08-03-2024

MODULE-II
(Convection)

Dittus -Bolter Correlation

𝑁 = 0.023 𝑅𝑒 . 𝑃𝑟
(i) Turbulent flow 𝑅𝑒 ≥ 10000
(ii) Fully developed flow ( No entrance effect) (i) n = 0.4 for heating a fluid (𝑇 > 𝑇 )
(iii) Smooth Pipe ( No roughness) (ii) n = 0.3 for cooling a fluid (𝑇 < 𝑇 )
o 0.7 ≤ 𝑃𝑟 ≤ 160
o < 0.1
o Moderate temperature difference (tube wall (vs) bulk fluid)
(i.e No viscosity correction)

Sieder & Tate Correlation

.
(i) 𝑁 = 0.027 𝑅𝑒 . (𝑃𝑟) .

𝜇 & 𝜇 − Viscosity of fluid @ bulk temp, Viscosity of fluid @ wall temp

0.7 ≤ 𝑃𝑟 ≤ 16700
𝑅𝑒 ≥ 10000
𝑑
≤ 0.1
𝐿
Evaluate Physical Properties of bulk fluid @ arithmetic mean bulk temp
2
 08-03-2024

Physical significance of the groups

𝐼𝑛𝑒𝑟𝑡𝑖𝑎 𝑓𝑜𝑟𝑐𝑒𝑠
𝑅𝑒𝑦𝑛𝑜𝑙𝑑𝑠 𝑁𝑢𝑚𝑏𝑒𝑟 =
𝑉𝑖𝑠𝑐𝑜𝑢𝑠 𝑓𝑜𝑟𝑐𝑒𝑠

𝐵𝑢𝑜𝑦𝑎𝑛𝑐𝑦 𝑓𝑜𝑟𝑐𝑒 ∗ 𝐼𝑛𝑒𝑟𝑡𝑖𝑎 𝑓𝑜𝑟𝑐𝑒𝑠

𝐺𝑟𝑎𝑠ℎ𝑜𝑓 𝑁𝑢𝑚𝑏𝑒𝑟 =
(𝑉𝑖𝑠𝑐𝑜𝑢𝑠 𝑓𝑜𝑟𝑐𝑒𝑠)

𝑚𝑜𝑙𝑒𝑐𝑢𝑙𝑎𝑟 𝑑𝑖𝑓𝑓𝑢𝑠𝑖𝑣𝑖𝑡𝑦 𝑜𝑓 𝑚𝑜𝑚𝑒𝑛𝑡𝑢𝑚

𝑃𝑟𝑎𝑛𝑑𝑡𝑙 𝑁𝑢𝑚𝑏𝑒𝑟 =
𝑚𝑜𝑙𝑒𝑐𝑢𝑙𝑎𝑟 𝑑𝑖𝑓𝑓𝑢𝑠𝑖𝑣𝑖𝑡𝑦 𝑜𝑓 ℎ𝑒𝑎𝑡

𝑊𝑎𝑙𝑙 ℎ𝑒𝑎𝑡 𝑡𝑟𝑎𝑛𝑠𝑓𝑒𝑟 𝑟𝑎𝑡𝑒

𝑁𝑢𝑠𝑠𝑒𝑙𝑡 𝑁𝑢𝑚𝑏𝑒𝑟 =
𝐻𝑒𝑎𝑡 𝑡𝑟𝑎𝑛𝑠𝑓𝑒𝑟 𝑏𝑦 𝑐𝑜𝑛𝑑𝑢𝑐𝑡𝑖𝑜𝑛

𝑁 = = 𝑁 = = 𝑃𝑖𝑝𝑒

∆ ∆
𝑁 = 𝑁 = = (𝑝𝑖𝑝𝑒)
/
𝑁 = = = = 3
/

Summary (dimensionless groups in convective heat transfer)

Name Expression Physical significance
Nusselt number ℎ𝑙 Wall temprature gradient
𝑁 =
𝑘 Temprature gradient across the fluid in the pipe
Reynolds number 𝑙𝑉𝜌 Inertial force
𝑅 =
𝜇 Viscous force
Prandtl number 𝑐 𝜇 Momentum diffusity
𝑃 =
𝑘 Thermal diffusity
Stanton number ℎ Rate of heat transfer by convection
𝑆 =
𝑉𝜌𝑐 Rate of heat transfer by bulk flow
/
= =
( / )( / ) ( )( )

