13.2 Forced/Natural Convection, Radiation & Heat Exchanger LMTD/NTU
Key Takeaways
- Convective heat transfer relies on dimensionless parameters: Nu = h*L/k_fluid (enhancement over pure conduction), Re = rho*V*L/mu (inertia vs viscous), Pr = nu/alpha (momentum vs thermal diffusivity), and Gr = g*beta*Delta_T*L^3/nu^2 (buoyancy vs viscous).
- Flow regime classification governs correlation selection: forced convection dominates when Gr/Re^2 << 1, natural convection dominates when Gr/Re^2 >> 1, and mixed convection occurs when Gr/Re^2 approx 1; turbulent tube flow follows Dittus-Boelter Nu_D = 0.023*Re_D^0.8*Pr^n (n=0.4 heating, n=0.3 cooling).
- Radiation heat transfer between grey bodies obeys the Stefan-Boltzmann law E_b = sigma*T^4, Wien's displacement law lambda_max*T = 2898 micron*K, and view factor algebra (sum F_ij = 1, A_i*F_ij = A_j*F_ji); n radiation shields reduce net radiative transfer by a factor of 1/(n+1).
- Heat exchanger rating uses LMTD = (Delta_T1 - Delta_T2)/ln(Delta_T1/Delta_T2); counter-flow geometry yields LMTD_counter > LMTD_parallel for identical terminal temperatures, requiring less surface area and allowing cold fluid outlet temperature to exceed hot fluid outlet temperature.
- The Effectiveness-NTU method evaluates heat exchangers where NTU = UA/C_min and epsilon = Q_act/Q_max; for phase-change components (condensers/evaporators) where capacity ratio C_r = C_min/C_max = 0, effectiveness simplifies to epsilon = 1 - exp(-NTU) for all flow configurations.
Forced/Natural Convection, Radiation & Heat Exchanger LMTD/NTU
Thermal energy exchange in industrial mining complexes encompasses convective cooling of heavy mining haul truck engines, radiation inside coal-fired power boilers, and heat recovery via shell-and-tube heat exchangers. This section establishes the theoretical foundations and empirical correlations for convective transport, radiative enclosures, and heat exchanger thermal design.
1. Convective Heat Transfer Fundamentals & Non-Dimensional Numbers
Convection heat transfer occurs between a solid surface and a moving fluid governed by Newton's Law of Cooling:
Where $h\ (\text{W/m}^2\cdot\text{K})$ is the local or average convective heat transfer coefficient. Because $h$ is a complex function of fluid properties ($\rho, \mu, c_p, k$), flow velocity ($V$), and surface geometry ($L$), dimensional analysis groups these variables into foundational dimensionless numbers.
+-----------------------------------------------------------------------------------------+
| FOUNDATIONAL NON-DIMENSIONAL NUMBERS |
| |
| Dimensionless Group Definition Physical Significance |
| -------------------- ----------------------- --------------------------------- |
| Nusselt Number (Nu) Nu = h * L / k_fluid Ratio of convective to conductive |
| heat transfer across fluid layer. |
| Reynolds Number (Re) Re = rho * V * L / mu Ratio of dynamic inertia forces |
| = V * L / nu to viscous shear forces. |
| Prandtl Number (Pr) Pr = nu / alpha Ratio of momentum diffusivity to |
| = mu * cp / k_fluid thermal diffusivity. |
| Grashof Number (Gr) Gr = g*beta*DeltaT*L^3/nu^2 Ratio of natural buoyancy forces |
| to viscous hydrodynamic forces. |
| Rayleigh Number (Ra) Ra = Gr * Pr Governs laminar-to-turbulent |
| transition in natural convection. |
+-----------------------------------------------------------------------------------------+
Boundary Layer Mechanics & The Prandtl Number ($\text{Pr}$)
When a fluid flows over a surface, two distinct boundary layers develop:
- Hydrodynamic Boundary Layer ($\delta$): Region where velocity transitions from $u=0$ at the wall (no-slip condition) to $0.99 V_{\infty}$.
- Thermal Boundary Layer ($\delta_t$): Region where fluid temperature transitions from $T_s$ at the wall to $0.99 T_{\infty}$.
