8.3 Fluid Flow in Pipes, Friction Losses, and Pipe Networks

8.3 Fluid Flow in Pipes, Friction Losses, and Pipe Networks

Water distribution systems, sewer force mains, transmission aqueducts, and penstocks rely on closed conduit pressurized pipe hydraulics. In the Philippine CELE board exam, conduit analysis requires calculating major frictional resistance, evaluating minor fitting losses, balancing parallel and branching pipe configurations, and performing Hardy Cross network iterations.

---

1. Flow Regimes and the Reynolds Number

In 1883, Osborne Reynolds demonstrated that fluid flow in circular pipes transitions between two fundamentally distinct physical states based on the dimensionless Reynolds number ($Re$): $Re = \frac{\rho v D}{\mu} = \frac{v D}{\nu}$ where $v$ is average velocity ($\text{m/s}$), $D$ is internal diameter ($\text{m}$), $\rho$ is mass density ($\text{kg/m}^3$), $\mu$ is dynamic viscosity ($\text{Pa}\cdot\text{s}$), and $\nu = \mu / \rho$ is kinematic viscosity ($\text{m}^2/\text{s}$). For standard water at $20^\circ\text{C}$, $\nu \approx 1.00 \times 10^{-6}\text{ m}^2/\text{s}$.

Critical Limits and Velocity Profiles

1. Laminar Flow ($Re < 2,000$): Viscous forces dominate inertial forces. Fluid particles move in straight, parallel streamlines without radial mixing. The velocity profile is a true paraboloid: $u(r) = u_{\max} \left(1 - \frac{r^2}{R^2}\right), \quad u_{\max} = 2 v_{\text{avg}}$ The boundary shear stress at the pipe wall is $\tau_0 = \frac{8 \mu v}{D}$.
2. Critical / Transitional Zone ($2,000 \le Re \le 4,000$): Flow is intermittently laminar and turbulent, sensitive to external vibrations and boundary roughness.
3. Turbulent Flow ($Re > 4,000$): Inertial forces overwhelm viscous damping. Chaotic eddy currents induce intense momentum exchange across the conduit. The velocity profile flattens into a logarithmic or power-law distribution: $u(r) = u_{\max} \left(1 - \frac{r}{R}\right)^{1/n} \quad (n \approx 7), \quad u_{\max} \approx 1.20\text{–}1.25 v_{\text{avg}}$

---

2. Major Friction Head Loss: The Darcy-Weisbach Equation

The Darcy-Weisbach equation is the theoretically exact, dimensionally homogeneous equation for major friction loss in circular pipes: $h_f = f \frac{L}{D} \frac{v^2}{2g}$ Substituting continuity $v = \frac{4Q}{\pi D^2}$ yields the standard PRC SI discharge formulation: $h_f = f \frac{L}{D} \frac{16 Q^2}{2 g \pi^2 D^4} = \frac{8 f L Q^2}{\pi^2 g D^5} \approx \frac{0.0826 f L Q^2}{D^5}$

Determination of the Friction Factor ($f$)

1. Laminar Flow ($Re < 2,000$): Derived analytically from the Hagen-Poiseuille equation. The friction factor is strictly a function of Reynolds number, independent of wall roughness: $h_f = \frac{32 \mu L v}{\gamma D^2} = \frac{64}{Re} \frac{L}{D} \frac{v^2}{2g} \implies f = \frac{64}{Re}$

2. Turbulent Flow ($Re > 4,000$): The friction factor depends on $Re$ and the relative pipe roughness $\frac{\epsilon}{D}$ (where $\epsilon$ is absolute equivalent sand-grain roughness height):

   • Colebrook-White Implicit Equation: The universal standard embodied in the Moody Diagram: $\frac{1}{\sqrt{f}} = -2.0 \log_{10} \left(\frac{\epsilon}{3.7 D} + \frac{2.51}{Re \sqrt{f}}\right)$
   • Swamee-Jain Explicit Approximation: Solves the Colebrook equation directly within $1\%$ accuracy for $10^{-6} \le \frac{\epsilon}{D} \le 10^{-2}$ and $5,000 \le Re \le 10^8$: $f = \frac{0.25}{\left[\log_{10}\left(\frac{\epsilon}{3.7 D} + \frac{5.74}{Re^{0.9}}\right)\right]^2}$
   • Wholly Turbulent (Rough Pipe) Flow: At very high $Re$, the laminar sublayer vanishes completely, and $f$ becomes independent of $Re$ (von Kármán equation): $\frac{1}{\sqrt{f}} = 2.0 \log_{10} \left(\frac{3.7 D}{\epsilon}\right)$

