Why are the Navier-Stokes equations inconsistent in this case?

In summary, the conversation discusses the case of a one-dimensional incompressible, non-viscous fluid flowing down a vertical pipe under the influence of gravity. The Navier Stokes equations in this case consist of three separate conservation equations - conservation of mass, momentum, and energy. Under the assumption of incompressibility, the equations yield inconsistent results, leading to the conclusion that the assumption of incompressibility may be incorrect. However, it is possible that the equations can be made consistent by including the potential energy of the fluid in the internal energy term. Further examination of this possibility is needed to determine the correctness of the assumption.
  • #1
chandrahas
72
2
Consider the case of a one-dimensional incompressible, non-viscous fluid flowing down a vertical pipe under the influence of gravity. Since we assume the flow is constant along the cross section of the pipe from the one dimensional assumption, let us denote the velocity of the fluid down the pipe to be ##v##, with all other velocity components zero. Similarly, let us orient our axis such that the positive x axis points down the pipe. The Navier Stokes equations in this case consist of three separate conservation equations - conservation of mass, momentum, and energy. Under the assumption of incompressibility, conservation of mass simply gives:

$$
\begin{equation} \tag{1}
\frac{dv}{dx} = 0
\end{equation}
$$

The momentum equation gives us the following equation:

$$
\begin{equation}\tag{2}
\rho v \frac{dv}{dx} = -\frac{dp}{dx} + \rho g
\end{equation}
$$

Where ##\rho## is the fluid density, ##p## the fluid pressure, and ##g## the gravitational acceleration.
Finally, the energy equation relates the internal energy per unit mass, kinetic energy per unit mass, and the work done against pressure, and by the gravitational body force. We have the following relationship:

$$
\begin{equation}\tag{3}
\rho v \frac{de}{dx} = -p \frac{dv}{dx}
\end{equation}
$$

Where ##e## is the internal energy per unit mass of the fluid. This equation was obtained from the application of equation ##(2.55)##, from [this article](https://www.eng.auburn.edu/~tplacek/courses/fluidsreview-1.pdf) to the present case assuming no viscosity, no thermal heat transfer, and no heat source.

Using the equation of state for an ideal gas, we have the following relationship between the internal energy of a gas and its temperature:
$$ U = \frac{3}{2} NkT $$
Where we assumed the fluid to be a mono-atomic gas, with number of molecules ##N##, and temperature ##T##. however, we also have the following relationship for an ideal gas:

$$ p = (\frac{N}{V})kT = nkT$$

Where ##n## is the number density of the fluid.

Now, the internal energy density per mass of the fluid ##e## can be written as:
$$ e = \frac{U}{\rho V} = \frac{3}{2 \rho}nkT = \frac{3}{2} \frac{p}{\rho}$$

Finally, under the assumption of incompressibility and equation ##(1)##, equation ##(2)## simply becomes:

$$
\begin{equation}\tag{3}
\frac{dp}{dx} = \rho g
\end{equation}
$$

Similarly, plugging in ##e## into equation ##(3)## and setting the right hand side to zero yields the following:

$$
\begin{equation}\tag{5}
\frac{dp}{dx} = 0
\end{equation}
$$

It seems like these two equations for the pressure of the fluid are inconsistent. Yet, I can't find any assumption that is incorrect. Which leads me to believe that the assumption of incompressibility is incorrect and it is not possible to have an incompressible fluid under a constant cross section with a body force acting on it. The fluid necessarily needs to be compressed.

However, I am unsure and I might have made a mistake in my assumption or have a flaw in my understanding. **Would the system have been consistent if the fluid were compressible and why?**
 
Physics news on Phys.org
  • #2
An ideal gas is compressible which is an inconsistency.
 
  • Like
Likes hutchphd and vanhees71
  • #3
Frabjous said:
An ideal gas is compressible which is an inconsistency.
Incorrect. A gas is compressible, but that doesn't mean the flow of a gas is a compressible flow. They are two different but related concepts.
 
  • Like
  • Informative
Likes wrobel and Juanda
  • #4
chandrahas said:
...

