Darcy–Weisbach Friction Factor

Updated 21 April 2026

Darcy–Weisbach friction factor is a dimensionless parameter that relates head loss to pipe length, diameter, flow velocity, and wall friction.
It exhibits regime-specific behaviors, with explicit expressions for laminar (64/Re), transitional, and turbulent flows adjusted for surface roughness.
Modern methods, including data-driven and continuous crossover formulations, enhance its predictive accuracy and computational efficiency.

The Darcy–Weisbach friction factor, denoted $f$ , is a central dimensionless parameter quantifying energy loss due to wall shear in internal and open-channel flows, linking the pressure drop or head loss to the mean velocity, geometry, and, critically, fluid–wall interaction. It forms the basis for the Darcy–Weisbach equation, which is universally adopted in both analytical models and simulation frameworks across hydraulic, environmental, and industrial fluid mechanics.

1. Fundamental Definitions and Canonical Formulations

The classical Darcy–Weisbach equation connects the head loss per unit length or total pressure drop in a conduit to the mean flow parameters via

$h_f = f\,\frac{L}{D}\,\frac{V^2}{2g}, \quad \text{or equivalently,} \quad \Delta p = f\,\frac{L}{D}\,\frac{\rho V^2}{2}$

where $f$ is the Darcy–Weisbach friction factor, $L$ is the pipe length, $D$ diameter, $V$ mean velocity, and $g$ gravitational acceleration. For general geometries and flow regimes, modifications appear (wetted perimeter, hydraulic radius, effective velocities), but the core definition remains.

The friction factor $f$ encapsulates all wall-shear effects and depends on Reynolds number $Re$ , relative roughness $\epsilon/D$ , and, for non-Newtonian fluids, the rheological response. In laminar (fully developed) flow of Newtonian fluids in a circular pipe, $h_f = f\,\frac{L}{D}\,\frac{V^2}{2g}, \quad \text{or equivalently,} \quad \Delta p = f\,\frac{L}{D}\,\frac{\rho V^2}{2}$ 0 adopts the exact form $h_f = f\,\frac{L}{D}\,\frac{V^2}{2g}, \quad \text{or equivalently,} \quad \Delta p = f\,\frac{L}{D}\,\frac{\rho V^2}{2}$ 1, where $h_f = f\,\frac{L}{D}\,\frac{V^2}{2g}, \quad \text{or equivalently,} \quad \Delta p = f\,\frac{L}{D}\,\frac{\rho V^2}{2}$ 2, and for turbulent regimes or open channels, it is determined empirically or via semi-theoretical correlations (Basse, 2016, Cerbus et al., 2021, Kirstetter et al., 2016).

2. Regime-Specific Correlations: Laminar, Transitional, and Turbulent Laws

Explicit expressions and regime boundaries for $h_f = f\,\frac{L}{D}\,\frac{V^2}{2g}, \quad \text{or equivalently,} \quad \Delta p = f\,\frac{L}{D}\,\frac{\rho V^2}{2}$ 3 depend on flow state, wall roughness, and geometry. For Newtonian fluids in pipes:

Laminar (Re $h_f = f\,\frac{L}{D}\,\frac{V^2}{2g}, \quad \text{or equivalently,} \quad \Delta p = f\,\frac{L}{D}\,\frac{\rho V^2}{2}$ 4): $h_f = f\,\frac{L}{D}\,\frac{V^2}{2g}, \quad \text{or equivalently,} \quad \Delta p = f\,\frac{L}{D}\,\frac{\rho V^2}{2}$ 5 (Hagen–Poiseuille law). This holds for both smooth and rough pipes if $h_f = f\,\frac{L}{D}\,\frac{V^2}{2g}, \quad \text{or equivalently,} \quad \Delta p = f\,\frac{L}{D}\,\frac{\rho V^2}{2}$ 6 remains below the transition, with roughness effects negligible due to the dominance of viscous sublayers (Senturk et al., 2019, Malik et al., 2019).
Turbulent (Re $h_f = f\,\frac{L}{D}\,\frac{V^2}{2g}, \quad \text{or equivalently,} \quad \Delta p = f\,\frac{L}{D}\,\frac{\rho V^2}{2}$ 7), smooth/rough:
- Blasius law for hydraulically smooth pipes: $h_f = f\,\frac{L}{D}\,\frac{V^2}{2g}, \quad \text{or equivalently,} \quad \Delta p = f\,\frac{L}{D}\,\frac{\rho V^2}{2}$ 8 ( $h_f = f\,\frac{L}{D}\,\frac{V^2}{2g}, \quad \text{or equivalently,} \quad \Delta p = f\,\frac{L}{D}\,\frac{\rho V^2}{2}$ 9) (Basse, 2016).
- General turbulent regime: Implicit relations, e.g., Colebrook–White,
$f$ 0

or explicit forms such as the Haaland approximation (Cerbus et al., 2021). Roughness-limited flows ( $f$ 1) asymptote to $f$ 2 (“Strickler regime”).
Transitional regime ( $f$ 3): Laboratory measurements reveal that $f$ 4 does not interpolate linearly between laminar and turbulent values, but follows reproducible, protocol-dependent curves depending on driving conditions (see Table 1 below) (Cerbus et al., 2021).

$f$ 5	$f$ 6 (gravity-driven)	$f$ 7 (mass-flux)
1600	0.0530	0.0660
2000	0.0380	0.0510
2400	0.0310	0.0440
2700	0.0280	—

The appropriate regime and correlation must be selected based on the specific $f$ 8, relative roughness, and flow control protocol (Cerbus et al., 2021).

