User Forum of Software BASEMENT

chri_and

User

Registered: 2017-07-27

Posts: 34

simulation does not work mit Morphology Block

Hi

I'm running Basement V3. The simulation for the initial conditions (same Q over time) and the following one with the hydrograph (Initial = continue) work fine. When I introduce the morphology block I get a mass balance problem and the the delta time gets very large, the simulation ends witihin a few seconds and after the first step the results are all NaNs.

How can I solve this?

Cheers

BASEMENT v3.0.2
BASEplane OpenMP binary
v3.0.2
Copyright: (C) 2020 ETH Zurich / Laboratory of hydraulics, hydrology and glaciology (VAW)

 -> composing the model from archive
 -> found aggregate <hydraulics_aggr>
 -> found aggregate <morphology_aggr>
 -> found aggregate <hyd_boundary_aggr>
 -> found aggregate <mor_boundary_aggr>
 -> found aggregate <mor_direction_ikeda_aggr>
 -> found aggregate <hyd_friction_aggr>
 -> running on 1 threads

 Initializing OP2
Writing '/RESULTS/CellsAll/HydState/0000000' to file 'G:/2)Geschiebe/results.h5'
Writing '/RESULTS/CellsAll/BottomEl/0000000' to file 'G:/2)Geschiebe/results.h5'
 -> Progress: 0/100, dt = 0.02 [s], RTS = 0
MASS BALANCE PROBLEM: h = -0.00208824 < 0.0
MASS BALANCE PROBLEM: h = -0.00208824 < 0.0
MASS BALANCE PROBLEM: h = -0.0106359 < 0.0
MASS BALANCE PROBLEM: h = -0.0108442 < 0.0
MASS BALANCE PROBLEM: h = -0.00232838 < 0.0
MASS BALANCE PROBLEM: h = -0.00208824 < 0.0
MASS BALANCE PROBLEM: h = -0.0106359 < 0.0
MASS BALANCE PROBLEM: h = -0.0108442 < 0.0
MASS BALANCE PROBLEM: h = -0.00208824 < 0.0
MASS BALANCE PROBLEM: h = -0.00208824 < 0.0
MASS BALANCE PROBLEM: h = -0.00208824 < 0.0
MASS BALANCE PROBLEM: h = -0.00208824 < 0.0
MASS BALANCE PROBLEM: h = -0.00208824 < 0.0
MASS BALANCE PROBLEM: h = -0.00208824 < 0.0
MASS BALANCE PROBLEM: h = -0.00208824 < 0.0
 -> Progress: 1/100, dt = 95 [s], RTS = 320.123
Writing '/RESULTS/CellsAll/HydState/0000001' to file 'G:/2)Geschiebe/results.h5'
Writing '/RESULTS/CellsAll/BottomEl/0000001' to file 'G:/2)Geschiebe/results.h5'
 -> Progress: 2/100, dt = 95 [s], RTS = 591.763
 -> Progress: 3/100, dt = 95 [s], RTS = 839.893
Writing '/RESULTS/CellsAll/HydState/0000002' to file 'G:/2)Geschiebe/results.h5'
Writing '/RESULTS/CellsAll/BottomEl/0000002' to file 'G://2)Geschiebe/results.h5'
 -> Progress: 4/100, dt = 95 [s], RTS = 1089.18
...

chri_and

User

Registered: 2017-07-27

Posts: 34

Re: simulation does not work mit Morphology Block

I forgot the last part

 -> Progress: 100/100, dt = 50 [s], RTS = 12129
Writing '/RESULTS/CellsAll/HydState/0000060' to file 'G:/2)Geschiebe/results.h5'
Writing '/RESULTS/CellsAll/BottomEl/0000060' to file 'G:/results.h5'

-> simulation finished! (cpu-time = 4.96186s)

  count   plan time     MPI time(std)        time(std)           GB/s      GB/s   kernel name 
 -------------------------------------------------------------------------------------------
      1;    0.0000;    0.0000(  0.0000);    0.0000(  0.0000);       inf;         ;   InitCellHydFeatures 
      1;    0.0000;    0.0000(  0.0000);    0.0000(  0.0000);       inf;         ;   InitFrictionFeatures 
      1;    0.0000;    0.0000(  0.0000);    0.0000(  0.0000);       inf;         ;   InitCellMorFeatures 
    908;    0.0000;    0.0000(  0.0000);    0.0000(  0.0000);      inf;      inf;   NodeStringStateCalc 
    908;    0.0000;    0.0000(  0.0000);    0.0000(  0.0000);      inf;      inf;   EdgeDataInterpolation 
    908;    0.0000;    0.0000(  0.0000);    0.0000(  0.0000);      inf;      inf;   EdgeDataModification 
    908;    0.0000;    0.0000(  0.0000);    0.3290(  0.0000);   5.7345;  11.3986;   TimeStepCalculation 
    908;    0.0000;    0.0000(  0.0000);    0.0000(  0.0000);      inf;      inf;   HydBoundaryFluxes 
    907;    0.0000;    0.0000(  0.0000);    0.0000(  0.0000);      inf;      inf;   MorBoundaryFluxes 
    908;    0.0000;    0.0000(  0.0000);    0.0000(  0.0000);      inf;      inf;   HydWallBFluxes 
    907;    0.0000;    0.0000(  0.0000);    0.0000(  0.0000);      inf;      inf;   MorWallBFluxes 
    908;    0.0000;    0.0000(  0.0000);    1.3140(  0.0000);   1.8932;   3.4257;   HydInternalFluxes 
    907;    0.0000;    0.0000(  0.0000);    0.4980(  0.0000);    3.4370;         ;   MorBedloadTransport 
    907;    0.0000;    0.0000(  0.0000);    0.7520(  0.0000);   3.9494;   8.5176;   MorInternalFluxes 
    907;    0.0000;    0.0000(  0.0000);    0.0470(  0.0000);  26.0178;  36.5881;   CalcBedGradient 
    908;    0.0000;    0.0000(  0.0000);    0.2990(  0.0000);  11.2819;  19.0048;   Update 
    908;    0.0000;    0.0000(  0.0000);    0.7920(  0.0000);    1.2020;         ;   HydFriction 
Total plan time:   0.0470

------------------------------------------------------------
 BASEMENT successfully processed all necessary jobs!

Matteo Facchini

Developer

From: Trento

Registered: 2014-09-05

Posts: 284

Re: simulation does not work mit Morphology Block

I have similar problems when I use inner boundary conditions.
I gues it is due to the mass balance problem, the speed up in the simulation is because everything becomes NaN and so the calculation is very fast.

What i usually do is to run debug simulations, i.e. a simulation which starts very close to the problematic time step and gives one output per second. By doing this I can check what is the process triggering the problem.

As for the solution: you might want to check the cells slope and shape.

Note that if you are trying to use an inner boundary condition without a hole in the mesh you might run into troubles due to mass conservation issues.

Hoe this helps,
Matteo

Matthias Bürgler

Developer

From: Zürich

Registered: 2019-04-04

Posts: 171

Re: simulation does not work mit Morphology Block

Hi Chris

Have you tried to run the simulation without the DIRECTION (Ikeda) block?

Best
Matthias

Matteo Facchini

Developer

From: Trento

Registered: 2014-09-05

Posts: 284

Re: simulation does not work mit Morphology Block

Have you tried to run the simulation without the DIRECTION (Ikeda) block?

Hi Matthias,

that totally worked for me!

Cheers,
Matteo