Peclet number 𝜌𝑉𝑙𝑐 𝜌𝑉𝑙 𝑐 𝜇 Rate of heat transfer by bulk flow

𝑃 = =
𝑘 𝜇 𝑘 Rate of heat transfer by conductiion
=𝑅 𝑃
Graetz number 𝐺 = 𝑃 𝑑/𝐿 = 𝑅 𝑃 𝑑/𝐿 Similar to Peclet number (used in connection with analysis
4
of heat transfer in laminar flow in pipes)
 08-03-2024

Thermal entrance region (NU-Graetz problem)

 Objective : To calculate temp profile & local NU
(i.e before the temp profile becomes fully developed)
Assumption : Velocity profile is fully developed before heating
𝜕 𝑇 1𝜕 𝜕𝑇 𝑟 𝜕𝑇
𝐾 + 𝑟 = 𝜌𝐶 𝑉 1−
𝜕𝑧 𝑟 𝜕𝑟 𝜕𝑟 𝑅 𝜕𝑧
 Neglect axial conduction term (vs radial conduction )

 BCs : (i) @ z=0, @ any r, T=T0

(ii) @ all z, @ r=R,
𝜕𝑇 𝑞
𝑘
𝜕𝑟
=
𝐴
(constant wall heat flux)

𝑇=𝑇 (constant wall temp)

Avg NU over a length “z” from beginning of heating =

f(dimensionless axial distance)
𝜈 𝐶 𝜇
(z/D) 𝑃 = =
Z* = 𝛼 𝑘 5

Thermal entrance region (NU-Graetz problem) in a tube

Constant wall heat flux
Local NU
𝑁𝑈 = 1.302 𝑧 ∗ − 0.5 𝑓𝑜𝑟 0.00005 < 𝑧 ∗ < 0.0015

∗ 𝑓𝑜𝑟 𝑧 ∗ > 0.0015

= 4.364 + 8.68 1000𝑧 ∗ . 𝑒

Velocity & Temp profiles develop

simultaneously
Avg. NU
𝑓𝑜𝑟 𝑧 ∗ < 0.03
= 1.953 𝑧 ∗

0.0722 𝑓𝑜𝑟 𝑧 ∗ > 0.03

= 4.364 +
𝑧∗
𝐿
Thermal entrance length = 0.04305 𝑅𝑒 𝑃𝑟
𝐷

𝑧 ∗ = 𝐷𝑖𝑚𝑒𝑛𝑠𝑖𝑜𝑛𝑙𝑒𝑠𝑠 𝑎𝑥𝑖𝑎𝑙 𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒 = 𝑧⁄𝐷 ⁄𝑅𝑒 6𝑃𝑟

 08-03-2024

Constant wall temperature:

𝑁𝑢 = 1.077 𝑧 ∗ ⁄
− 0.7 𝑓𝑜𝑟 𝑧 ∗ < 0.01
Local NU,

∗
𝑁𝑢 = 3.657 + 6.874 1000𝑧 ∗ . 𝑒 .
𝑓𝑜𝑟 𝑧 ∗ > 0.01

Average NU 𝑁𝑢 = 1.615 𝑧 ∗ ⁄
− 0.2 𝑓𝑜𝑟 0.005 < 𝑧 ∗ < 0.03

0.0499 𝑓𝑜𝑟 𝑧 ∗ > 0.03

𝑁𝑢 = 3.675 +
𝑧∗

𝐿
Thermal entrance length = 0.03347𝑅𝑒 𝑃𝑟
𝐷

Sieder & Tate correlation

Velocity & temp profiles develop simultaneously

.
𝑅𝑒 𝑃𝑟 𝜇
𝑁𝑢 = 1.86
𝐿 ⁄𝐷 𝜇

Evaluate all properties @ temp (Tmi+ Tm0)/2

Evaluate µw @ temp Tw,

0.48 < Pr < 16700 Note

 If the 3rd condition is not satisfied,
𝜇  Fully-developed conditions will occur, Use NU = 3.657
0.0044 < < 9.75
𝜇 (reasonable accuracy).

𝑅𝑒 𝑃𝑟 𝜇
. To obtain “h” over the length L, multiply avg “h” (from S-D) by
>2 inside area of the tube & mean ΔT,
𝐿 ⁄𝐷 𝜇

8
 08-03-2024

Turbulent Flow Heat Transfer in a Pipe

Local heat transfer coefficient (smooth circular pipes) under fully developed conditions
Dittus-Boelter equation (1930)
.
𝑁𝑢 = 0.023 𝑅𝑒 𝑃𝑟𝑛
.
ℎ𝐷 𝜌𝑉𝐷 𝐶 𝜇
= 0.023
𝑘 𝜇 𝐾
 n = 0.4 (heating) & n = 0.3 (cooling)
 Evaluate properties : k, p, µ & CP @ local bulk mean temp

