CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   SU2 Shape Design (https://www.cfd-online.com/Forums/su2-shape-design/)
-   -   Pitching NACA64A010 Test Case Error (https://www.cfd-online.com/Forums/su2-shape-design/215400-pitching-naca64a010-test-case-error.html)

Buxwax March 4, 2019 08:21

Pitching NACA64A010 Test Case Error
 
2 Attachment(s)
Hi there,


I'm currently trying to run the pitching NACA64A010 euler optimization test case. I am using the config and mesh files from the GitHub repository and all options were left at the default settings - see config copy attached. The case was run using command:
"shape_optimization.py -f pitching_NACA64A010.cfg"



The problem I'm having is that the solver is exiting out after only 1 Design iteration with the following error:



NIT FC OBJFUN GNORM
Traceback (most recent call last):
File "/usr/local/bin/shape_optimization.py", line 187, in <module>
main()
File "/usr/local/bin/shape_optimization.py", line 118, in main
options.nzones )
File "/usr/local/bin/shape_optimization.py", line 163, in shape_optimization
SU2.opt.SLSQP(project,x0,xb,its,accu)
File "/usr/local/bin/SU2/opt/scipy_tools.py", line 143, in scipy_slsqp
epsilon = eps )
File "/usr/lib/python2.7/dist-packages/scipy/optimize/slsqp.py", line 207, in fmin_slsqp
constraints=cons, **opts)
File "/usr/lib/python2.7/dist-packages/scipy/optimize/slsqp.py", line 390, in _minimize_slsqp
g = append(fprime(x),0.0)
File "/usr/lib/python2.7/dist-packages/scipy/optimize/optimize.py", line 292, in function_wrapper
return function(*(wrapper_args + args))
File "/usr/local/bin/SU2/opt/scipy_tools.py", line 410, in obj_df
dobj_list = project.obj_df(x)
File "/usr/local/bin/SU2/opt/project.py", line 241, in obj_df
return self._eval(konfig, func,dvs)
File "/usr/local/bin/SU2/opt/project.py", line 205, in _eval
vals = design._eval(func,*args)
File "/usr/local/bin/SU2/eval/design.py", line 146, in _eval
vals = eval_func(*inputs)
File "/usr/local/bin/SU2/eval/design.py", line 379, in obj_df
grad = su2grad(this_obj,grad_method,config,state)
File "/usr/local/bin/SU2/eval/gradients.py", line 105, in gradient
grads = adjoint( func_name, config, state )
File "/usr/local/bin/SU2/eval/gradients.py", line 280, in adjoint
info = su2run.adjoint(config)
File "/usr/local/bin/SU2/run/adjoint.py", line 89, in adjoint
SU2_CFD(konfig)
File "/usr/local/bin/SU2/run/interface.py", line 122, in CFD
run_command( the_Command )
File "/usr/local/bin/SU2/run/interface.py", line 302, in run_command
raise exception(message)
SU2.EvaluationFailure: Path = /home/buxwax/Documents/SU2/SU2-6.2.0/TestCases/optimization_euler/pitching_naca64a010/DESIGNS/DSN_001/ADJOINT_DRAG/,
Command = /bin/SU2_CFD_AD config_CFD_AD.cfg
SU2 process returned error '1'






When I check the log_Adjoint.out it says the following:




Error in "void CConfig::SetPostprocessing(short unsigned int, short unsigned int, short unsigned int)":
-------------------------------------------------------------------------
Dynamic mesh movement currently not supported for the discrete adjoint solver.
------------------------------ Error Exit -------------------------------


--------------------------------------------------------------------------
MPI_ABORT was invoked on rank 0 in communicator MPI_COMM_WORLD
with errorcode 1.

NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on
exactly when Open MPI kills them.
--------------------------------------------------------------------------





The log_direct.out and log_geometry.out seem fine and exit successfully.



It's worth noting that I have built and compiled the SU2 code with AD Support. The problem seems like it has a simple solution so I hope someone can help.