The relative thickness of these two boundary layers is governed directly by the Prandtl number:
- Liquid Metals ($\text{Pr} \sim 0.005\text{--}0.03$): $\delta_t \gg \delta$. Thermal conduction outpaces momentum diffusion; thermal boundary layer extends far into free stream.
- Gases / Air ($\text{Pr} \sim 0.70\text{--}0.80$): $\delta_t \approx \delta$. Velocity and temperature profiles develop at nearly identical rates.
- Water ($\text{Pr} \sim 2\text{--}10$): $\delta > \delta_t$. Viscous hydrodynamic boundary layer is thicker than the thermal layer.
- Heavy Oils ($\text{Pr} \sim 100\text{--}100,000$): $\delta \gg \delta_t$. High viscosity suppresses momentum while thermal energy diffuses slowly.
Regime Selection: Forced vs. Natural vs. Mixed Convection
The relative dominance of natural versus forced convection is determined by the ratio $\frac{\text{Gr}}{\text{Re}^2}$:
- $\frac{\text{Gr}}{\text{Re}^2} \ll 1$: Forced convection dominates; buoyancy effects are entirely negligible.
- $\frac{\text{Gr}}{\text{Re}^2} \gg 1$: Natural (free) convection dominates; forced velocity effects are negligible.
- $\frac{\text{Gr}}{\text{Re}^2} \approx 1$: Mixed convection; both buoyancy and forced inertia forces must be accounted for ($Nu_{\text{combined}} = [Nu_{\text{forced}}^n \pm Nu_{\text{natural}}^n]^{1/n}$).
2. Forced Convection Empirical Correlations
1. Fully Developed Laminar Flow in Circular Tubes ($\text{Re}_D < 2300$)
For fully developed internal laminar pipe flow, the Nusselt number is a constant value independent of Reynolds and Prandtl numbers:
2. Fully Developed Turbulent Flow in Circular Tubes ($\text{Re}_D > 10,000$)
The classical Dittus-Boelter Equation applies for smooth circular tubes under moderate temperature differences ($0.6 \le \text{Pr} \le 160$, $L/D > 10$):
[!NOTE] Property Evaluation: In the Dittus-Boelter correlation, all fluid physical properties ($\rho, \mu, c_p, k$) are evaluated at the bulk mean fluid temperature $T_b = (T_{m,i} + T_{m,o})/2$.
When large temperature differences cause significant radial variation in fluid dynamic viscosity, the Sieder-Tate Correlation is applied:
Where $\mu_b$ is evaluated at bulk temperature and $\mu_w$ is evaluated at the tube wall temperature.
3. Natural (Free) Convection Correlations
Natural convection is driven by density gradients in the presence of a body force (gravity $g$). The volumetric coefficient of thermal expansion $\beta$ is defined as:
Vertical Plates and Cylinders of Height $L$
- Laminar Flow ($\text{Ra}_L < 10^9$): $\overline{\text{Nu}}_L = 0.59 , \text{Ra}_L^{1/4}$
- Turbulent Flow ($\text{Ra}_L > 10^9$): $\overline{\text{Nu}}_L = 0.10 , \text{Ra}_L^{1/3}$
[!IMPORTANT] Height Independence in Turbulent Natural Convection: In the turbulent regime where $\overline{\text{Nu}}_L \propto \text{Ra}_L^{1/3} \propto (L^3)^{1/3} \propto L$, we find $\frac{h L}{k} \propto L \implies h = \text{constant}$. The average convective heat transfer coefficient $h$ becomes completely independent of vertical plate height $L$ in turbulent free convection.
4. Radiation Fundamentals & Governing Laws
Thermal radiation is electromagnetic radiation emitted by matter as a result of its temperature, propagating at the speed of light ($c_0 = 3 \times 10^8\text{ m/s}$) across wavelengths $\lambda = 0.1\text{--}100\ \mu\text{m}$ (spanning ultraviolet, visible, and infrared spectra).