Valid for
2300 < 𝑅𝑒 < 1.2 𝑋 10
0.7 < 𝑃𝑟 < 100
 Applicable : Constant heat flux & Constant wall temp BC

 Local heat transfer coefficient (thermal entrance region) > fully-developed region
 asymptotically approaches the fully-developed value over a short distance(< 20D in length)
9

For fully-developed turbulent flows,

Well-known correlations : Colburn, Prandtl, Petukhov and Gnielinski

Based on
 exp data for better understanding of turbulent flow.
 Development of analogy between HT & MT
𝑓
𝑅𝑒 − 1000 𝑃𝑟
2
𝑁𝑢 =
2300 < 𝑅𝑒 < 5 𝑋 10
𝑓 /
1 + 12.7 𝑃𝑟 −1
2 0.5 < 𝑃𝑟 < 2000
 Valid : constant wall heat flux & constant wall temp BC ;
 smooth & rough tubes
 Evaluate properties @ local bulk mean temp For rough surfaces
 Obtain friction factor (moody chart) 1 2e −16.2426
= 3.48 − 1 ⋅ 7372 𝑙𝑛 𝑙𝑛 𝐴
𝑓 𝐷 𝑅𝑒
For smooth surfaces
. 4 × 10 < 𝑅𝑒 < 10
𝑓 = 0.079 𝑅𝑒
.
. 2 𝑒 ⁄𝐷 . 7.149
𝑓 = 0.046 𝑅𝑒 3 × 10 < 𝑅𝑒 < 10 𝐴 = +
6.0983 𝑅𝑒
1
𝑓= 1.82𝑙𝑜𝑔𝑅𝑒 − 1.64 10 < 𝑅𝑒 < 10 10
4 4000 < 𝑅𝑒 < 10 ; 10 ≤ ( ) ≤ 0.05
 08-03-2024

Low Prandtl number fluids (fully-developed turbulent flows)

 Gnielinski eqn : valid only for fluids Pr > 0.5.

 Not applicable to fluids like liquid metals (low Pr).

 Recommended correlations (Notter and Sleicher ) for local Nusselt number (Pr< 0.1 ),

𝑁𝑢 = 4.8 + 0.0156 𝑅𝑒 .
𝑃𝑟 . (constant wall temp BC) 10 < 𝑅𝑒 < 10

. . (constant heat flux BC) 0.004 < 𝑃𝑟 < 0.1

𝑁𝑢 = 6.3 + 0.0167 𝑅𝑒 𝑃𝑟

Evaluate properties @ local bulk mean temp.

11

Laminar flow over a flat plate

12
 08-03-2024

𝑑 𝜏 𝜇 𝜕𝑉
(𝑉 − 𝑉 𝑑𝑦 = = (1) (integral momentum eqn of BL : constant P conditions)
𝑑𝑥 𝜌 𝜌 𝜕𝑦

Velocity distribution (polynomial of the form) BCs

(i) 𝑦 = 0, 𝑉 = 0
(ii) 𝑦 = 0, = 0 (𝑉 = 0, 𝑉 = 0 @ 𝑦 = 0)
(2)
𝑉 = 𝑎 + 𝑎 𝑦 + 𝑎 𝑦 +𝑎 𝑦
(iii) 𝑦 = 𝛿, 𝑉 = 𝑉
(iv) 𝑦 = 𝛿, =0

Evaluate 𝑎 , 𝑎 , 𝑎 & 𝑎 using BCs

3𝑉 −𝑉
𝑎 = 0, 𝑎 = ,𝑎 = 0 & 𝑎 =
2𝛿 2𝛿
3 𝑦 1 𝑦
𝑉 =𝑉 − (3)
2 𝛿 2 𝛿
13
This is applicable 0 < y < δ & 𝑦 > 𝛿, 𝑉 = 𝑉

𝛿 4.64
=
𝑥 𝜌𝑢 𝑥
𝜇

𝛿 4.64 Shear stress @ surface,

= (relation b/w δ of velocity BL & x)
𝑥 𝑁 ,
𝜕𝑉
𝜏 =𝜇
𝜌𝑢 𝑥 𝑢 𝑥 𝜕𝑦
𝑊ℎ𝑒𝑟𝑒 , 𝑁 , = =
𝜇 𝜗 =

𝐶,
= 0.323 𝑁 ,
2

(local drag coefficient)

14
 08-03-2024

Exact soln (BL)

𝐶,
= 0.332 𝑁 ,
2

1
𝐶 = 𝐶 , 𝑑𝑥 = 1.328 𝑁 , 𝐶 : avg friction coefft b/w x=0 & x=L (drag coefft)
𝐿

𝜌𝑢
𝐹 =𝐶 𝐴
2

(As - surface area)