If you require any more information about the case let me know.


Thanks.

Buxwax March 4, 2019 09:42

Update
 
1 Attachment(s)
I have found after looking at the config_log in the SU2_AD folder that some of the Adjoint features were not enabled during the preconfiguration - namely SU2_DEF, SU2_MSH SU2_SOL and SU2_GEO. See output:


Build Configuration Summary:

Source code location: /home/buxwax/Documents/SU2/SU2-6.2.0
Install location:
Version: 6.2.0
C++ Compiler: mpicxx
C Compiler: mpicc
Preprocessor flags: -DHAVE_MPI
Compiler flags: -O3 -Wall
Linker flags:
MPI support: yes
Metis support: yes
Parmetis support: yes
TecIO support: yes
CGNS support: yes
Mutation++ support: no
MKL support: no
Datatype support:
double no
codi_reverse yes
codi_forward no

External includes: -DHAVE_CGNS -I$(top_srcdir)/externals/cgns -I$(top_srcdir)/externals/cgns/adf -DHAVE_PARMETIS -I$(top_srcdir)/externals/parmetis/include -DHAVE_METIS -I$(top_srcdir)/externals/metis/include -DHAVE_TECIO -I$(top_srcdir)/externals/tecio/teciosrc
External libs: $(top_builddir)/externals/cgns/libcgns.a $(top_builddir)/externals/parmetis/libparmetis.a $(top_builddir)/externals/metis/libmetis.a $(top_builddir)/externals/tecio/libtecio.a
External libpthreads: -lpthread

Build SU2_CFD: yes
Build SU2_DOT: yes
Build SU2_MSH: no
Build SU2_DEF: no
Build SU2_SOL: no
Build SU2_GEO: no
Build Py Wrapper: no



How do I configure the build so that these features are enabled? What commands are required?



Note: I followed instructions on AD Build SU2 instructions page when configuring initially.

Buxwax March 7, 2019 15:24

Update
 
It is also worth mentioning that when I try to run the Test Case using CONTINUOUS ADJOINT, I receive the following in the log_Adjoint.out file:



WARNING: The adjoint solver should use a non-dimensional flow solution.
WARNING: The adjoint solver should use a non-dimensional flow solution.
WARNING: The adjoint solver should use a non-dimensional flow solution.

-------------------------------------------------------------------------
| ___ _ _ ___ |
| / __| | | |_ ) Release 6.2.0 "Falcon" |
| \__ \ |_| |/ / |
| |___/\___//___| Suite (Computational Fluid Dynamics Code) |
| |
-------------------------------------------------------------------------
| The current SU2 release has been coordinated by the |
| SU2 International Developers Society <www.su2devsociety.org> |
| with selected contributions from the open-source community. |
-------------------------------------------------------------------------
| The main research teams contributing to the current release are: |
| - Prof. Juan J. Alonso's group at Stanford University. |
| - Prof. Piero Colonna's group at Delft University of Technology. |
| - Prof. Nicolas R. Gauger's group at Kaiserslautern U. of Technology. |
| - Prof. Alberto Guardone's group at Polytechnic University of Milan. |
| - Prof. Rafael Palacios' group at Imperial College London. |
| - Prof. Vincent Terrapon's group at the University of Liege. |
| - Prof. Edwin van der Weide's group at the University of Twente. |
| - Lab. of New Concepts in Aeronautics at Tech. Inst. of Aeronautics. |
-------------------------------------------------------------------------
| Copyright 2012-2019, Francisco D. Palacios, Thomas D. Economon, |
| Tim Albring, and the SU2 contributors. |
| |
| SU2 is free software; you can redistribute it and/or |
| modify it under the terms of the GNU Lesser General Public |
| License as published by the Free Software Foundation; either |
| version 2.1 of the License, or (at your option) any later version. |
| |
| SU2 is distributed in the hope that it will be useful, |
| but WITHOUT ANY WARRANTY; without even the implied warranty of |
| MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| Lesser General Public License for more details. |
| |
| You should have received a copy of the GNU Lesser General Public |
| License along with SU2. If not, see <http://www.gnu.org/licenses/>. |
-------------------------------------------------------------------------