Finally, under the assumption of incompressibility and equation ##(1)##, equation ##(2)## simply becomes:

$$
\begin{equation}\tag{3}
\frac{dp}{dx} = \rho g
\end{equation}
$$

Similarly, plugging in ##e## into equation ##(3)## and setting the right hand side to zero yields the following:

$$
\begin{equation}\tag{5}
\frac{dp}{dx} = 0
\end{equation}
$$

It seems like these two equations for the pressure of the fluid are inconsistent. Yet, I can't find any assumption that is incorrect. Which leads me to believe that the assumption of incompressibility is incorrect and it is not possible to have an incompressible fluid under a constant cross section with a body force acting on it. The fluid necessarily needs to be compressed.

However, I am unsure and I might have made a mistake in my assumption or have a flaw in my understanding. **Would the system have been consistent if the fluid were compressible and why?**
You left out the body force term in your energy equation, which would account for gravity.
 
  • #5
boneh3ad said:
You left out the body force term in your energy equation, which would account for gravity.
The equation I used from the source I linked claimed that the body force term cancels out with the material derivative term of fluid velocity squared. I have also followed through with the derivation and it does indeed cancel out.

However, I am wondering if I should include the potential energy of the fluid in the "internal energy" term. That would fix the equation. But my original scenario includes a body force that cannot be written in the form of a potential. So I am not sure if this is the correct approach.
 
  • #6
Frabjous said:
An ideal gas is compressible which is an inconsistency.

I believe this might be a possible inconsistency. This is the answer I have received from multiple sources. But it still seems to me that I should be able to approximate it as incompressible for small velocities. But this would in-turn mean a small force, so I guess both equations agree approximately
 
  • #7
chandrahas said:
I believe this might be a possible inconsistency. This is the answer I have received from multiple sources. But it still seems to me that I should be able to approximate it as incompressible for small velocities. But this would in-turn mean a small force, so I guess both equations agree approximately
That's not the inconsistency. Unless you have a massive change in altitude or the flow is moving faster than Mach 0.3, you will not have meaningful changes in density and the flow is incompressible.
 
  • #8
chandrahas said:
uation gives us the following equation:

(2)ρvdvdx=−dpdx+ρg

Where ρ is the fluid density, p the fluid pressure, and g the
you have lost ##v_t##
 
  • #9
wrobel said:
you have lost ##v_t##
I have assumed steady state
 
  • #10
I think the problem is that you use an equation of state (ideal gas/fluid) that is inconsistent with the incompressibility assumption, which simply uses the equation of state, ##\rho=\rho_0=\text{const}##. In that case the Navier-Stokes equations for ##\vec{v}=0## simply reads
$$\vec{\nabla} p=\rho_0 \vec{g}=\text{const} \; \Rightarrow\; p=\rho_0 \vec{g} \cdot \vec{x},$$
as it should be.

For a compressible fluid you have an equation of state in addition to the Navier Stokes and continuity equations. For ideal fluids you often get a "polytrope" equation,
$$p=p_0 (\rho/\rho_0)^n,$$
where ##n## depends on the kind of thermodynamic process during the motion of the fluid. For ideal fluid dynamics the change should be taken to be adiabatic. Then ##n=1+2/f##, where ##f=3## for monatomic, ##f=5## for diatomic, and ##f=6## for all other molecules; ##(\rho_0,p_0)## denotes a reference state.

The solution of of the statics equation of motion, which holds true for the compressible case too then is
$$p=p_0 \left (1+\frac{n-1}{n} \frac{\rho_0}{p_0}\vec{g} \cdot \vec{x} \right )^{n/(n-1)}.$$
This also holds in the limit ##n \rightarrow 1##, which refers to isothermal conditions. Then
$$p=p_0 \exp \left (\frac{\rho_0 \vec{g} \cdot \vec{x}}{p_0} \right).$$
From the ideal-gas law,
$$N k T=p V \; \Rightarrow \; p=\frac{\rho k T}{m} \; \rightarrow \frac{\rho}{p} =\frac{m}{k T}=\text{const},$$
and you get the barometric formula for an isothermal atmosphere
$$p=p_0 \exp \left (\frac{m \vec{g} \cdot \vec{x}}{k T} \right)=p_0 \exp[-V(\vec{x})/(k T)],$$
as to be expected from statistical thermodynamics.
 
  • #11
chandrahas said:
I have assumed steady state
that is a source of inconsistency
 
  • #12
I gues the OP meant hydrostatics. Since when is this a source of inconsistency?
 