+-----------------------------------------------------------------------------------------+
| FUNDAMENTAL RADIATION LAWS |
| |
| Law Mathematical Formula Physical Meaning |
| -------------------- --------------------------------- ----------------------- |
| Stefan-Boltzmann Law E_b = sigma * T^4 Total emissive power of |
| sigma = 5.67e-8 W/(m^2*K^4) an ideal black body. |
| Planck's Law E_b_lambda(lambda, T) Spectral distribution |
| = C1 / [lambda^5 * (exp(C2/lambda*T) - 1)] of black body radiation.|
| Wien's Displacement lambda_max * T = 2898 micron*K Peak emission wavelength|
| = 2.898e-3 m*K shifts inversely with T.|
| Kirchhoff's Law alpha_lambda = epsilon_lambda Monochromatic absorptivity|
| alpha = epsilon (Grey Body) equals emissivity at Teq|
+-----------------------------------------------------------------------------------------+
Surface Radiation Properties
For any incident radiative irradiation $G\ (\text{W/m}^2)$ striking a surface:
Where $\alpha$ is absorptivity, $\rho_r$ is reflectivity, and $\tau_r$ is transmissivity.
- Opaque Body ($\tau_r = 0$): $\alpha + \rho_r = 1$.
- Black Body ($\alpha = 1, \rho_r = 0, \tau_r = 0$): Perfect absorber and perfect emitter ($\epsilon = 1$).
- Grey Body: A surface whose monochromatic emissivity $\epsilon_{\lambda}$ and absorptivity $\alpha_{\lambda}$ are independent of wavelength across the spectral band: $\epsilon = \alpha = \text{const} < 1$.
5. View Factor (Shape Factor) Algebra & Enclosure Theory
The view factor (or configuration/shape factor) $F_{ij}$ is defined as the fraction of diffuse radiation leaving surface $i$ that is directly intercepted by surface $j$:
+-----------------------------------------------------------------------------------------+
| VIEW FACTOR ALGEBRA RULES |
| |
| 1. Reciprocity Relation: A_i * F_ij = A_j * F_ji |
| 2. Summation Rule (Enclosure): sum_{j=1}^N F_ij = 1 |
| 3. Flat / Convex Surface: F_ii = 0 (A surface cannot see itself) |
| 4. Concave Surface: F_ii > 0 (A concave cavity radiates to itself) |
| 5. Superposition Rule: F_1(2+3) = F_12 + F_13 |
| F_(2+3)1 = (A_2*F_21 + A_3*F_31) / (A_2 + A_3) |
+-----------------------------------------------------------------------------------------+
Standard 2-Surface View Factor Geometries
- Hemisphere Cavity (Surface 1) closed by Flat Disc Base (Surface 2):
- Base 2 is flat: $F_{22} = 0 \implies F_{21} = 1$.
- By Reciprocity: $A_1 F_{12} = A_2 F_{21} \implies (2\pi R^2) F_{12} = (\pi R^2)(1) \implies F_{12} = 0.5$.
- By Summation: $F_{11} + F_{12} = 1 \implies F_{11} = 1 - 0.5 = 0.5$.
- Infinite Concentric Cylinders (Inner 1, Outer 2):
- Inner cylinder is convex: $F_{11} = 0 \implies F_{12} = 1$.
- By Reciprocity: $A_1 F_{12} = A_2 F_{21} \implies F_{21} = \frac{A_1}{A_2} = \frac{r_1}{r_2}$.
- By Summation: $F_{22} = 1 - F_{21} = 1 - \frac{r_1}{r_2}$.
6. Radiative Heat Exchange Between Grey Surfaces & Radiation Shields
Using the radiation electrical network analogy, each real surface has a surface resistance and each pair of surfaces shares a space resistance:
+-----------------------------------------------------------------------------------------+
| TWO-SURFACE RADIATION NETWORK CIRCUIT |
| |
| [E_b1 = sigma*T1^4] [E_b2 = sigma*T2^4] |
| o---/\/\/\/\/\/---[J1]---/\/\/\/\/\/---[J2]---/\/\/\/\/\/---o |
| R_surface,1 R_space R_surface,2 |
| (1-eps1)/(eps1*A1) 1/(A1*F12) (1-eps2)/(eps2*A2) |
| |
| Net Heat Exchange Rate: |
| q_12 = sigma * (T1^4 - T2^4) / [ (1-eps1)/(eps1*A1) + 1/(A1*F12) + (1-eps2)/(eps2*A2) ]|
+-----------------------------------------------------------------------------------------+
1. Two Infinite Parallel Grey Plates ($A_1 = A_2 = A, F_{12} = 1$)
2. Radiation Shields
A radiation shield is a thin, highly reflective sheet placed between two radiating surfaces to introduce additional surface and space resistances without generating or absorbing work.
