Elsevier

Chemical Engineering Science

Volume 72, 16 April 2012, Pages 142-154
Chemical Engineering Science

Shear layers in the turbulent pipe flow of drag reducing polymer solutions

https://doi.org/10.1016/j.ces.2011.12.044Get rights and content

Abstract

A range of high molecular weight polymers (polyethylene oxide) was dissolved at very low concentrations – in the order of few wppm – in a solvent (water). The Newtonian character of the polymer solutions was confirmed by rheological measurements. The polymer solutions were then pumped through a long horizontal pipe section in fully developed turbulent conditions. The flow experienced a reduction in frictional drag when compared to the drag experienced by the equivalent flow of the pure solvent. Specifically, drag reduction was measured at Reynolds numbers ranging from 3.5×104 to 2.1×105 in a pressure driven flow facility with a circular tube section of internal diameter 25.3 mm. The turbulent flow was visualized by Particle Image Velocimetry and the resulting data were used to investigate the effect of the drag reducing additives on the turbulent pipe flow. Close attention was paid to the mean and instantaneous velocity fields, as well as the two-dimensional vorticity and streamwise shear strain rate. The results indicate that drag reduction is accompanied by the appearance of “shear layers” (i.e. thin filament-like regions of high spatial velocity gradients) that act as interfaces separating low-momentum flow regions near the pipe wall and high-momentum flow regions closer to the centerline. The shear layers are not stationary. They are continuously formed close to the wall at a random frequency and move towards the pipe centerline until they eventually disappear, thus occupying or existing within a “shear layer region”. It is found that the mean thickness of the shear layer region is correlated with the measured level of drag reduction. The shear layer region thickness is increased by the presence of polymer additives when compared to the pure solvent, in a similar way to the thickening of the buffer layer. The results provide valuable insights into the characteristics of the turbulent pipe flow of a solvent containing drag reducing polymers that can be used to further our understanding of the role of polymers on the mechanism of drag reduction and to develop advanced drag reduction models.

Highlights

► Drag reduction of polymer solutions was measured in a turbulent pipe flow. ► PIV was used to assess the effect of the additives on the instantaneous flow. ► Flows with additives underwent a separation into low- and high-momentum regions. ► The two regions were separated by a thin layer of intense vorticity and strain rate. ► The thickness of these regions correlates with the measured level of drag reduction.

Introduction

It is well known that the frictional resistance caused by turbulent flow of a Newtonian fluid (solvent) in a pipe can be considerably reduced by the addition of a polymer to the solvent (Toms, 1948). An important consequence of this phenomenon, known as drag reduction, is that a fluid (liquid) solution containing the polymer additive will exhibit a lower pressure drop in a pipe flow compared to the pure solvent at the same flow-rate. The phenomenon of drag reduction occurs exclusively in turbulent flow and is of great industrial importance. Specifically, it is relevant in a broad array of applications that involve liquid transport through pipelines (Burger et al., 1980), ranging from fire fighting to field irrigation (Singh et al., 1995) and flows in urban sewage networks (Sellin and Ollis, 1980), hydraulic fracturing (Lucas et al., 2009, Morgan and McCormick, 1990), oil pipeline systems and secondary oil well operations.

Polymers have been identified as the most efficient drag reducers with respect to other drag reduction additives such as surfactants or bubbles (Bismarck et al., 2005). The amount of polymer additive required to alter the turbulent flow structure is in the order of few parts per million by weight (wppm). Drag reduction was for example observed for polymer concentrations as low as 0.02 wppm (20 wppb) by Oliver and Bakhtiyarov (1983).

Virk et al. (1967) were the first to propose an important universal asymptote of maximum drag reduction, sometimes called Virk's asymptote, which is independent of experimental set-up or polymer additive. In a follow-up paper Virk (1975) established common characteristics of velocity profiles associated with the turbulent flow of polymeric solutions. An increasing presence of polymer additives was found to be associated with a thickening of the buffer layer and a shift of the log-law region away from the Newtonian law of the wall, or viscous sublayer. At maximum drag reduction the buffer layer was found to extend to the centerline and the log-law region disappeared.

Drag reduction has been studied extensively since its discovery (Brostow, 2008, Lumley and Blossey, 1998, Virk, 1975, White and Mungal, 2008), but direct phenomenological insight into the effect of polymer additives on the turbulent flow intensity and structure was lacking until recently when non-intrusive flow visualization techniques were used (den Toonder et al., 1997, Ptasinski et al., 2001). The advancement and employment of planar flow measurement techniques, such as Particle Image Velocimetry (PIV), have been a key development that has provided valuable insight into the instantaneous turbulent flow structure (Willert and Gharib, 1991). Warholic et al. (2001) used the PIV technique to identify turbulent structures close to the wall that the authors designated as being typical of flows involving Newtonian solvents. These structures were characterized by the ejection of low-momentum fluid to the outer velocity region, or defect region, and by quasi-streamwise vortices. Such structures were recognized as locations of large Reynolds stresses. For high measured drag reduction the authors observed reduction or elimination of the ejections from the wall (Warholic et al., 2001). Liberatore et al. (2004) observed that the presence of polymers lead to a decrease in the frequency and the intensity of large-scale ejections when compared to a Newtonian solvent. The polymer induced drag reduction effect was also found to reduce the small-scale fluctuations, as determined from spectral functions (Warholic et al., 2001), and to reduce the magnitude and frequency of the small-scale eddies (Liberatore et al., 2004). Additionally, large regions of almost unidirectional fluctuating velocity vectors were observed for solutions exhibiting high levels of drag reduction (Liberatore et al., 2004, Warholic et al., 2001).