15

Forced convection : Thermal Boundary Layer 𝜕 𝑇 𝜕 𝑇 𝜕𝑇 𝜕𝑇

𝐾 + = 𝜌𝐶 𝑉 +𝑉
𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦

𝜕 𝑇 𝜕𝑇 𝜕𝑇
k = 𝜌𝐶 𝑉 +𝑉
𝜕𝑦 𝜕𝑥 𝜕𝑦

𝜕𝑇 𝑑
𝛼 = 𝑉 𝑇−𝑇 𝑑𝑦
𝜕𝑦 𝑑𝑥

(integral eqn of BL : constant properties & constant 𝑇 )

𝑘
Integral has a finite value within TBL (but it is zero outside it) 𝛼=
𝜌𝐶
16
 08-03-2024

Solution : Integral eqn of BL 𝜕𝑇 𝑑 (1)

𝛼 = 𝑉 𝑇−𝑇 𝑑𝑦
𝜕𝑦 𝑑𝑥
Temp distribution : polynomial of the form

𝑇 = 𝑎 +𝑎 𝑦+𝑎 𝑦 +𝑎 𝑦
𝑎 =𝑇
3 𝑇 − 𝑇
Evaluate 𝑎 , 𝑎 , 𝑎 & 𝑎 using BCs 𝑎 =
2𝛿
𝑎 =0
BCs
3 𝑇 − 𝑇
𝑎 = −
2𝛿
𝜕 𝑇
𝑖 𝑦 = 0 ,𝑇 = 𝑇 , =0 𝑉 =𝑉 =0 @ 𝑦= 0
𝜕𝑦

𝜕𝑇
𝑖𝑖 𝑦 = 𝛿 , 𝑇 = 𝑇 , =0
𝜕𝑦

(2)
𝑇 −𝑇 3 𝑦 1 𝑦
= 1− +
𝑇 −𝑇 2 𝛿 2 𝛿
17

3𝛼 𝑑 3 𝑦 1 𝑦 3 𝑦 1 𝑦
𝑇 −𝑇 = 𝑉 − 𝑇 − 𝑇 1− + 𝑑𝑦
2𝑉 𝛿 𝑑𝑥 2 𝛿 2 𝛿 2 𝛿 2 𝛿
.

3𝛼 3 𝑑 𝛿 1 𝛿
= 1−
2𝑉 𝛿 20 𝑑𝑥 𝛿 14 𝛿 (3)

𝛿
Pr = 1 , =1
𝛿

𝛿
Pr > 1 , <1
𝛿
𝛿 𝑁 >1, thickness of hydrodynamic BL is bigger
𝛿
Case : Pr ≥ 1 (neglect compared to unity, 2nd term)
14

10𝛼 𝑑 𝛿 𝛿 4.64
(4) = 𝜗𝑥
= 𝛿 𝑥 𝜌𝑉 𝑥 ⸫ 𝛿 = 4.64 (5)
𝑉 𝑑𝑥 𝛿 𝑉
𝜇
18
 08-03-2024

𝑑 𝛿 10𝛼 𝜇𝐶 𝜌 𝜇 𝜌𝐶 𝜗
𝛿 = ⸫𝑁 = = =
𝑑𝑥 𝜗𝑥 𝑉 𝑘 𝜌 𝜌 𝑘 𝛼
4.64
𝑉
/ / /
92.8 𝜗𝑥 4.52 𝜗𝑥
𝛿 = =
𝑑 𝛿 4.64𝑋10𝛼 𝜗 𝑁 𝑉 𝑁 𝑉
𝛿 = (6)
𝑑𝑥 𝑥 𝑉 𝑉

𝜗𝑥
𝛿 = 4.64
Form of the soln of this differential eqn :𝛿 = 𝐶 𝑥 (7) 𝑢

𝐶 = constant 𝛿 4.52 1
=
𝛿 4.64 𝑁
Sub this soln in eqn (6)
𝛿 1
/ =
20𝛼 280𝑣 92.8𝛼 𝜗 𝛿 1.026 𝑁
𝐶 = 𝐶 =
𝑉 13𝑉 𝑉 𝑉

