Skip to main content

Chapter 1 · 9 hours

Pipe flow

IOE past exam questions

Past questions and answers

28 questions set from this chapter, 1 of them more than once. Most repeated first.

  • Asked 2 times
  • 2081 Chaitra · 2+3+3 marks
  • 2071 Magh · 2+3+3 marks

Explain Prandtl's mixing length theory. Show that the velocity distribution in pipe for turbulent flow is logarithmic. Derive an expression for head loss due to sudden expansion of pipe.

Answer

Prandtl's mixing length theory

In turbulent flow, fluid lumps move across the layers and carry momentum with them. Prandtl assumed that a lump keeps its identity over a distance ll (the mixing length) and then mixes with the new layer, like the mean free path in kinetic theory.

  • Velocity difference between layers a distance ll apart: u′≈l dudyu' \approx l\,\dfrac{du}{dy}
  • The transverse fluctuation v′v' is of the same order as u′u', so v′≈l dudyv' \approx l\,\dfrac{du}{dy}
  • Turbulent (Reynolds) shear stress: τt=−ρ u′v′‾=ρ l2(dudy)2\tau_t = -\rho\,\overline{u'v'} = \rho\, l^2 \left(\dfrac{du}{dy}\right)^2

Near a wall the eddy size grows with distance from the wall, so Prandtl took l=κyl = \kappa y, where κ≈0.4\kappa \approx 0.4 is von Karman's constant. Total shear stress is τ=μdudy+ρl2(dudy)2\tau = \mu\dfrac{du}{dy} + \rho l^2\left(\dfrac{du}{dy}\right)^2; away from the wall the viscous part is negligible.

Logarithmic velocity distribution in a pipe

Near the wall the shear stress is nearly constant and equal to the wall shear τ0\tau_0. Let u∗=τ0/ρu_* = \sqrt{\tau_0/\rho} be the shear velocity and yy the distance from the wall.

τ0=ρ κ2y2(dudy)2  ⇒  dudy=u∗κ y\tau_0 = \rho\,\kappa^2 y^2\left(\frac{du}{dy}\right)^2 \;\Rightarrow\; \frac{du}{dy} = \frac{u_*}{\kappa\, y}

Integrating:

u=u∗κln⁡y+Cu = \frac{u_*}{\kappa}\ln y + C

so velocity varies with the logarithm of the distance from the wall. The constant CC is fixed by the wall condition:

  • Smooth pipe: uu∗=5.75log⁡10u∗yν+5.5\dfrac{u}{u_*} = 5.75\log_{10}\dfrac{u_* y}{\nu} + 5.5
  • Rough pipe: uu∗=5.75log⁡10yk+8.5\dfrac{u}{u_*} = 5.75\log_{10}\dfrac{y}{k} + 8.5

Here 1κln⁡y=2.3030.4log⁡10y=5.75log⁡10y\dfrac{1}{\kappa}\ln y = \dfrac{2.303}{0.4}\log_{10} y = 5.75\log_{10} y. Measured velocities in pipes follow this law over almost the whole cross-section, and for y=Ry = R at the axis, umax−uu∗=5.75log⁡10Ry\dfrac{u_{max}-u}{u_*} = 5.75\log_{10}\dfrac{R}{y}.

Head loss due to sudden expansion

   1          2            Section 1: A1, V1, p1 (small pipe)
  ____       _______       Section 2: A2, V2, p2 (large pipe)
 |    |_____|      
 |  V1   ~~~~ eddies  V2  Control volume between 1 and 2
 |____       _______      (eddy zone pressure = p1 on the annulus)
      |_____|

Take a control volume from section 1 (just before the expansion) to section 2 (where the jet has spread over the full area). Eddies in the corner zone keep the pressure on the annular area (A2−A1)(A_2 - A_1) equal to p1p_1. Friction on the short length is neglected.

Momentum equation (flow direction):

p1A2−p2A2=ρQ(V2−V1),Q=A2V2p_1 A_2 - p_2 A_2 = \rho Q (V_2 - V_1), \qquad Q = A_2 V_2 p1−p2γ=V2(V2−V1)g\frac{p_1 - p_2}{\gamma} = \frac{V_2(V_2 - V_1)}{g}

Energy equation between 1 and 2 (same level), with head loss heh_e:

p1γ+V122g=p2γ+V222g+he  ⇒  he=p1−p2γ+V12−V222g\frac{p_1}{\gamma} + \frac{V_1^2}{2g} = \frac{p_2}{\gamma} + \frac{V_2^2}{2g} + h_e \;\Rightarrow\; h_e = \frac{p_1-p_2}{\gamma} + \frac{V_1^2 - V_2^2}{2g}

Substituting the momentum result:

he=V22−V1V2g+V12−V222g=2V22−2V1V2+V12−V222gh_e = \frac{V_2^2 - V_1V_2}{g} + \frac{V_1^2 - V_2^2}{2g} = \frac{2V_2^2 - 2V_1V_2 + V_1^2 - V_2^2}{2g} he=(V1−V2)22g\boxed{h_e = \frac{(V_1 - V_2)^2}{2g}}

Using continuity V2=V1A1/A2V_2 = V_1A_1/A_2, this is also he=(1−A1A2)2V122gh_e = \left(1 - \dfrac{A_1}{A_2}\right)^2\dfrac{V_1^2}{2g}.

  • 2082 Kartik · 2+4 marks

What are HGL (Hydraulic Gradient Line) and TEL (Total Energy Line)? Draw HGL and TEL showing elevation, pressure and kinetic energy heads for uniform and non-uniform pipes with real fluid flow.

Answer

Definitions

  • Hydraulic Gradient Line (HGL) is the line joining the levels to which water would rise in piezometer tubes along the pipe. Its height above datum is the piezometric head pγ+z\dfrac{p}{\gamma} + z.
  • Total Energy Line (TEL) (also called EGL) is the line showing the total head pγ+z+V22g\dfrac{p}{\gamma} + z + \dfrac{V^2}{2g}. It is above the HGL by the velocity head V22g\dfrac{V^2}{2g}.

Features for a real fluid

  • TEL always falls in the direction of flow because of head losses. Its slope equals the energy gradient Sf=hf/LS_f = h_f/L.
  • In a uniform pipe VV is constant, so the HGL is parallel to the TEL, a fixed distance V2/2gV^2/2g below it.
  • Minor losses (entrance, expansion, valve, bend) show as sudden vertical drops of the TEL.
  • Where the HGL is below the pipe axis the pressure is below atmospheric (negative); where above, it is positive.
  • A pump gives a sudden rise of both lines; a turbine gives a drop.
  • At a free outlet to the atmosphere the HGL meets the pipe axis (gauge pressure zero); at a reservoir both lines coincide with the free surface (V = 0).

Uniform pipe (constant diameter)

 reservoir          entrance loss
 TEL=HGL ---.
 (surface)   \_
              `-._  TEL (slope = hf/L)
        0.5V²/2g  `-._
               ___  `-._
          HGL ..  `-.__ `-._
        V²/2g gap const `-._`-._
 pipe axis ======================== outlet

TEL and HGL are two parallel straight sloping lines; the gap is constant at V2/2gV^2/2g.

Non-uniform pipe (diameter changes from D1D_1 to D2>D1D_2 > D_1)

 TEL ---.\_
         \ `-._ steeper (small pipe, high V)
          \     `-._
           |        `-.|  <- drop (V1-V2)²/2g
           |  HGL        |`-._ flatter slope
           | rises?      |     `-._ TEL
   ==========[small D1]==[large D2]=======
  • In the smaller pipe the velocity head is large, so the HGL is far below the TEL and the slope hf/Lh_f/L is steeper.
  • At the sudden expansion the TEL drops by (V1−V2)2/2g(V_1 - V_2)^2/2g; as VV falls, the velocity head V2/2gV^2/2g falls and the HGL can rise (pressure recovery) even though energy is lost.
  • In the larger pipe the TEL is flatter and the gap between the lines is smaller.
  • In a gradual contraction/expansion the lines change slope smoothly instead of jumping.
  • 2076 Baisakh · 8 marks

Draw the hydraulic gradient line (HGL) and total energy line (TEL) for the following, considering all losses of head. The pipe is laid horizontal and discharges freely into the atmosphere. Take coefficient of friction f = 0.01. [Figure: a reservoir with water level 8 m above the pipe axis; a horizontal pipe of D1 = 0.15 m and length 25 m followed by a pipe of D2 = 0.3 m and length 15 m discharging to the atmosphere]

Answer

Given. Reservoir level 8 m above the pipe axis (datum = pipe axis). Pipe 1: D1=0.15D_1 = 0.15 m, L1=25L_1 = 25 m. Pipe 2: D2=0.3D_2 = 0.3 m, L2=15L_2 = 15 m, f=0.01f = 0.01. The pipe discharges freely, so the outlet velocity head stays as kinetic energy (no exit loss term, the HGL ends at the axis).

Velocities

Continuity: V2=V1(D1D2)2=V1/4V_2 = V_1\left(\dfrac{D_1}{D_2}\right)^2 = V_1/4.

Energy equation from the reservoir surface to the outlet (all losses):

H=0.5V122g+fL1D1V122g+(V1−V2)22g+fL2D2V222g+V222gH = 0.5\frac{V_1^2}{2g} + f\frac{L_1}{D_1}\frac{V_1^2}{2g} + \frac{(V_1-V_2)^2}{2g} + f\frac{L_2}{D_2}\frac{V_2^2}{2g} + \frac{V_2^2}{2g} 8=V122g[0.5+0.01(250.15)+(1−0.25)2+0.01(150.3)(0.25)2+(0.25)2]=2.8229 V122g8 = \frac{V_1^2}{2g}\left[0.5 + 0.01\left(\frac{25}{0.15}\right) + (1-0.25)^2 + 0.01\left(\frac{15}{0.3}\right)(0.25)^2 + (0.25)^2\right] = 2.8229\,\frac{V_1^2}{2g} V1=2(9.81)(8)2.8229=7.457 m/s,V2=1.864 m/sV_1 = \sqrt{\frac{2(9.81)(8)}{2.8229}} = 7.457\ \text{m/s},\qquad V_2 = 1.864\ \text{m/s}

Velocity heads: V12/2g=2.834V_1^2/2g = 2.834 m, V22/2g=0.177V_2^2/2g = 0.177 m. Discharge =131.8= 131.8 l/s.

Head losses

LossHead (m)
Entrance, 0.5V12/2g0.5V_1^2/2g1.417
Friction in pipe 14.723
Sudden expansion, (V1−V2)2/2g(V_1-V_2)^2/2g1.594
Friction in pipe 20.089
Exit velocity head (kinetic)0.177
Total8.000

The total equals the 8 m available head.

