Hostname: page-component-78c5997874-j824f Total loading time: 0 Render date: 2024-11-17T18:11:50.392Z Has data issue: false hasContentIssue false

New formulations for the mean wall-normal velocity and Reynolds shear stress in a turbulent boundary layer under zero pressure gradient

Published online by Cambridge University Press:  10 August 2023

Tie Wei*
Affiliation:
Department of Mechanical Engineering, New Mexico Tech, Socorro, NM 87801, USA
Zhaorui Li
Affiliation:
Department Of Engineering, Texas A&M University-Corpus Christi, 6300 Ocean Drive, Corpus Christi, TX 78412, USA
Yanxing Wang
Affiliation:
Department of Mechanical Engineering, New Mexico State University, 1780 E University Ave, Las Cruces, NM 88003, USA
*
Email address for correspondence: [email protected]

Abstract

In this study, a novel analytical solution is derived rigorously for the mean wall-normal velocity in a turbulent boundary layer (TBL) under zero pressure gradient (ZPG), without relying on any a priori assumptions regarding the mean streamwise velocity. By neglecting the higher-order terms in the exact solution, an approximate formulation for the mean wall-normal velocity is obtained. The accuracy of this approximation is then validated through a comparison with data from direct numerical simulations (DNS). Drawing upon the distinct behaviours of force balance in the inner and outer regions of the ZPG TBL, simplified approximate mean momentum equations are derived by decomposing the Reynolds shear stress into inner and outer parts. The inner and outer Reynolds shear stresses are then obtained through integration of the corresponding approximate mean momentum equations. The outer Reynolds shear stress is revealed to exhibit a strong connection to the mean wall-normal velocity. Specifically, it is observed that $R_{uv-{out}}/u^2_\tau \approx 1 - UV/(U_e V_e)$, where $u_\tau$ represents the friction velocity, and $U_e$ and $V_e$ denote the mean streamwise and wall-normal velocities at the boundary layer edge, respectively. Finally, an approximate formulation for the Reynolds shear stress across the entire TBL is developed by synthesizing the inner and outer parts. Extensive validation against both DNS and experimental data, spanning a wide range of Reynolds numbers, demonstrates the excellent agreement achieved.

Type
JFM Papers
Copyright
© The Author(s), 2023. Published by Cambridge University Press

1. Introduction

Reynolds shear stress has been a key focus in the study of turbulent flows, and various turbulence models have been proposed to effectively describe its behaviour (Monin & Yaglom Reference Monin and Yaglom1965; Tennekes & Lumley Reference Tennekes and Lumley1972; Schlichting Reference Schlichting1979; Pope Reference Pope2000). This is because predicting accurately the behaviour of the Reynolds shear stress is essential for accurate modelling of turbulent flows. In contrast, the mean wall-normal velocity has received relatively less attention in previous research, and our understanding of its behaviour remains limited. In the present study, we derive an analytical equation for the mean wall-normal velocity profile, and reveal a close relationship between the Reynolds shear stress and the mean wall-normal velocity. The new results provide valuable insights for the development of new turbulence models and the understanding of turbulent boundary layer flows.

The scaling of the mean wall-normal velocity in a zero pressure gradient (ZPG) turbulent boundary layer (TBL) was studied by Wei & Klewicki (Reference Wei and Klewicki2016). Unlike the inner and outer scales required to describe the mean streamwise velocity or Reynolds shear stress, they found that the mean wall-normal velocity across the entire boundary layer can be characterized by a single length scale $\delta _e$ and a single velocity scale $V_e$, where $\delta _e$ is the boundary layer thickness, and $V_e$ is the mean wall-normal velocity at the the boundary layer edge. Wei & Klewicki (Reference Wei and Klewicki2016) demonstrated the importance of the mean wall-normal or transverse flow in understanding TBLs. Furthermore, the mean transverse flow has been recognized as a fundamental aspect in scaling and comprehending free shear turbulent flows, including jets (Wei & Livescu Reference Wei and Livescu2021), wakes (Wei, Livescu & Liu Reference Wei, Livescu and Liu2022b) and mixing layers (Wei, Li & Livescu Reference Wei, Li and Livescu2022a).

Since the publication of Wei & Klewicki (Reference Wei and Klewicki2016), there has been a growing interest in understanding the role of mean wall-normal flow in TBLs. The relationship between the mean wall-normal velocity and Reynolds shear stress was investigated by Kumar & Dey (Reference Kumar and Dey2019). By neglecting the viscous effect in the outer region of a ZPG TBL, Kumar & Dey (Reference Kumar and Dey2019) derived a linear relation between the Reynolds shear stress and the mean wall-normal velocity. This finding further confirms the importance of the mean wall-normal flow in TBLs and its impact on the overall behaviour of the flow. In a subsequent study, Kumar & Mahesh (Reference Kumar and Mahesh2022) expanded their analysis to include the viscous effect. They derived an approximate equation for the sum of viscous shear stress and Reynolds shear stress in a ZPG TBL, providing a more comprehensive understanding of the total shear stress in the boundary layer. While the significance of the mean wall-normal velocity in predicting Reynolds shear stress has been recognized in these studies, neither of them derived explicitly a rigorous equation for the mean wall-normal velocity. Instead, they used a less accurate approximation for mean wall-normal velocity when solving the mean momentum equation, which may lead to less accurate predictions of Reynolds shear stress.

The aim of this work is to derive new analytical equations for the mean wall-normal velocity and Reynolds shear stress in a ZPG TBL, and establish the close relationship between these two quantities. In § 2, an analytical equation for the mean wall-normal velocity is derived from the integral of the mean continuity equation, and its accuracy is verified through comparison with direct numerical simulations (DNS) data. The Reynolds shear stress is decomposed into two parts: an inner part and an outer part. Approximate equations for both parts are derived from the integral of the approximate mean momentum equation. The interdependence of the Reynolds shear stress, mean streamwise velocity and mean wall-normal velocity is also analysed and explained in this section. Finally, the results of the study are summarized in § 3.

