Seperate adapt_sources for density and pressure terms #8
Conversation
hermes-2.cxx
Outdated
| @@ -878,6 +899,10 @@ int Hermes::rhs(BoutReal t) { | |||
| sheath_model = 0; | |||
| } | |||
|
|
|||
| if (ramp_j_diamag < 1.0) { | |||
| ramp_j_diamag = ramp_j_diamag_generator->generate(0, 0, 0, 20 * t/NOUT); | |||
There was a problem hiding this comment.
Here t will be the simulation time, not the output step; NOUT is the number of steps, but t will go to NOUT * TIMESTEP. I'd prefer to just remove NOUT, since any normalisation like this can be done in the input file expression. I suggest here just have generate(0, 0, 0, t); and then change the expression in BOUT.inp.
hermes-2.cxx
Outdated
|
|
||
| // Options::root()["mesh"]["paralleltransform"].as<std::string>() | ||
| // TIMESTEP = Options::root()["TIMESTEP"].as<BoutReal>() * 1e-6; | ||
| NOUT = Options::root()["NOUT"].as<BoutReal>(); |
There was a problem hiding this comment.
I think this (NOUT) can be removed; NOUT (and TIMESTEP) can be used in BOUT.inp expressions if needed.
hermes-2.hxx
Outdated
|
|
||
|
|
||
| BoutReal TIMESTEP; | ||
| BoutReal NOUT; |
There was a problem hiding this comment.
Here TIMESTEP and NOUT can be removed
makefile
Outdated
| @@ -1,6 +1,8 @@ | |||
|
|
|||
| BOUT_TOP = ../.. | |||
|
|
|||
| BOUT_GIT_COMMIT:=$(shell git --git-dir=$(BOUT_TOP)/.git rev-parse --short HEAD >> BOUT_commit) | |||
There was a problem hiding this comment.
It looks like this isn't used anywhere; can you see how STORM gets the commit into the executable?
There was a problem hiding this comment.
In STORM we have
CXXFLAGS += -DSTORM_VERSION=\"$(GIT_VERSION)\"
in the makefile.
Although this looks like it's getting the commit for BOUT rather than HERMES, but BOUT saves it's commit to the log and dump files already.
I'd suggest
| BOUT_GIT_COMMIT:=$(shell git --git-dir=$(BOUT_TOP)/.git rev-parse --short HEAD >> BOUT_commit) | |
| GIT_VERSION := $(shell git describe --abbrev=40 --dirty --always --tags) | |
| CXXFLAGS += -DHERMES_VERSION=\"$(GIT_VERSION)\" |
or similar (depending how you like your git hash to be formatted).
hermes-2.cxx
Outdated
| if (slab_radial_buffers || radial_buffers) { | ||
| // Need to set the sources in the radial buffer regions to zero | ||
|
|
||
| if ((mesh->getGlobalXIndex(mesh->xstart) - mesh->xstart) < radial_inner_width) { |
There was a problem hiding this comment.
This only does the inner buffer region; do we need something for the outer region too?
|
Hi Ben, looks like this branch is pulling from |
|
Hi @johnomotani sorry no I was a bit naughty and pushed changes to master... see the last few commits |
|
I've added a rule to protect |
…lag used to add it to the radiaiton sink
…me of the numerical diffusion terms to be variable instead of just 1; also have re-introduced slab_radial_buffers because the 3D sims seem to require this ... need to figure out why more thorughly
…tes and included the option to reverse the Curl(b/B) vector. Also had some small code tidy ups
…plied in a weird finite volume way); also minor fix to the split impurity fraction - wrong name was used before
…splits in further dimensions
…ence delp2 dissipation terms in the radial buffers - delp2 operation outputs a field3D so way too expensive to keep it in a loop
Want to alter behaviour of the PI controller to ensure a constant power entering the SOL when performing sim scans over density