Levels of TEL and HGL (m above the pipe axis)

LocationTELHGL = TEL −V2/2g-V^2/2g
Reservoir surface8.0008.000
Just inside pipe 1 (after entrance loss)6.5833.749
End of pipe 1 (before expansion)1.860-0.974
Start of pipe 2 (after expansion)0.2660.089
Outlet0.1770.000
 8.0 ___                        TEL
     |  |\ 0.5V²/2g
     |  | \__ TEL falls steeply in pipe 1 (small D)
     |  |    `-.__
     |  |  HGL     `-.  | drop (V1-V2)²/2g
     |  |  (well below) |`.__ TEL falls gently in pipe 2
     |  |               |    `-.__ HGL
 0 --|__|===[D1=0.15]===[D2=0.3]===> free outlet

Draw to scale: TEL: 8.0, 6.58 (entrance drop), 1.86 (end of pipe 1), 0.27 (after expansion drop), 0.18 (outlet). HGL is lower by 2.83 m in pipe 1 and by 0.18 m in pipe 2, and meets the pipe axis at the outlet. The HGL rises slightly at the expansion (from -0.97 to 0.09 m) because the velocity head falls.

Answer: V1=7.46V_1 = 7.46 m/s, V2=1.86V_2 = 1.86 m/s, TEL ends 0.177 m above the axis at the outlet and the HGL ends at the axis.

  • 2071 Bhadra · 1.5+1.5 marks

Draw HGL and EGL diagram for the flow system shown in the figure considering all major and minor losses. [Figure: Reservoir A connected to Reservoir B by three pipes in series of diameters d1, d2 and d3, with d1 < d3 < d2]

Answer

Reservoir A (higher) feeds reservoir B through three pipes in series, with d1<d3<d2d_1 < d_3 < d_2. Both reservoir surfaces are at rest, so HGL and EGL start at the water surface of A and end at the water surface of B.

Rules used

  • EGL falls in the direction of flow: slope =fV2/(2gd)= f V^2/(2gd), so it is steepest in the smallest pipe 1 and flattest in pipe 2.
  • HGL is below the EGL by V2/2gV^2/2g; the gap is largest in pipe 1 (highest velocity) and smallest in pipe 2.
  • Sudden drops of the EGL: entrance loss 0.5V12/2g0.5V_1^2/2g; expansion from d1d_1 to d2d_2: (V1−V2)2/2g(V_1 - V_2)^2/2g; contraction from d2d_2 to d3d_3: 0.5(1−A3A2)V32/2g0.5\left(1-\dfrac{A_3}{A_2}\right)V_3^2/2g (approximately); exit loss into B: V32/2gV_3^2/2g.
  • At the exit, the whole velocity head is lost, so the EGL falls to the surface of B and both lines end at the water level of B.
 A __                                           
    |\ entrance                 
    | \___ EGL                              
    |     `-.__ (steep, d1)                  
    |  HGL     `-.   | expansion drop        
    |  (gap     `-.__|`-.__ (flat, d2)     
    |   V1²/2g)         |  `-.__|contraction
    |                    HGL     |`-.__ (d3)
    |==[d1]=======[d2 big]=======[d3]=====> 
                                       exit drop
                                          \__ B
  • The HGL rises after the expansion (velocity falls) and falls again after the contraction (velocity rises).
  • Total of all drops (friction + minor losses) equals the difference of the two reservoir levels, H=hf1+hf2+hf3+hentry+hexp+hcontr+hexitH = h_{f1} + h_{f2} + h_{f3} + h_{entry} + h_{exp} + h_{contr} + h_{exit}.
  • 2082 Kartik · 4 marks

Galvanized iron pipes can be assumed to have equivalent roughness of 0.15 mm. What minimum size of pipe will be hydrodynamically smooth at Reynolds number 2×1052 \times 10^5?

Answer

Criterion. A pipe is hydrodynamically smooth when the roughness is buried in the laminar sub-layer, i.e. when the roughness Reynolds number (Nikuradse) satisfies

u∗kν≤4\frac{u_* k}{\nu} \le 4

with u∗=Vf/8u_* = V\sqrt{f/8} and k=0.15k = 0.15 mm.

Step 1: friction factor of a smooth pipe at Re=2×105Re = 2\times10^5. Prandtl's smooth-pipe law 1f=2log⁡10(Ref)−0.8\dfrac{1}{\sqrt f} = 2\log_{10}(Re\sqrt f) - 0.8 gives (by iteration)

f=0.01564,f/8=0.04421f = 0.01564, \qquad \sqrt{f/8} = 0.04421

Step 2: express the criterion in terms of DD. Since Re=VD/νRe = VD/\nu,

u∗kν=Vf/8 kν=Re f8 kD≤4\frac{u_* k}{\nu} = \frac{V\sqrt{f/8}\,k}{\nu} = Re\,\sqrt{\frac{f}{8}}\,\frac{k}{D} \le 4 D≥Re kf/84=(2×105)(0.00015)(0.04421)4=0.332 mD \ge \frac{Re\,k\sqrt{f/8}}{4} = \frac{(2\times10^5)(0.00015)(0.04421)}{4} = 0.332\ \text{m}

Answer: Minimum diameter ≈0.332\approx 0.332 m (about 330 mm). Any smaller pipe at this Reynolds number is not smooth. (Using the limit u∗k/ν=5u_*k/\nu = 5 instead of 4 gives 0.265 m.)

  • 2082 Kartik · 2+4 marks

An oil with density 900 kg/m³ and kinematic viscosity 0.0002 m²/s flows through an inclined pipe as shown in figure. Assuming steady laminar flow, verify that the flow is upward and find the flow rate. [Figure: inclined pipe of diameter d = 6 cm and length 10 m, inclined at 40° to the horizontal; section 1 at the lower end with p1p_1 = 350,000 Pa and z1z_1 = 0; section 2 at the upper end with p2p_2 = 250,000 Pa]

Answer

Given. ρ=900\rho = 900 kg/m³, ν=0.0002\nu = 0.0002 m²/s so μ=ρν=0.18\mu = \rho\nu = 0.18 Pa s; d=0.06d = 0.06 m, L=10L = 10 m, inclination 40°. At section 1: p1=350 000p_1 = 350\,000 Pa, z1=0z_1 = 0. At section 2: p2=250 000p_2 = 250\,000 Pa, z2=Lsin⁡40∘=6.428z_2 = L\sin 40^\circ = 6.428 m.

Direction of flow

Flow goes from high to low piezometric pressure p∗=p+ρgzp^* = p + \rho g z, not from high to low pp.

p1∗=350 000+0=350 000 Pa,p2∗=250 000+900(9.81)(6.428)=306752 Pap_1^* = 350\,000 + 0 = 350\,000\ \text{Pa}, \qquad p_2^* = 250\,000 + 900(9.81)(6.428) = 306752\ \text{Pa}

Since p1∗>p2∗p_1^* > p_2^*, the flow is from 1 to 2, i.e. upward, as required. The driving pressure drop is

Δp∗=p1∗−p2∗=43248 Pa\Delta p^* = p_1^* - p_2^* = 43248\ \text{Pa}

Flow rate (Hagen-Poiseuille, laminar)

Q=πd4 Δp∗128 μL=π(0.06)4(43248)128(0.18)(10)=0.007643 m3/s=7.64 l/sQ = \frac{\pi d^4\,\Delta p^*}{128\,\mu L} = \frac{\pi (0.06)^4 (43248)}{128(0.18)(10)} = 0.007643\ \text{m}^3/\text{s} = 7.64\ \text{l/s}

Check that the flow is laminar

V=Qπd2/4=2.703 m/s,Re=Vdν=2.703(0.06)0.0002=811<2000V = \frac{Q}{\pi d^2/4} = 2.703\ \text{m/s},\qquad Re = \frac{Vd}{\nu} = \frac{2.703(0.06)}{0.0002} = 811 < 2000

So the laminar assumption is valid.

Answer: The flow is upward; Q=0.007643Q = 0.007643 m³/s (7.647.64 l/s), Re≈811Re \approx 811.

  • 2079 Chaitra · 8 marks

Derive an expression for the loss due to sudden enlargement in pipe flow and therefrom deduce the loss due to sudden contraction.

Answer

Loss due to sudden enlargement

  1                2
 _______          _________    Pipe area A1 -> A2 (A2 > A1)
        |  ~ ~ ~ |             Eddies form in the corner;
  V1,p1 |  eddy  |  V2,p2      the pressure on the annular face
 _______|  zone  |_________    equals p1.

Consider section 1 (just before the enlargement) and section 2 (downstream, where the flow has spread over the full area A2A_2). Assume the pipe is horizontal and neglect wall friction between 1 and 2. The pressure on the annular shoulder is taken as p1p_1 (eddy zone).

Momentum equation:

p1A1+p1(A2−A1)−p2A2=ρQ(V2−V1),Q=A2V2p_1A_1 + p_1(A_2 - A_1) - p_2A_2 = \rho Q (V_2 - V_1), \qquad Q = A_2V_2 p1−p2γ=V2(V2−V1)g⋯(1)\frac{p_1 - p_2}{\gamma} = \frac{V_2(V_2 - V_1)}{g} \quad \cdots (1)

Energy equation (head loss heh_e):

p1γ+V122g=p2γ+V222g+he⇒he=p1−p2γ+V12−V222g\frac{p_1}{\gamma} + \frac{V_1^2}{2g} = \frac{p_2}{\gamma} + \frac{V_2^2}{2g} + h_e \quad\Rightarrow\quad h_e = \frac{p_1-p_2}{\gamma} + \frac{V_1^2 - V_2^2}{2g}

Substituting (1):

he=V22−V1V2g+V12−V222g=V12−2V1V2+V222gh_e = \frac{V_2^2 - V_1V_2}{g} + \frac{V_1^2 - V_2^2}{2g} = \frac{V_1^2 - 2V_1V_2 + V_2^2}{2g} he=(V1−V2)22g=(1−A1A2)2V122g\boxed{h_e = \frac{(V_1 - V_2)^2}{2g} = \left(1 - \frac{A_1}{A_2}\right)^2\frac{V_1^2}{2g}}

Loss due to sudden contraction (deduced)

  1      c         2       Jet contracts to the vena contracta c
 ________                  (area Cc A2), then expands to fill
          \ ~~~ _          the smaller pipe A2.
  V1       \___|___ V2     Loss is only in the expansion c -> 2.
 ________/  Vc

In a sudden contraction from A1A_1 to A2A_2 the stream contracts to a vena contracta of area Ac=CcA2A_c = C_c A_2, then expands suddenly to fill the pipe A2A_2. The only significant loss is this expansion from cc to 2. Applying the enlargement result between cc and 2:

hc=(Vc−V2)22g,Vc=A2V2Ac=V2Cch_c = \frac{(V_c - V_2)^2}{2g}, \qquad V_c = \frac{A_2V_2}{A_c} = \frac{V_2}{C_c} hc=(1Cc−1)2V222g=kcV222g\boxed{h_c = \left(\frac{1}{C_c} - 1\right)^2\frac{V_2^2}{2g} = k_c\frac{V_2^2}{2g}}

With Cc≈0.62C_c \approx 0.62, kc=(1/0.62−1)2≈0.375k_c = (1/0.62 - 1)^2 \approx 0.375; a common value used is kc≈0.5k_c \approx 0.5 for a large area ratio.