2. Governing equation for the mean flow

The governing equations for a statistically steady two-dimensional TBL over a flat plate are (see Townsend (Reference Townsend1956) or Tennekes & Lumley (Reference Tennekes and Lumley1972), for example)

(2.1a)$$\begin{gather} 0 = \frac{\partial U}{\partial x} + \frac{\partial {V}}{\partial y}, \end{gather}$$
(2.1b)$$\begin{gather}0 ={-}U\,\frac{\partial U}{\partial x} - V\, \frac{\partial U}{\partial y}+ \frac{\partial R_{uv}}{\partial y} + \nu\,\frac{\partial^2 U }{\partial y^2} + \frac{\partial (R_{uu}-R_{vv})}{\partial x}. \end{gather}$$

Here, uppercase letters $U$ and $V$ denote the mean velocity component in the streamwise $x$-direction and wall-normal $y$-direction, respectively, while lowercase letters $u$ and $v$ denote the corresponding velocity fluctuations. The fluid kinematic viscosity is denoted by $\nu$. The kinematic Reynolds shear stress is denoted as $R_{uv}=-\langle uv \rangle$, and the kinematic Reynolds normal stresses in the streamwise and wall-normal directions are denoted as $R_{uu}=-\langle uu \rangle$ and $R_{vv}=-\langle vv \rangle$, respectively, with the angle brackets denoting the Reynolds averaging operator. The term $\partial (R_{uu}-R_{vv})/\partial x$ is neglected in the analysis as it is assumed to have a negligible impact on the balance of the mean momentum equation in the ZPG TBL (Townsend Reference Townsend1956). The corresponding boundary conditions at the wall and the boundary layer edge are provided in table 1.

Table 1. Boundary conditions.

2.1. Analytical and approximate equations for the mean wall-normal velocity

The mean wall-normal velocity plays a significant role in TBLs, as it characterizes the rate at which fluid is transported perpendicular to the wall. In this study, we derive an analytical equation for the mean wall-normal velocity from the integration of the dimensionless mean continuity equation. Previous research by Wei & Klewicki (Reference Wei and Klewicki2016) has identified the proper length and velocity scales for the mean wall-normal velocity in a ZPG TBL as $\delta _e$ and $V_e$, respectively. The variables in the mean continuity equation are normalized as

(2.2ad)\begin{equation} x^* \mathop{=}\limits^{\rm def} \frac{x}{L}, \quad y^- \mathop{=}\limits^{\rm def} \frac{y}{\delta_e(x)}, \quad U^- \mathop{=}\limits^{\rm def} \frac{U_e - U(x,y)}{U_e}, \quad V^- \mathop{=}\limits^{\rm def} \frac{V(x,y)}{V_e(x)}, \end{equation}

where $L$ is a length scale in the $x$-direction, e.g. the computational domain size in the $x$-direction or the length of test section in a wind tunnel. Note that in a ZPG TBL, ${\rm d}U_e/{{\rm d}\kern 0.06em x}=0$ because $U_e$ is a constant, but $\delta _e$ and $V_e$ vary in the $x$-direction. It is worth highlighting that the dimensional velocity components $U$ and $V$ are functions of $x$ and $y$. In contrast, the normalized velocity components $U^-$ and $V^-$ are dependent on the normalized spatial variables $x^*$ and $y^-$.

Using the normalized variables $x^*$, $y^-$ and $U^-$, the $\partial U/\partial x$ term in the mean continuity equation can be written as

(2.3)\begin{equation} \frac{\partial U}{\partial x} = \frac{\partial (U_e - U_e U^-)}{\partial x} ={-}\frac{U_e}{L}\, \frac{\partial U^-}{\partial x^*} + \frac{U_e}{\delta_e}\,\frac{{\rm d}\delta_e}{{\rm d}\kern 0.06em x}\, y^-\,\frac{\partial U^-}{\partial y^-}. \end{equation}

Note that in the derivation of Wei & Klewicki (Reference Wei and Klewicki2016), an implicit self-similarity assumption is made that omits the $\partial U^-/\partial x^*$ term in (2.3). The importance of the streamwise derivative term has been recognized by Perry, Marusic & Jones (Reference Perry, Marusic and Jones2002), who incorporated it into the second term in their (2.9) (with $\zeta$). Disregarding this term is similar to assuming a constant Coles wake factor ($\varPi$, as defined by Coles Reference Coles1956) for ZPG TBLs. Although this assumption is a reasonable approximation, it is not strictly valid at finite Reynolds numbers.

By substituting the normalized variables in (2.2ad) and (2.3) into (2.1a), the mean continuity equation becomes

(2.4)\begin{equation} 0 ={-}\frac{U_e}{L}\,\frac{\partial U^-}{\partial x^*} + \frac{U_e}{\delta_e}\, \frac{{\rm d}\delta_e}{{\rm d}\kern 0.06em x}\,y^-\,\frac{\partial U^-}{\partial y^-} + \frac{V_e}{\delta_e}\, \frac{\partial V^-}{\partial y^-}. \end{equation}

To obtain a dimensionless continuity equation, we multiply both sides of (2.4) by $\delta _e/V_e$:

(2.5)\begin{equation} 0 ={-}\frac{U_e}{V_e}\,\frac{\delta_e}{L}\, \frac{\partial U^-}{\partial x^*} + \frac{U_e}{V_e}\,\frac{{\rm d}\delta_e}{{\rm d}\kern 0.06em x} \left( \frac{\partial (y^- U^-)}{\partial y^-} - U^- \right) + \frac{\partial V^-}{\partial y^-}. \end{equation}

An analytical equation for the mean wall-normal velocity in a ZPG TBL can be obtained by integrating (2.5) with respect to $y^-$, starting from the wall:

(2.6)\begin{equation} V^- = \frac{U_e}{V_e}\,\frac{{\rm d}\delta_e}{{\rm d}\kern 0.06em x} \left( \int_0^{y^-} U^- {{\rm d} y}^- - y^-U^- \right) + \int_0^{y^-} \frac{U_e}{V_e}\,\frac{\delta_e}{L}\,\frac{\partial U^-}{\partial x^*}\,{{\rm d} y}^-. \end{equation}

