundred or more in turbulent flows, reflecting the enormous enhancement that fluid motion provides.

Rayleigh Number

For natural convection, the product $\mathrm{Ra} = \mathrm{Gr} \cdot \mathrm{Pr}$ combines buoyancy and fluid property effects into a single parameter. Correlations of the form $\mathrm{Nu} = C \, \mathrm{Ra}^n$ are widely used for standard geometries such as vertical plates, horizontal cylinders, and heated enclosures.

Common Empirical Correlations

Forced convection over a flat plate (laminar)

For a laminar boundary layer ($\mathrm{Re}_L < 5 \times 10^5$), the average Nusselt number is:

$${\displaystyle \overline{\mathrm{Nu}}_L = 0.664 \; \mathrm{Re}_L^{1/2} \; \mathrm{Pr}^{1/3}}$$

Forced convection over a flat plate (turbulent)

For a fully turbulent boundary layer ($\mathrm{Re}_L > 5 \times 10^5$):

$${\displaystyle \overline{\mathrm{Nu}}_L = 0.037 \; \mathrm{Re}_L^{4/5} \; \mathrm{Pr}^{1/3}}$$

Natural convection on a vertical plate

Churchill and Chu's widely used correlation covers the entire Rayleigh number range:

$${\displaystyle \overline{\mathrm{Nu}}_L = \left[ 0.825 + \frac{0.387 \; \mathrm{Ra}_L^{1/6}}{\left(1 + \left(0.492 / \mathrm{Pr}\right)^{9/16}\right)^{8/27}} \right]^2}$$

These correlations allow engineers to estimate convective coefficients without full-scale numerical simulation, though for complex geometries or flow conditions, computational thermal analysis using CFD offers far greater accuracy and spatial resolution.

Engineering Applications

1. Electronics Cooling

Modern processors dissipate power densities exceeding 100 W/cm² in some areas. Removing this heat efficiently is critical to reliability and performance. Heat sinks with extended fins increase the surface area $A_s$, while fans drive forced convection to raise $h$. Advanced systems use liquid cooling loops—exploiting water's high thermal capacity and convective performance—to push heat transfer rates even further. The design of these systems requires careful optimisation of fin geometry, flow rates, and pressure drops.

2. Automotive Thermal Management

An internal combustion engine converts roughly one-third of fuel energy into waste heat that must be removed by the cooling system. Coolant circulates through passages in the engine block (forced convection from metal to liquid), then passes through a radiator where a fan drives air over finned tubes (forced convection from liquid to air). Each stage involves convective heat transfer, and optimising the overall system demands a thorough understanding of flow regimes, pressure losses, and heat exchanger effectiveness.

3. HVAC and Building Climate Control

Heating, ventilation, and air conditioning systems rely almost entirely on convective heat transfer. Warm or cool air is distributed through ducts and delivered to rooms, where it mixes with ambient air through natural and forced convection. The sizing of heat exchangers, ductwork, and diffusers all depend on convective heat transfer calculations. Natural convection also plays a role in passive building design strategies such as solar chimneys and Trombe walls.

4. Heat Exchangers

From shell-and-tube exchangers in petrochemical plants to plate heat exchangers in food processing, these devices are purpose-built to maximise convective heat transfer between two fluid streams. Their design hinges on correlating flow velocities, turbulence levels, and surface geometries to achieve the required heat duty within acceptable pressure drop limits. The effectiveness-NTU method and LMTD approach are standard analysis tools, but modern designs increasingly benefit from high-fidelity CFD simulation that resolves the conjugate heat transfer between fluid and solid domains.

Case Study: Forced Air Cooling of a Heat Sink

Consider a finned aluminium heat sink with a total exposed area of $A_s = 0.05 \; \mathrm{m^2}$ attached to a processor dissipating $\dot{Q} = 65 \; \mathrm{W}$. A fan drives air at 25 °C over the fins, and the estimated convective coefficient is $h = 45 \; \mathrm{W/(m^2 \cdot K)}$.

From Newton's law of cooling, the surface temperature rise above ambient is:

$${\displaystyle \Delta T = T_s - T_{\infty} = \frac{\dot{Q}}{h \, A_s} = \frac{65}{45 \times 0.05} = \frac{65}{2.25} \approx 28.9 \; \mathrm{°C}}$$

So the heat sink surface reaches approximately 54 °C—well within the typical operating range for most processors. Increasing the airflow (raising $h$ to, say, 80 W/(m²·K)) would reduce the temperature rise to about 16 °C, demonstrating the direct leverage that forced convection provides in thermal management.

Note that this simplified analysis assumes uniform $h$ across all surfaces. In practice, the convective coefficient varies significantly along the fin surfaces and between the fin channels. Accurate prediction of the full temperature field typically requires detailed thermal simulation using conjugate heat transfer models that resolve both the solid and fluid domains simultaneously.

Mixed and Combined Convection

In many situations, natural and forced convection coexist. Consider warm air rising along a heated vertical wall (natural convection) while a ventilation system blows air across the same wall (forced convection). Whether the two effects reinforce or oppose each other depends on their relative direction. The Richardson number, $\mathrm{Ri} = \mathrm{Gr} / \mathrm{Re}^2$, serves as the criterion: when $\mathrm{Ri} \ll 1$, forced convection dominates; when $\mathrm{Ri} \gg 1$, natural convection prevails; and when $\mathrm{Ri} \approx 1$, both mechanisms must be considered together.

Real-world convection problems also frequently involve simultaneous thermal radiation, especially at elevated temperatures. The total surface heat loss then combines convective and radiative contributions—a coupling that adds complexity but is essential for accurate thermal predictions in applications ranging from solar receivers to engine exhaust manifolds.

Conclusion

Convection is the dominant mode of heat transfer in countless engineering systems, from the smallest chip-level heat sink to the largest industrial heat exchanger. Its richness stems from the coupling of fluid dynamics with thermal transport, governed by dimensionless parameters like the Reynolds, Nusselt, Prandtl, and Grashof numbers. Whether the flow is natural or forced, laminar or turbulent, the convective heat transfer coefficient $h$ remains the critical quantity—and determining it accurately is both an art and a science. Empirical correlations serve well for standard geometries, but modern engineering increasingly demands the spatial resolution and physical fidelity that only computational methods can provide. For projects where thermal performance is critical, professional thermal analysis services combining CFD and FEA offer the insight needed to optimise designs, avoid thermal failures, and push the boundaries of what is possible.