I would like to have an equation that says "The spatial change of the linear momentum vector is equal to the sum of these terms...", or written as an equation: $$\mathrm{grad}\left(\rho \mathbf{u} \right) = \mathbf{A} + \mathbf{B} + \mathbf{C} + \dots$$ where $\mathbf{A}$, $\mathbf{B}$, $\mathbf{C}$ (and so on) are terms that depend on density $\rho$, fluid velocity $\mathbf{u}$, pressure $p$ and/or pressure gradient $\mathrm{grad}\left(p\right)$, and the stress tensor $\mathbf{\tau}$, its gradient or divergence...

I have all the quantities, they are a result of a computational fluid dynamics simulation. But I want to show that, for example, $\partial\,\rho u_i / \partial\,x_j$ (i.e. one component of the linear momentum vector gradient) is the result of (for example) the increase in pressure (positive pressure gradient) or something similar.

This is how far I have come while trying to solve this problem:

Start with the total differential of the linear momentum vector $\rho \mathbf{u}$ (material derivative): $$\frac{\mathrm{d}\,\rho \mathbf{u}}{\mathrm{d}\,t} = \underbrace{\frac{\partial \, \rho \mathbf{u}}{\partial \, t}}_{=0} + \mathrm{grad}\left(\rho \mathbf{u} \right) \cdot \mathbf{u}$$ The total differential $\mathrm{d}\,\rho \mathbf{u}/\mathrm{d}\,t$ is the infinitesimal change in linear momentum as the fluid particle travels an infinitesimal distance in the direction of its velocity vector $\mathbf{u}$.

Next, look at the local form of the (steady state) linear momentum balance in the Navier-Stokes-Equations: $$\mathrm{div}\left(\rho \mathbf{u} \otimes \mathbf{u} \right) = \mathrm{grad}\left(\rho \mathbf{u} \right) \cdot \mathbf{u} + \rho \mathbf{u} \, \mathrm{div}\left(\mathbf{u}\right) = -\mathrm{grad}\left(p\right) + \mathrm{div}\left(\tau\right)$$ Solve for $\mathrm{grad}\left(\rho \mathbf{u} \right) \cdot \mathbf{u}$ and thus $$\frac{\mathrm{d}\,\rho \mathbf{u}}{\mathrm{d}\,t} = \mathrm{grad}\left(\rho \mathbf{u} \right) \cdot \mathbf{u} = -\rho \mathbf{u}\, \mathrm{div}\left(\mathbf{u}\right) -\mathrm{grad}\left(p\right) + \mathrm{div}\left(\tau\right)$$

This explains the change of the linear momentum vector along its direction of travel, but not along a chosen direction (i.e. $x_1$, $x_2$ or $x_3$ ...). These are three equations, but I need 9: The linear momentum gradient has 9 components and is not necessarily symmetric.

Does someone know how to solve this?


1 Answer 1


Nevermind, i figured it out myself. It was just an error in reasoning.

Taking the linear momentum in the $x_1$ direction as an example: $$\frac{\partial\,\rho u_1}{\partial\,x_1} u_1 = - \frac{\partial\,\rho u_1}{\partial\,x_2} u_2 - \frac{\partial\,\rho u_1}{\partial\,x_3} u_3 - \rho u_1 \mathrm{div}\left(\mathbf{u}\right) - \frac{\partial\,p}{\partial\,x_1} + \sum_{i=1}^3 \frac{\partial\,\tau_{1i}}{\partial\,x_i}$$ Linear momentum in $x_1$ direction can be

  1. converted/redirected into linear momentum in another direction ($u_2 \partial\,\rho u_1/\partial\,x_2 \neq 0$ and/or $u_3 \partial\,\rho u_1/\partial\,x_3\neq 0$)
  2. "stored" or "released" when the divergence of the velocity field is positive or negative ($\rho u_1 \mathrm{div}\left(\mathbf{u}\right)\neq 0$)
  3. increased or decreased by a pressure gradient ($\frac{\partial\,p}{\partial\,x_1} \neq 0$)
  4. increased or decreased by stresses ($\sum_{i=1}^3 \frac{\partial\,\tau_{1i}}{\partial\,x_i}\neq0$)

Thus, there are only three equations that can be rearranged in 9 ways to give the desired information.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.