/
92.8𝛼 𝜗 / 𝑘 𝜇𝐶
𝛿 = .𝑥 𝛼= ; 𝑁 =
𝑉 𝑉 𝜌𝐶 𝑘 19

Local HTC @ a distance from the leading edge,

.
ℎ = =

𝜕𝑇 3 𝑇 −𝑇
=
𝜕𝑦 2𝛿
3𝑘
ℎ =
2𝛿

Sub value of 𝛿
/
3. 𝑘 𝑁 ,
ℎ = 𝑁
2𝑋4.52 𝑥

ℎ 𝑥 /
= 0.332 𝑁 𝑁 ,
𝑘

/
Local Nusselt number, 𝑁 , =ℎ .
𝑁 = 0.332 𝑁 𝑁 20
, ,
 08-03-2024

Avg 𝑵𝑵𝒖,𝒙 (entire heated length of plate L) /

1 𝑐 𝑐 /
𝑐 𝑥
ℎ= /
𝑑𝑥 = 𝑥 𝑑𝑥 =
ℎ𝐿 𝐿 𝑥 𝐿 𝐿 1/2
𝑁 =
𝑘

h : avg HTC for a plate of length L (mean value b/w x=0 & x = 𝐿 / = = 2ℎ = 2ℎ
/
=L), is

Avg value HTC : Twice of Local HTC @ plate end

∫ ℎ 𝑑𝑥1
ℎ= = ℎ 𝑑𝑥
∫ 𝑑𝑥 𝐿 /
𝑁 = 0.664 𝑁 𝑁 ,

ℎ 𝑥 𝑥𝑢 /
= 0.332 𝑁
𝑘 𝜗 𝐿𝑢
𝑊ℎ𝑒𝑟𝑒, 𝑁 , =𝑁 , =
𝜗
(@ end of the plate /entire length)

⸫ℎ = ⁄ /

Constant ‘C’ – containing all the factors other than ℎ & x

21

Analogy : heat (vs) momentum transfer [Reynolds -Colburn analogy]

𝐶.
 heat transfer coefficient can be deduced from measurements of drag. 𝑆𝑡𝑥 𝑃𝑟 =
2
o mechanism of HT ≈ mechanism of MT (Colburn analogy).
(Colburn analogy eqn)
 On an average basis,
𝐶
𝑆𝑡𝑥 𝑃𝑟 =
2
Colburn analogy (external flows)
 Valid : laminar BL flow over a flat plate because the pressure gradient term ( dp / dx) = 0.
 Not valid : laminar flow over curved surfaces (or) laminar flows inside tubes
o pressure gradient in the flow direction is non-zero.
 For turbulent flow : conditions are less sensitive to pressure gradients.
 Colburn analogy eqn : approximately valid for local coefficients (e.g turbulent flow over a flat plate)
 Corresponding eqn on an average basis is also valid if the BL is turbulent from the leading edge.
Colburn analogy (internal flows)
Fully-developed turbulent flow in a circular pipe
 Equivalent form of the Colburn analogy (friction factor replaces the drag coefficient )
𝑓
𝑆𝑡𝑥 𝑃𝑟 = =
2 22
 08-03-2024

Heat Transfer in Turbulent Flow Over a Flat Plate

𝐶 = 0.0592 𝑅𝑒𝑥 / (FM)
.

 Use Colburn analogy & local drag coefft/friction factor , calculate Local NU (constant wall temp)

⁄ .
𝑁 = 0.0292 𝑃𝑟 𝑅𝑒 (𝑽𝒂𝒍𝒊𝒅 ∶ 𝑅𝑒 < 𝑅𝑒 < 10 ; 0.7 < 𝑃𝑟 < 100)

 Evaluate properties @ mean film temp

Flow over a flat plate (length L) : LBL upto x = Xc & turbulent BL for the remaining length.

1
ℎ= ℎ 𝑑𝑥 + ℎ 𝑑𝑥 /
𝐿 𝑉
ℎ = 0.332 𝑘 𝑃𝑟
𝜗𝑥
ℎ𝐿 ⁄ .
𝑁 = = 0.0366 𝑃𝑟 𝑅𝑒 −𝐶
𝑘 𝐶 = f (value of the transition Re)

Where, 𝐶 = 0 ; 𝑅𝑒 = 0 ( BL is turbulent from the leading edge)

𝐶 = 14,500 (𝑅𝑒 = 3 × 10 ) 23

Flow Across a Cylinder

 Drag coefficient (CD ) for flow across a cylinder
𝐹
𝑐 =
𝜌𝐴 𝑉 ⁄2
Where, A : projected area of the cylinder normal to the flow.