  • 2080 Chaitra · 4 marks

Prove that velocity distribution in the case of laminar flow in circular pipe is parabolic.

Answer

Consider steady laminar flow in a horizontal circular pipe of radius RR. Take a coaxial cylinder of fluid of radius rr and length LL.

        <-- tau (on surface)
   p1 ->  ____________  <- p2
         |   cylinder |  radius r, length L
   ------|------------|-------- axis
         |____________|

Force balance (no acceleration): pressure force on the ends = shear force on the curved surface.

(p1−p2)πr2=τ (2πrL)⇒τ=(p1−p2)Lr2=−dpdxr2(p_1 - p_2)\pi r^2 = \tau\,(2\pi r L) \quad\Rightarrow\quad \tau = \frac{(p_1 - p_2)}{L}\frac{r}{2} = -\frac{dp}{dx}\frac{r}{2}

So the shear stress varies linearly from zero at the axis to τ0\tau_0 at the wall.

Newton's law of viscosity. With y=R−ry = R - r, τ=μdudy=−μdudr\tau = \mu\dfrac{du}{dy} = -\mu\dfrac{du}{dr}:

−μdudr=−dpdxr2⇒dudr=12μdpdx r-\mu\frac{du}{dr} = -\frac{dp}{dx}\frac{r}{2} \quad\Rightarrow\quad \frac{du}{dr} = \frac{1}{2\mu}\frac{dp}{dx}\,r

Integrate:

u=14μdpdx r2+Cu = \frac{1}{4\mu}\frac{dp}{dx}\,r^2 + C

Boundary condition (no slip): u=0u = 0 at r=Rr = R, so C=−14μdpdxR2C = -\dfrac{1}{4\mu}\dfrac{dp}{dx}R^2.

u=−14μdpdx(R2−r2)\boxed{u = -\frac{1}{4\mu}\frac{dp}{dx}\left(R^2 - r^2\right)}

This is the equation of a parabola in rr (uu falls as r2r^2), with its maximum at the axis:

umax=−14μdpdxR2,uumax=1−r2R2u_{max} = -\frac{1}{4\mu}\frac{dp}{dx}R^2, \qquad \frac{u}{u_{max}} = 1 - \frac{r^2}{R^2}

The mean velocity is V=umax/2V = u_{max}/2, so the velocity distribution is parabolic. (For an inclined pipe, replace pp by the piezometric pressure p+γzp + \gamma z.)

  • 2080 Chaitra · 4 marks

A 120 cm diameter pipe has equivalent roughness height of 1 mm. What will be the velocity at which the roughness causes the flow to become fully turbulent?

Answer

Criterion. Flow is fully turbulent (hydrodynamically rough) when the roughness Reynolds number (Nikuradse) is

u∗kν≥100\frac{u_* k}{\nu} \ge 100

Take ν=1.0×10−6\nu = 1.0\times10^{-6} m²/s (water at about 20 °C), k=1k = 1 mm =0.001= 0.001 m, D=1.2D = 1.2 m.

Step 1: rough-pipe friction factor (von Karman-Prandtl, D/k=1200D/k = 1200):

1f=2log⁡10Dk+1.14  ⇒  f=0.01877,f/8=0.04844\frac{1}{\sqrt f} = 2\log_{10}\frac{D}{k} + 1.14 \;\Rightarrow\; f = 0.01877, \qquad \sqrt{f/8} = 0.04844

Step 2: shear velocity at the limit

u∗=100 νk=100(1.0×10−6)0.001=0.100 m/su_* = \frac{100\,\nu}{k} = \frac{100(1.0\times10^{-6})}{0.001} = 0.100\ \text{m/s}

Step 3: mean velocity

V=u∗f/8=0.1000.04844=2.06 m/sV = \frac{u_*}{\sqrt{f/8}} = \frac{0.100}{0.04844} = 2.06\ \text{m/s}

Answer: The flow becomes fully turbulent (rough) for V≥2.06V \ge 2.06 m/s, about 2 m/s. (If the limit u∗k/ν=70u_*k/\nu = 70 is used, the velocity is 1.45 m/s.)

  • 2078 Chaitra · 6 marks

Consider steady, incompressible, laminar flow of a Newtonian fluid in an infinitely long round pipe of diameter D or radius R = D/2 inclined at angle α\alpha as shown in the figure. The fluid flows down the pipe due to gravity alone. Consider the coordinate system shown, with X down the axis of pipe. Derive an expression for the X-component of velocity u as a function of radius r and the other parameters of the problem. [Figure: inclined pipe wall at angle α\alpha to the horizontal, fluid of density ρ\rho and viscosity μ\mu, radial coordinate r, x measured down the pipe axis, gravity g vertical]

Answer

Set-up. Pipe of radius RR inclined at angle α\alpha below the horizontal; the fluid flows down the pipe under gravity alone, so there is no externally imposed pressure gradient: ∂p∂x=0\dfrac{\partial p}{\partial x} = 0 (the pressure along the pipe is atmospheric/constant). Flow is steady, laminar, fully developed, so u=u(r)u = u(r) only, and vr=vθ=0v_r = v_\theta = 0.

        x (down the pipe)
     ___ \
    /     \  alpha
   / pipe  \
  /_________\ ____ horizontal
  gravity component along x: g sin(alpha)

Momentum equation along xx (Navier-Stokes in cylindrical coordinates, reduced for this case):

0=ρgsin⁡α+μ 1rddr(rdudr)0 = \rho g\sin\alpha + \mu\,\frac{1}{r}\frac{d}{dr}\left(r\frac{du}{dr}\right)

(Equivalent force balance on a cylinder of radius rr: weight component ρgsin⁡α πr2L\rho g \sin\alpha\,\pi r^2 L is balanced by the shear force −μdudr 2πrL-\mu\dfrac{du}{dr}\,2\pi r L.)

Integrate.

ddr(rdudr)=−ρgsin⁡αμ r\frac{d}{dr}\left(r\frac{du}{dr}\right) = -\frac{\rho g\sin\alpha}{\mu}\,r rdudr=−ρgsin⁡α2μ r2+C1r\frac{du}{dr} = -\frac{\rho g\sin\alpha}{2\mu}\,r^2 + C_1

C1=0C_1 = 0, since du/drdu/dr must be finite at r=0r = 0. Then

dudr=−ρgsin⁡α2μ r⇒u=−ρgsin⁡α4μ r2+C2\frac{du}{dr} = -\frac{\rho g\sin\alpha}{2\mu}\,r \quad\Rightarrow\quad u = -\frac{\rho g\sin\alpha}{4\mu}\,r^2 + C_2

Boundary condition. No slip at the wall: u=0u = 0 at r=Rr = R, so C2=ρgsin⁡α4μR2C_2 = \dfrac{\rho g\sin\alpha}{4\mu}R^2.

u(r)=ρgsin⁡α4μ(R2−r2)\boxed{u(r) = \frac{\rho g\sin\alpha}{4\mu}\left(R^2 - r^2\right)}

The profile is parabolic with umax=ρgsin⁡α4μR2u_{max} = \dfrac{\rho g\sin\alpha}{4\mu}R^2 at the axis. With R=D/2R = D/2, u=ρgsin⁡α16μ(D2−4r2)u = \dfrac{\rho g\sin\alpha}{16\mu}\left(D^2 - 4r^2\right). The mean velocity is V=umax/2=ρgD2sin⁡α32μV = u_{max}/2 = \dfrac{\rho g D^2\sin\alpha}{32\mu}.

  • 2078 Chaitra · 4 marks

Kerosene (density = 800 kg/m³) flows in a 20 cm diameter pipe with a mean velocity of 5 m/s. The pipe has an equivalent sandgrain roughness of 0.5 mm. If the friction factor f = 0.02. Is the pipe behaving as rough, smooth or in transition?

Answer

The type of boundary is decided by the roughness Reynolds number Re∗=u∗kνRe_* = \dfrac{u_* k}{\nu} (Nikuradse): smooth Re∗<4Re_* < 4, transition 4≤Re∗≤1004 \le Re_* \le 100, rough Re∗>100Re_* > 100.

Data. D=0.2D = 0.2 m, V=5V = 5 m/s, k=0.5k = 0.5 mm, f=0.02f = 0.02, ρ=800\rho = 800 kg/m³. The viscosity is not given, so it is assumed that for kerosene at about 20 °C, μ≈1.9×10−3\mu \approx 1.9\times10^{-3} Pa s, giving ν=μ/ρ=2.375×10−6\nu = \mu/\rho = 2.375\times10^{-6} m²/s.

Shear velocity

u∗=Vf8=50.028=0.250 m/su_* = V\sqrt{\frac{f}{8}} = 5\sqrt{\frac{0.02}{8}} = 0.250\ \text{m/s}

Roughness Reynolds number

Re∗=u∗kν=(0.250)(0.0005)2.375×10−6=52.6Re_* = \frac{u_*k}{\nu} = \frac{(0.250)(0.0005)}{2.375\times10^{-6}} = 52.6

Check: laminar sub-layer thickness δ′=11.6νu∗=0.110\delta' = \dfrac{11.6\nu}{u_*} = 0.110 mm, which is smaller than k=0.5k = 0.5 mm, but k/δ′≈4.5k/\delta' \approx 4.5 lies in the transition range (0.25 to 6).

Since 4<Re∗<1004 < Re_* < 100 (about 53), the pipe is neither smooth nor fully rough.

Answer: The pipe behaves in the transition regime (Re=421053Re = 421053, k/D=0.0025k/D = 0.0025, Re∗≈53Re_* \approx 53). The conclusion is the same for any realistic kerosene viscosity (ν\nu between 1.9 and 2.8 ×10−6\times10^{-6} m²/s).

  • 2077 Chaitra · 2+3+3 marks

What is the pressure needed to drive a viscous fluid upslope through a 12 cm diameter pipe? The length of the pipe is 50 m, slope is 20 deg. At the end of the pipe, the pressure is 1 bar. Oil of specific gravity 0.85 and viscosity 0.01 kg/m-s, and the target flow rate is 0.25 cubic meter per minute. (i) What is the pressure needed to drive the flow if there were no slope? (ii) The supplier is out of 12 cm pipes and your manager wonders if multiple 8 cm diameter pipes can be used side by side to achieve the same total flow rate at the same driving pressure, determine how many 8 cm pipes are needed. Round up to nearest integer. [Figure: pipe of length 50 m inclined at 20° to the horizontal, driving pressure P at the lower end, 1 bar at the upper end]

Answer

Data. ρ=850\rho = 850 kg/m³ (s.g. 0.85), μ=0.01\mu = 0.01 Pa s, D=0.12D = 0.12 m, L=50L = 50 m, slope 20°, outlet pressure 1 bar =105= 10^5 Pa, Q=0.25Q = 0.25 m³/min =0.004167= 0.004167 m³/s. The fluid is viscous, so Hagen-Poiseuille (laminar) flow is assumed.