------------------------ Physical Case Definition -----------------------
Continuous Euler adjoint equations.
Mach number: 0.796.
Angle of attack (AoA): 0 deg, and angle of sideslip (AoS): 0 deg.
Performing a dynamic mesh simulation: rigid mesh motion.
No restart solution, use the values at infinity (freestream).
Read flow solution from: restart_flow.dat.
Dimensional simulation.
The reference area is 1 m^2.
The semi-span will be computed using the max y(3D) value.
The reference length is 1 m.
Reference origin for moment evaluation is (0.248, 0, 0).
Surface(s) where the force coefficients are evaluated: airfoil.

Surface(s) where the objective function is evaluated: airfoil.
Surface(s) plotted in the output file: airfoil.
Surface(s) affected by the design variables: airfoil.
Input mesh file name: mesh_optimization.su2

----------------------- Design problem definition -----------------------
CD objective function.

---------------------- Space Numerical Integration ----------------------
Jameson-Schmidt-Turkel scheme for the adjoint inviscid terms.
JST viscous coefficients (1st, 2nd, & 4th): 0.15, 0, 0.02.
The method includes a grid stretching correction (p = 0.3).
Second order integration.
The reference sharp edge distance is: 0.3.
Gradient Computation using weighted Least-Squares method.

---------------------- Time Numerical Integration -----------------------
Unsteady simulation, dual time stepping strategy (second order in time).
Unsteady time step provided by the user (s): 0.023555.
Total number of internal Dual Time iterations: 110.
Euler implicit method for the adjoint equations.
W Multigrid Cycle, with 3 multigrid levels.
Damping factor for the residual restriction: 0.9.
Damping factor for the correction prolongation: 0.9.
No CFL adaptation.
+-------------------------------------------+
| MG Level| Presmooth|PostSmooth|CorrectSmo|
+-------------------------------------------+
| 0| 1| 0| 0|
| 1| 2| 0| 0|
| 2| 3| 0| 0|
| 3| 3| 0| 0|
+-------------------------------------------+
Courant-Friedrichs-Lewy number: 5

------------------------- Convergence Criteria --------------------------
Maximum number of iterations: 99999.
Reduce the adjoint density residual 3 orders of magnitude.
The minimum value for the adjoint density residual is 10^(-10).

-------------------------- Output Information ---------------------------
Writing a solution file every 250 iterations.
Writing the convergence history every 1 iterations.
Writing the dual time flow solution every 1 iterations.
Writing the dual time convergence history every 10 iterations.
The output file format is Tecplot ASCII (.dat).
Convergence history file name: history_adjoint.
Forces breakdown file name: forces_breakdown.dat.
Adjoint solution file name: solution_adj.dat.
Restart adjoint file name: restart_adj.dat.
Adjoint variables file name: adjoint.
Surface adjoint coefficients file name: surface_adjoint.

------------------- Config File Boundary Information --------------------
+-----------------------------------------+
| Marker Type| Marker Name|
+-----------------------------------------+
| Euler wall| airfoil|
+-----------------------------------------+
| Far-field| farfield|
+-----------------------------------------+

---------------------- Read Grid File Information -----------------------
Two dimensional problem.
8606 points.
2 surface markers.
+------------------------------------+
| Index| Marker| Elements|
+------------------------------------+
| 0| airfoil| 200|
| 1| farfield| 75|
+------------------------------------+
16937 triangles.

