Darcy Friction Factor: Colebrook-White, Swamee-Jain & Churchill Compared

1. Introduction: The Hydraulic Foundation of Pipe Pressure Drop

In chemical process engineering, petroleum refining, and water distribution networks, the calculation of frictional energy loss in closed conduits is the single most frequent hydraulic calculation performed. Frictional pressure drop directly determines pump head requirements, line pipe diameter sizing, compressor station spacing, and piping stress wall thicknesses.

The universal relationship governing frictional head loss is the Darcy-Weisbach Equation, first formulated by Henry Darcy and Julius Weisbach in the mid-19th century:

$$h_f = f_D \cdot \frac{L}{D} \cdot \frac{v^2}{2g} \quad \iff \quad \Delta P_f = f_D \cdot \frac{L}{D} \cdot \frac{\rho v^2}{2}$$

Where:

• $h_f$ = Frictional head loss (m or ft)
• $\Delta P_f$ = Frictional pressure loss (Pa, bar, or psi)
• $f_D$ = Dimensionless Darcy-Weisbach friction factor (also known as the Darcy or Moody friction factor)
• $L$ = Pipe length (m)
• $D$ = Pipe internal diameter (m)
• $v$ = Bulk mean flow velocity (m/s)
• $\rho$ = Fluid density ($ \text{kg/m}^3$)
• $g$ = Gravitational acceleration ($9.80665 \text{ m/s}^2$)

2. Flow Regimes & The Friction Factor Spectrum

2.1 Laminar Flow Regime ($Re < 2,100$)

In laminar flow, fluid elements slide past each other in concentric cylindrical streamlines without macroscopic radial mixing. Momentum transfer is purely molecular, governed strictly by Newton's law of viscosity.

By integrating the Navier-Stokes equations for steady, fully developed laminar flow in a circular conduit (the Hagen-Poiseuille Law):

$$\Delta P = \frac{32 \mu L v}{D^2} = \left(\frac{64}{Re}\right) \cdot \frac{L}{D} \cdot \frac{\rho v^2}{2}$$

Equating this directly to the Darcy-Weisbach formulation reveals that the Darcy friction factor in laminar flow depends exclusively on the Reynolds number, completely independent of pipe wall roughness ($\varepsilon$):

$$f_D = \frac{64}{Re} \quad \left(f_{fan} = \frac{16}{Re}\right)$$

2.2 The Critical Transition Zone ($2,100 \le Re \le 4,000$)

In this zone, the laminar boundary layer destabilizes. Intermittent turbulent eddies erupt and decay unpredictably. The friction factor cannot be described by a single deterministic equation. Experimental measurements exhibit significant scatter. Process engineers should avoid designing steady-state systems in this window.

2.3 Fully Developed Turbulent Flow ($Re > 4,000$)

In turbulent flow, inertial momentum transport overwhelms viscous dissipation. The flow field divides into three sub-layers:

1. Viscous Sublayer: A microscopically thin film adhering to the wall where viscous shear dominates ($\delta_v \approx 5 u / u^*$).
2. Buffer Zone: Transition zone between viscous and eddy transport.
3. Turbulent Core: Vigorous three-dimensional turbulent eddies govern momentum exchange; velocity profile follows the classic logarithmic law of the wall.

Depending on the ratio of wall roughness asperities ($\varepsilon$) to the viscous sublayer thickness ($\delta_v$), turbulent flow is further categorized into:

• Hydraulically Smooth Pipes ($\varepsilon \ll \delta_v$): Roughness peaks are submerged inside the quiescent viscous sublayer. Friction factor depends only on $Re$.
• Transitional Turbulent Flow ($\varepsilon \approx \delta_v$): Roughness peaks protrude through the sublayer, creating parasitic form drag. $f_D$ depends on both $Re$ and relative roughness $\varepsilon / D$.
• Fully Rough / Wholly Turbulent Regime ($\varepsilon \gg \delta_v$): Sublayer is completely obliterated. Viscous forces become negligible; pressure drop scales strictly with $v^2$ and $f_D$ becomes independent of $Re$ (the horizontal plateau on the Moody diagram).

3. Governing Equations: Implicit vs. Explicit Formulations

3.1 The Colebrook-White Equation (1939)

Cyril Frank Colebrook and Cedric M. White combined Prandtl's smooth pipe theory and von K??rm??n's rough pipe experiments into an overarching semi-empirical formulation that forms the mathematical basis of the universal Moody Diagram:

$$\frac{1}{\sqrt{f_D}} = -2 \log_{10}\left(\frac{\varepsilon / D}{3.7} + \frac{2.51}{Re \sqrt{f_D}} \right)$$

Because $f_D$ appears on both sides of the equation inside and outside the logarithm, it is implicit and cannot be resolved analytically. In computer algorithms, it is solved using Newton-Raphson iteration:

$$F(x) = x + 2 \log_{10}\left(\frac{\varepsilon / D}{3.7} + \frac{2.51 x}{Re} \right) = 0 \quad \text{where } x = \frac{1}{\sqrt{f_D}}$$ $$F'(x) = 1 + \frac{2}{\ln(10)} \cdot \frac{2.51 / Re}{\frac{\varepsilon / D}{3.7} + \frac{2.51 x}{Re}}$$ $$x_{new} = x - \frac{F(x)}{F'(x)}$$

With an initial estimate from the Swamee-Jain formula, the Newton-Raphson method converges to machine precision ($10^{-9}$) within 2 to 3 iterations.