(i) Pressure needed if there were no slope

Mean velocity and Reynolds number:

V=QπD2/4=0.3684 m/s,Re=ρVDμ=3758V = \frac{Q}{\pi D^2/4} = 0.3684\ \text{m/s}, \qquad Re = \frac{\rho V D}{\mu} = 3758

(This is above 2000, but in the critical zone; the laminar formula is used because the fluid is described as viscous and the flow is steady and smooth. A turbulent estimate would give a friction drop of the order of 1 kPa, which is still negligible against the 142 kPa static term, so the overall pressure is almost unchanged.)

Friction pressure drop (Hagen-Poiseuille):

Δpf=128μLQπD4=128(0.01)(50)(0.004167)π(0.12)4=409.3 Pa\Delta p_f = \frac{128\mu L Q}{\pi D^4} = \frac{128(0.01)(50)(0.004167)}{\pi(0.12)^4} = 409.3\ \text{Pa}

Pressure at the inlet without slope: P=105+409.3=100409.3P = 10^5 + 409.3 = 100409.3 Pa ≈100.41\approx 100.41 kPa.

Pressure needed on the actual slope

Extra pressure to lift the oil through the height Lsin⁡20∘=17.10L\sin 20^\circ = 17.10 m:

Δpz=ρgLsin⁡20∘=850(9.81)(50)(0.3420)=142597 Pa\Delta p_z = \rho g L \sin 20^\circ = 850(9.81)(50)(0.3420) = 142597\ \text{Pa} P=105+409.3+142597=243006 Pa≈243.0 kPaP = 10^5 + 409.3 + 142597 = 243006\ \text{Pa} \approx 243.0\ \text{kPa}

(ii) Number of 8 cm pipes in parallel

At the same driving pressure the friction drop per pipe is the same, and in laminar flow Q∝D4Q \propto D^4 (the slope term is the same for any pipe size). Flow per pipe ratio:

Q8Q12=(812)4=0.1975\frac{Q_8}{Q_{12}} = \left(\frac{8}{12}\right)^4 = 0.1975 n=Q12Q8=(128)4=5.0625  ⇒  n=6 pipes (rounded up)n = \frac{Q_{12}}{Q_8} = \left(\frac{12}{8}\right)^4 = 5.0625 \;\Rightarrow\; n = 6 \text{ pipes (rounded up)}

Check with 6 pipes: each carries Q/6=0.000694Q/6 = 0.000694 m³/s, V=0.138V = 0.138 m/s, Re=939Re = 939 (laminar), and the friction drop 345.4345.4 Pa is less than 409.3 Pa, so the driving pressure is not exceeded.

Answer: (i) P≈100.41P \approx 100.41 kPa without slope (Δpf=409\Delta p_f = 409 Pa); on the 20° slope P≈243.0P \approx 243.0 kPa; (ii) 6 pipes of 8 cm diameter.

  • 2077 Chaitra · 2+4+2 marks

Describe Moody's Chart. What are the basis to draw such a diagram? State its uses.

Answer

Description

Moody's chart is a graph of the Darcy-Weisbach friction factor ff against the Reynolds number ReRe for flow in circular pipes, with a family of curves for different relative roughness k/Dk/D (equivalent sand roughness / diameter). Both axes are logarithmic: ReRe (from about 10310^3 to 10810^8) on the horizontal axis, ff (about 0.008 to 0.1) on the vertical axis, and the relative roughness is marked on the right-hand scale.

 f
 0.1 |\ laminar     critical     k/D curves
     | \  f=64/Re   zone    _________ 0.05
 0.05|  \          |      __/   ______ 0.01
     |   \_        |  ___/ ___/ ______ 0.001
 0.02|     \______ |_/ ___/   _______  0.0001
     |  smooth pipe curve ___/  (rough: flat lines)
 0.01|                   transition|  fully rough
     +--------------------------------- Re (log)
     10^3   10^4    10^5   10^6    10^7

Regions of the chart:

RegionRef depends on
Laminar< 2000ReRe only, f=64/Ref = 64/Re (straight line)
Critical zone2000 to 4000Unstable, not reliable
Smooth turbulent> 4000ReRe only (lowest curve)
Transition (partly rough)middleboth ReRe and k/Dk/D (Colebrook-White)
Fully roughright of dashed linek/Dk/D only (horizontal lines)

Basis of the chart

  1. Laminar flow: Hagen-Poiseuille law gives f=64/Ref = 64/Re.
  2. Smooth pipes: Blasius (f=0.316/Re0.25f = 0.316/Re^{0.25}, Re<105Re < 10^5) and Prandtl's universal law 1f=2log⁡10(Ref)−0.8\dfrac{1}{\sqrt f} = 2\log_{10}(Re\sqrt f) - 0.8.
  3. Rough pipes: von Karman-Prandtl law from Nikuradse's sand-roughened pipe experiments, 1f=2log⁡10Dk+1.14\dfrac{1}{\sqrt f} = 2\log_{10}\dfrac{D}{k} + 1.14.
  4. Transition region: the Colebrook-White equation, which joins the smooth and rough laws:
1f=−2log⁡10(k/D3.7+2.51Ref)\frac{1}{\sqrt f} = -2\log_{10}\left(\frac{k/D}{3.7} + \frac{2.51}{Re\sqrt f}\right)

Moody plotted this equation using the equivalent roughness of commercial pipes (Nikuradse's artificial roughness corrected to commercial pipe behaviour).

Uses

  • To find ff for a given ReRe and k/Dk/D, then the head loss hf=fLV22gDh_f = \dfrac{fLV^2}{2gD}.
  • To identify the flow regime (laminar, smooth, transition, rough) of a pipe.
  • To solve the three types of pipe problems (find hfh_f, find QQ, find DD), by iteration.
  • To get the equivalent roughness kk of an old or commercial pipe from test data.
  • To compare materials and judge how much capacity is lost as a pipe ages and becomes rough.
  • 2076 Baisakh · 8 marks

A straight smooth pipe 100 mm diameter and 60 m long is inclined at 10° to the horizontal. A liquid of relative density 0.9 and kinematic viscosity 120 mm²/s is to be pumped through it into a reservoir at the upper end where the gauge pressure is 120 kPa. The pipe friction factor f is given by 64/Re for laminar flow and by 0.32 Re−1/40.32\,Re^{-1/4} for turbulent flow when Re < 10510^5. Determine (a) the maximum pressure at the lower, inlet, end of the pipe if the mean shear stress at the pipe wall is not to exceed 200 Pa; (b) the corresponding rate of flow.

Answer

Data. D=0.1D = 0.1 m, L=60L = 60 m, inclination 10°, ρ=900\rho = 900 kg/m³ (R.D. 0.9), ν=120\nu = 120 mm²/s =1.2×10−4= 1.2\times10^{-4} m²/s, outlet gauge pressure p2=120p_2 = 120 kPa, maximum mean wall shear τ0=200\tau_0 = 200 Pa.

Which flow regime?

Mean wall shear in terms of ff: τ0=fρV28\tau_0 = \dfrac{f\rho V^2}{8}.

  • If laminar: τ0=8μVD\tau_0 = \dfrac{8\mu V}{D}, i.e. V=τ0D8μ=200(0.1)8(900)(1.2×10−4)=23.15V = \dfrac{\tau_0 D}{8\mu} = \dfrac{200(0.1)}{8(900)(1.2\times10^{-4})} = 23.15 m/s, giving Re=19290Re = 19290, which is not laminar. So the flow is turbulent.
  • Turbulent: f=0.32 Re−1/4f = 0.32\,Re^{-1/4} (valid since Re<105Re < 10^5).

(b) Flow rate

τ0=ρV28(0.32)(νVD)1/4=0.04 ρ(νD)1/4V7/4\tau_0 = \frac{\rho V^2}{8}(0.32)\left(\frac{\nu}{VD}\right)^{1/4} = 0.04\,\rho\left(\frac{\nu}{D}\right)^{1/4} V^{7/4} 200=0.04(900)(1.2×10−40.1)1/4V1.75=6.700 V1.75200 = 0.04(900)\left(\frac{1.2\times10^{-4}}{0.1}\right)^{1/4} V^{1.75} = 6.700\,V^{1.75} V=(2006.700)1/1.75=6.963 m/sV = \left(\frac{200}{6.700}\right)^{1/1.75} = 6.963\ \text{m/s}

Check: Re=6.963(0.1)1.2×10−4=5803<105Re = \dfrac{6.963(0.1)}{1.2\times10^{-4}} = 5803 < 10^5, so the formula applies; f=0.32Re−1/4=0.03666f = 0.32Re^{-1/4} = 0.03666 and fρV28=200.0\dfrac{f\rho V^2}{8} = 200.0 Pa.

Q=π4D2V=π4(0.1)2(6.963)=0.05469 m3/s=54.69 l/sQ = \frac{\pi}{4}D^2V = \frac{\pi}{4}(0.1)^2(6.963) = 0.05469\ \text{m}^3/\text{s} = 54.69\ \text{l/s}

(a) Maximum inlet pressure

Force balance on the liquid in the pipe along the incline (steady, uniform flow):

(p1−p2)πD24=τ0 πDL+ρg πD24Lsin⁡θ(p_1 - p_2)\frac{\pi D^2}{4} = \tau_0\,\pi D L + \rho g\,\frac{\pi D^2}{4}L\sin\theta p1−p2=4τ0LD+ρgLsin⁡θ=4(200)(60)0.1+900(9.81)(60)sin⁡10∘p_1 - p_2 = \frac{4\tau_0L}{D} + \rho g L\sin\theta = \frac{4(200)(60)}{0.1} + 900(9.81)(60)\sin10^\circ p1−p2=480000+91988=571988 Pap_1 - p_2 = 480000 + 91988 = 571988\ \text{Pa} p1=120 000+480000+91988=691988 Pa≈692 kPa (gauge)p_1 = 120\,000 + 480000 + 91988 = 691988\ \text{Pa} \approx 692\ \text{kPa (gauge)}

Answer: (a) p1,max≈692p_{1,max} \approx 692 kPa gauge; (b) Q=54.7Q = 54.7 l/s (V=6.96V = 6.96 m/s).