The global integral of the mean continuity equation can be obtained by evaluating (2.6) at the boundary layer edge:

(2.7)\begin{equation} 1 = \frac{U_e}{V_e}\,\frac{{\rm d}\delta_e}{{\rm d}\kern 0.06em x} \int_0^{1} U^- {{\rm d} y}^- + \frac{U_e}{V_e}\,\frac{\delta_e}{L} \int_0^{1} \frac{\partial U^-}{\partial x^*}\,{{\rm d} y}^-. \end{equation}

Note that by definition, $\int _0^{1} U^-\,{{\rm d} y}^- =\delta _1/\delta _e$, where $\delta _1$ is the mass displacement thickness (see Schlichting Reference Schlichting1979). Therefore, the global integral equation (2.7) can be rewritten as

(2.8)\begin{equation} 1 = \frac{U_e}{V_e} \left( \frac{{\rm d}\delta_e}{{\rm d}\kern 0.06em x}\,\frac{\delta_1}{\delta_e} + \frac{\delta_e}{L} \int_0^{1} \frac{\partial U^-}{\partial x^*}\,{{\rm d} y}^- \right). \end{equation}

By applying the Leibniz integral rule to $\int _0^1 ({\partial U^-}/{\partial x^*})\,{{\rm d} y}^-$, an exact equation for $V_e$ is obtained from (2.8) as follows:

(2.9)\begin{equation} V_e = U_e\,\frac{\mathrm{d} \delta_1}{\mathrm{d}\kern 0.06em x}. \end{equation}

In many experimental studies, determining accurately $\mathrm {d} \delta _e/\mathrm{d}\kern 0.06em x$ or $\partial U^-/\partial x^*$ is often challenging due to the limited resolution in the streamwise direction. However, reliable and accurate determination of $\delta _e$, especially $\delta _1$, can be obtained. In the case of numerical simulation, two-dimensional mean field data must be available in order to calculate $\mathrm {d} \delta _e/\mathrm{d}\kern 0.06em x$ or $\partial U^-/\partial x^*$. Certain archived direct numerical simulation data include only mean profiles at discrete Reynolds numbers, which do not provide the necessary information to calculate $x$-derivatives. To obtain an approximate equation that can be determined more easily, the exact (2.6) is rearranged as

(2.10)\begin{align} V^- & = \frac{\delta_e}{\delta_1} \int_0^{y^-} U^-\,{{\rm d} y}^-- \frac{\delta_e}{\delta_1}\,y^-U^- \nonumber\\ &\quad + \left\{ \left( \frac{\mathrm{d}\delta_e/\mathrm{d}\kern 0.06em x}{\mathrm{d} \delta_1/\mathrm{d}\kern 0.06em x} - \frac{\delta_e}{\delta_1} \right) \left(\int_0^{y^-} U^-\,{{\rm d} y}^-- y^-U^- \right) + \frac{\delta_e/L}{\mathrm{d} \delta_1/\mathrm{d}\kern 0.06em x} \int_0^{y^-} \frac{\partial U^-}{\partial x^*}\,{{\rm d} y}^- \right\}. \end{align}

It is worth noting that the first two functions on the right-hand side of (2.10) can be determined using only the $U$ profile data at a single $x$-location, which is readily available in experimental or archived numerical simulation data. Figure 1(a) illustrates these two functions along with the DNS-computed $V^-$ values. For clarity, figure 1(a) presents the DNS data from Schlatter & Örlü (Reference Schlatter and Örlü2010) at a single Reynolds number $Re_\tau =974$ or $Re_{\delta _2}=3030$. Here, $Re_\tau$ represents the Reynolds number defined by the friction velocity $u_\tau$ and the boundary layer thickness $\delta _e$ as $Re_\tau = u_\tau \delta _e/\nu$, while $Re_{\delta _2}$ represents the Reynolds number defined by the free-stream velocity $U_e$ and the momentum thickness $\delta _2$ as $Re_{\delta _2} = U_e \delta _2/\nu$. Both of these Reynolds numbers are commonly used when presenting data for ZPG TBLs. Figure 1(a) shows that the function $\int _0^{y^-} U^-\,{{\rm d} y}^-$ approaches a constant value towards the boundary layer edge, but the function $y^- U^-$ approaches zero towards the boundary layer edge.

Figure 1. (a) Shapes of the mean wall-normal velocity $V^-$, $y^- U^-$ and $\int _0^{y^-} U^-\,{{\rm d} y}^-$. The data are from the DNS of Schlatter & Örlü (Reference Schlatter and Örlü2010) at $Re_\tau = 974$ or $Re_{\delta _2}=3030$. To avoid clutter, only the fifth data point is plotted. (b) Comparing (2.6) with the mean wall-normal velocity. The DNS data ‘SO’ are from Schlatter & Örlü (Reference Schlatter and Örlü2010), and DNS data ‘SJHM’ are from Simens et al. (Reference Simens, Jimenez, Hoyas and Mizuno2009).

Figure 1(a) demonstrates a small difference between the sum of the first two functions on the right-hand side of (2.10) and the DNS data of $V^-$. To assess the influence of Reynolds number, figure 1(b) displays the summation of the first two functions on the right-hand side of (2.10) over a range of Reynolds numbers ($359 \leq Re_\tau \leq 1847$ or $1000 \leq Re_{\delta _2} \leq 6000$) represented by the filled symbols. The open symbols represent the difference between the summation and the DNS data of $V^-$. Notably, figure 1(b) reveals a small difference between $V^-$ and the summation of the first two functions on the right-hand side of (2.10), without a discernible dependence on Reynolds number. Therefore, based on the DNS data, it can be inferred that a reasonable approximation for $V^-$ in the ZPG TBL can be achieved by neglecting the terms on the second line of (2.10) as follows:

(2.11)\begin{equation} V^- \approx \frac{\delta_e}{\delta_1} \int_0^{y^-} U^-\,{{\rm d} y}^{-} - \frac{\delta_e}{\delta_1}\,y^-U^-. \end{equation}

To further examine the influence of Reynolds number on the approximate (2.11), figure 2 presents numerical and experimental data spanning a wide range of Reynolds numbers. In figure 2(a), DNS data of relatively moderate Reynolds numbers are presented, exhibiting similar shapes to those in figure 1(a). It is well-known that the mean streamwise velocity profiles $U$ in a ZPG TBL are not ‘self-similar’, but vary with Reynolds number (see e.g. Schlichting Reference Schlichting1979). However, figure 2(a) demonstrates that the shapes of $V^-$ versus $y^-$ remain consistent across different Reynolds numbers, indicating that the $V^-$ profile in a ZPG TBL exhibits self-similarity.