3.2 Swamee-Jain Explicit Approximation (1976)

To eliminate iterative loops in spreadsheets and embedded microcontrollers, P.K. Swamee and A.K. Jain formulated an explicit approximation accurate within $\pm 1.0\%$ for $5,000 \le Re \le 10^8$ and $10^{-6} \le \varepsilon/D \le 10^{-2}$:

$$f_D = \frac{0.25}{\left[ \log_{10}\left(\frac{\varepsilon / D}{3.7} + \frac{5.74}{Re^{0.9}} \right) \right]^2}$$

Swamee-Jain is widely used in standard pipeline calculation routines due to its exceptional computational efficiency and negligible error relative to Colebrook-White.

3.3 Churchill Universal Equation (1977)

Stuart W. Churchill developed a remarkable continuous equation that spans all flow regimes???seamlessly transitioning through laminar ($Re < 2,100$), critical transition ($2,100 \le Re \le 4,000$), and fully turbulent ($Re > 4,000$):

$$f_D = 8 \left[ \left(\frac{8}{Re}\right)^{12} + \frac{1}{(A + B)^{1.5}} \right]^{1/12}$$

Where auxiliary variables $A$ and $B$ are defined as:

$$A = \left[ -2.457 \ln\left(\left(\frac{7}{Re}\right)^{0.9} + 0.27 \frac{\varepsilon}{D} \right) \right]^{16}$$ $$B = \left(\frac{37,530}{Re} \right)^{16}$$

The Churchill correlation is the industry benchmark for transient dynamic simulation software (e.g., Aspen HYSYS Dynamics, OLGA) because it avoids discontinuous derivative crashes across regime boundaries.

4. Pipe Roughness Reference Table (Commercial Materials)

The table below provides standard design values for absolute surface roughness ($\varepsilon$) per Crane Technical Paper No. 410, Cameron Hydraulic Data, and ISO 5167.

5. Step-by-Step Industrial Worked Example

Design Problem: A refinery transfer pump circulates crude oil ($\rho = 850 \text{ kg/m}^3$, dynamic viscosity $\mu = 4.5 \text{ cP} = 4.5 \times 10^{-3} \text{ Pa}\cdot \text{s}$) through a DN150 (6-inch) Schedule 40 carbon steel pipeline ($D = 154.1 \text{ mm} = 0.1541 \text{ m}$) over a straight length of $L = 500 \text{ m}$. The volumetric flow rate is $Q = 120 \text{ m}^3/ \text{h}$. Absolute roughness is $\varepsilon = 0.0457 \text{ mm}$.

Step 1: Calculate Velocity & Reynolds Number

$$A = \frac{\pi}{4} D^2 = \frac{\pi}{4} (0.1541 \text{ m})^2 = 0.01865 \text{ m}^2$$ $$v = \frac{Q}{A} = \frac{120 / 3600 \text{ m}^3/ \text{s}}{0.01865 \text{ m}^2} = 1.787 \text{ m/s}$$ $$Re = \frac{\rho v D}{\mu} = \frac{(850 \text{ kg/m}^3)(1.787 \text{ m/s})(0.1541 \text{ m})}{0.0045 \text{ Pa}\cdot \text{s}} = \frac{234.07}{0.0045} = 52,016$$

Regime Check: $Re = 52,016 > 4,000$. Flow is fully turbulent.

Step 2: Determine Relative Roughness

$$\frac{\varepsilon}{D} = \frac{0.0457 \text{ mm}}{154.1 \text{ mm}} = 0.0002965$$

Step 3: Solve Friction Factor via Swamee-Jain

$$f_{SJ} = \frac{0.25}{\left[ \log_{10}\left(\frac{0.0002965}{3.7} + \frac{5.74}{(52,016)^{0.9}} \right) \right]^2} = \frac{0.25}{\left[ \log_{10}(0.00008015 + 0.0003264) \right]^2}$$ $$f_{SJ} = \frac{0.25}{[-3.3908]^2} = \frac{0.25}{11.4975} = 0.02174$$

Step 4: Verify via Iterative Colebrook-White

Starting with $x_0 = 1/\sqrt{0.02174} = 6.782$:

• Iteration 1: $x_1 = 6.7844 \implies f_D = 1 / (6.7844)^2 = 0.02173$

The Swamee-Jain value ($0.02174$) matches the exact Colebrook-White value ($0.02173$) within 0.04%.

Step 5: Calculate Hydraulic Pressure Drop & Pump Head Required

$$\Delta P_f = f_D \cdot \frac{L}{D} \cdot \frac{\rho v^2}{2} = (0.02173) \left(\frac{500 \text{ m}}{0.1541 \text{ m}}\right) \left(\frac{(850)(1.787)^2}{2}\right)$$ $$\Delta P_f = (70.51) \cdot (1357.2) = 95,696 \text{ Pa} = 0.957 \text{ bar} = 95.7 \text{ kPa}$$ $$h_f = \frac{\Delta P_f}{\rho g} = \frac{95,696}{(850)(9.80665)} = 11.48 \text{ m of crude oil}$$

6. Common Industrial Pitfalls & Engineering Best Practices