  • 2076 Bhadra · 6 marks

Test on a 500 mm diameter commercial pipe indicated that head loss in 100 m length of pipe at different discharge of water is as given below.
Q (l/s)40200400800
hfh_f (m)0.010.2100.8203.27
(i) Determine the equivalent sand grain roughness of the pipe. (ii) What is the maximum water discharge at which this pipe will act as smooth pipe? (iii) What is the maximum discharge at which this pipe will act as rough pipe?

Answer

Data. D=0.5D = 0.5 m, L=100L = 100 m, water with ν=1.0×10−6\nu = 1.0\times10^{-6} m²/s. Area A=0.19635A = 0.19635 m².

Friction factor from the test

f=hf 2gDLV2f = \dfrac{h_f\,2gD}{LV^2} with V=Q/AV = Q/A:

Q (l/s)V (m/s)Rehfh_f (m)f
400.2041.019e+050.010.02364
2001.0195.093e+050.210.01986
4002.0371.019e+060.820.01938
8004.0742.037e+063.270.01932

At high flows ff levels off at about 0.0193, which means the pipe is fully rough there.

(i) Equivalent sand-grain roughness

Use the highest flow (QQ = 800 l/s, Re=2037183Re = 2037183, f=0.01932f = 0.01932) in the Colebrook-White equation:

1f=−2log⁡10(k/D3.7+2.51Ref)\frac{1}{\sqrt f} = -2\log_{10}\left(\frac{k/D}{3.7} + \frac{2.51}{Re\sqrt f}\right)

Solving, kD=0.000903\dfrac{k}{D} = 0.000903, so

k=0.000903(500)=0.45 mmk = 0.000903(500) = 0.45\ \text{mm}

Check: Colebrook with this kk gives ff = 0.0218, 0.0198, 0.0195 and 0.0193 for the four flows, which agree with the test values.

(ii) Maximum discharge for smooth behaviour

The pipe is smooth while u∗kν≤4\dfrac{u_*k}{\nu} \le 4:

u∗=4νk=4(10−6)0.45×10−3=0.0089 m/su_* = \frac{4\nu}{k} = \frac{4(10^{-6})}{0.45\times10^{-3}} = 0.0089\ \text{m/s}

With smooth-pipe ff (Prandtl law) at the limiting ReRe (found by iteration, Re≈92619Re \approx 92619, f=0.0183f = 0.0183):

V=u∗f/8=0.185 m/s,Q=AV=36.4 l/sV = \frac{u_*}{\sqrt{f/8}} = 0.185\ \text{m/s},\qquad Q = AV = 36.4\ \text{l/s}

(iii) Discharge for fully rough behaviour

The pipe is fully rough when u∗kν≥100\dfrac{u_*k}{\nu} \ge 100, i.e. u∗=100νk=0.2214u_* = \dfrac{100\nu}{k} = 0.2214 m/s. For the rough pipe, 1f=2log⁡10Dk+1.14\dfrac{1}{\sqrt f} = 2\log_{10}\dfrac{D}{k} + 1.14 gives f=0.01914f = 0.01914:

V=u∗f/8=4.526 m/s,Q=889 l/sV = \frac{u_*}{\sqrt{f/8}} = 4.526\ \text{m/s}, \qquad Q = 889\ \text{l/s}

This is the limit of the rough regime: for QQ above about 889 l/s (it is a minimum, not a maximum, flow for rough behaviour) the pipe is fully rough and ff no longer depends on ReRe.

Answer: (i) k≈0.45k \approx 0.45 mm; (ii) smooth for Q≲36Q \lesssim 36 l/s; (iii) fully rough for Q≳889Q \gtrsim 889 l/s (transition in between).

  • 2076 Bhadra · 6 marks

A fluid of constant density ρ\rho enters a horizontal pipe of radius R with uniform velocity V and pressure p1p_1. At a downstream section the pressure is p2p_2 and the velocity varies with radius r according to the equation u=2V{1−(r2/R2)}u = 2V\{1 - (r^2/R^2)\}. Show that the friction force at the pipe walls from the inlet to the section considered is given by πR2(p1−p2−ρV23)\pi R^2\left(p_1 - p_2 - \frac{\rho V^2}{3}\right).

Answer

Control volume. Take the fluid in the pipe between the inlet section 1 (uniform velocity VV, pressure p1p_1) and the section 2 (velocity u=2V(1−r2/R2)u = 2V(1 - r^2/R^2), pressure p2p_2). The pipe is horizontal. Let FF be the friction force exerted by the wall on the fluid (equal and opposite to the friction force of the fluid on the wall).

Continuity check. Mean velocity at section 2:

uˉ=1πR2∫0R2V(1−r2R2)2πr dr=4VR2[R22−R24]=V\bar u = \frac{1}{\pi R^2}\int_0^R 2V\left(1 - \frac{r^2}{R^2}\right)2\pi r\,dr = \frac{4V}{R^2}\left[\frac{R^2}{2} - \frac{R^2}{4}\right] = V

So the discharge is the same at both sections.

Momentum flux at section 2. The momentum flux is ρ∫u2 dA\rho\int u^2\,dA, since uu is not uniform:

∫0Ru2 2πr dr=4V2⋅2π∫0R(1−r2R2)2r dr=8πV2[R22−R22+R26]=43πR2V2\int_0^R u^2\,2\pi r\,dr = 4V^2\cdot 2\pi\int_0^R\left(1 - \frac{r^2}{R^2}\right)^2 r\,dr = 8\pi V^2\left[\frac{R^2}{2} - \frac{R^2}{2} + \frac{R^2}{6}\right] = \frac{4}{3}\pi R^2V^2

(using ∫0R(r−2r3/R2+r5/R4) dr=R2/2−R2/2+R2/6=R2/6\int_0^R (r - 2r^3/R^2 + r^5/R^4)\,dr = R^2/2 - R^2/2 + R^2/6 = R^2/6).

Momentum flux at section 1 (uniform): ρV2 πR2\rho V^2\,\pi R^2.

Momentum equation (flow direction). Net force = rate of change of momentum flux:

p1πR2−p2πR2−F=ρ(43πR2V2−πR2V2)=ρπR2V23p_1\pi R^2 - p_2\pi R^2 - F = \rho\left(\frac{4}{3}\pi R^2V^2 - \pi R^2V^2\right) = \frac{\rho \pi R^2V^2}{3} F=πR2(p1−p2−ρV23)\boxed{F = \pi R^2\left(p_1 - p_2 - \frac{\rho V^2}{3}\right)}

This is the friction force of the wall on the fluid, and by Newton's third law it is also the friction drag on the pipe walls from the inlet to the section. The term ρV2/3\rho V^2/3 is the pressure drop used to build up the parabolic profile (increase in momentum flux), not lost by friction.

  • 2075 Bhadra · 8 marks

One meter diameter pipe is to carry a water discharge of 1.0 m³/s at the minimum loss of energy. What will be the permissible height of surface roughness?

Answer

Idea. A pipe has the minimum energy loss when it is hydrodynamically smooth, since then ff depends only on ReRe and the roughness adds nothing. The pipe stays smooth as long as the roughness is buried in the laminar sub-layer:

u∗kν≤4⇒k≤4νu∗\frac{u_*k}{\nu} \le 4 \quad\Rightarrow\quad k \le \frac{4\nu}{u_*}

Data. D=1D = 1 m, Q=1.0Q = 1.0 m³/s, ν=1.0×10−6\nu = 1.0\times10^{-6} m²/s (water at about 20 °C).

Step 1: velocity and Reynolds number

V=QπD2/4=1.2732 m/s,Re=VDν=1273240V = \frac{Q}{\pi D^2/4} = 1.2732\ \text{m/s},\qquad Re = \frac{VD}{\nu} = 1273240

Step 2: smooth-pipe friction factor (Prandtl: 1f=2log⁡10(Ref)−0.8\dfrac{1}{\sqrt f} = 2\log_{10}(Re\sqrt f) - 0.8)

f=0.01118f = 0.01118

Step 3: shear velocity

u∗=Vf8=1.27320.011188=0.04759 m/su_* = V\sqrt{\frac{f}{8}} = 1.2732\sqrt{\frac{0.01118}{8}} = 0.04759\ \text{m/s}

Step 4: permissible roughness

k≤4νu∗=4(1.0×10−6)0.04759=0.0840 mmk \le \frac{4\nu}{u_*} = \frac{4(1.0\times10^{-6})}{0.04759} = 0.0840\ \text{mm}

Answer: The surface roughness height must not exceed about 0.0840.084 mm (0.08 mm) for minimum loss. With the looser limit u∗k/ν≤5u_*k/\nu \le 5, the value is 0.105 mm.

  • 2074 Bhadra · 8 marks

Petrol of kinematic viscosity 0.6 mm²/s is to be pumped at the rate of 0.8 m³/s through a horizontal pipe 500 mm diameter. However, to reduce pumping costs a pipe of different diameter is suggested. Assuming that the absolute roughness of the walls would be the same for a pipe of slightly different diameter, and that, for Re > 10610^6, f is approximately proportional to the cube root of the roughness, determine the diameter of pipe for which the pumping costs would be halved. Neglect all head losses other than pipe friction. How are the running costs altered if n pipes of equal diameter are used in parallel to give the same total flow rate at the same Reynolds number as for a single pipe?

Answer

Data. ν=0.6×10−6\nu = 0.6\times10^{-6} m²/s, Q=0.8Q = 0.8 m³/s, D=0.5D = 0.5 m, same absolute roughness kk for the new pipe.

V=4QπD2=4.074 m/s,Re=VDν=3395305>106V = \frac{4Q}{\pi D^2} = 4.074\ \text{m/s},\qquad Re = \frac{VD}{\nu} = 3395305 > 10^6

So f∝(k/D)1/3f \propto (k/D)^{1/3} may be used.

Part 1: Diameter for half the pumping cost

Pumping cost is proportional to the power P=ρgQhfP = \rho g Q h_f. With QQ fixed, cost ∝hf\propto h_f (neglecting all losses except friction).

hf=fLV22gD,V=4QπD2  ⇒  hf∝fD5h_f = \frac{fLV^2}{2gD}, \qquad V = \frac{4Q}{\pi D^2} \;\Rightarrow\; h_f \propto \frac{f}{D^5}

For constant kk: f∝(kD)1/3∝D−1/3f \propto \left(\dfrac{k}{D}\right)^{1/3} \propto D^{-1/3}, so

hf∝D−1/3 D−5=D−16/3h_f \propto D^{-1/3}\,D^{-5} = D^{-16/3}

Cost halved: (D′D)−16/3=12\left(\dfrac{D'}{D}\right)^{-16/3} = \dfrac{1}{2} gives

D′D=23/16=1.1388  ⇒  D′=0.5(1.1388)=0.569 m≈569 mm\frac{D'}{D} = 2^{3/16} = 1.1388 \;\Rightarrow\; D' = 0.5(1.1388) = 0.569\ \text{m} \approx 569\ \text{mm}

Check: new velocity V′=3.142V' = 3.142 m/s and Re′=2981506>106Re' = 2981506 > 10^6, so the assumption holds.

