Merge remote-tracking branch 'origin/main' into mfournier01/tracer

Author: pgrete

.github/workflows/ci.yml

@@ -1,6 +1,10 @@
 name: CI
 
-on: [pull_request]
+on:
+  # when triggered manually
+  workflow_dispatch:
+  # when auto merge is enabled (hack to make sure it's run before merging)
+  pull_request:
 
 # Cancel "duplicated" workflows triggered by pushes to internal
 # branches with associated PRs.

CHANGELOG.md

@@ -20,12 +20,14 @@ To enable, set `do_coalesced_comms=true` in the `<parthenon/mesh>` block of the
 by adding an optional string as last argument to any `ParameterInput` `Get` or `GetOrAdd` call.
 
 ### Added (new features/APIs/variables/...)
+- [[PR 157]](https://github.com/parthenon-hpc-lab/athenapk/pull/157) Support injection of blobs with density/temp contrast in turbulence simulations
 
 ### Changed (changing behavior/API/variables/...)
 
 ### Fixed (not changing behavior/API/variables/...)
 
 ### Infrastructure
+- [[PR 149]](https://github.com/parthenon-hpc-lab/athenapk/pull/149) Allow triggering of pipelines manually
 - [[PR 156]](https://github.com/parthenon-hpc-lab/athenapk/pull/156) Bump formatters to clang-format-20 and black 25.12
 - [[PR 146]](https://github.com/parthenon-hpc-lab/athenapk/pull/146) Bump Parthenon 25.12 and Kokkos 4.7.02
 

docs/turbulence.md

@@ -88,6 +88,64 @@ in the input file.
 Alternatively, wavemodes can be chosen/defined manually, e.g., if not all wavemodes are desired or
 only individual modes should be forced.
 
+### Blob injection
+
+In order to study the evolution of (cold) clouds in a turbulent environment,
+blobs can be injected to the simulation.
+The injection is a one off mechanism controlled by the following parameters
+
+```
+<problem/turbulence>
+# only one of the following three conditions can be set at a given time
+inject_once_at_time = -1.0
+inject_once_at_cycle = -1
+inject_once_on_restart = false
+inject_n_blobs = -1 # number of blob to inject
+# then for the given number of blobs follow parameters need to be given (starting to count with 0)
+inject_blob_radius_0 = ... # float, in code length units, no default value
+inject_blob_loc_0 = ...,...,... # location vector of three comma-separated floats, in code length units, no default value
+inject_blob_chi_0 = ... # float, dimensionless, no default value, density ratio to existing value
+inject_blob_radius_1 = ...
+...
+```
+
+In practice, this will result in blobs being seeded at a given time, cycle, or upon restart
+by adjusting the density within the blob's radius by a factor of $\chi$ and
+at the same time adjusting the temperature by a factor of $1/\chi$ so that the
+blob remain in pressure equilibrium.
+
+While this is an action that is performed once, it can be repeated upon restart (or a later
+time) by resetting the variables.
+
+A current restriction is that the blobs cannot be seeded across a domain boundary (i.e.,
+the periodicity of the box is not taken into account).
+
+### Rescaling **not recommended*
+
+*The rescaling described in the following is generally not recommended, as it result in a
+state that is not naturally reached.
+Moreover, given the artificial nature of a hard reset, some time after the rescaling is
+required for the system to readjust.
+
+For non-isothermal simulations, the plasma will eventually heat up over time due to dissipation.
+One possibility to remove that extra heat (or add heat), is to rescale the temperature in the simulation.
+This can be done via the following parameters:
+
+```
+<problem/turbulence>
+# only one of the following three conditions can be set at a given time
+rescale_once_at_time = -1.0
+rescale_once_at_cycle = -1
+rescale_once_on_restart = false
+rescale_to_rms_Ms = -1.0
+```
+
+As the parameters suggest, rescaling is a one off action (though it can be repeated when
+the parameters are set again for a subsequent restart).
+The density and velocity field are not changed, only the specific internal energy is
+adjusted so that the volume-weighted root mean squared sonic Mach number matches
+the target value.
+
 ## Typical results
 
 The results shown here are obtained from running simulations with the parameters given in the next section.