The aim of this work is to study the effect of drag reducing polymers on the instantaneous structure of turbulent pipe flow and to extend our knowledge to practically relevant conditions with measurements at Reynolds numbers up to Re=210000. Three different molecular lengths, and thus three different corresponding weights, of the same linear drag reducing polymer (polyethylene oxide, or PEO) have been studied. The Newtonian character of the polymer solutions was first verified by rheological measurements. A pressure-driven horizontal pipe flow apparatus was then used to generate the turbulent flow, and to measure pressure drops over a range of flow-rates. The turbulent flow was characterized with the use of a PIV system.

In the sections below, firstly, the rheological and drag reduction measurements are reported for the various polymers, at different concentrations, over the investigated range of Reynolds numbers. We then present corresponding profiles of the mean flow velocity and of the root mean square (rms) of its fluctuations and, based on the former, measures of the thickness of the buffer layers. This is followed by a presentation of instantaneous images of the turbulent flow with and without the polymer additives. The main contribution of the current study is the uncovering of the presence of two distinct momentum regions of turbulent pipe flow for solutions containing polymer additives, separated by a thin interface layer that is associated with high strain (and shear). It will be shown that these flow structures (i.e. thin layers) propagate far from the wall towards the center of the pipe flow, within an overall region that is larger than that which is investigated typically in turbulent drag reduction studies (e.g. by numerical simulations due to the computational cost). Finally, the spanwise extent of the regions within which these layers are found is correlated to the level of independently measured drag reduction, and also compared to the extent of the velocity buffer layer.

Section snippets

Drag reducing polymers

Three different molecular lengths and thus also weights of polyethylene oxide (PEO) were chosen for this investigation. PEO is a linear non-ionic water soluble polymer. All three PEOs were purchased from Sigma-Aldrich, Inc. (Steinhelm, Germany). The molecular weights as given by the manufacturer are 2×106, 4×106 and 8×106kgmol1 and the corresponding abbreviations used in the following text are PEO2, PEO4 and PEO8, respectively. Tap water was used throughout as the solvent.

During the

Rheology

A rheological characterization was performed in order to assess the Newtonian character of the polymer solutions. All solutions show a linear, flat (zero-gradient) dependency of apparent shear viscosity on strain rate, which is an indicator of Newtonian behavior. As an indicative result, the apparent shear viscosity μ as a function of strain rate γ˙ for solutions with different concentrations c of PEO8 from 10 to 250 wppm are shown in Fig. 4(a). Fig. 4(b) shows the dependence of the mean

Conclusions

The drag reduction efficiency of polyethylene oxide was measured in a turbulent pipe flow over 122 conditions with varying polymer molecular weight, concentration and Reynolds numbers up to 210 000. Different levels of drag reduction were observed with a maximum of 72%. Particle Image Velocimetry was used to assess the effect of the polymer additives on the instantaneous turbulent pipe flow. The presence of the drag reducing polymers was associated with a thickening of the buffer layer and a

Acknowledgments

The authors gratefully acknowledge the financial support of Halliburton Energy Services.

References (34)

  • J.M.J. den Toonder et al.

    Drag reduction by polymer additives in a turbulent pipe flow: numerical and laboratory experiments

    J. Fluid Mech.

    (1997)
  • F. Durst et al.

    LDA measurements in the near-wall region of a turbulent pipe flow

    J. Fluid Mech.

    (1995)
  • B.R. Elbing et al.

    Degradation of homogeneous polymer solutions in high shear turbulent pipe flow

    Exp. Fluids

    (2009)
  • Hassan, Y., 1980. Experimental and Modeling Studies of Two-Point Stochastic Structure in Turbulent Pipe Flow. Ph.D....
  • K.C. Kim et al.

    Very large-scale motion in the outer layer

    Phys. Fluids

    (1999)
  • K. Kim et al.

    Wall versus centerline polymer injection in turbulent channel flows

    Flow Turbul. Combust.

    (2007)
  • K. Kim et al.

    Effect of macromolecular polymer structures on drag reduction in a turbulent channel flow

    Phys. Fluids

    (2004)
  • Cited by (31)

    • Enhance flows of waxy crude oil in offshore petroleum pipeline: A review

      2022, Journal of Petroleum Science and Engineering
      Citation Excerpt :

      The first type uses a polymer as the DRA. The main factors of concentration, pipe size and geometry, molecular weight, chain flexibility, and flow rate (Al-Sarkhi, 2010; Cao et al., 2018; Yousif, 2018) all work to reduce turbulence in the crude oil (Mucharam et al., 2017; Zadrazil et al., 2012). In the second type, adjacent fibers interact and entangle in the suspension and form three-dimensional structures or networks at moderate concentrations to change the transfer characteristics of the suspension and reduce flow resistance (Marmy, 2012; Salehudin and Ridha, 2016).

    • DRAG REDUCTION FOR SINGLE-PHASE WATER FLOW IN AND AROUND 180<sup>O</sup> BENDS

      2021, Journal of Non-Newtonian Fluid Mechanics
      Citation Excerpt :

      The curves show reduced pressure drops in all the sections of the loop considered. The curve for fully developed flow before the bend agrees with that of Zadrazil and co. [54]. Comparison of P-K plots for flow without DRP and flow with 10-ppm HPAM at different locations are shown in Figure 7.

    • Separated oil-water flows with drag reducing polymers

      2019, Experimental Thermal and Fluid Science
      Citation Excerpt :

      Particle image velocimetry (PIV) in particular offers whole field, instantaneous velocity measurements and has been used extensively in single and in some cases in two phase flows [20,40,15,16,5,18,59,67]. These techniques have also been applied in some cases to the investigation of velocity and turbulence profiles in flows with drag reducing polymers [29,62,31,52]. Wei and Willmarth [58] used LDV to study velocity fields in channel flows with PEO added in water.

    View all citing articles on Scopus
    View full text