Part 2: n equal pipes in parallel at the same Reynolds number

The same Re=VD/νRe = VD/\nu for the same fluid means V′D′=VDV'D' = VD. Each of the nn pipes carries Q/nQ/n:

Qn=π4D′2V′  ⇒  π4D2V=nπ4D′2V′=nπ4D′ (VD)  ⇒  D′=Dn,V′=nV\frac{Q}{n} = \frac{\pi}{4}D'^2V' \;\Rightarrow\; \frac{\pi}{4}D^2V = n\frac{\pi}{4}D'^2V' = n\frac{\pi}{4}D'\,(VD)\;\Rightarrow\; D' = \frac{D}{n},\quad V' = nV

Relative roughness rises to k/D′=n(k/D)k/D' = n(k/D), so f′=n1/3ff' = n^{1/3}f. Head loss of each pipe (all pipes have the same hfh_f in parallel):

hf′=f′LV′22gD′=n1/3f L n2V22g(D/n)=n3+1/3hf=n10/3hfh_f' = \frac{f'LV'^2}{2gD'} = \frac{n^{1/3}f\,L\,n^2V^2}{2g(D/n)} = n^{3+1/3}h_f = n^{10/3}h_f

The discharge is the same, so the power and the running cost increase by the factor

cost′cost=n10/3\boxed{\frac{\text{cost}'}{\text{cost}} = n^{10/3}}

For example, two pipes (n=2n = 2) cost 210/3=10.12^{10/3} = 10.1 times as much. Using one large pipe is much cheaper than several small ones.

Answer: D′≈569D' \approx 569 mm (about 570 mm); running cost rises by n10/3n^{10/3} for nn parallel pipes.

  • 2073 Bhadra · 8 marks

In a hydrodynamically rough pipe of 100 mm diameter, the ratio of velocities at 10 mm and 30 mm from the pipe wall is 0.838. Determine the average height of the wall roughness, shear stress at the wall and mean velocity of flow if velocity at 30 mm is 1.90 m/s.

Answer

Data. D=100D = 100 mm so R=50R = 50 mm. Distances from the wall: y1=10y_1 = 10 mm, y2=30y_2 = 30 mm. u2=1.90u_2 = 1.90 m/s at 30 mm; u1/u2=0.838u_1/u_2 = 0.838. Water, ρ=1000\rho = 1000 kg/m³.

Velocity distribution in a hydrodynamically rough pipe:

uu∗=5.75log⁡10yk+8.5\frac{u}{u_*} = 5.75\log_{10}\frac{y}{k} + 8.5

Shear velocity

u1=0.838(1.90)=1.5922u_1 = 0.838(1.90) = 1.5922 m/s. Subtracting the equation at y1y_1 from the one at y2y_2 (the constants and kk cancel):

u2−u1u∗=5.75log⁡10y2y1=5.75log⁡103=2.7434\frac{u_2 - u_1}{u_*} = 5.75\log_{10}\frac{y_2}{y_1} = 5.75\log_{10}3 = 2.7434 u∗=1.90−1.59222.7434=0.30782.7434=0.1122 m/su_* = \frac{1.90 - 1.5922}{2.7434} = \frac{0.3078}{2.7434} = 0.1122\ \text{m/s}

(i) Average roughness height

At y2=30y_2 = 30 mm: 1.900.1122=16.935=5.75log⁡1030k+8.5\dfrac{1.90}{0.1122} = 16.935 = 5.75\log_{10}\dfrac{30}{k} + 8.5

log⁡1030k=16.935−8.55.75=1.4669  ⇒  k=30101.4669=1.024 mm\log_{10}\frac{30}{k} = \frac{16.935 - 8.5}{5.75} = 1.4669 \;\Rightarrow\; k = \frac{30}{10^{1.4669}} = 1.024\ \text{mm}

Check: u∗kν=0.1122(1.024×10−3)10−6=115>100\dfrac{u_*k}{\nu} = \dfrac{0.1122(1.024\times10^{-3})}{10^{-6}} = 115 > 100, so the pipe is hydrodynamically rough, as given.

(ii) Wall shear stress

τ0=ρu∗2=1000(0.1122)2=12.59 Pa\tau_0 = \rho u_*^2 = 1000(0.1122)^2 = 12.59\ \text{Pa}

(iii) Mean velocity

For rough pipes Vu∗=5.75log⁡10Rk+4.75\dfrac{V}{u_*} = 5.75\log_{10}\dfrac{R}{k} + 4.75, with R/k=501.024=48.84R/k = \dfrac{50}{1.024} = 48.84:

V=0.1122[5.75log⁡10(48.84)+4.75]=1.622 m/sV = 0.1122\left[5.75\log_{10}(48.84) + 4.75\right] = 1.622\ \text{m/s}

Answer: k=1.024k = 1.024 mm (about 1.0 mm); τ0=12.59\tau_0 = 12.59 Pa (about 12.6 Pa); V=1.622V = 1.622 m/s.

  • 2073 Magh · 8 marks

Determine the size of steel pipe required to carry water at 30 l/s if the permissible energy gradient is 0.05. Will the boundary act as smooth or in transition?

Answer

Data. Q=0.03Q = 0.03 m³/s, permissible energy gradient S=hf/L=0.05S = h_f/L = 0.05. Commercial steel pipe: k=0.046k = 0.046 mm (Moody table). Water at 20 °C, ν=1.0×10−6\nu = 1.0\times10^{-6} m²/s. This is a "find the diameter" problem, solved by trial.

Working equation

S=fDV22g,V=4QπD2  ⇒  S=8fQ2π2gD5  ⇒  D=(8fQ2π2gS)1/5S = \frac{f}{D}\frac{V^2}{2g},\quad V = \frac{4Q}{\pi D^2} \;\Rightarrow\; S = \frac{8fQ^2}{\pi^2 g D^5} \;\Rightarrow\; D = \left(\frac{8fQ^2}{\pi^2 g S}\right)^{1/5}

Trials (f from the Colebrook-White equation at each new D)

Trialf usedD (mm)
00.02000124.4
10.01740120.9
20.01742121.0
30.01742121.0

The values converge to D=0.1210D = 0.1210 m. Final check: V=2.610V = 2.610 m/s, Re=315734Re = 315734, k/D=0.000380k/D = 0.000380, f=0.01742f = 0.01742 (Colebrook), and S=fV2/(2gD)=0.0500S = fV^2/(2gD) = 0.0500, which matches the required 0.05.

Diameter required: D≈121D \approx 121 mm. In practice a 125 mm commercial size would be adopted (it gives SS = 0.0424 < 0.05 at VV = 2.44 m/s).

Smooth or transition?

u∗=Vf8=2.6100.017428=0.1218 m/s,Re∗=u∗kν=0.1218(0.046×10−3)10−6=5.60u_* = V\sqrt{\frac{f}{8}} = 2.610\sqrt{\frac{0.01742}{8}} = 0.1218\ \text{m/s},\qquad Re_* = \frac{u_*k}{\nu} = \frac{0.1218(0.046\times10^{-3})}{10^{-6}} = 5.60

Since 4<Re∗<1004 < Re_* < 100, the boundary acts in the transition zone (neither smooth nor fully rough), as the Colebrook-White equation assumes.

Answer: D≈121D \approx 121 mm (use 125 mm); the boundary is in the transition regime.

  • 2072 Asoj · 4+2+2 marks

A horizontal pipe 60 mm in diameter carries oil of specific gravity 0.8. The pressure difference between two sections 5 km apart is found to be 200 kPa. The oil flowing through the pipe is collected in a tank. It is found that 1962 N of oil is collected in 4 minutes. Compute the dynamic viscosity of the oil. Assume the flow to be laminar and verify it. Also, find the velocity at a distance of 20 mm from the pipe wall.

Answer

Data. D=0.06D = 0.06 m (R=0.03R = 0.03 m), s.g. 0.8 so ρ=800\rho = 800 kg/m³, Δp=200\Delta p = 200 kPa over L=5L = 5 km. Weight collected: 1962 N in 4 min = 240 s.

Discharge

Weight flow rate =1962/240=8.175= 1962/240 = 8.175 N/s, so

Q=Wρg=8.175800(9.81)=0.001042 m3/s=1.042 l/sQ = \frac{W}{\rho g} = \frac{8.175}{800(9.81)} = 0.001042\ \text{m}^3/\text{s} = 1.042\ \text{l/s}

Dynamic viscosity (laminar flow, Hagen-Poiseuille)

Δp=128μLQπD4  ⇒  μ=πD4 Δp128LQ=π(0.06)4(200 000)128(5000)(0.001042)=0.01221 Pa s\Delta p = \frac{128\mu L Q}{\pi D^4} \;\Rightarrow\; \mu = \frac{\pi D^4\,\Delta p}{128 L Q} = \frac{\pi(0.06)^4(200\,000)}{128(5000)(0.001042)} = 0.01221\ \text{Pa s}

Verification that the flow is laminar

V=QπD2/4=0.3684 m/s,Re=ρVDμ=800(0.3684)(0.06)0.01221=1448<2000V = \frac{Q}{\pi D^2/4} = 0.3684\ \text{m/s},\qquad Re = \frac{\rho VD}{\mu} = \frac{800(0.3684)(0.06)}{0.01221} = 1448 < 2000

So the flow is laminar and the assumption is correct.

Velocity at 20 mm from the wall

Radius at that point: r=R−0.020=0.010r = R - 0.020 = 0.010 m. For laminar flow

u=14μΔpL(R2−r2)=14(0.01221)(200 0005000)(0.032−0.012)=0.655 m/su = \frac{1}{4\mu}\frac{\Delta p}{L}\left(R^2 - r^2\right) = \frac{1}{4(0.01221)}\left(\frac{200\,000}{5000}\right)\left(0.03^2 - 0.01^2\right) = 0.655\ \text{m/s}

(The centre-line velocity is 0.737 m/s, twice the mean velocity.)

Answer: μ=0.0122\mu = 0.0122 Pa s (about 0.0122 Pa s); Re≈1448Re \approx 1448 (laminar); uu at 20 mm from the wall =0.655= 0.655 m/s.

  • 2070 Bhadra · 8 marks

Show that for turbulent flow in rough pipes VV∗=5.75log⁡(RK)+4.75\frac{V}{V_*} = 5.75\log\left(\frac{R}{K}\right) + 4.75, where V = mean velocity, V∗V_* = shear velocity, R = radius of pipe, K = average height of surface protrusions.