  • #13
vanhees71 said:
I gues the OP meant hydrostatics. Since when is this a source of inconsistency?
Why must a fluid in the gravity field obey the hydrostatic\steady state hypothesis?

chandrahas said:
Consider the case of a one-dimensional incompressible, non-viscous fluid flowing down a vertical pipe under the influence of gravity.

With the same success, you may plug v=const into the equations of a thrown stone
 
Last edited:
  • #14
wrobel said:
Why must a fluid in the gravity field obey the hydrostatic\steady state hypothesis?
It doesn't need to, but it can. I still don't understand your objection. Another important example of (general-relativistic) hydrostatics is the description of stars. It's an important subtopic of fluid dynamics/kinetic theory.
 
  • #15
My objection is very simple. Consider a dynamical system ##m\dot v=mg##. And that is what OP does. He says assume that our process is steady state: v=const. Then he asks: why do I get an inconsistency? ##0=mg##. What can I say? That's why!

vanhees71 said:
but it can.
in this concrete problem, it can't;

once again:

chandrahas said:
flowing down a vertical pipe under the influence of gravity.
 
  • #16
In the concrete problem you have no static but a stationary flow, i.e., not ##\vec{v}=0## but ##\vec{v}=\text{const}##. By a Galilei boost you can make it a static problem. Of course, here you need in addition the friction, i.e., the viscous in the NS equation. Of course this vanishes in the Galilei boosted frame, vhere ##\vec{v}'=0##. In this sense the problem should be solvable as a hydrostatic problem.
 
  • #17
vanhees71 said:
In the concrete problem you have no static but a stationary flow, i.e., not v→=0 but v→=const. By a Galilei boost you can make it a static problem. Of course, here you need in addition the friction, i.e., the viscous in the NS equation
great! do that for a trivial example: ##m\dot v=-mg-\gamma v,\quad \gamma>0##
 
  • #18
Yes, the trivial example of couse admits a stationary solution, ##v=\text{const}##. With the EoM you get ##v=-mg/\gamma##.
 
  • Like
Likes wrobel
  • #19
Oh, now I see what is going on this thread thanx,
but if OP cannot do the same for the Navier-Stokes it is reasonable to suspect that in the initial problem there are no stationary solutions
 
  • Like
Likes vanhees71

FAQ: Why are the Navier-Stokes equations inconsistent in this case?

What are the Navier-Stokes equations?

The Navier-Stokes equations are a set of partial differential equations that describe the motion of viscous fluid substances, such as liquids and gases. These equations express the balance of forces acting at any given region of the fluid, encompassing aspects like velocity, pressure, density, and viscosity.

Why are the Navier-Stokes equations considered inconsistent in some cases?

The inconsistency arises because the mathematical properties of the Navier-Stokes equations are not fully understood in all scenarios. Specifically, it is not known whether solutions always exist, are unique, and remain smooth (i.e., free of singularities or discontinuities) for all initial and boundary conditions in three dimensions. This lack of understanding can lead to situations where the equations appear to be inconsistent or fail to provide valid physical predictions.

What are singularities in the context of the Navier-Stokes equations?

Singularities refer to points or regions in the fluid where certain properties, like velocity or pressure, become infinite or undefined. These singularities indicate a breakdown in the mathematical description provided by the Navier-Stokes equations, leading to potential inconsistencies in the solutions.

How do numerical methods deal with potential inconsistencies in the Navier-Stokes equations?

Numerical methods often involve discretizing the equations and solving them on a computational grid. While these methods can provide approximate solutions, they may not capture singularities or other subtle mathematical issues. Techniques like regularization, adaptive mesh refinement, and stabilization methods are used to mitigate these problems, but they do not fully resolve the underlying theoretical inconsistencies.

What is the significance of the Navier-Stokes existence and smoothness problem?

The Navier-Stokes existence and smoothness problem is one of the seven Millennium Prize Problems for which the Clay Mathematics Institute has offered a prize of one million dollars. Solving this problem would mean proving or disproving the existence of smooth, globally defined solutions for the Navier-Stokes equations in three dimensions. Resolving this question would have profound implications for both mathematics and fluid dynamics, potentially leading to a deeper understanding of fluid behavior and more accurate predictive models.

Similar threads

Replies
4
Views
2K
Replies
6
Views
896
Replies
6
Views
2K
Replies
5
Views
1K
Replies
1
Views
5K
Back
Top