Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

docs: fix equation errors #2631

Merged
merged 1 commit into from
Feb 6, 2024
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
39 changes: 22 additions & 17 deletions docs/src/edmf_equations.md
Original file line number Diff line number Diff line change
Expand Up @@ -60,10 +60,10 @@ We make use of the following operators

## Auxiliary and derived quantities

* ``\tilde{\boldsymbol{u}^j}`` is the mass-weighted reconstruction of velocity at the interfaces:
* ``\tilde{\boldsymbol{u}}^j`` is the mass-weighted reconstruction of velocity at the interfaces:
by interpolation of contravariant components
```math
\tilde{\boldsymbol{u}^j} = WI^f(\rho^j J, \boldsymbol{u}_h) + \boldsymbol{u}_v^j
\tilde{\boldsymbol{u}}^j = WI^f \left( \rho^j J, \boldsymbol{u}_h \right) + \boldsymbol{u}_v^j.
```
Technically, from mass conservation, the weighting factor should be ``\hat{\rho}^j J``.
However, in order to avoid issues coming from close to zero sub-domain area fractions,
Expand All @@ -72,24 +72,24 @@ we can instead use ``\rho^j J`` or even ``\rho J``.
* ``\bar{\boldsymbol{u}}^j`` is the reconstruction of velocity at cell-centers,
carried out by linear interpolation of the covariant vertical component:
```math
\bar{\boldsymbol{u}}^j = \boldsymbol{u}_h + I_{c}(\boldsymbol{u}_v^j)
\bar{\boldsymbol{u}}^j = \boldsymbol{u}_h + I_{c}(\boldsymbol{u}_v^j),
```

* ``\boldsymbol{b}^j`` is the reduced gravitational acceleration
```math
\boldsymbol{b}^j = - \frac{\rho^j - \rho_{\text{ref}}}{\rho^j} \nabla \Phi
\boldsymbol{b}^j = - \frac{\rho^j - \rho_{\text{ref}}}{\rho^j} \nabla \Phi,
```
* ``K^j = \tfrac{1}{2} \|\boldsymbol{u}^j\|^2 `` is the specific kinetic energy (J/kg), reconstructed at cell centers by
```math
K^j = \tfrac{1}{2} (\boldsymbol{u}_{h}^j \cdot \boldsymbol{u}_{h}^j + 2 \boldsymbol{u}_{h}^j \cdot I_{c} (\boldsymbol{u}_{v}^j) + I_{c}(\boldsymbol{u}_{v}^j \cdot \boldsymbol{u}_{v}^j)),
K^j = \tfrac{1}{2} \left(\boldsymbol{u}_{h}^j \cdot \boldsymbol{u}_{h}^j + 2 \boldsymbol{u}_{h}^j \cdot I_{c} (\boldsymbol{u}_{v}^j) + I_{c}(\boldsymbol{u}_{v}^j \cdot \boldsymbol{u}_{v}^j) \right),
```
where ``\boldsymbol{u}_{h}^j`` is defined on cell-centers, ``\boldsymbol{u}_{v}^j`` is defined on cell-faces, and ``I_{c} (\boldsymbol{u}_{v})`` is interpolated using covariant components.

* ``\nu_u``, ``\nu_h``, and ``\nu_\chi`` are hyperdiffusion coefficients, and ``c`` is the divergence damping factor.

* No-flux boundary conditions are enforced by requiring the third contravariant component of the face-valued velocity at the boundary, ``\boldsymbol{\tilde{u}}^{v,j}``, to be zero. The vertical covariant velocity component is computed as
```math
\tilde{u}_{v}^j = - \tfrac{u_{1}g^{31} + u_{2}g^{32}}{g^{33}}.
\tilde{u}_{v}^j = - \frac{u_{1}g^{31} + u_{2}g^{32}}{g^{33}}.
```

## Equations and discretizations
Expand All @@ -98,20 +98,20 @@ we can instead use ``\rho^j J`` or even ``\rho J``.

Follows the continuity equation
```math
\frac{\partial}{\partial t} \hat{\rho}^j = - \nabla \cdot (\hat{\rho}^j \boldsymbol{u}^j) + RHS
\frac{\partial}{\partial t} \hat{\rho}^j = - \nabla \cdot (\hat{\rho}^j \boldsymbol{u}^j) + RHS.
```

This is discretized using the following
```math
\frac{\partial}{\partial t} \hat{\rho}^j
= - D_h[ \hat{\rho}^j (\boldsymbol{u}_h + I^c(\boldsymbol{u}_v^j))] - D^c_v \left[WI^f( J, \hat{\rho}^j) \tilde{\boldsymbol{u}^j} \right] + RHS
= - D_h \left[ \hat{\rho}^j (\boldsymbol{u}_h + I^c(\boldsymbol{u}_v^j)) \right] - D^c_v \left[WI^f( J, \hat{\rho}^j) \tilde{\boldsymbol{u}^j} \right] + RHS.
```

