Fluidsim: Animated inflow (and outflow) never worked #30605

New Issue

Daniel Genrich · 2012-03-20T14:59:26+01:00

Daniel Genrich commented

2012-03-20 14:59:26 +01:00

%%%I'm just reporting this so this issue doesn't get lost. Was initially reported by Daniel Salazar (ZanQdo).

Problem:

Basicly if an inflow got more than 1 keyframe and "Volume initialisation" is set to either "VOLUME" or "BOTH", only the shell is initialized and only from the shell is new inflow generated.

My bug hunting outcome so far:

The relevant code seems to be in solver_init.cpp (blender/intern/elbeem/intern).

a) Whenever a fluid inflow object got more than 1 keyframe, the function "getIsAnimated()" returns "true". If an inflow is detected to be animated, another set of code is used to initialize the object.

b) The interesting functions to investigate seem to be "LbmFsgrSolver::initGeometryFlags()" and "LbmFsgrSolver::initMovingObstacles(bool staticInit)". Look for "FGI_MBNDINFLOW".

c) When I comment line 1871 in solver_init.cpp "ntype = CFInvalid;" and change it to "rhomass = 1.0; ntype = CFFluid;", the inflow gets initialized correctly BUT this does not fix the problem: New inflow is only generated from the shell. So basicly this "fix" only treats 1 symptom, not the real problem. BUT it also means that the problem + solution is in those functions (I think!).

d) To see how the correct initialisation should look like, you can delete the second keyframe in "AnimInflow-bad-active2.blend".

I am also stuck at this point, if someone else knows what to do, feel free to try your luck. A fresh look can never hurt.

Thanks
Daniel / Genscher
%%%

%%%I'm just reporting this so this issue doesn't get lost. Was initially reported by Daniel Salazar (ZanQdo). Problem: ------------------------ Basicly if an inflow got more than 1 keyframe and "Volume initialisation" is set to either "VOLUME" or "BOTH", only the shell is initialized and only from the shell is new inflow generated. My bug hunting outcome so far: ------------------------------------------------------ 0) The relevant code seems to be in solver_init.cpp (blender/intern/elbeem/intern). a) Whenever a fluid inflow object got more than 1 keyframe, the function "getIsAnimated()" returns "true". If an inflow is detected to be animated, another set of code is used to initialize the object. b) The interesting functions to investigate seem to be "LbmFsgrSolver::initGeometryFlags()" and "LbmFsgrSolver::initMovingObstacles(bool staticInit)". Look for "FGI_MBNDINFLOW". c) When I comment line 1871 in solver_init.cpp "ntype = CFInvalid;" and change it to "rhomass = 1.0; ntype = CFFluid;", the inflow gets initialized correctly BUT this does not fix the problem: New inflow is only generated from the shell. So basicly this "fix" only treats 1 symptom, not the real problem. BUT it also means that the problem + solution is in those functions (I *think*!). d) To see how the correct initialisation should look like, you can delete the second keyframe in "AnimInflow-bad-active2.blend". ----------------------------------------------------- I am also stuck at this point, if someone else knows what to do, feel free to try your luck. A fresh look can never hurt. Thanks Daniel / Genscher %%%

Daniel Genrich commented

2012-03-20 14:59:26 +01:00

Changed status to: 'Open'

Ton Roosendaal commented

2012-03-21 16:36:25 +01:00

%%%I would only add bugs in the tracker when it's an open issue; i.e. our team has it on the todo to fix it.

If you like to create a todo item for future reference, put this in the wiki?
http://wiki.blender.org/index.php/Dev:2.5/Source/Development/Todo/Tools#Fluid_Simulation%%%

%%%I would only add bugs in the tracker when it's an open issue; i.e. our team has it on the todo to fix it. If you like to create a todo item for future reference, put this in the wiki? http://wiki.blender.org/index.php/Dev:2.5/Source/Development/Todo/Tools#Fluid_Simulation%%%

Daniel Genrich commented

2012-03-21 17:03:16 +01:00

%%%Hey,

where is the difference between a TODO-item and a bug? Since the object moves and spits out fluid, it basicly works. But it's buggy. So I guess thought this is a valid bug which nobody fixed but should be fixed since moving inflows are pretty useless so far.%%%

%%%Hey, where is the difference between a TODO-item and a bug? Since the object moves and spits out fluid, it basicly works. But it's buggy. So I guess thought this is a valid bug which nobody fixed but should be fixed since moving inflows are pretty useless so far.%%%

Sergey Sharybin commented

2012-03-23 09:07:30 +01:00

%%%Basically in this case bug means we've got a developer who's familiar with code in which some crap happens and able to fix that crappiness. In case of simulations.. We don't have active developer here. If you think you can look into the issue and propose some fix, feel free to look into it, otherwise i'd prefer this issue be moved to TODO list.%%%

Sergey Sharybin commented

2012-03-25 18:09:51 +02:00

%%%Added not to a TODO wiki there http://wiki.blender.org/index.php/Dev:2.5/Source/Development/Todo/Tools#Fluid_Simulation%%%

Sergey Sharybin commented

2012-03-25 18:09:51 +02:00

Changed status from 'Open' to: 'Archived'

Sergey Sharybin closed this issue

2012-03-25 18:09:51 +02:00

Sign in to join this conversation.

No Label

Download

What's New

Blender Studio

Manual

Developers Blog

Documentation

Benchmark

Blender Conference

Development Fund

One-time Donations

Fluidsim: Animated inflow (and outflow) never worked #30605

Problem:

My bug hunting outcome so far: