Inclined Plate

SOLUTION: Gravity-driven viscous ow over an inclined plate
Conguration
We consider the situation represented in Fig.1. Due to gravity, a thin liquid lm (assumed of constant thickness h) ows steadily and at low speed (a laminar ow can be assumed) along an inclined plate (inclination angle ). The uid properties (density , dynamic viscosity ) are known and constant.
Figure 1: Gravity-driven viscous ow over an inclined plate.
2
2.1
Questions and answers

Velocity eld
Determine the velocity prole u(y ). We are in a typical situation where we may employ the simplest formulation of the Navier-Stokes equations, since many assumptions have been made: the ow is steady, so that all time derivatives will disappear; 1
the ow is incompressible (liquid); the liquid is Newtonian (constant dynamic viscosity ); the ow is indeed two-dimensional, since no specic information is given concerning the third spatial direction, perpendicular to the gure. Only the directions x and y will be considered, neglecting all possible changes along z . In order to fully determine this ow, we just need to determine the x and y components of the velocity, written respectively u and v , and the pressure p (the density is constant and therefore not a variable). In principle, the three variables of interest u, v and p all depend on x and y . What are our starting conservation equations? The generic conservation equations for mass and momentum, simplied using the above assumptions. For mass, one gets v = 0, here: u v + =0 x y (1)
Now, what are the boundary conditions for the y -component v ? The no-slip boundary condition at the wall reads v (y = 0) = 0. Since the lm thickness h is constant, the ow velocity at the contact point with the surrounding atmosphere must necessarily be parallel to the wall, leading again to v (y = h) = 0. Starting from zero, going back to 0, and without any particular reason to appear, the y -component of the velocity is therefore zero everywhere! This is the rst, major result: v=0 (2) Plugging back this result in the mass conservation equation, Eq.(1), one automatically: u =0 x Hence, the x-component of the velocity is constant along the x-direction changes only with y : u = u(y ) gets (3) and (4)
What are the boundary conditions for u? Again, along the plate, the no-slip boundary conditions reads u(y = 0) = 0 (5) like for v . Now, at the contact surface with the atmosphere, the exchange of momentum at the molecular level between the molecules of the liquid near the surface and the air molecules near the surface lead to the same no-slip condition: the velocity in the liquid must be the same as the velocity in the gas near the contact surface. 2
What is the consequence? The air near the surface of the liquid lm will be entrained by the movement of the lm and will ow locally at the same velocity (at present unknown). How can we translate mathematically this property? We know that friction is the result of a velocity dierence. Therefore, if the velocity in liquid and gas are identical, there are no friction forces along the lm surface. The only component of the friction force along the lm surface, obtained by simplifying the friction u . Finally, the tensor for the conditions considered here, is proportional to y boundary condition we have to enforce at the lm surface reads: u (y = h) = 0 y (6)
Let us now write the standard Navier-Stokes equations taking into account the specic assumptions of this problem. It comes: u 1 p u u +v = gx + x y x v v 1 p u +v = gy + x y y 2u 2u + x2 y 2 2v 2v + x2 y 2 (7) (8)
Since v = 0, the second equation just leads to p = gy = g cos y (9)
The boundary condition for the pressure reads classically at the contact surface between liquid and gas (10) p(y = h) = pa This is the same at every x-position. Considering this condition and the previous equation (9), one obtains directly: p = p(y ) = pa + g (h y ) cos (11)
This is indeed the same as hydrostatics (the pressure increases linearly with the depth (h y ) within the uid), which is not a surprise, since the ow is completely perpendicular to the pressure gradient. Note that the change in pressure will be indeed extremely small! The overpressure at the wall is gh, which for = 1000 kg/m3 and h = 1 cm would mean about 100 Pa, or about 1/1000 of the atmospheric pressure pa . Therefore, you understand why we would rightly assume p = pa = constant when analyzing globally this ow in the light of the Bernoulli equation!
Now, taking into account this (small) pressure variation, the only ow velocity component u can be easily obtained by considering what remains of Eq.(7): g sin 2u = . 2 y (12)
Since the right-hand side is a constant value, this can be easily integrated twice as g sin y 2 (13) + C1 y + C2 u(y ) = 2 The two integration constants C1 and C2 are easily determined by enforcing the two boundary conditions, Eq.(5) and (6), leading to: C1 = C2 gh sin = 0 (14) (15)
Finally, the x-component of the velocity is now perfectly known: u = u(y ) = g sin y y h 2 (16)
Of course, for = 0 (a horizontal plate), there is no gravity-driven ow (u = 0), as expected!
2.2
Volumetric ow rate
Determine the resulting volumetric ow-rate q . The volumetric ow-rate is obtained by integrating through the depth of the lm, taking into account the ow length b perpendicular to the gure: q :=
y =h y =0
bu(y )dy
(17)
Using the known solution for u, Eq.(16), one becomes successively: q =

y =h bg sin 1 h ydy 2 y =0 bg sin h2 1 h3 h = 2 2 3 y =h y =0
y 2 dy
(18) (19)
Finally, the answer reads: q = bg sin 3 h 3 (20)
2.3
Connection to Bernoulli
If we would try to solve this problem in a Bernoulli-like manner, then we would also obtain similar informations, but of course not at the same level of details! We would just write as usual p = pa throughout the ow. As we have seen, this is almost right (typically 0.1% error) for the present ow. We would not get the detailed velocity eld, but we could write the Bernoulli equation (only valid along streamlines, since the vorticity g sin v u = (y h) x y (21)
is not zero). By expressing that the ow velocity (assumed, rightly, to be only in the x-direction) cannot change with the x-coordinate (same as stating that the lm thickness h is constant), one would get that the change in potential energy is exactly what is needed to compensate the friction loss in the ow. In this manner, it is indirectly possible to get back the mean ow velocity in the lm.

Inclined Plate

Uploaded by

Copyright:

Available Formats

Inclined Plate

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Inclined Plate

Uploaded by

Copyright:

Available Formats

SOLUTION: Gravity-driven viscous ow over an inclined plate

Figure 1: Gravity-driven viscous ow over an inclined plate.

Questions and answers

Since v = 0, the second equation just leads to p = gy = g cos y (9)

Of course, for = 0 (a horizontal plate), there is no gravity-driven ow (u = 0), as expected!

Using the known solution for u, Eq.(16), one becomes successively: q =

Finally, the answer reads: q = bg sin 3 h 3 (20)

You might also like