Figure 2. Evaluating the properties of (2.11). (a) DNS data at moderate Reynolds numbers. (b) Experimental data over a wider range of Reynolds number. The experimental data (VHS) are from Vallikivi, Hultmark & Smits (Reference Vallikivi, Hultmark and Smits2015). The lines (solid and dashed) are from DNS data at $Re_{\delta _2}=3030$ from Schlatter & Örlü (Reference Schlatter and Örlü2010). The red symbol represents the DNS data of $V^-$ at $Re_{\delta _2}=3030$ as in figure 1.

In experimental studies, typically direct measurement of the mean wall-normal velocity is not feasible. Figure 2(b) illustrates the two terms on the right-hand side of (2.11) and their sum, calculated from experimental data across a wide range of Reynolds numbers. Remarkably, the sum obtained from the higher Reynolds number experimental data remains consistent with the moderate Reynolds number DNS data. This indicates that the approximate (2.11) can be a valuable tool for estimating the mean wall-normal velocity in situations where direct measurements are not feasible.

In the outer region, figure 2(b) shows a subtle difference between the high Reynolds number measurements and the simulation data at moderate Reynolds numbers. The exact cause of this difference is not clear, as it could be attributed to the Reynolds number effect or uncertainties in determining $\delta _e$. In DNS data, $\delta _e$ is determined based on the Reynolds shear stress profile, specifically as the location where the Reynolds shear stress reaches 1 % of its maximum value (as described in Wei & Knopp Reference Wei and Knopp2023). In physical experiments where the Reynolds shear stress distribution is measured, the spatial resolution in the outer region is often coarse. Depending on the available resolution, $\delta _e$ is determined using 2 % or 5 % of the maximum Reynolds shear stress. In physical experiments where the Reynolds shear stress is not measured, such as in Vallikivi et al. (Reference Vallikivi, Hultmark and Smits2015), $\delta _e$ is based on the $U$ profile (location of $99\,\% U_e$). The difference in the determination of $\delta _e$ between DNS data and physical experiments may contribute to the observed discrepancies between high Reynolds number measurements and moderate Reynolds number simulation data.

The approximate (2.11) for $V^-$ can also be obtained by neglecting the $\partial /\partial x^*$ term in both (2.6) and (2.8). In the case of laminar boundary layer flow over a flat plate, $U^-$ exhibits self-similarity with respect to $y^-$, resulting in the $\partial U^-/\partial x^*$ term being zero in (2.6) and (2.8). Therefore, (2.11) is exact for flat-plate laminar boundary layer flows. For a ZPG TBL, the $\partial U^-/\partial x^*$ term becomes negligible at sufficiently high Reynolds numbers, as demonstrated by DNS and experimental data shown in figures 1 and 2. However, at transitional Reynolds numbers, the $\partial U^-/\partial x^*$ term becomes less negligible, which can result in larger errors when utilizing the approximate (2.11) to predict $V^-$.

2.2. Approximate equations for the Reynolds shear stress

The mean momentum equation (2.1b) for a ZPG TBL represents the balance of three forces: a viscous force ${F_{vis}} = \nu \,\partial ^2 U/\partial y^2$, an advective force ${F_{adv}} = U\,\partial V/\partial y - V\, \partial U/\partial y$, and a turbulent force ${F_{tur}} = \partial R_{uv}/\partial y$. Analysing the force balance in the mean momentum equation offers valuable insights into the behaviour of the boundary layer. The examination of force ratios in figure 3 provides an understanding of the relative significance of these forces in different regions of the boundary layer (e.g. Wei et al. Reference Wei, Fife, Klewicki and McMurtry2005). In the near-wall region, figure 3(a) illustrates that the force balance is primarily between the viscous and turbulent forces. However, in the outer region, as shown in figure 3(b), the viscous force becomes negligible, and the mean momentum equation (2.1b) is balanced between the advective and turbulent forces.

Figure 3. Ratio of viscous to turbulent forces and advective to turbulent forces in a ZPG TBL. (a) Inner scaled wall-normal distance, showing the near-wall region. (b) Outer scaled wall-normal distance, showing the entire boundary layer. Data are the DNS of Schlatter & Örlü (Reference Schlatter and Örlü2010) at $Re_{\tau }=974$.

In the present work, we decompose the Reynolds shear stress into inner and outer parts. The inner part, $R_{uv-{in}}$, represents the Reynolds shear stress close to the wall, and the outer part, $R_{uv-{out}}$, represents the Reynolds shear stress in the outer region. The decomposition of the Reynolds shear stress into inner and outer parts allows us to analyse the contribution of the different parts to the total Reynolds shear stress. Using the decomposed Reynolds shear stress, the mean momentum equation in the inner and outer regions can be approximated as

(2.12)\begin{gather} 0 \approx \underbrace{\nu\,\frac{\partial^2 U}{\partial y^2} + \frac{\partial R_{uv-{in}}}{\partial y}}_{{Inner\ region}}, \end{gather}
(2.13)\begin{gather}0 \approx \underbrace{U\,\frac{\partial V}{\partial y} - V\,\frac{\partial U}{\partial y} + \frac{\partial R_{uv-{out}}}{\partial y}}_{{Outer\ region}}. \end{gather}

Based on the four-layer structure proposed by Wei et al. (Reference Wei, Fife, Klewicki and McMurtry2005), the approximate momentum equation (2.12) holds in layer II, known as the stress gradient balance layer, which extends up to $y^+ \approx 1.6 \sqrt {Re_\tau }$, where $y^+= y u_\tau /\nu$ is the inner scaled wall-normal distance. The approximate momentum equation (2.13) is applicable in layer IV, covering the range $2.6 \sqrt {Re_\tau } < y^+ < \delta ^+_e$. In the meso-layer III, centred around the peak location of the Reynolds shear stress, the gradient of the Reynolds shear stress changes sign, and all three forces contribute to the balance of the mean momentum equation.

