I'm trying to derive the fully compressible Euler-Momentum equation for the given Lagrangian.

We wish to derive $$\rho\frac{D\boldsymbol{u}}{Dt} + \nabla P + \rho\nabla\phi = 0 $$ from the Lagrangian $$ \mathcal{L} = \rho \left( \frac{|\boldsymbol{u}|^{2}}{2} -e(\rho,s) - \phi \right) $$

I've gotten as far as

$$\frac{\partial}{\partial t}(\rho\boldsymbol{u}) + \nabla\cdot(\rho\boldsymbol{u}\boldsymbol{u}) + \rho\nabla\frac{|\boldsymbol{u}|^{2}}{2} -\rho\nabla \left(\frac{|\boldsymbol{u}|^{2}}{2}-e-\frac{P}{\rho}-\phi \right) - \rho T\nabla s = 0 $$ which has been verified as correct. I've recognized the following relationships, $$\frac{\partial}{\partial t}(\rho\boldsymbol{u}) + \nabla\cdot(\rho\boldsymbol{u}\boldsymbol{u}) = -\nabla P $$ $$\frac{D\boldsymbol{u}}{Dt} = -\nabla \frac{P}{\rho} $$ $$ \rho T\nabla s = \rho\nabla e - \frac{P}{\rho}\nabla\rho $$ which are the Euler-Momentum equation in conservative form, the Euler-Momentum equation, and the 2nd Law of Thermodynamics, respectively.

but after implementing these I arrive at:

$$\frac{\partial}{\partial t}(\rho\boldsymbol{u}) + \nabla\cdot(\rho\boldsymbol{u}\boldsymbol{u}) + \rho\nabla\frac{|\boldsymbol{u}|^{2}}{2} - \rho\nabla\frac{|\boldsymbol{u}|^{2}}{2} + \rho\nabla e + \rho\nabla\frac{P}{\rho} + \rho\nabla\phi - \rho\nabla e + \frac{P}{\rho}\nabla\rho = 0 $$ $$ \Rightarrow -\nabla P - \rho \frac{D\boldsymbol{u}}{Dt} + \rho\nabla\phi + \frac{P}{\rho}\nabla\rho = 0 $$ which is "close" but incorrect. Where am I going wrong here?

  • $\begingroup$ Welcome to Physics SE! Please do not post formulae as images, but use MathJax instead. $\endgroup$ Feb 22, 2021 at 2:43
  • 2
    $\begingroup$ @NiharKarve Thanks! Roger that. $\endgroup$ Feb 22, 2021 at 3:12
  • $\begingroup$ Did you derive $\frac{D\boldsymbol{u}}{Dt} = -\nabla \frac{P}{\rho}$ from the equation above it? If so, it is wrong and should be $\frac{D\boldsymbol{u}}{Dt} = -\frac{1}{\rho}\nabla P$ unless the flow is incompressible. Given the parts which are deviating i suspect this will clear up the incorrect terms $\endgroup$
    – nluigi
    Feb 22, 2021 at 21:12
  • $\begingroup$ @nluigi I did not, got that from the definition of the nonlinear Eularian momentum eqn. $\endgroup$ Feb 23, 2021 at 1:18
  • $\begingroup$ Which variational principle do you use to go from the second equation (the Lagrangian density) to the third one (your result)? $\endgroup$
    – Quillo
    Nov 24, 2021 at 11:02

1 Answer 1


It works out just fine

Starting from your equation: $$\frac{\partial}{\partial t}(\rho\boldsymbol{u})+\nabla\cdot(\rho\boldsymbol{u}\boldsymbol{u})+\rho\nabla\frac{|\boldsymbol{u}|^{2}}{2}-\rho\nabla\left(\frac{|\boldsymbol{u}|^{2}}{2}-e-\frac{P}{\rho}-\phi\right)-\rho T\nabla s=0$$ it can be simplified: $$\frac{\partial}{\partial t}(\rho\boldsymbol{u})+\nabla\cdot(\rho\boldsymbol{u}\boldsymbol{u})+\rho\nabla\left(e+\frac{P}{\rho}+\phi\right)-\rho T\nabla s=0$$ by canceling kinetic energy terms.

Then applying the 2nd law of thermodynamics: $$\rho T\nabla s=\rho\nabla e-\frac{P}{\rho}\nabla\rho$$ substituting, cancelling and rearranging we get: $$\frac{\partial}{\partial t}(\rho\boldsymbol{u})+\nabla\cdot(\rho\boldsymbol{u}\boldsymbol{u})+\rho\nabla\frac{P}{\rho}+\frac{P}{\rho}\nabla\rho+\rho\nabla\phi=0$$

The pressure terms can be combined using the product rule: $$\nabla P = \nabla \left[\rho\frac{P}{\rho}\right]=\rho\nabla\frac{P}{\rho}+\frac{P}{\rho}\nabla\rho$$ to yield: $$\frac{\partial}{\partial t}(\rho\boldsymbol{u})+\nabla\cdot(\rho\boldsymbol{u}\boldsymbol{u})+\nabla P+\rho\nabla\phi=0$$

Then using the continuity equation: $$\frac{\partial\rho}{\partial t}+\nabla\cdot(\rho\boldsymbol{u})=0$$ we simplify the first two terms on the left: $$\frac{\partial}{\partial t}(\rho\boldsymbol{u})+\nabla\cdot(\rho\boldsymbol{u}\boldsymbol{u})=\rho\left[\frac{\partial\boldsymbol{u}}{\partial t}+\boldsymbol{u}\nabla\cdot\boldsymbol{u}\right]+\left[\frac{\partial\rho}{\partial t}+\nabla\cdot(\rho\boldsymbol{u})\right]\boldsymbol{u}=\rho\frac{D\boldsymbol{u}}{Dt}$$

Finally we get to what you wished to derive: $$\rho\frac{D\boldsymbol{u}}{Dt}+\nabla P+\rho\nabla\phi=0$$

  • $\begingroup$ Oh perfect, thank you! $\endgroup$ Feb 24, 2021 at 22:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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