+-----------------------------------------------------------------------------------------+
| RADIATION SHIELD ATTENUATION |
| |
| For n identical thin shields (emissivity eps_s = eps_1 = eps_2) inserted between |
| two large parallel plates: |
| |
| (q_12)_with_n_shields = [ 1 / (n + 1) ] * (q_12)_without_shield |
| |
| - 1 Shield (n=1): Reduces radiative heat transfer by 50% (Factor = 1/2) |
| - 2 Shields (n=2): Reduces radiative heat transfer by 66.7% (Factor = 1/3) |
| - 3 Shields (n=3): Reduces radiative heat transfer by 75% (Factor = 1/4) |
+-----------------------------------------------------------------------------------------+
7. Heat Exchanger Analysis: Log Mean Temperature Difference (LMTD)
Heat exchangers facilitate thermal transfer between two fluids separated by a solid wall. The fundamental rating equation is:
Where $U$ is the overall heat transfer coefficient, $A$ is the heat transfer area, and $\Delta T_m$ is the Log Mean Temperature Difference (LMTD).
+-----------------------------------------------------------------------------------------+
| PARALLEL FLOW VS COUNTER FLOW PROFILES |
| |
| PARALLEL FLOW: COUNTER FLOW: |
| Temp Temp |
| ^ T_h,in ^ T_h,in |
| | * | * |
| | \ Hot Fluid | \ Hot Fluid T_c,out |
| | \---------> T_h,out | \-------------> * |
| | * | \ Cold Fluid |
| | /---------> T_c,out | /-----------------\ |
| | / Cold Fluid | / * T_c,in |
| | * T_c,in | * T_h,out |
| +----------------------> Area +----------------------> Area |
| DeltaT1 DeltaT2 DeltaT1 DeltaT2 |
| (Inlet) (Outlet) (Left End) (Right End) |
+-----------------------------------------------------------------------------------------+
- Parallel Flow: $\Delta T_1 = T_{h,\text{in}} - T_{c,\text{in}}$, $\Delta T_2 = T_{h,\text{out}} - T_{c,\text{out}}$.
- Counter Flow: $\Delta T_1 = T_{h,\text{in}} - T_{c,\text{out}}$, $\Delta T_2 = T_{h,\text{out}} - T_{c,\text{in}}$.
[!IMPORTANT] Why Counter-Flow is Thermodynamically Superior:
- For identical inlet and outlet temperatures, $\text{LMTD}{\text{counter}} > \text{LMTD}{\text{parallel}}$, meaning a counter-flow heat exchanger requires less surface area ($A$) to transfer the same thermal duty.
- In counter-flow, the cold fluid exit temperature can exceed the hot fluid exit temperature ($T_{c,\text{out}} > T_{h,\text{out}}$), which is physically impossible in parallel flow.
Special Mathematical Cases of LMTD
- Balanced Counter-Flow ($C_h = C_c \implies \Delta T_1 = \Delta T_2 = \Delta T$): The temperature difference is constant throughout the exchanger. Direct evaluation of LMTD yields $0/0$; applying L'Hôpital's rule gives:
- Phase Change Components (Condensers and Boilers/Evaporators): One fluid changes phase at constant saturation temperature ($C_{\text{max}} \to \infty$). The temperature profile is identical regardless of flow direction:
8. Heat Exchanger Analysis: The Effectiveness-NTU ($\epsilon\text{-NTU}$) Method
When fluid outlet temperatures are unknown, the LMTD method requires tedious iterative trials. The Effectiveness-NTU method solves performance directly from known inlet conditions.