Answer

Start. For a hydrodynamically rough pipe the velocity at distance yy from the wall, from Prandtl's mixing length theory and Nikuradse's experiments, is

uu∗=5.75log⁡10yK+8.5⋯(1)\frac{u}{u_*} = 5.75\log_{10}\frac{y}{K} + 8.5 \quad\cdots(1)

where u∗=V∗=τ0/ρu_* = V_* = \sqrt{\tau_0/\rho} is the shear velocity and KK is the roughness height. It is assumed to hold over the whole section. Mean velocity is

V=1πR2∫0Ru 2πr drV = \frac{1}{\pi R^2}\int_0^R u\,2\pi r\,dr

Put y=R−ry = R - r, so r=R−yr = R - y and dr=−dydr = -dy:

V=2R2∫0Ru (R−y) dyV = \frac{2}{R^2}\int_0^R u\,(R - y)\,dy

Split (1) as uu∗=5.75log⁡10RK+8.5+5.75log⁡10yR\dfrac{u}{u_*} = 5.75\log_{10}\dfrac{R}{K} + 8.5 + 5.75\log_{10}\dfrac{y}{R}. The first two terms are constant, and for them 2R2∫0R(R−y) dy=1\dfrac{2}{R^2}\displaystyle\int_0^R (R-y)\,dy = 1. For the last term use log⁡10x=0.4343ln⁡x\log_{10}x = 0.4343\ln x and let η=y/R\eta = y/R:

2R2∫0R(R−y)ln⁡yR dy=2∫01(1−η)ln⁡η dη\frac{2}{R^2}\int_0^R (R - y)\ln\frac{y}{R}\,dy = 2\int_0^1 (1 - \eta)\ln\eta\,d\eta

Using ∫01ln⁡η dη=−1\int_0^1 \ln\eta\,d\eta = -1 and ∫01ηln⁡η dη=−14\int_0^1 \eta\ln\eta\,d\eta = -\dfrac14:

2[−1−(−14)]=2(−34)=−322\left[-1 - \left(-\frac14\right)\right] = 2\left(-\frac34\right) = -\frac32

Hence the area-weighted average of 5.75log⁡10(y/R)5.75\log_{10}(y/R) is

5.75(0.4343)(−32)=2.5(−32)=−3.755.75(0.4343)\left(-\frac32\right) = 2.5\left(-\frac32\right) = -3.75

Result.

VV∗=5.75log⁡10RK+8.5−3.75\frac{V}{V_*} = 5.75\log_{10}\frac{R}{K} + 8.5 - 3.75 VV∗=5.75log⁡10(RK)+4.75\boxed{\frac{V}{V_*} = 5.75\log_{10}\left(\frac{R}{K}\right) + 4.75}

(The same integral applied to the smooth-pipe law gives VV∗=5.75log⁡10V∗Rν+1.75\dfrac{V}{V_*} = 5.75\log_{10}\dfrac{V_*R}{\nu} + 1.75.) The equation also gives ff for rough pipes: since V∗=Vf/8V_* = V\sqrt{f/8}, 1f=2.03log⁡10RK+1.68\dfrac{1}{\sqrt f} = 2.03\log_{10}\dfrac{R}{K} + 1.68, close to the von Karman-Prandtl form.

  • 2070 Magh · 8 marks

Show that in both smooth and rough pipes for turbulent flow u−Vv∗=5.75log⁡(yR)+3.75\frac{u-V}{v_*} = 5.75\log\left(\frac{y}{R}\right) + 3.75, where V = mean velocity; u = point velocity at distance y from boundary; v∗v_* = shear velocity; R = radius of pipe.

Answer

Idea. Write the log law for the point velocity uu and for the mean velocity VV, then subtract. The wall constants (and the roughness kk or ν\nu) cancel, so one law describes both smooth and rough pipes. Let v∗v_* be the shear velocity and yy the distance from the wall.

Point velocity (from Prandtl's mixing length theory and experiments)

  • Smooth pipe: uv∗=5.75log⁡10v∗yν+5.5\dfrac{u}{v_*} = 5.75\log_{10}\dfrac{v_*y}{\nu} + 5.5 ... (1)
  • Rough pipe: uv∗=5.75log⁡10yk+8.5\dfrac{u}{v_*} = 5.75\log_{10}\dfrac{y}{k} + 8.5 ... (2)

Mean velocity

V=1πR2∫0Ru 2πr dr=2R2∫0Ru (R−y) dyV = \frac{1}{\pi R^2}\int_0^R u\,2\pi r\,dr = \frac{2}{R^2}\int_0^R u\,(R - y)\,dy

Write (1) as uv∗=5.75log⁡10v∗Rν+5.5+5.75log⁡10yR\dfrac{u}{v_*} = 5.75\log_{10}\dfrac{v_*R}{\nu} + 5.5 + 5.75\log_{10}\dfrac{y}{R}. The first two terms are constant. For the last term, with η=y/R\eta = y/R and log⁡10η=0.4343ln⁡η\log_{10}\eta = 0.4343\ln\eta:

2∫01(1−η)ln⁡η dη=2[−1+14]=−322\int_0^1 (1-\eta)\ln\eta\,d\eta = 2\left[-1 + \frac14\right] = -\frac32

so its average is 5.75(0.4343)(−1.5)=−3.755.75(0.4343)(-1.5) = -3.75. Hence

  • Smooth: Vv∗=5.75log⁡10v∗Rν+1.75\dfrac{V}{v_*} = 5.75\log_{10}\dfrac{v_*R}{\nu} + 1.75 ... (3)
  • Rough (same steps with (2)): Vv∗=5.75log⁡10Rk+4.75\dfrac{V}{v_*} = 5.75\log_{10}\dfrac{R}{k} + 4.75 ... (4)

Subtract to get the velocity defect law

Smooth pipe, (1) −- (3):

u−Vv∗=5.75[log⁡10v∗yν−log⁡10v∗Rν]+5.5−1.75=5.75log⁡10yR+3.75\frac{u - V}{v_*} = 5.75\left[\log_{10}\frac{v_*y}{\nu} - \log_{10}\frac{v_*R}{\nu}\right] + 5.5 - 1.75 = 5.75\log_{10}\frac{y}{R} + 3.75

Rough pipe, (2) −- (4):

u−Vv∗=5.75[log⁡10yk−log⁡10Rk]+8.5−4.75=5.75log⁡10yR+3.75\frac{u - V}{v_*} = 5.75\left[\log_{10}\frac{y}{k} - \log_{10}\frac{R}{k}\right] + 8.5 - 4.75 = 5.75\log_{10}\frac{y}{R} + 3.75

Both give the same result:

u−Vv∗=5.75log⁡10(yR)+3.75\boxed{\frac{u - V}{v_*} = 5.75\log_{10}\left(\frac{y}{R}\right) + 3.75}

Because the viscosity and roughness have dropped out, the shape of the velocity profile, measured as the defect from the mean velocity, is the same in smooth and rough pipes. At the axis (y=Ry = R): umax−Vv∗=3.75\dfrac{u_{max} - V}{v_*} = 3.75.

  • 2069 Bhadra · 8 marks

Describe with appropriate expressions (a) Prandtl's mixing length theory (b) Hagen-Poiseuille equation (c) Nikuradse's experiments and (d) Colebrook-White equation.

Answer

(a) Prandtl's mixing length theory

Turbulent eddies move lumps of fluid across the layers; a lump keeps its identity over the mixing length ll before mixing. The velocity fluctuations are u′≈v′≈l dudyu' \approx v' \approx l\,\dfrac{du}{dy}, so the turbulent shear stress is

τt=−ρu′v′‾=ρ l2(dudy)2,l=κy, κ≈0.4\tau_t = -\rho\overline{u'v'} = \rho\,l^2\left(\frac{du}{dy}\right)^2, \qquad l = \kappa y,\ \kappa \approx 0.4

The total stress is τ=μdudy+ρl2(dudy)2\tau = \mu\dfrac{du}{dy} + \rho l^2\left(\dfrac{du}{dy}\right)^2. For constant τ=τ0\tau = \tau_0 near the wall, it leads to the logarithmic velocity law uu∗=5.75log⁡10u∗yν+5.5\dfrac{u}{u_*} = 5.75\log_{10}\dfrac{u_*y}{\nu} + 5.5 (smooth) and 5.75log⁡10yk+8.55.75\log_{10}\dfrac{y}{k} + 8.5 (rough), where u∗=τ0/ρu_* = \sqrt{\tau_0/\rho}.

(b) Hagen-Poiseuille equation

For steady laminar flow in a circular pipe of diameter DD and length LL the discharge and head loss are

Q=πD4 Δp128 μL,hf=32μLVγD2Q = \frac{\pi D^4\,\Delta p}{128\,\mu L}, \qquad h_f = \frac{32\mu L V}{\gamma D^2}

The velocity is parabolic, u=Δp4μL(R2−r2)u = \dfrac{\Delta p}{4\mu L}(R^2 - r^2), with umax=2Vu_{max} = 2V. Comparing with Darcy-Weisbach gives f=64/Ref = 64/Re. It is valid for Re<2000Re < 2000.

(c) Nikuradse's experiments

Nikuradse (1933) measured head loss in pipes artificially roughened with uniform sand grains of size kk, for relative roughness k/Dk/D from 1/30 to 1/1014 and ReRe from 10310^3 to 10610^6. Plotting log⁡f\log f against log⁡Re\log Re he found:

  • Laminar flow: all pipes lie on f=64/Ref = 64/Re, independent of roughness.
  • Smooth turbulent: ff follows Blasius/Prandtl smooth curve, depends on ReRe only.
  • Transition: ff depends on both ReRe and k/Dk/D.
  • Fully rough: ff becomes constant for given k/Dk/D, depends on k/Dk/D only; 1f=2log⁡10Dk+1.14\dfrac{1}{\sqrt f} = 2\log_{10}\dfrac{D}{k} + 1.14.

(d) Colebrook-White equation

For commercial pipes, Colebrook and White combined the smooth and rough laws into one transition formula:

1f=−2log⁡10(k3.7D+2.51Ref)\frac{1}{\sqrt f} = -2\log_{10}\left(\frac{k}{3.7D} + \frac{2.51}{Re\sqrt f}\right)
  • When k→0k \to 0 (smooth): 1f=2log⁡10(Ref)−0.8\dfrac{1}{\sqrt f} = 2\log_{10}(Re\sqrt f) - 0.8.
  • When Re→∞Re \to \infty (rough): 1f=2log⁡10Dk+1.14\dfrac{1}{\sqrt f} = 2\log_{10}\dfrac{D}{k} + 1.14. The equation is implicit in ff and is solved by iteration; it is the basis of the Moody chart.
  • 2069 Poush · 4 marks

Measurement in a fully developed turbulent flow in pipe indicate that velocity midway between the pipe wall and the pipe centerline is 0.9 times the centerline velocity. Determine the expression for the average velocity in multiples of maximum velocity. What is the value of e/D or K/D (relative roughness) if pipe acts as rough pipe?

