Error use multiple time dependent boundary conditions

I am trying to simulate a time dependent heat transfer with time dependent heat sources.  The time dependent heat sources are at boundaries 36, 39, 45.



The simulation does not have any issues if the heat sources are fixed values, time independent, at these boundaries.

The simulation does not have any issues if the heat source at boundary 36 is time dependent but causes an error if any of the other boundaries (39, 45) have a time dependent source.

The error that the tool runs into is: "Deletion requires an existing variable"

Any ideas?

Thanks.

https://forum.featool.com/IMPORTANT-Issue-and-bug-reporting-guidelines-tp1735.html

^ Please include the model file(s) and system information with all technical issues.

Hi,
Additional information.

System Info:
App Id: FEATool Multiphysics
App Version: 1.17.2 Build 24.11.332
App License: PDE0DDK17299QDC16187B
App Path: C:\Users\pthia\AppData\Roaming\MathWorks\MATLAB Add-Ons\Apps\FEAToolMultiphysics
App Data: C:\Users\pthia\OneDrive\Documents\MATLAB\.featool
User Home: C:\Users\pthia
Runtime: R2024a en_US
System: PCWIN64-L
OS: Win 11
Id: C4C04F4E5A33 (31)
Cpu: 13th Gen Intel(R) Core(TM) i9-13900F
Geom Engine: geomtool 2.0.4
Triangle Binary: C:\Users\pthia\OneDrive\Documents\MATLAB\.featool\triangle.exe
FEniCS Version: Not found. Could not detect any Python installation.
OpenFOAM: Not found
SU2 Binary: C:\Users\pthia\AppData\Roaming\MathWorks\MATLAB Add-Ons\Apps\FEAToolMultiphysics\lib\su2\SU2_CFD.exe
SU2 Version: v7.5.1 Blackbird

Matlab m-script attached

Steps to create error:
Modify boundary conditions in lines  120 and/or lines 121 to:
fea.bdr.n{1,1}{38} = '(rho_ht*cp_ht*(nx*u_ht+ny*v_ht)*T)+(20*(t<300e-6))';
fea.bdr.n{1,1}{45} = '(rho_ht*cp_ht*(nx*u_ht+ny*v_ht)*T)+(20*(t<300e-6))';

Error information:
Deletion requires an existing variable.

Error in featool

Error in featool

Error in featool

Error in featool

Error in solvetime (line 44)
if ~(nargin||nargout),help mfilename,return,end,varargout=cell(1,nargout);[varargout{:}]=featool('feval','solvetime',varargin{:});if ~nargout,clear varargout,end

Error in Multi_Bar_Thermal (line 123)
[fea.sol.u,fea.sol.t] = solvetime( fea, ...

Thanks.Multi_Bar_Thermal.m

Thank you for submitting the additional info. It seems to be an internal issue with how multiple subdomains are handled. Please try to use the bugfix version linked below until the next official release:

https://drive.google.com/file/d/1npcg9kvlejXPLp-88ZoyZekZNgO_4-w5/view?usp=sharing

Thank you for fixing the boundary condition issue.  The solver works without any errors now.

However the solver seems to terminate prematurely.  The final solve time in the model was t=0.01 but looks like the solver terminates at t=0.00085.  Here is the end of the log:
 
---------------------------------
 Time step   86 at time 0.00085   |
---------------------------------
    it |    def-T  |     rel-T  |
---------------------------------
     1 |  7.1e-09  |     1e-06  |
---------------------------------
 t-chg             |     1e-06  |
---------------------------------
 
--------------------------------------------------
          Solver Statistics (solvetime)
--------------------------------------------------
 Number of dependent variables
--------------------------------------------------
 T : 9579
--------------------------------------------------
 Simulation timings
--------------------------------------------------
 t-asm_m : 4.1e-02 \   0%
 t-asm_a : 1.3e-02 \   0%
 t-asm_f : 6.1e-02 \   0%
 t-bdr   : 6.2e-01 \   6%
 t-lc_mv : 9.2e-02 \   0%
 t-sol   : 7.1e+00 \  68%
 t-tot   : 1.0e+01 \ 100%
--------------------------------------------------

Do you know what could be causing this?

Probably the "tstop" parameter that detects no or small changes. Try reducing it.

https://github.com/precise-simulation/featool-multiphysics/blob/master/solvetime.m#L12-L13

Thank you.  This fixed it.

Hi,
With the updated simulator release I am seeing issues with the solution.  I have attached all the relevant files.

When I run the simulation with a time dependent heat source (ie., 20*(t<300e-6)) the solution does not look right.  See the screen shot of the output (Screenshot_Single_Step.png)

So I split the solution into 2 steps: first step has a fixed heat source (ie., 20) for tmax = 300e-6 and use the output as the initial condition for the second step that has a heat source set to 0 for tmax = 0.01.  The output (Screenshot_Two_Step.png) from this looks as what I would expect.  

So it looks like the time dependent heatsource is not set properly in the simulation.

Can you kindly look at this issue.

Thanks.

Multi_Bar_Thermal_Single_Step.m
Multi_Bar_Thermal_Two_Step.m
Screenshot_Single_Step_.png
Screenshot_Two_Step_.png

Thank you for reporting the issue, and sorry for the delayed reply due to the holiday break.

This specifically seems to be due to the expression parser erroneously removing the parenthesis around the "20*(t<300e-6)" expression. This issue will be fixed in the next update/release. In the meantime you can either modify the expression and add a "(...)^1" to prevent this, that is "20*(t<300e-6)^1". Or as you already have found out split the solving in several steps.

Thank you for looking into this and providing a work around.

Please let me know when an updated release is available.

The fix should now be available in version 1.17.3.