Turbulent Flow

Key Takeaways

• The Fanning friction factor (f) in turbulent pipe flow is significantly higher than in laminar flow and varies with Reynolds number and pipe roughness according to the Colebrook-White equation (for smooth pipes: f = 0.046 * Re^(-0.2) for 3x10^4 less than Re less than 10^6), meaning that the frictional pressure drop per unit length in turbulent flow scales approximately as v^1.75 rather than the v^1.0 relationship of Hagen-Poiseuille laminar flow; this nonlinear scaling has profound implications for drilling hydraulics because doubling the pump rate in turbulent flow (which is common in the turbulent annular flow regime needed for hole cleaning) requires approximately 3.4 times the pressure drop rather than only twice the pressure, increasing equivalent circulating density (ECD) at depth and risking formation fracture if the turbulent cleaning is applied in a well with a narrow drilling window between pore pressure and fracture pressure; the Metzner-Reed Reynolds number extension for non-Newtonian fluids (which replaces Newtonian viscosity with an effective viscosity from the power-law or Bingham plastic model) allows the transition criterion to be applied to drilling muds, yielding a critical pump rate for the laminar-turbulent transition that is higher for higher-viscosity fluids, so that viscosity manipulation can be used to maintain laminar flow in the annulus when ECD margins are tight.
• Hole cleaning efficiency in deviated and horizontal wells depends critically on the flow regime in the annulus: in near-vertical wells (less than 30 degrees from vertical), cuttings settle gravitationally toward the borehole axis and can be transported upward by laminar or turbulent flow as long as the annular velocity exceeds the terminal settling velocity of the cuttings (typically 0.1 to 0.5 m/s for medium sand-sized cuttings in water-based mud); in highly deviated wells (50 to 90 degrees from vertical), cuttings settle to the low side of the annulus and form a stationary or slowly migrating cuttings bed that turbulent flow must mechanically erode and re-suspend; the critical annular velocity for turbulent erosion of a cuttings bed in a horizontal well depends on the cuttings size, density, and bed packing, the mud density and viscosity, and the borehole geometry, with empirical correlations (Ford-Peden-Osgouei, Larsen-Pilehvari-Azar) suggesting that turbulent annular velocities of 1.2 to 2.0 m/s are required for effective cuttings transport in horizontal wells drilled in medium-density mud; when pump rate is insufficient to maintain these velocities (which is common in extended-reach wells with long horizontal sections where pipe friction limits the available pressure for annular flow), mechanical agitation methods (drill string rotation, reciprocation, or wiper trips) are used to mechanically re-suspend the cuttings bed so that even a low-velocity turbulent flow can carry the re-suspended cuttings to surface.
• Turbulent flow in perforated completions and near-wellbore reservoir flow (non-Darcy flow) occurs when the pressure gradient near the wellbore or in the perforations is large enough that the inertial forces on the flowing fluid exceed the viscous forces, invalidating the linear Darcy's law relationship (q proportional to delta-P) and creating a rate-dependent skin (also called turbulent skin, Darcy-Forchheimer inertial effect, or non-Darcy flow effect) that adds an additional pressure drop proportional to the square of the flow rate; the Forchheimer equation (delta-P = alpha*mu*q + beta*rho*q^2, where alpha is the Darcy term and beta is the inertial (turbulence) coefficient) describes the combined viscous and inertial pressure drop in high-rate wells, with the turbulent component typically becoming significant when the Reynolds number in the perforation tunnels or near-wellbore rock exceeds approximately 1 to 10 (at pore-scale); gas wells are far more susceptible to turbulent near-wellbore pressure drop than oil wells because gas has much lower viscosity (lower alpha contribution) and is produced at much higher linear velocity through perforations at the same volumetric rate, with turbulent skin factors of 0.5 to 5 or more contributing equivalent pressure drops to skin factors of 5 to 50 in high-rate gas wells.
• Turbulent flow in surface facilities (pipelines, separators, process vessels, and heat exchangers) causes higher than expected erosion rates at bends, elbows, tees, and constrictions where turbulent velocity fluctuations create locally high wall shear stresses that abrade metal surfaces at a rate that scales with the velocity raised to the power 2 to 3 (the erosion model exponent), making flow velocity the most important single parameter for managing erosion in produced fluid systems; the API RP 14E recommended erosional velocity limit (Ve = C/sqrt(rho), where C is a constant of 100 for continuous service in clean fluids and 125 to 150 for intermittent service) provides a simplified erosion-prevention criterion, though more detailed models (DNV RP O501, NORSOK P-003) account for particle size, particle hardness, pipe material, and bend geometry; in gas wells producing small amounts of sand, even moderate turbulent velocities (5 to 15 m/s) can cause significant erosion of choke beams, pipe elbows, and separator inlet nozzles when the turbulent eddies repeatedly impact sand particles against the metal wall, making sand monitoring (sand probes, acoustic sand detectors, sample pot analysis) and velocity management part of the routine operation of any sand-producing well.
• Turbulent flow detection in wellbore interpretation is performed using production logging tools (PLT), particularly the spinner flowmeter (which measures the rotational speed of an impeller in the flowing fluid, calibrated to flow velocity), the fullbore spinner, and the gradiomanometer (which measures fluid density to identify flow composition changes); in gas wells, turbulent flow in the production tubing can cause the spinner to over-read relative to the true volumetric flow rate because the turbulent velocity fluctuations create a non-parabolic velocity profile that the spinner integrates differently than the calibration profile, requiring multi-pass PLT data and appropriate turbulence corrections in the flow model to interpret the true flow contribution of each producing zone; two-phase turbulent flow identification using the combination of spinner velocity, holdup (water holdup tool, capacitance probe), and gradiomanometer density helps diagnose flow assurance problems including annular mist flow in high-rate gas wells, slug flow in gas-lifted wells, and dispersed bubble flow in high-water-cut oil producers, each of which produces characteristic PLT signatures that guide artificial lift optimization and zonal allocation of production.