2.2.1. Approximate equation for the Reynolds shear stress in the inner region

The variables for the inner layer of the ZPG TBL are normalized by the traditional inner scales $u_\tau$ and $\nu /u_\tau$ (see e.g. Tennekes & Lumley Reference Tennekes and Lumley1972). These scales represent the characteristics of the turbulent flow in the near-wall region, where the viscous forces play an important role. Using the traditional inner scales, the variables in the near-wall region are normalized as follows:

(2.14ac)\begin{equation} y^+ \mathop{=}\limits^{\rm def} \frac{y}{\nu/u_\tau}, \quad U^+ \mathop{=}\limits^{\rm def} \frac{U}{u_\tau}, \quad R^+_{uv-{in}} \mathop{=}\limits^{\rm def} \frac{R_{uv-{in}}}{u^2_\tau}. \end{equation}

Substituting the normalized variables in (2.14ac) into (2.12), the approximate mean momentum equation for the inner region of the ZPG TBL can be written as

(2.15)\begin{equation} 0 \approx \frac{u^3_\tau}{\nu}\,\frac{\partial^2 U^+}{\partial (y^+)^2} + \frac{u^3_\tau}{\nu}\, \frac{\partial R^+_{uv-{in}}}{\partial y^+}. \end{equation}

Multiplying by $\nu /u^3_\tau$ yields the mean momentum equation in dimensionless form as

(2.16)\begin{equation} 0 \approx \underbrace{\frac{\partial^2 U^+}{\partial (y^+)^2} + \frac{\partial R^+_{uv-{in}}}{\partial y^+} }_{{Inner\ region}}.\end{equation}

An equation for the inner part of the Reynolds shear stress $R_{uv-{in}}$ can be obtained by integrating (2.16) with respect to $y^+$, starting from the wall:

(2.17)\begin{equation} R^+_{uv-{in}} \approx 1 - \frac{\partial U^+}{\partial y^+}. \end{equation}

The approximate Reynolds shear stress equation (2.17) for the inner layer is well known (see e.g. Chauhan Reference Chauhan2007).

Figure 4 illustrates a comparison between the inner Reynolds shear stress calculated using the approximate (2.17) (represented by filled symbols) and the DNS data of $R^+_{uv}$ (represented by open symbols) across a range of Reynolds numbers. The use of a log-linear scale in figure 4 facilitates the visualization of the rapid increase of the Reynolds shear stress in the near-wall region. The agreement between the approximate (2.17) and the DNS data is observed clearly from the wall up to near the peak value of the Reynolds shear stress, highlighting the accurate depiction of the Reynolds shear stress behaviour within this region.

Figure 4. Approximate (2.17) for the Reynolds shear stress in the inner region $R_{uv-{in}}$. Data from SO and SJHM as in figure 1.

2.2.2. Approximate equation for the Reynolds shear stress in the outer region

The variables for the outer layer of the ZPG TBL are normalized by the boundary layer thickness $\delta _e$, and the mean velocity at the boundary layer edge, $U_e$ and $V_e$. The outer scaled variables are denoted as follows:

(2.18ad)\begin{equation} y^- \mathop{=}\limits^{\rm def} \frac{y}{\delta_e}, \quad U^- \mathop{=}\limits^{\rm def} \frac{U_e - U}{U_e}, \quad V^- \mathop{=}\limits^{\rm def} \frac{V}{V_e}, \quad R^-_{uv-{out}} \mathop{=}\limits^{\rm def} \frac{R_{uv-{out}}}{U_e V_e}. \end{equation}

The scale proposed by Wei & Maciel (Reference Wei and Maciel2018) is used as the reference scale for the Reynolds shear stress.

Substituting the normalized variables in (2.18ad) into (2.13), the approximate mean momentum equation for the outer region can be written as

(2.19)\begin{equation} 0 \approx (U_e-U_e U^-)\,\frac{V_e}{\delta_e}\, \frac{\partial V^-}{\partial y^-} - V_e V^-\left(-\frac{U_e}{\delta_e}\,\frac{\partial U^-}{\partial y^-}\right) + \frac{U_e V_e}{\delta_e}\, \frac{\partial R^-_{uv-{out}}}{\partial y^-}. \end{equation}

In the majority of experimental or numerical studies on TBLs, the Reynolds shear stress data are commonly reported as $R^+_{uv} = R_{uv}/u^2_\tau$. Additionally, in physical experimental studies, the quantity $V_e$ is typically not measured directly. To derive an approximate equation for $R^+_{uv-{out}}$, (2.19) is multiplied by $\delta _e/u^2_\tau$, thereby obtaining the mean momentum equation in a dimensionless form as follows:

(2.20)\begin{equation} 0 \approx \underbrace{ \frac{U_e V_e}{u^2_\tau} \left( (1 - U^-)\,\frac{\partial V^-}{\partial y^-} + V^-\,\frac{\partial U^-}{\partial y^-} \right) + \frac{\partial R^+_{uv-{out}}}{\partial y^-} }_{{Outer\ region}}.\end{equation}

Note that by definition, $({U_e V_e}/{u^2_\tau }) ({\partial R^-_{uv}}/{\partial y^-}) = {\partial R^+_{uv}}/{\partial y^-}$. Using a thin boundary layer assumption, Wei & Klewicki (Reference Wei and Klewicki2016) demonstrated the approximate equality $U_e V_e/u_\tau ^2 \approx H_{12}$ in a ZPG TBL, where $H_{12}=\delta _1/\delta _2$ is the shape factor. An equation for the outer part of the Reynolds shear stress $R_{uv-{out}}$ can be obtained by integrating (2.20) with respect to $y^-$, starting from the boundary layer edge:

(2.21)\begin{equation} R^+_{uv-{out}} \approx \int_1^{y^-} -H_{12} \left( (1-U^-)\,\frac{\partial V^-}{\partial y^-} + V^-\,\frac{\partial U^-}{\partial y^-} \right) {{\rm d} y}^- . \end{equation}

Substituting (2.11) for $V^-$ and (2.5) for $\partial V^-/\partial y^-$ (neglecting the $x$-derivative term), the integrand in (2.21) can be written as a function of $U^-$ as

(2.22)\begin{align} & {-H_{12}} \left( (1-U^-)\,\frac{\partial V^-}{\partial y^-} + V^-\,\frac{\partial U^-}{\partial y^-} \right) \nonumber\\ &\quad \approx{-}\frac{\delta_e}{\delta_2} \left( (U^-- (U^-)^2) + \frac{\partial}{\partial y^-} \left( U^- \int_0^{y^-} U^-\,{{\rm d} y}^-- y^- U^- \right) \right). \end{align}

Therefore, an approximate equation for the outer Reynolds shear stress can be obtained from (2.21) as

(2.23)\begin{equation} R^+_{uv-{out}} \approx \frac{\delta_e}{\delta_2} \left\{ \int_1^{y^-} ( (U^-)^2 - U^-)\,{{\rm d} y}^- + \left[ y^- U^-- U^- \int_0^{y^-} U^-\,{{\rm d} y}^- \right] \right\}. \end{equation}

Although somewhat complicated, (2.23) requires only the data of $U^-$ at a single $x$-location to obtain an approximation of the Reynolds shear stress in the outer layer of the ZPG TBL. Empirically, it has been observed that the integrand in (2.21) can be simplified further. To begin, the integrand in (2.21) is rearranged as follows:

(2.24)\begin{align} & {-H_{12}} \left( (1-U^-)\,\frac{\partial V^-}{\partial y^-} + V^-\,\frac{\partial U^-}{\partial y^-} \right) \nonumber\\ &\quad ={-}H_{12} \left( \frac{\partial [(1-U^-)V^-]}{\partial y^-} - 2 V^-\, \frac{\partial (1-U^-)}{\partial y^-} \right) \nonumber\\ &\quad ={-} \frac{\partial [(1-U^-)V^-]}{\partial y^-} + (1-H_{12})\,\frac{\partial [(1-U^-)V^-]}{\partial y^-} + 2 H_{12} V^- \frac{\partial (1-U^-)}{\partial y^-}. \end{align}

The three terms in (2.24) are then calculated from DNS data, as illustrated in figure 5(a) using the DNS data at $Re_{\tau }=974$. According to the DNS data, the sum of the last two terms in (2.24) is nearly equal to zero. Figure 5(b) displays the first term and the sum of the last two terms of (2.24) for a range of Reynolds numbers, showing that the sum of the last two terms is close to zero for all Reynolds numbers considered. Therefore, the integrand in (2.21) can be approximated as follows:

(2.25)\begin{equation} -H_{12} \left( (1-U^-)\,\frac{\partial V^-}{\partial y^-} + V^-\,\frac{\partial U^-}{\partial y^-} \right) \approx{-} \frac{\partial [(1-U^-)V^-]}{\partial y^-}. \end{equation}

Figure 5. (a) Terms in (2.24). The DNS data are from Schlatter & Örlü (Reference Schlatter and Örlü2010) at $Re_{\tau }=974$. (b) Terms in (2.24) at different Reynolds numbers. Data from SO and SJHM as in figure 1.

By substituting (2.25) into (2.21), the integration can be simplified, leading to a simpler approximation for the Reynolds shear stress in the outer layer:

(2.26)\begin{equation} R^+_{uv-{out}} \approx 1 - (1-U^-) V^- = 1 - \frac{UV}{U_e V_e} = 1- \frac{U^+V^+}{H_{12}}, \end{equation}

where $V^+=V/u_\tau$, and $1-U^- = U/U_e$ by the definition of $U^-$. The approximate (2.26) for $R^+_{uv-{out}}$ is compared with DNS data of $R^+_{uv}$ over a range of Reynolds numbers in figure 6, showing good agreement in the outer region.

Figure 6. Approximate (2.26) for the Reynolds shear stress in the outer region $R_{uv-{out}}$. Data from SO and SJHM as in figure 1.

To further validate the approximate (2.26), a plot of $U^+V^+/H_{{12}}$ versus $R^+_{uv}$ can be generated, as shown in figure 7. In the outer region, the profile is well captured by a linear function, which indicates that the approximate equation is valid for the Reynolds shear stress in the outer layer.

Figure 7. Evaluation of the approximate (2.26) as ${U^+V^+}/{H_{12}} \approx 1- R^+_{uv-{out}}$. Data from SO and SJHM as in figure 1.

Substituting (2.11) for $V^-$ into (2.26), we get an equation to approximate the Reynolds shear stress as a function of the mean streamwise velocity only as

(2.27)\begin{equation} R^+_{uv-{out}} \approx 1 - (1-U^-)\, \frac{\delta_e}{\delta_1} \left( \int_0^{y^-} U^-\, {{\rm d} y}^-- y^-U^- \right). \end{equation}

2.2.3. Approximate equation for the Reynolds shear stress across the entire boundary layer

By observing that the second term in (2.17) tends to approach zero in the outer region (as depicted in figure 4) and the second term in (2.27) becomes negligible in the inner region (as shown in figure 6), an approximate equation that effectively describes the Reynolds shear stress across the entire ZPG TBL can be synthesized as follows:

(2.28)\begin{equation} R^+_{uv} \approx 1 - \frac{\delta_e}{\delta_1} (1- U^-) \left( \int_0^{y^-} U^- {{\rm d} y}^-- y^-U^- \right) - \frac{\partial U^+}{\partial y^+}. \end{equation}

Within the inner region, the contribution of the second term in (2.28) is negligible, leading to a reduction of the equation to the inner Reynolds shear stress as described by (2.17). On the other hand, in the outer region, the impact of the last term becomes negligible, resulting in the simplification of (2.28) to the outer Reynolds shear stress equation given by (2.27).