---

3. Empirical Formulations: The Hazen-Williams Equation

While Darcy-Weisbach applies to all fluids, the Hazen-Williams equation is an empirical power-law formula tailored specifically for water flow at ordinary temperatures ($10^\circ\text{C}$ to $25^\circ\text{C}$): $v = 0.8492 C R_h^{0.63} S^{0.54} \quad (\text{SI Metric})$ For a circular pipe flowing full ($R_h = D/4$, $S = h_f/L$): $h_f = \frac{10.67 L Q^{1.852}}{C^{1.852} D^{4.87}}$ where $Q$ is in $\text{m}^3/\text{s}$, $L$ and $D$ are in $\text{m}$, and $C$ is the dimensionless Hazen-Williams roughness coefficient.

---

4. Minor (Local) Head Losses

Minor losses ($h_m$) result from localized streamline disruption, boundary layer separation, and secondary eddy generation across fittings, valves, contractions, and bends: $h_m = K \frac{v^2}{2g}$ where $K$ is the dimensionless minor loss coefficient.

The Equivalent Length Method

A minor loss can be modeled as an equivalent length ($L_e$) of straight pipe that produces the identical major head loss: $K \frac{v^2}{2g} = f \frac{L_e}{D} \frac{v^2}{2g} \implies L_e = \frac{K D}{f}$ The total effective length used in Darcy-Weisbach is $L_{\text{total}} = L_{\text{actual}} + \sum L_e$.

---

5. Multi-Pipe Systems: Series, Parallel, and Branching

Loading diagram...

1. Pipes in Series

Conduits of different diameters are connected end-to-end:

• Discharge is constant: $Q_1 = Q_2 = \dots = Q$.
• Head losses are additive: $h_{L, \text{total}} = h_{L1} + h_{L2} + \dots + h_{Ln}$.
• Equivalent Single Pipe ($D_e, L_e$): Assuming uniform friction factor $f$: $\frac{0.0826 f L_e Q^2}{D_e^5} = \sum_{i=1}^n \frac{0.0826 f L_i Q^2}{D_i^5} \implies \frac{L_e}{D_e^5} = \sum_{i=1}^n \frac{L_i}{D_i^5}$

2. Pipes in Parallel

A single conduit branches into two or more parallel lines that rejoin downstream:

• Head loss across every branch is identical: $h_{L1} = h_{L2} = \dots = h_L = \Delta H$.
• Discharges are additive: $Q_{\text{total}} = Q_1 + Q_2 + \dots + Q_n$.
• Flow Distribution Ratio (Darcy-Weisbach): For two parallel branches with equal $f$: $h_{L1} = h_{L2} \implies \frac{0.0826 f L_1 Q_1^2}{D_1^5} = \frac{0.0826 f L_2 Q_2^2}{D_2^5} \implies \frac{Q_1}{Q_2} = \left(\frac{D_1}{D_2}\right)^{5/2} \sqrt{\frac{L_2}{L_1}}$

3. Branching Pipes: The Three-Reservoir Problem

Three open reservoirs at surface elevations $z_A > z_B > z_C$ connect to a common underground junction $J$ at elevation $z_J$:

1. Define the junction piezometric head: $H_J = z_J + \frac{P_J}{\gamma}$.
2. Flow in pipe $A$ always leaves Reservoir A: $h_{fA} = z_A - H_J$, so $Q_A = \sqrt{\frac{(z_A - H_J) D_A^5}{0.0826 f_A L_A}}$.
3. Flow in pipe $C$ always enters Reservoir C: $h_{fC} = H_J - z_C$, so $Q_C = \sqrt{\frac{(H_J - z_C) D_C^5}{0.0826 f_C L_C}}$.
4. Flow in the intermediate reservoir pipe $B$ depends on whether $H_J$ is above or below $z_B$:
   • If $H_J > z_B$: Water flows into reservoir B: $Q_A = Q_B + Q_C$.
   • If $H_J < z_B$: Water flows out of reservoir B: $Q_A + Q_B = Q_C$.
   • If $H_J = z_B$: No flow occurs in pipe B ($Q_B = 0$), so $Q_A = Q_C$.
5. Solve iteratively by adjusting $H_J$ until continuity $\sum Q = 0$ at the junction is satisfied within tolerance.