### Momentum

Uses the advective form equation
```math
\frac{\partial}{\partial t} \boldsymbol{u}^j = - (2 \boldsymbol{\Omega} + \nabla \times \boldsymbol{u}^j) \times \boldsymbol{u}^j - \frac{1}{\rho^j} \nabla (p - p_{\text{ref}}) + \boldsymbol{b}^j - \nabla K^j + RHS
\frac{\partial}{\partial t} \boldsymbol{u}^j = - (2 \boldsymbol{\Omega} + \nabla \times \boldsymbol{u}^j) \times \boldsymbol{u}^j - \frac{1}{\rho^j} \nabla (p - p_{\text{ref}}) + \boldsymbol{b}^j - \nabla K^j + RHS.
```
By breaking the curl and cross product terms into horizontal and vertical contributions, and removing zero terms (e.g. ``\nabla_v \times \boldsymbol{u}_v = 0``), we obtain
the vertical momentum equation. The horizontal momentum equation is only solved in the grid-mean.
Expand All @@ -133,18 +133,18 @@ The ``(\nabla_v \times \boldsymbol{u}_h + \nabla_h \times \boldsymbol{u}_v^j) \t
```math
(C^f_v[\boldsymbol{u}_h] + C_h[\boldsymbol{u}_v^j]) \times I^f(\boldsymbol{u}^h) ,
```
and the ``-\frac{1}{\rho^j} \nabla_v (p - p_{\text{ref}}) - \frac{\rho^j - \rho_{\text{ref}}}{\rho^j} \nabla_v \Phi - \nabla_v K^j`` term as
and the ``-\frac{1}{\rho^j} \nabla_v (p - p_{\text{ref}}) - \frac{\rho^j - \rho_{\text{ref}}}{\rho^j} \nabla_v \Phi - \nabla_v K^j`` terms as
```math
-\frac{1}{I^f(\rho^j)} G^f_v[p - p_{\text{ref}}] - \frac{I^f(\rho^j - \rho_{\text{ref}})}{I^f(\rho^j)} G^f_v[\Phi] - G^f_v[K^j] ,
```

The hyperviscosity term is
```math
- \nu_u \hat{\mathcal{D}}_h (\mathcal{G}_h (\psi) )
- \nu_u \hat{\mathcal{D}}_h (\mathcal{G}_h (\psi) ),
```
where
```math
\psi = \mathcal{P} \left[ \hat{\mathcal{D}}_h \left( \mathcal{G}_h (w^j)\right) \right]
\psi = \mathcal{P} \left[ \hat{\mathcal{D}}_h \left( \mathcal{G}_h (w^j)\right) \right].
```

### Total energy
Expand All @@ -154,15 +154,20 @@ where
```
which is stabilized with the addition of a 4th-order hyperdiffusion term on total enthalpy:
```math
- \nu_h \nabla \cdot \left( \hat{\rho}^j \nabla^3 \left(\frac{\rho^j e^j + p}{\rho^j} \right)\right)
- \nu_h \nabla \cdot \left( \hat{\rho}^j \nabla^3 \left(\frac{\rho^j e^j + p}{\rho^j} \right)\right).
```

is discretized using
The equation is discretized as
```math
\frac{\partial}{\partial t} \hat{\rho}^j e^j \approx
- D_h[ (\hat{\rho^j} e^j + \frac{\hat{\rho^j}}{\rho^j}p) (\boldsymbol{u}_h + I^c(\boldsymbol{u}_v^j))]
- D^c_v \left[ WI^f(J,\hat{\rho}^j) \, \tilde{\boldsymbol{u}}^j \, I^f \left(\frac{\hat{\rho^j} e^j + \frac{\hat{\rho^j}}{\rho^j}p}{\hat{\rho}^j} \right)
\right] - \frac{p}{\rho} \frac{\partial}{\partial t} \hat{\rho}^j - \nu_h \hat{\mathcal{D}}_h( \rho \mathcal{G}_h(\psi^j) ) + RHS .
- D_h \left[
\left( \hat{\rho}^j e^j + \frac{\hat{\rho}^j}{\rho^j}p \right)
\left( \boldsymbol{u}_h + I^c(\boldsymbol{u}_v^j) \right)
\right]
- D^c_v \left[
WI^f(J,\hat{\rho}^j) \, \tilde{\boldsymbol{u}}^j \, I^f \left(\frac{\hat{\rho}^j e^j + \frac{\hat{\rho}^j}{\rho^j}p}{\hat{\rho}^j} \right)
\right]
- \frac{p}{\rho} \frac{\partial}{\partial t} \hat{\rho}^j - \nu_h \hat{\mathcal{D}}_h( \rho \mathcal{G}_h(\psi^j) ) + RHS .
```
where
```math
Expand Down
6 changes: 3 additions & 3 deletions docs/src/equations.md
Original file line number Diff line number Diff line change
Expand Up @@ -137,9 +137,9 @@ where ``\boldsymbol{u}^h`` and ``\boldsymbol{u}^v`` are the horizontal and verti

This is stabilized with the addition of 4th-order vector hyperviscosity
```math
-\nu_u \, \nabla_h^2 (\nabla_h^2(\boldsymbol{\overbar{u}})),
-\nu_u \, \nabla_h^2 (\nabla_h^2(\boldsymbol{\overline{u}})),
```
projected onto the first two contravariant directions, where ``\nabla_{h}^2(\boldsymbol{v})`` is the horizontal vector Laplacian. For grid scale hyperdiffusion, ``\boldsymbol{v}`` is identical to ``\boldsymbol{\overbar{u}}``, the cell-center valued velocity vector.
projected onto the first two contravariant directions, where ``\nabla_{h}^2(\boldsymbol{v})`` is the horizontal vector Laplacian. For grid scale hyperdiffusion, ``\boldsymbol{v}`` is identical to ``\boldsymbol{\overline{u}}``, the cell-center valued velocity vector.
```math
\nabla_h^2(\boldsymbol{v}) = \nabla_h(\nabla_{h} \cdot \boldsymbol{v}) - \nabla_{h} \times (\nabla_{h} \times \boldsymbol{v}).
```
Expand Down Expand Up @@ -192,7 +192,7 @@ with the latter treated implicitly.

This is stabilized with the addition of 4th-order vector hyperviscosity
```math
-\nu_u \, \nabla_h^2 (\nabla_h^2(\boldsymbol{\overbar{u}})),
-\nu_u \, \nabla_h^2 (\nabla_h^2(\boldsymbol{\overline{u}})),
```
projected onto the third contravariant direction.

Expand Down
Loading