In many experimental or numerical studies, the inner scaled mean flow variables are commonly presented. For convenience, (2.29) presents the approximation function for the Reynolds shear stress using only the inner scaled variables:

(2.29)\begin{equation} R^+_{uv} \approx 1 - \frac{1}{\delta^+_1}\, \frac{U^+}{(U^+_e)^2} \left( \int_0^{y^+} (U^+_e - U^+)\,{{\rm d} y}^+ - y^+(U^+_e - U^+) \right) - \frac{\partial U^+}{\partial y^+}.\end{equation}

Figure 8 provides a comparison between the DNS data of the Reynolds shear stress and the approximate (2.28). Both the inner scaled wall-normal distance (figure 8a) and the outer scaled wall-normal distance (figure 8b) are used for this comparison. In both plots, the open symbols represent the DNS data, while the filled symbols depict values calculated from the approximate (2.28). The use of log-linear axes in figure 8(a) enhances the visualization of the variation in the Reynolds shear stress and the approximation equation within the near-wall region. The figure demonstrates notable agreement between the approximate equation and the DNS data, with only minor deviations observed for the lowest Reynolds number data.

Figure 8. Comparison of the approximate (2.28) and DNS data. (a) Inner scaled wall-normal distance. (b) Outer scaled wall-normal distance. The open symbols represent the DNS data of the Reynolds shear stress, and the filled symbols represent the approximate equation. Data from SO and SJHM as in figure 1.

The DNS data presented in figure 8 correspond to relatively moderate Reynolds numbers. In order to investigate the influence of Reynolds number on the accuracy of the approximate (2.28), figure 9 showcases experimental data encompassing a broader range of Reynolds numbers. For comparison purposes, the DNS data from Schlatter & Örlü (Reference Schlatter and Örlü2010) at $Re_\tau =974$ are also included in the figure. As expected, the experimental data exhibit larger variability compared to the DNS data. However, despite the increased scatter, the deviation between the measured Reynolds shear stress and the approximate (2.28) remains relatively small and does not demonstrate a distinct dependence on Reynolds number. The larger discrepancy observed in the near-wall region can be attributed to the limited availability of experimental measurements close to the wall. Nevertheless, the overall agreement between the experimental data and the approximate equation suggests that it offers a reasonable estimation of the Reynolds shear stress, especially in the outer region of the boundary layer.

Figure 9. Comparison of the approximate (2.28) with the experimental data from Baidya et al. (Reference Baidya, Philip, Hutchins, Monty and Marusic2021) (B). The DNS data from Schlatter & Örlü (Reference Schlatter and Örlü2010) (SO) at $Re_\tau =974$ are included for reference.

Figures 8 and 9 provide strong evidence supporting the effectiveness of the approximate (2.28) in accurately capturing the distribution of Reynolds shear stress around its peak. In other words, the approximate (2.28) predicts reliably the Reynolds shear stress distribution within the meso-layer region of the four-layer structure proposed by Wei et al. (Reference Wei, Fife, Klewicki and McMurtry2005).

3. Summary

This study makes two significant contributions to our understanding of ZPG TBLs.

First, it presents a novel analytical solution for the mean wall-normal velocity $V$. By neglecting terms involving $\partial U^-/\partial x^*$, a highly accurate approximate equation is derived for $V$, comprising two functions: $\int _0^{y^-} U^- \,{{\rm d} y}^-$ and $y^- U^-$. The validity of this approximation is demonstrated through a comprehensive comparison with DNS data. This analytical solution offers an alternative method for determining the mean wall-normal velocity profile in experimental studies of ZPG TBLs. By utilizing measurements of the mean streamwise velocity, researchers can obtain the mean wall-normal velocity profile, eliminating the need for direct measurements. This approach proves advantageous for guiding experimental measurements and establishing a reference point for comparing experimental findings with numerical simulations.

Second, the study focuses on the force balance behaviours in the inner and outer regions of a ZPG TBL, leading to the development of approximate mean momentum equations. The Reynolds shear stress is decomposed into inner and outer parts, denoted as $R_{uv-{in}}$ and $R_{uv-{out}}$, respectively. Approximate equations for both parts are derived through integration of the approximate mean momentum equations. The accuracy of these approximation equations is validated rigorously using DNS and experimental data. Notably, an approximate equation is obtained for the distribution of the Reynolds shear stress across the entire ZPG TBL, relying solely on the measurement of the streamwise mean velocity, which is easily accessible through experimental measurements. This approximation equation serves as a practical tool for studying and characterizing ZPG TBL, bridging the gap between experimental measurements and theoretical predictions. It enhances our understanding of flow dynamics in ZPG TBLs, and facilitates more accurate predictions for practical applications.

The mean wall-normal velocity has been neglected previously in many turbulence models, but a direct relation between the Reynolds shear stress and the mean wall-normal velocity is established here as $R_{uv-{out}}/u^2_\tau \approx 1 - (UV)/(U_eV_e)$. Overall, this study advances our understanding of turbulent flows by highlighting the importance of the mean wall-normal velocity and providing analytical solutions and approximation equations that contribute to the development of more accurate and reliable turbulence models.

Acknowledgements

We express our gratitude to Dr Jimenez, Dr Marusic, Dr Schlatter, Dr Vallikivi and their research groups for generously sharing their simulation and experimental data.

Declaration of interests

The authors report no conflict of interest.