 Based on the exp data ( Schlichting), correlations between CD &

ReD (valid upto 2 x 105 )
Unseparated flow regime ∶ 𝑐 = 10.41 𝑅𝑒 . 0.1 < 𝑅𝑒 < 4

Separated flow region : 𝑐 = 5.67 𝑅𝑒 . 4 < 𝑅𝑒 < 1000

=1 1000 < 𝑅𝑒 < 5000
(laminar BL in the front portion, separation occurs & formation of vortices)

𝑐 = 0.310 𝑅𝑒 . 5000 < 𝑅𝑒 < 10

Separation with highly turbulent eddies :
= 1.10 10 < 𝑅𝑒 < 2 × 10

24
 08-03-2024

Note
 When ReD around 2 x 105 : Sharp reduction in the CD.
(BL becomes turbulent→ separation point moves downstream & separated flow region reduces in size)

 At these ReD , most of the drag is contributed by this region.

 Reduction in the size of the region →reduction in the drag coefficient.

Heat transfer correlations for the cylinder (cross flow) , Hilpert(1933)

Air flowing past cylinders of various diameters, 𝑁 = 𝐶𝑅𝑒

(valid for the constant wall temp BC ) & for other gases also since the Prandtl number is almost the same.
C & n : constants

𝑹𝒆𝑫 C n
40 – 4000 0.615 0.466
4000 – 40,000 0.174 0.618
40,000 – 400,000 0.0239 0.805
25

Churchill & Bernstein

⁄
⁄ ⁄ ⁄
0.62 𝑅𝑒 𝑃𝑟 𝑅𝑒
𝑁 = 0.3 + ⁄
1+
1 + 0.4⁄𝑃 𝑟 ⁄ 282,000

 Valid for (i) constant wall temperature (ii) over a wide range of Pr, (iii) ReD <107 ; ReD Pr > 0.2.

 For 20,000 < ReD < 400,000 ,

⁄
⁄ ⁄
 Modify the last term 1 + ,
by 1 +
,

 Evaluate properties @ mean film temp

 Also, can be used for the constant wall heat flux,

(𝑞/𝐴)
ℎ=
𝑇 −𝑇

26
 08-03-2024

Empirical correlations for flow across cylinders & Spheres

Flow across cylinders (external flow)
Flow across cylinders & spheres is frequently encountered in practice.
 tubes in a shell-and-tube heat exchanger : both internal flow through the tubes & external flow over the tubes,
 both flows must be considered in the analysis of the HE
 characteristic length for a circular cylinder or sphere : external diameter (D)
 Thus, Re = (u- uniform velocity of the fluid as it approaches the cylinder or sphere.)
 Avg HTC,

⁄ ⁄
𝑁 =𝑐 𝑁 =𝑐 𝑁 𝑁

o C & n : Constants
o Evaluate all the fluid properties @ film temp @ ; Characteristic dimension (NNu) & NRe (dia of
cylinder)
o Applicable for gases & liquids
Constants for flow across cylinder
NRe 0.4 -4 4-40 40-4000 4000-40000 40000-400000
C 0.989 0.911 0.683 0.193 0.027
27
n 0.330 0.385 0.466 0.618 0.805

 A Whitaker correlation for calculating avg HTC,

/
/ / .
𝜇
𝑁 = 0.4 𝑁 + 0.06 𝑁 𝑁
𝜇

o Valid for 40 < 𝑁 < 10 , 0.65 < 𝑁 < 300, 025 < < 5.2
o A characteristic dimension (characteristic length) : Diameter of cylinder
o Evaluate all properties @ free stream temperature (𝑇 ) except 𝜇 (evalute @ wall/temperature)

28
 08-03-2024

Non-circular cylinders
 Avg HTC is given by, 𝑁 =𝑐 𝑁 𝑁
o C & n : Constants
Constants for non-circular cylinders

29

Spheres
 Correlation for 𝑄̇ from spheres to flowing gases (McAdams),
.
ℎ𝐷 𝑢 𝐷
= 0.37
𝑘 𝜗

. for 17 < 𝑁 < 70000

i.e , 𝑁 = 0.37 𝑁

 For air with NPr= 0.71

 Recommended correlation for wide range of NRe

/ /
𝑁 = 2 + 0.25 + 3𝑋10 𝑁 for 100 < 𝑁 < 3𝑋10

 For flow of liquids past spheres (Kramer’s correlation)

.
.
𝑢 𝐷
𝑁 .𝑁 = 0.97 + 0.68
𝜗 for 1 < 𝑁 < 2000
 Evaluate properties of fluid @ film temperature
30
 08-03-2024

Film coefft in pipes : Laminar flow (Empirical correlation : Sider-Tate eqn)

