Schneider flow

For laminar or turbulent jets and for laminar plumes, the volumetric entertainment rate per unit axial length is constant as can be seen from the solution of Schlichting jet and Yih plume. Thus, the jet or plume can be considered as a link sink, as was first done by G. I. Taylor and this sink will be drive the fluid outside the jet or plume. Prior to Schneider, it was assumed that this outer fluid motion is also a large Reynolds number flow, hence the outer fluid motion is assumed to be a potential flow solution, which was solved by G. I. Taylor in 1958. For turbulent plume, the entrainment is not constant, nevertheless, the outer fluid is still governed by Taylors solution.

Though Taylor's solution is still true for turbulent jet, for laminar jet or laminar plume, the effective Reynolds number for outer fluid is found to be of order unity since the entertainment by the sink in these cases is such that the flow is not inviscid. In this case, full Navier-Stokes equations has to be solved for the outer fluid motion and at the same time, since the fluid is bounded from the bottom by a solid wall, the solution has to satisfy the non-slip condition. Schneider obtained a self-similar solution for this outer fluid motion, which naturally reduced to Taylor's potential flow solution as the entrainment rate by the line sink is increased.

Suppose a conical wall of semi-angle ${\displaystyle \alpha }$ with polar axis along the cone-axis and assume the vertex of the solid cone sits at the origin of the spherical coordinates ${\displaystyle (r,\theta ,\phi )}$ extending along the negative axis. Now, put the line sink along the positive side of the polar axis. Set this way, ${\displaystyle \alpha =\pi /2}$ represents the common case of flat wall with jet or plume emerging from the origin. The case ${\displaystyle \alpha =\pi }$ corresponds to jet/plume issuing from a thin injector. The flow is axisymmetric with zero azimuthal motion, i.e., the velocity components are ${\displaystyle (v_{r},v_{\theta },0)}$. The usual technique to study the flow is to introduce the Stokes stream function ${\displaystyle \psi }$ such that

${\displaystyle v_{r}={\frac {1}{r^{2}\sin \theta }}{\frac {\partial \psi }{\partial \theta }},\quad v_{\theta }=-{\frac {1}{r\sin \theta }}{\frac {\partial \psi }{\partial r}}.}$

Introducing ${\displaystyle \xi =\cos \theta }$ as the replacement for ${\displaystyle \theta }$ and introducing the self-similar form ${\displaystyle \psi =K\nu rf(\xi )}$ into the axisymmetric Navier-Stokes equations, we obtain[5]

${\displaystyle K^{-1}[(1-\xi ^{2})f''''-4\xi f''']-ff'''-3f'f''=0.}$

where the constant ${\displaystyle K}$ is such that the volumetric entrainment rate per unit axial length is equal to ${\displaystyle 2\pi K\nu }$. For laminar jet, ${\displaystyle K=4}$ and for laminar plume, it depends on the Prandtl number ${\displaystyle Pr}$, for example with ${\displaystyle Pr=1}$, we have ${\displaystyle K=6}$ and with ${\displaystyle Pr=2}$, we have ${\displaystyle K=4}$. For turbulent jet, this constant is the order of the jet Reynolds number, which is a large number.

The above equation can easily be reduced to a Riccati equation by integrating thrice, a procedure that is same as in the Landau–Squire jet (main difference between Landau-Squire jet and the current problem are the boundary conditions). The boundary conditions on the conical wall ${\displaystyle \xi =\xi _{w}=\cos \alpha }$ become

${\displaystyle f(\xi _{w})=f'(\xi _{w})=0}$

and along the line sink ${\displaystyle \xi =1}$, we have

${\displaystyle f(1)=1,\quad \lim _{\xi \rightarrow 1}(1-\xi )^{1/2}f''\rightarrow 0.}$

The problem has been solved numerically from here.

The exact solution of the Navier-Stokes solutions was verified experimentally by Zauner in 1985.[6] Further analysis[7][8] showed that the axial momentum flux decays slowly along the axis unlike the Schlichting jet solution and it is found that the Schneider flow becomes invalid when distance from the origin increases to a distance of the order exponential of square of the jet Reynolds number, thus the domain of validity of Schneider solution increases with increasing jet Reynolds number.

The presence of swirling motion, i.e., ${\displaystyle v_{\phi }\neq 0}$ is shown not to influence the axial motion given by ${\displaystyle \psi =K\nu rf(\xi )}$ provided ${\displaystyle K\sim O(1)}$. If ${\displaystyle K}$ is very large, the presence of swirl completely alters the motion on the axial plane. For ${\displaystyle K\sim O(1)}$, the azimuthal solution can be solved in terms of the circulation ${\displaystyle 2\pi \Gamma }$, where ${\displaystyle \Gamma =r\sin \theta v_{\phi }}$. The solution can be described in terms of the self-similar solution of the second kind, ${\displaystyle \Gamma =Ar^{\lambda }\Lambda (\xi )}$, where ${\displaystyle A}$ is an unknown constant and ${\displaystyle \lambda }$ is an eigenvalue. The function ${\displaystyle \Lambda (\xi )}$ satisfies

${\displaystyle K^{-1}[(1-\xi ^{2})\Lambda ''+\lambda (\lambda -1)\Lambda ]-f\Lambda '+\lambda f'\Lambda =0}$

subjected to the boundary conditions ${\displaystyle \Lambda (\xi _{w})=0}$ and ${\displaystyle (1-\xi )^{1/2}\Lambda '\rightarrow 0}$ as ${\displaystyle \xi \rightarrow 1}$.

• Landau–Squire jet
• Schlichting jet