References

Baidya, R., Philip, J., Hutchins, N., Monty, J.P. & Marusic, I. 2021 Spanwise velocity statistics in high-Reynolds-number turbulent boundary layers. J. Fluid Mech. 913, A35.CrossRefGoogle Scholar
Chauhan, K.A. 2007 Study of canonical wall-bounded turbulent flows. PhD thesis, Illinois Institute of Technology, Chicago, IL.Google Scholar
Coles, D. 1956 The law of the wake in the turbulent boundary layer. J. Fluid Mech. 1 (2), 191226.CrossRefGoogle Scholar
Kumar, P. & Dey, J. 2019 Shape factor of the turbulent boundary layer on a flat plate and the Reynolds shear stress in the outer region. Phys. Rev. Fluids 4 (2), 024605.CrossRefGoogle Scholar
Kumar, P. & Mahesh, K. 2022 A method to determine wall shear stress from mean profiles in turbulent boundary layers. Exp. Fluids 63 (1), 118.CrossRefGoogle Scholar
Monin, A.S. & Yaglom, A.M. 1965 Statistical Fluid Mechanics, Volume I: Mechanics of Turbulence, vol. 1. MIT Press.Google Scholar
Perry, A.E., Marusic, I. & Jones, M.B. 2002 On the streamwise evolution of turbulent boundary layers in arbitrary pressure gradients. J. Fluid Mech. 461, 6191.CrossRefGoogle Scholar
Pope, S.B. 2000 Turbulent Flows. Cambridge University Press.CrossRefGoogle Scholar
Schlatter, P. & Örlü, R. 2010 Assessment of direct numerical simulation data of turbulent boundary layers. J. Fluid Mech. 659, 116126.CrossRefGoogle Scholar
Schlichting, H. 1979 Boundary-Layer Theory. McGraw-Hill.Google Scholar
Simens, M.P., Jimenez, J., Hoyas, S. & Mizuno, Y. 2009 A high-resolution code for turbulent boundary layers. J. Comput. Phys. 228 (11), 42184231.CrossRefGoogle Scholar
Tennekes, H. & Lumley, J.L. 1972 A First Course in Turbulence. MIT Press.CrossRefGoogle Scholar
Townsend, A.A. 1956 The Structure of Turbulent Shear Flow. Cambridge University Press.Google Scholar
Vallikivi, M., Hultmark, M. & Smits, A.J. 2015 Turbulent boundary layer statistics at very high Reynolds number. J. Fluid Mech. 779, 371389.CrossRefGoogle Scholar
Wei, T., Fife, P., Klewicki, J. & McMurtry, P. 2005 Properties of the mean momentum balance in turbulent boundary layer, pipe and channel flows. J. Fluid Mech. 522, 303327.CrossRefGoogle Scholar
Wei, T. & Klewicki, J. 2016 Scaling properties of the mean wall-normal velocity in zero-pressure-gradient boundary layers. Phys. Rev. Fluids 1 (8), 082401.CrossRefGoogle Scholar
Wei, T. & Knopp, T. 2023 Outer scaling of the mean momentum equation for turbulent boundary layers under adverse pressure gradient. J. Fluid Mech. 958, A9.CrossRefGoogle Scholar
Wei, T., Li, Z. & Livescu, D. 2022 a Scaling patch analysis of planar turbulent mixing layers. Phys. Fluids 34, 115120.CrossRefGoogle Scholar
Wei, T. & Livescu, D. 2021 Scaling of the mean transverse flow and Reynolds shear stress in turbulent plane jet. Phys. Fluids 33 (3), 035142.CrossRefGoogle Scholar
Wei, T., Livescu, D. & Liu, X. 2022 b Scaling patch analysis of planar turbulent wake. Phys. Fluids 34, 065116.CrossRefGoogle Scholar
Wei, T. & Maciel, Y. 2018 Derivation of Zagarola–Smits scaling in zero-pressure-gradient turbulent boundary layers. Phys. Rev. Fluids 3 (1), 012601.CrossRefGoogle Scholar
Wei, T., Maciel, Y. & Klewicki, J. 2017 Integral analysis of boundary layer flows with pressure gradient. Phys. Rev. Fluids 2 (9), 092601.CrossRefGoogle Scholar
Figure 0

Table 1. Boundary conditions.

Figure 1

Figure 1. (a) Shapes of the mean wall-normal velocity $V^-$, $y^- U^-$ and $\int _0^{y^-} U^-\,{{\rm d} y}^-$. The data are from the DNS of Schlatter & Örlü (2010) at $Re_\tau = 974$ or $Re_{\delta _2}=3030$. To avoid clutter, only the fifth data point is plotted. (b) Comparing (2.6) with the mean wall-normal velocity. The DNS data ‘SO’ are from Schlatter & Örlü (2010), and DNS data ‘SJHM’ are from Simens et al. (2009).

Figure 2

Figure 2. Evaluating the properties of (2.11). (a) DNS data at moderate Reynolds numbers. (b) Experimental data over a wider range of Reynolds number. The experimental data (VHS) are from Vallikivi, Hultmark & Smits (2015). The lines (solid and dashed) are from DNS data at $Re_{\delta _2}=3030$ from Schlatter & Örlü (2010). The red symbol represents the DNS data of $V^-$ at $Re_{\delta _2}=3030$ as in figure 1.

Figure 3

Figure 3. Ratio of viscous to turbulent forces and advective to turbulent forces in a ZPG TBL. (a) Inner scaled wall-normal distance, showing the near-wall region. (b) Outer scaled wall-normal distance, showing the entire boundary layer. Data are the DNS of Schlatter & Örlü (2010) at $Re_{\tau }=974$.

Figure 4

Figure 4. Approximate (2.17) for the Reynolds shear stress in the inner region $R_{uv-{in}}$. Data from SO and SJHM as in figure 1.

Figure 5

Figure 5. (a) Terms in (2.24). The DNS data are from Schlatter & Örlü (2010) at $Re_{\tau }=974$. (b) Terms in (2.24) at different Reynolds numbers. Data from SO and SJHM as in figure 1.

Figure 6

Figure 6. Approximate (2.26) for the Reynolds shear stress in the outer region $R_{uv-{out}}$. Data from SO and SJHM as in figure 1.

Figure 7

Figure 7. Evaluation of the approximate (2.26) as ${U^+V^+}/{H_{12}} \approx 1- R^+_{uv-{out}}$. Data from SO and SJHM as in figure 1.

Figure 8

Figure 8. Comparison of the approximate (2.28) and DNS data. (a) Inner scaled wall-normal distance. (b) Outer scaled wall-normal distance. The open symbols represent the DNS data of the Reynolds shear stress, and the filled symbols represent the approximate equation. Data from SO and SJHM as in figure 1.

Figure 9

Figure 9. Comparison of the approximate (2.28) with the experimental data from Baidya et al. (2021) (B). The DNS data from Schlatter & Örlü (2010) (SO) at $Re_\tau =974$ are included for reference.