3. Dependency on Surface Roughness and Extension to Arbitrary Geometries

Wall roughness effects are incorporated via $f$ 9 or equivalent sand-grain roughness $L$ 0, modifying $L$ 1 and shifting regime boundaries:

In turbulent pipe flow, measured friction factors for rough pipes deviate systematically from smooth curves as $L$ 2 increases, saturating at a roughness-controlled asymptote (Basse, 2016, Ünal et al., 19 Feb 2026).
In laminar regimes, the scaling $L$ 3 typically persists even with periodic/bar roughness, but the prefactor increases quadratically with roughness height and decreases linearly with pitch (Senturk et al., 2019):

$L$ 4

Valid for pipes with periodic square-bar roughness, 200 $L$ 5 Re $L$ 6 2000.

For open-channel and shallow flows (e.g., rainfall-driven overland flow), analogs of the Darcy–Weisbach law are derived, with the friction factor adapted to local Reynolds numbers and flow depth (Kirstetter et al., 2016).

4. Extensions: Non-Newtonian Fluids and Multiphase Systems

The Darcy–Weisbach friction factor is explicitly generalized for non-Newtonian, non-single phase, and viscoelastic flows:

Non-Newtonian turbulent pipe flow is governed by a generalized Reynolds number ( $L$ 7) and wall-consistency index, yielding an implicit log-law (Trinh, 2010):

$L$ 8

which must be solved iteratively for $L$ 9 for given rheological parameters.
Laminar viscoelastic fluids (e.g., FENE-P model): The friction factor is strictly bounded by the values for pure solvent and total zero-shear viscosity

$D$ 0

with the reduction below $D$ 1 due to strain-rate redistribution (shear-thinning) (Malik et al., 2019).
Multiphase/aerated flows: In high-Froude, air–water mixtures, the effective friction factor is decomposed into uniform, spatial, and temporal (evolving) contributions,

$D$ 2

incorporating Favre (density-weighted) averaging and vertical structure corrections for momentum and pressure terms (Kramer, 9 Jan 2026).

5. Modern Computational and Symbolic Approaches to Friction-Factor Correlations

The need for accurate, explicit, and robust friction factor expressions for both routine engineering and simulation code integration has driven recent developments:

Continuous crossover formulations (e.g., Churchill equation) are implemented in modern open-source solvers like openKARST, blending laminar and turbulent regimes seamlessly and removing the need for iterative solution of implicit equations such as Colebrook–White (Kordilla et al., 28 Mar 2025).
Data-based symbolic regression: Physics-informed symbolic regression, constrained by order-of-magnitude momentum and turbulence scaling, yields explicit expressions matching reference data (Nikuradse, Superpipe) to $D$ 31% RMS error across $D$ 4 and all roughness levels. These models encode both viscous and turbulent mechanisms, recover the characteristic inflection (“minimum”) in $D$ 5 for rough pipes, and impose physically justified sensitivity exponents for velocity, roughness, viscosity, and density (Ünal et al., 19 Feb 2026).

Example explicit expression (“Candidate 1”): $D$ 6 where each $D$ 7 is an explicit function of $D$ 8 and $D$ 9, designed to recover the appropriate asymptotic scaling in all regimes (Ünal et al., 19 Feb 2026).

6. Sensitivity, Physical Constraints, and Influence of Protocol

Quantitative sensitivity envelopes—velocity, roughness, viscosity, and density exponents—serve as physical priors for correlation development and validation, ensuring that $V$ 0 behaves consistently with underlying momentum and energy balances (Ünal et al., 19 Feb 2026). In transitional flows, laboratory evidence has established that friction factor curves depend on protocol (constant pressure vs. constant mass flux), initial conditions, and flow history, with transitional $V$ 1 curves occupying a distinct region in $V$ 2– $V$ 3 space not interpolated by classical laminar–turbulent laws (Cerbus et al., 2021).

Moreover, in flows with geometric or physical complexity (e.g., helical pipes), rigorous upper bounds on $V$ 4 can be constructed using variational “background” methods, which account for curvature and torsion effects but may be quantitatively loose relative to direct numerical/experimental estimates at high $V$ 5 (Kumar, 2020).

7. Practical Applications, Implementation, and Limitations

The Darcy–Weisbach friction factor underpins pressure-loss, head-loss, and flow-capacity computations across hydraulics, pipeline engineering, subsurface fracture propagation, open-channel design, and even transient multi-regime network solvers. Successive generations of correlations—empirical, semi-theoretical, and data-driven—have systematically expanded $V$ 6’s applicability, precision, and computational convenience. However, accurate prediction still depends on correct identification of the flow regime, careful characterization of wall properties (roughness, geometry), protocol adaptation (especially in the transitional regime), and suitable consideration of fluid rheology and multiphase effects.

Current best practice is to employ explicit, physics-consistent friction-factor formulas validated on the target flow regime and geometry, and to employ modern continuous or data-driven models where legacy iterative formulas (e.g., Colebrook–White) introduce numerical instability or excess computational cost (Kordilla et al., 28 Mar 2025, Ünal et al., 19 Feb 2026).

References:

(Kirstetter et al., 2016, Basse, 2016, Cerbus et al., 2021, Senturk et al., 2019, Malik et al., 2019, Trinh, 2010, Kramer, 9 Jan 2026, Kordilla et al., 28 Mar 2025, Ünal et al., 19 Feb 2026, Kumar, 2020, Zolfaghari et al., 2016)