---

6. Pipe Networks: The Hardy Cross Method

In looped municipal distribution networks, pipe flows must satisfy two fundamental laws analogous to Kirchhoff's circuit laws:

1. Junction Law (Continuity): $\sum Q = 0$ at every pipe node.
2. Loop Law (Energy Conservation): $\sum H_L = 0$ around every closed loop.

Head loss in any pipe is expressed as $H_L = r Q |Q|^{n-1}$, where $n = 2.0$ for Darcy-Weisbach ($r = \frac{0.0826 f L}{D^5}$) and $n = 1.852$ for Hazen-Williams ($r = \frac{10.67 L}{C^{1.852} D^{4.87}}$).

Derivation of the Hardy Cross Loop Correction ($\Delta Q$)

Let initial assumed pipe flows satisfying junction continuity be $Q_0$. Let $\Delta Q$ be the uniform counter-clockwise flow correction applied to the loop: $\sum r (Q_0 + \Delta Q)^n = 0 \approx \sum r Q_0^n + n \Delta Q \sum r Q_0^{n-1} = 0$ $\Delta Q = - \frac{\sum H_L}{n \sum \left|\frac{H_L}{Q}\right|}$

• Sign Convention: Clockwise flows and head losses are positive (+); counter-clockwise flows are negative (-).
• Denominator: The denominator $n \sum |H_L / Q|$ is always positive because it represents the derivative of head loss with respect to flow rate.
• Shared Pipes: Pipes common to two loops receive corrections from both: $Q_{\text{new}} = Q_{\text{old}} + \Delta Q_I - \Delta Q_{II}$.

---

7. Worked Example: Parallel Pipe Discharge Distribution

Problem Statement: A main transmission pipeline conveys $0.250\text{ m}^3/\text{s}$ ($250\text{ L/s}$) of water. To cross a river gorge, the pipeline splits into two parallel pipes A and B that rejoin on the opposite side:

• Pipe A: Length $L_A = 1,000\text{ m}$, internal diameter $D_A = 300\text{ mm}$ ($0.300\text{ m}$), $f_A = 0.020$.
• Pipe B: Length $L_B = 1,000\text{ m}$, internal diameter $D_B = 200\text{ mm}$ ($0.200\text{ m}$), $f_B = 0.020$.

Neglecting minor losses, calculate (a) the flow rate carried by Pipe A, (b) the flow rate carried by Pipe B, and (c) the total friction head loss across the crossing in meters.

Step-by-Step Solution:

1. Equate head losses across parallel branches: $h_{fA} = h_{fB} \implies \frac{0.0826 f_A L_A Q_A^2}{D_A^5} = \frac{0.0826 f_B L_B Q_B^2}{D_B^5}$ Since $f_A = f_B = 0.020$ and $L_A = L_B = 1,000\text{ m}$: $\frac{Q_A^2}{D_A^5} = \frac{Q_B^2}{D_B^5} \implies \frac{Q_A}{Q_B} = \left(\frac{D_A}{D_B}\right)^{5/2} = \left(\frac{0.300}{0.200}\right)^{2.5} = (1.5)^{2.5}$ $(1.5)^{2.5} = (1.5)^2 \times \sqrt{1.5} = 2.25 \times 1.224745 = 2.7557$ $Q_A = 2.7557 Q_B$

2. Apply continuity ($Q_{\text{total}} = Q_A + Q_B$): $2.7557 Q_B + Q_B = 0.250\text{ m}^3/\text{s} \implies 3.7557 Q_B = 0.250$ $Q_B = \frac{0.250}{3.7557} = 0.06657\text{ m}^3/\text{s} = 66.57\text{ L/s}$ $Q_A = 0.250 - 0.06657 = 0.18343\text{ m}^3/\text{s} = 183.43\text{ L/s}$

3. Compute the head loss across the parallel section: $h_{fA} = \frac{0.0826 (0.020)(1,000)(0.18343)^2}{(0.300)^5} = \frac{1.652 \times 0.033647}{0.00243} = \frac{0.055585}{0.00243} = 22.87\text{ m}$ (Verification via Pipe B: $h_{fB} = \frac{1.652 \times (0.06657)^2}{(0.200)^5} = \frac{1.652 \times 0.0044315}{0.00032} = \frac{0.007321}{0.00032} = 22.88\text{ m}$. Values match perfectly!)

---

8. CELE Exam Traps & Common Computational Errors