 For heating (or) cooling of viscous fluids, an empirical correlation.,

/ . / .
𝐷 𝜇 ℎ𝐷 𝜌𝑢𝐷 𝐶 𝜇 𝐷 𝜇
𝑁 = 1.86 𝑁 𝑁 (or) = 1.86 (1)
𝐿 𝜇 𝑘 𝜇 𝑘 𝐿 𝜇

o For calculation of HTC for laminar flow of fluids in horizontal tubes (or) pipes
o Valid for 𝑁 < 2100 & 𝑁 𝑁 > 100
.
o D- diameter of pipe; L- length of pipe; : Sider-Tate correction
o = ratio of μ of fluid @ bulk temp & μ of fluid @ wall temp ; D- inside dia of the pipe ; h-HTC.
o Flow in a tube can be laminar or turbulent, depending on the flow conditions

 All physical properties of the fluid are taken @ mean bulk temp of the flowing fluid (i.e., at arithmetic mean of
inlet & outlet temp of the fluid flowing through pipe (Ti + To)/2

 Evaluate μw @ avg wall temp

31

𝑚̇𝐶̇
Graetz number is defined by, 𝑁 = Peclet number = (Reynolds number) * (Prandtl number)
𝑘𝐿
𝜋
𝑚̇(mass flow rate) = 𝜌𝑢𝐴 = 𝜌𝑢𝐷
4
For thin fluid, 𝑁 𝑁 > 12 , the empirical eqn,
𝜋 𝐷 (2)
𝑁 = 𝑁 𝑁
4 𝐿 /
𝐷
𝑁 = 1.62 𝑁 𝑁 (3)
𝐿
Sub (2) in (1)
/
ℎ𝐷 𝑚̇
(3)=> = 1.75 𝐶
𝑘 𝑘𝐿
/ .
ℎ𝐷 𝑚̇ 𝜇
=2 𝐶
𝑘 𝑘𝐿 𝜇 (3)
For viscous liquids in laminar flow, (3) is applicable [variation of
viscosity with temp ]
Sider-Tade eqn for laminar/viscous flow = f(𝑁 )

32
 08-03-2024

Turbulent flow of gases

For organic liquids
For heating or cooling of gases in turbulent flow, (𝑁
~0.74), 𝑢 .
ℎ = 423 .
𝐷
An empirical correlation,

.
Mass velocity is given by
𝑁 = 0.02 𝑁
𝑚̇ 𝜌𝑢𝐴
𝐺= = = 𝜌𝑢
For turbulent flow in pipes, 𝐴 𝐴

For air @ 1 atm. pressure 𝜌𝑢𝐷 𝐺𝐷

⸫ 𝑁 = =
𝜇 𝜇
3.5𝑢 .
ℎ=
𝐷 .

For water in tubes

𝑢 .
h = 1063 1 − 10.00293𝑇 .
𝐷

T-avg bulk temp (K) 33

Flow in non-circular cross-sections Do = inner diameter of outer pipe (outside dia of annulus)
To use the Sider-Tate or Colburn eqn for cross sections Di = outer diameter of inner pipe (inside dia of annulus)
other than circular, Wetted perimeter for 𝑄̇ : 𝜋𝐷
 Replace the diameter D in both 𝑁 & 𝑁 by
equivalent diameter (De) ,
For a duct of rectangular c/s x by y, 𝐷 is ,
𝐷 = 4 𝑋 𝑟𝐻

4 𝑋 𝐶𝑟𝑜𝑠𝑠 − 𝑠𝑒𝑐𝑡𝑖𝑜𝑛𝑎𝑙 𝑎𝑟𝑒𝑎 4𝑥𝑦

𝐷 = 𝐷 =
𝑊𝑒𝑡𝑡𝑒𝑑 𝑝𝑒𝑟𝑖𝑚𝑒𝑡𝑒𝑟 2 𝑥+𝑦

For a circular pipe (𝐷 ),

𝑁 ,𝑁 & friction factor = f(hydraulic diameter)
4 𝜋
𝐷 = 4𝐷 = 𝐷
𝜋𝐷
𝐷 (𝐷 ) is defined such that it reduces to ordinary diameter D for
circular tubes

For an annulus of outer diameter (Do), inner diameter (Di) , 𝐷 is,

4 𝜋 4 𝐷 −𝐷 𝐷 −𝐷
𝐷 = = 34
𝜋𝐷 𝐷
 08-03-2024

Colburn analogy; Colburn j factor This eqn can be written in the j-factor form ,

For NRe between 5000 – 200000, friction factor (f) for a .