1. Fluid Heat Capacity Rates
2. Maximum Possible Heat Transfer Rate ($Q_{\text{max}}$)
3. Heat Exchanger Effectiveness ($\epsilon$)
4. Number of Transfer Units ($\text{NTU}$)
+-----------------------------------------------------------------------------------------+
| EFFECTIVENESS FORMULAS BY CONFIGURATION |
| |
| Configuration Capacity Ratio (Cr) Effectiveness Formula (epsilon) |
| --------------------- ------------------- ----------------------------------- |
| All Exchangers (Boiler/ C_r = 0 epsilon = 1 - exp( -NTU ) |
| Condenser phase change) (C_max -> infinity) |
| Counter Flow C_r = 1 (Balanced) epsilon = NTU / (1 + NTU) |
| Parallel Flow C_r = 1 epsilon = [ 1 - exp(-2*NTU) ] / 2 |
| Counter Flow (General) 0 < C_r < 1 epsilon = [1 - exp(-NTU*(1-Cr))] / |
| [1 - Cr*exp(-NTU*(1-Cr))] |
+-----------------------------------------------------------------------------------------+
9. Step-by-Step Worked Problems: Heat Exchanger Design & Radiation Enclosures
Worked Example 1: Mining Compressor Intercooler Sizing via $\epsilon\text{-NTU}$
An air compressor intercooler operates in counter-flow to cool compressed air from $T_{h,\text{in}} = 140^{\circ}\text{C}$ to $T_{h,\text{out}} = 40^{\circ}\text{C}$ using cooling water entering at $T_{c,\text{in}} = 20^{\circ}\text{C}$. The heat capacity rate of air is $C_h = 2.0\text{ kW/K}$ and cooling water is $C_c = 4.0\text{ kW/K}$. The overall heat transfer coefficient is $U = 250\text{ W/m}^2\cdot\text{K}$. Calculate the required heat transfer surface area $A$.
+-----------------------------------------------------------------------------------------+
| INTERCOOLER SIZING CALCULATION STEPS |
| |
| STEP 1: Determine Capacity Rates and Ratio |
| C_min = C_h = 2.0 kW/K = 2000 W/K |
| C_max = C_c = 4.0 kW/K = 4000 W/K |
| C_r = C_min / C_max = 2000 / 4000 = 0.50 |
| |
| STEP 2: Calculate Actual Heat Transfer Rate |
| Q_act = C_h * (T_h,in - T_h,out) = 2.0 kW/K * (140 - 40) = 200 kW |
| |
| STEP 3: Calculate Maximum Possible Heat Transfer & Effectiveness |
| Q_max = C_min * (T_h,in - T_c,in) = 2.0 kW/K * (140 - 20) = 240 kW |
| epsilon = Q_act / Q_max = 200 / 240 = 0.8333 (83.33%) |
| |
| STEP 4: Solve for NTU for Counter Flow (Cr = 0.50, epsilon = 0.8333) |
| NTU = [ 1 / (1 - C_r) ] * ln[ (1 - epsilon*C_r) / (1 - epsilon) ] |
| NTU = [ 1 / 0.50 ] * ln[ (1 - 0.8333*0.50) / (1 - 0.8333) ] |
| NTU = 2.0 * ln[ (1 - 0.41667) / 0.16667 ] = 2.0 * ln[ 0.58333 / 0.16667 ] |
| NTU = 2.0 * ln(3.50) = 2.0 * 1.25276 = 2.5055 |
| |
| STEP 5: Calculate Area A |
| NTU = U*A / C_min ===> A = (NTU * C_min) / U |
| A = (2.5055 * 2000 W/K) / (250 W/m^2*K) = 5011 / 250 = 20.04 m^2 |
+-----------------------------------------------------------------------------------------+
Water flows inside a circular tube under fully developed turbulent flow conditions (Re = 40,000, Pr = 4.5). If the fluid is being heated by the tube wall, which exponent n in the Dittus-Boelter correlation (Nu = 0.023·Re^0.8·Pr^n) must be selected, and what happens to the heat transfer coefficient h if the flow velocity is doubled?
A cylindrical combustion chamber of radius R and height H is enclosed by a flat base (Surface 1), a flat roof (Surface 2), and a cylindrical curved side wall (Surface 3). If the view factor between the flat base and flat roof is F_12 = 0.30, what is the self-view factor of the cylindrical side wall (F_33)? (Assume H = 2R, so A_1 = A_2 = piR^2 and A_3 = 2piRH = 4piR^2).
In an industrial steam condenser where exhaust steam condenses at a constant temperature of 50°C, cooling water enters at 20°C and leaves at 35°C. If the heat exchanger has an NTU value of 0.693, what is the thermal effectiveness (ε) of the condenser?