------------------------- Geometry Preprocessing ------------------------
Setting point connectivity.
Renumbering points (Reverse Cuthill McKee Ordering).
Recomputing point connectivity.
Setting element connectivity.
Checking the numerical grid orientation.
Identifying edges and vertices.
Computing centers of gravity.
Setting the control volume structure.
Area of the computational grid: 1.26e+03.
Finding max control volume width.
Searching for the closest normal neighbors to the surfaces.
Storing a mapping from global to local point index.
Compute the surface curvature.
Max K: 1.99e+03. Mean K: 17.3. Standard deviation K: 126.
Checking for periodicity.
Setting the multigrid structure.
+-------------------------------------------+
| MG Level| CVs|Aggl. Rate| CFL|
+-------------------------------------------+
| 0| 8606| 1/1.00| 5|
| 1| 2442| 1/3.52| 4|
| 2| 699| 1/3.49| 3.21|
| 3| 189| 1/3.7| 2.5|
+-------------------------------------------+
Wetted area = 2.02 m.
Area projection in the x-plane = 0.0996 m, y-plane = 0.999 m.
Max. coordinate in the x-direction = 1 m, y-direction = 0.0499 m.
Min coordinate in the x-direction = 0 m, y-direction = -0.0499 m.

------------------------- Driver information --------------------------
A Fluid driver has been instantiated.

------------------------ Iteration Preprocessing ------------------------
Zone 1: adjoint Euler/Navier-Stokes/RANS fluid iteration.

------------------------- Solver Preprocessing --------------------------
Inviscid flow: Computing density based on free-stream
temperature and pressure using the ideal gas law.
Force coefficients computed using MACH_MOTION.

-- Models:
+------------------------------------------------------------------------------+
| Viscosity Model| Conductivity Model| Fluid Model|
+------------------------------------------------------------------------------+
| -| -| STANDARD_AIR|
+------------------------------------------------------------------------------+
-- Fluid properties:
+------------------------------------------------------------------------------+
| Name| Dim. value| Ref. value| Unit|Non-dim. value|
+------------------------------------------------------------------------------+
| Gas Constant| 287.058| 1| N.m/kg.K| 287.058|
| Spec. Heat Ratio| -| -| -| 1.4|
+------------------------------------------------------------------------------+
-- Initial and free-stream conditions:
+------------------------------------------------------------------------------+
| Name| Dim. value| Ref. value| Unit|Non-dim. value|
+------------------------------------------------------------------------------+
| Static Pressure| 101325| 1| Pa| 101325|
| Density| 1.22498| 1| kg/m^3| 1.22498|
| Temperature| 288.15| 1| K| 288.15|
| Total Energy| 243476| 1| m^2/s^2| 243476|
| Velocity-X| 270.876| 1| m/s| 270.876|
| Velocity-Y| 0| 1| m/s| 0|
| Velocity Magnitude| 270.876| 1| m/s| 270.876|
+------------------------------------------------------------------------------+
| Mach Number| -| -| -| 0.796|
+------------------------------------------------------------------------------+
+------------------------------------------------------------------------------+
| Name| Dim. value| Ref. value| Unit|Non-dim. value|
+------------------------------------------------------------------------------+
-- Unsteady conditions
| Total Time| 0.59| 1| s| 0.59|
| Time Step| 0.023555| 1| s| 0.023555|
+------------------------------------------------------------------------------+
Initialize Jacobian structure (Euler). MG level: 0.
Initialize Jacobian structure (Adjoint Euler). MG level: 0.
Initialize Jacobian structure (Euler). MG level: 1.
Initialize Jacobian structure (Adjoint Euler). MG level: 1.
Initialize Jacobian structure (Euler). MG level: 2.
Initialize Jacobian structure (Adjoint Euler). MG level: 2.
Initialize Jacobian structure (Euler). MG level: 3.
Initialize Jacobian structure (Adjoint Euler). MG level: 3.


Error in "std::__cxx11::string CConfig::GetUnsteady_FileName(std::__cxx11::string , int)":
-------------------------------------------------------------------------
Requesting a negative iteration number for the restart file!!
------------------------------ Error Exit -------------------------------


--------------------------------------------------------------------------
MPI_ABORT was invoked on rank 0 in communicator MPI_COMM_WORLD
with errorcode 1.

NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on
exactly when Open MPI kills them.
--------------------------------------------------------------------------


All times are GMT -4. The time now is 06:37.