𝑗 = 0.023 𝑁
smooth pipe ,
.
𝜌𝑢𝐷
𝑓 = 0.046 For laminar flow region, we have
𝜇
/ .
Compare with Colburn eqn (𝑄̇ - turbulent flow inside pipe/tube) 𝑚̇ 𝜇
𝑁 =2 𝐶 (3)
ℎ 𝐶 𝜇
/
𝜇
.
𝜌𝑢𝐷
. 𝑘𝐿 𝜇
= 0.023
𝜌𝑢𝐶 𝑘 𝜇 𝜇
𝑚𝐶̇ 𝜋 𝐷
𝑁 = = 𝑁 𝑁
(Colburn eqn) 𝑘𝐿 4 𝐿
/ .
ℎ 𝐶 𝜇 𝜇 𝑓 / .
=𝑗 = (1) 𝜋 𝐷 𝜇 (4)
𝜌𝑢𝐶 𝑘 𝜇 2 (3)=> 𝑁 =2
4
𝑁 𝑁
𝐿 𝜇

(1)=> Colburn analogy between heat transfer and fluid friction. Multiply (4) by (1/NRe) (1/NPr),
/ . / /
The factor 𝑗 : Colburn j factor. ℎ 𝐶 𝜇 𝜇 𝐷 𝜌𝑢𝐷
= 𝑗 = 1.86
𝜇
. 𝜌𝑢𝐶 𝑘 𝜇 𝐿 𝜇
/ .
We have, 𝑁 𝑁 = 0.023 𝑁 35
𝜇

Film coefficients-in pipes (Transition Region)

/ . .
ℎ𝐷 𝜌𝑢𝐷 𝐶 𝜇 𝐷 𝜇 𝜌𝑢𝐷
= 1.86 𝑓 = 0.046
𝑘 𝜇 𝑘 𝐿 𝜇 valid for 𝑁 < 2100 𝜇

. Plot on the same coordinates → straight line ( slope

/ 𝜇 . 𝑁 >10,000 = (– 0.20) for NRe > 10,000)
𝑁 𝑁 = 0.023 𝑁
𝜇
Transition region: by curved lines (2100 > 𝑁 > 10000).
Range of 2100 > 𝑁 > 10000 : transition region
 For this, no simple equation is applicable
 Use graphical method to obtain h
 Colburn j factor (vs) NRe (for L/D - with L/D as
a parameter)
/ . / /
ℎ 𝐶 𝜇 𝜇 𝐷 𝜌𝑢𝐷
= 𝑗 = 1.86
𝜌𝑢𝐶 𝑘 𝜇 𝐿 𝜇

𝑗 = 0.023 𝑁 .

for each value of L/D, a logarithmic plot of L.H.S.,

i.e., jH (vs) NRe gives a straight line with a slope equal to (– 2/3).
36
The lines terminate@NRe = 2100.
 08-03-2024

Covers NRe 10000 >NRe > 30000

Beyond upper & lower limits : (For NRe < 1000 & NRe > 30000)
/ .
ℎ𝐷 𝜌𝑢𝐷 𝐶 𝜇 𝐷 𝜇
= 1.86
𝑘 𝜇 𝑘 𝐿 𝜇
.
/ 𝜇 .
𝑁 𝑁 = 0.023 𝑁
𝜇

Tube side heat transfer curves transition range

For the transition region, i.e., for 2100 < NRe < 10000, the empirical eqn to calculate the film HTC,

/ .
/ /
𝐷 𝜇
𝑁 = 𝑁 − 125 𝑁 1+
𝐿 𝜇

 Evaluate NRe from given data

 Obtain jH for the corresponding NRe & L/D
 value of the tube side HTC
37

For turbulent flow,

Wilson Plot .
. /
𝜇
Used for determine film HTC. 𝑁 𝑁 𝑁 = 0.023 𝑁 𝑁
𝜇
.
Shell & tube heat exchanger 𝑁 ∝ 𝑁
 Steam is condensing on the shell side
 Cold fluid is flowing through the tubes (turbulent flow ℎ ∝ 𝑁 .

region).
 Determine overall HTC by direct measurements of 𝑄̇ , overall ℎ ∝ 𝑈 .
ΔT & area (@various cold fluid velocities)
 Condensing steam side coefficient (ho) remains almost .
ℎ =𝑎𝑈
constant
 Resistance offered by the metal wall is constant. 1 1
 Clean tubes (assume) = . +𝐶
𝑈 𝑎𝑈
1 1 𝑥 1 'U’ -linear velocity of the cold fluid.
= + +
U ℎ 𝑘 ℎ
Plot of 1/U (vs) 1/(u)0.8
1 1  straight line (slope = 1/a) & intercept +
= +𝐶 (1)
𝑈 ℎ  Obtain ho from the intercept
 ‘a’ : value of film coefficient (ℎ )for a unit velocity of the
where 'C' is a constant cold fluid (Wilson plot) 38