Answer

Use the velocity defect law (valid for smooth and rough pipes): umax−uu∗=5.75log⁡10Ry\dfrac{u_{max} - u}{u_*} = 5.75\log_{10}\dfrac{R}{y}, and the mean velocity relation umax−Vu∗=3.75\dfrac{u_{max} - V}{u_*} = 3.75.

Step 1: use the midway reading

Midway between the wall and the centre line, y=R/2y = R/2 and u=0.9 umaxu = 0.9\,u_{max}:

umax−0.9umaxu∗=5.75log⁡10RR/2=5.75log⁡102=1.7309\frac{u_{max} - 0.9u_{max}}{u_*} = 5.75\log_{10}\frac{R}{R/2} = 5.75\log_{10}2 = 1.7309 0.1 umaxu∗=1.7309  ⇒  umax=17.309 u∗\frac{0.1\,u_{max}}{u_*} = 1.7309 \;\Rightarrow\; u_{max} = 17.309\,u_*

Step 2: average velocity

V=umax−3.75u∗=17.309u∗−3.75u∗=13.559 u∗V = u_{max} - 3.75u_* = 17.309u_* - 3.75u_* = 13.559\,u_* Vumax=13.55917.309=0.7834\frac{V}{u_{max}} = \frac{13.559}{17.309} = 0.7834 V=0.783 umax\boxed{V = 0.783\,u_{max}}

Step 3: relative roughness (rough pipe)

The velocity at the axis of a rough pipe: umaxu∗=5.75log⁡10Rk+8.5\dfrac{u_{max}}{u_*} = 5.75\log_{10}\dfrac{R}{k} + 8.5 (from y=Ry = R).

17.309=5.75log⁡10Rk+8.5  ⇒  log⁡10Rk=17.309−8.55.75=1.53217.309 = 5.75\log_{10}\frac{R}{k} + 8.5 \;\Rightarrow\; \log_{10}\frac{R}{k} = \frac{17.309 - 8.5}{5.75} = 1.532 Rk=34.04  ⇒  kD=12(34.04)=0.0147\frac{R}{k} = 34.04 \;\Rightarrow\; \frac{k}{D} = \frac{1}{2(34.04)} = 0.0147

Answer: V=0.783 umaxV = 0.783\,u_{max}; relative roughness k/D≈0.0147k/D \approx 0.0147 (about 1.5%).

  • 2069 Poush · 4 marks

Write down Colebrook and White equation. Show that this equation is also valid for variation of friction factor for turbulent rough as well as smooth pipes.

Answer

Colebrook-White equation for turbulent flow in commercial pipes:

1f=−2log⁡10(k3.7D+2.51Ref)\frac{1}{\sqrt f} = -2\log_{10}\left(\frac{k}{3.7D} + \frac{2.51}{Re\sqrt f}\right)

where ff is the Darcy friction factor, kk the equivalent roughness, DD the diameter and ReRe the Reynolds number.

Smooth pipes (k→0k \to 0, so k/D→0k/D \to 0)

The roughness term vanishes:

1f=−2log⁡102.51Ref=2log⁡10Ref2.51=2log⁡10(Ref)−2log⁡102.51\frac{1}{\sqrt f} = -2\log_{10}\frac{2.51}{Re\sqrt f} = 2\log_{10}\frac{Re\sqrt f}{2.51} = 2\log_{10}(Re\sqrt f) - 2\log_{10}2.51

Since 2log⁡102.51=0.802\log_{10}2.51 = 0.80:

1f=2log⁡10(Ref)−0.8\frac{1}{\sqrt f} = 2\log_{10}(Re\sqrt f) - 0.8

This is Prandtl's universal resistance law for smooth pipes.

Rough pipes (large ReRe, fully rough)

As Re→∞Re \to \infty the second term 2.51Ref→0\dfrac{2.51}{Re\sqrt f} \to 0:

1f=−2log⁡10k3.7D=2log⁡10Dk+2log⁡103.7\frac{1}{\sqrt f} = -2\log_{10}\frac{k}{3.7D} = 2\log_{10}\frac{D}{k} + 2\log_{10}3.7

Since 2log⁡103.7=1.142\log_{10}3.7 = 1.14:

1f=2log⁡10Dk+1.14\frac{1}{\sqrt f} = 2\log_{10}\frac{D}{k} + 1.14

This is the von Karman-Prandtl law for rough pipes, where ff is independent of ReRe.

For intermediate values both terms in the bracket matter, so the equation describes the transition zone, and it covers the whole turbulent range with one formula.

  • 2068 Bhadra · 2+6 marks

Explain Prandtl mixing length theory. Starting from the expression for turbulent shear stress derive the velocity distribution for turbulent flow near hydrodynamically smooth boundaries in the form uu∗=5.75log⁡10(u∗yν)+5.5\frac{u}{u_*} = 5.75\log_{10}\left(\frac{u_* y}{\nu}\right) + 5.5.

Answer

Prandtl's mixing length theory

In turbulent flow, lumps of fluid are moved across layers by eddies. Prandtl assumed a lump keeps its momentum over a distance ll (mixing length) before mixing. The velocity fluctuations are then u′≈v′≈l dudyu' \approx v' \approx l\,\dfrac{du}{dy}, and the turbulent shear stress is

τt=−ρu′v′‾=ρ l2(dudy)2\tau_t = -\rho\overline{u'v'} = \rho\,l^2\left(\frac{du}{dy}\right)^2

Near a wall the mixing length increases with distance from it: l=κyl = \kappa y with κ≈0.4\kappa \approx 0.4 (von Karman's constant).

Velocity distribution near a smooth boundary

Let u∗=τ0/ρu_* = \sqrt{\tau_0/\rho}. Close to the wall the shear stress is practically constant, τ=τ0\tau = \tau_0, and the viscous stress is negligible outside the laminar sub-layer:

τ0=ρκ2y2(dudy)2  ⇒  dudy=u∗κy\tau_0 = \rho\kappa^2y^2\left(\frac{du}{dy}\right)^2 \;\Rightarrow\; \frac{du}{dy} = \frac{u_*}{\kappa y}

Integrating:

u=u∗κln⁡y+C⋯(1)u = \frac{u_*}{\kappa}\ln y + C \quad\cdots(1)

Constant from the laminar sub-layer. In the laminar sub-layer of thickness δ′\delta', τ0=μu/y\tau_0 = \mu u/y, so u=τ0yμ=u∗2yνu = \dfrac{\tau_0y}{\mu} = \dfrac{u_*^2y}{\nu}. At its edge y=δ′y = \delta', experiments give δ′=11.6νu∗\delta' = \dfrac{11.6\nu}{u_*}, so

uδ′=u∗2ν⋅11.6νu∗=11.6 u∗u_{\delta'} = \frac{u_*^2}{\nu}\cdot\frac{11.6\nu}{u_*} = 11.6\,u_*

Applying (1) from y=δ′y = \delta' to yy:

u−11.6u∗=u∗κln⁡yδ′=u∗κln⁡u∗y11.6νu - 11.6u_* = \frac{u_*}{\kappa}\ln\frac{y}{\delta'} = \frac{u_*}{\kappa}\ln\frac{u_*y}{11.6\nu}

With κ=0.4\kappa = 0.4:

uu∗=11.6+2.5ln⁡u∗yν−2.5ln⁡11.6=2.5ln⁡u∗yν+11.6−6.13\frac{u}{u_*} = 11.6 + 2.5\ln\frac{u_*y}{\nu} - 2.5\ln 11.6 = 2.5\ln\frac{u_*y}{\nu} + 11.6 - 6.13

Convert ln⁡\ln to log⁡10\log_{10}: 2.5ln⁡x=2.5(2.303)log⁡10x=5.75log⁡10x2.5\ln x = 2.5(2.303)\log_{10}x = 5.75\log_{10}x:

uu∗=5.75log⁡10(u∗yν)+5.5\boxed{\frac{u}{u_*} = 5.75\log_{10}\left(\frac{u_*y}{\nu}\right) + 5.5}

(the exact constant 5.47 is rounded to 5.5, a value that agrees with Nikuradse's measurements).

  • 2068 Magh · 3+5 marks

Explain the experiment made by Nikuradse on resistance to artificially roughened pipes. Discuss the characteristic features of the result obtained.

Answer

The experiment

Nikuradse (1933) coated the inside of pipes with uniform sand grains of known size kk glued with varnish, giving relative roughness k/Dk/D = 1/30, 1/61, 1/120, 1/252, 1/504 and 1/1014. Water was passed through each pipe at a wide range of discharges (Reynolds number from about 5×1025\times10^2 to 10610^6). He measured the discharge and the pressure drop over a length of pipe, calculated the friction factor ff from hf=fLV22gDh_f = \dfrac{fLV^2}{2gD}, and plotted log⁡f\log f against log⁡Re\log Re for each k/Dk/D. He also measured velocity profiles to get the log laws.

Results (characteristic features)

 log f
   |  \ laminar              rough: f flat for each k/D
   |   \ f=64/Re   ____________________ k/D = 1/30
   |    \         /    ___________________ 1/61
   |     \   crit/   _/  ________________ 1/504
   |      \     /  _/ __/
   |       `-._/ _/  /  smooth (Blasius) curve
   |           `-----'
   +--------------------------------- log Re
RegionReynolds numberBehaviour of ff
LaminarRe<2000Re < 2000f=64/Ref = 64/Re for all roughnesses (a single straight line)
Critical2000 to 4000ff rises sharply; unstable
Smooth turbulentabove 4000 up to a limit depending on k/Dk/DAll pipes follow the smooth curve, f=f(Re)f = f(Re), e.g. Blasius 0.316/Re0.250.316/Re^{0.25}. The roughness is within the laminar sub-layer
TransitionintermediatePipe leaves the smooth curve; ff depends on both ReRe and k/Dk/D. The larger the k/Dk/D, the earlier it departs
Fully roughhigh ReReCurve is horizontal: ff depends only on k/Dk/D, 1f=2log⁡10Dk+1.14\dfrac{1}{\sqrt f} = 2\log_{10}\dfrac{D}{k} + 1.14

Other features:

  • Roughness has no effect on laminar flow.
  • Pipes with higher k/Dk/D leave the smooth curve at lower ReRe and reach higher constant ff values.
  • The classification by the roughness Reynolds number u∗kν\dfrac{u_*k}{\nu}: smooth below about 4, transition 4 to 100, fully rough above about 100 (some texts use 70).
  • Some curves show a slight dip below the smooth line before rising (in the transition for uniform roughness), a feature not seen in commercial pipes. The results formed the basis of the Moody chart and the Colebrook-White equation.

Questions from Old Question Collection (CE 555) (IOE Hydraulics (CE 555) exam papers from 2068 to 2082). Answers are written for this site; check them against your class notes.

Chapter titles and hours from the IOE syllabus ↗