[OVS]

Hello,

I'm running a RANS-SA calculation on a full aircraft but it doesn't seem to converge to acceptable residuals... the residuals go to a minimum before exploding into a neverending enveloppe (oscillatory solution).

I usually cannot see this until I finished the calculations and then the only solution file I have is the one corresponding to the last iteration.

How can I save the files every n iterations? I know I can specify the frequency of solution files writting, but what if I want to keep these files instead of overwritting them?

Doing it "by hand" is not an option since I'm running on a cluster. Is there an option in the configuration file to do this?

Thanks
Oliver

[fpalacios]

More than focus on the output, I think we need to focus in finding the problem. Could you please give us more details about the problem, grid, and configuration file?

Best,
Francisco

[OVS]

Hello,

Thanks for your reply.

The problem I'm trying to solve is a preliminary approach of a BWB aircraft on cruise configuration.
I cannot put the mesh here since it takes more than 100MB but I'll try to explain the problem (Convergence tendencies, cfg and log files are added at the end).

Physical problem:
height: 12500m
RE: 93e6
Mach: 0.6 (Tried 0.78 but I was recommended to lower the mach to minimize the possibilities of shockwaves).
AoA: 0 (my endgoal is to obtain the polar of the aircraft but I'm already having difficulties at 0 AoA).

The domain consists of half a BWB inside half a sphere.
The centerbody chord of the BWB is 25m
Thus the Farfield is 100 chords away (radius of 2500m centered at the middle of the centerbody chord).

The mesh-grid (ICEMCFD 14.5 -> CGNS) is an unstructured one with the following characteristics:

PARTS:

BWB: maximum size 0.1m; 30 prism layers (6e-6m initial height for y+ < 1); height ratio 1.2; Triangles
FARFIELD: Max size: 512m; Triangles
SYMMETRY: Max size: 512m; Triangles

FLUID: 4 220 332 cells (half of which correspond to the prism layers); Prisms, Pyramids and Tetras

Mesh size after ParMETIS partitionning:

4518314 interior elements including halo cells.
2152016 tetrahedra.
2359192 prisms.
7106 pyramids.
1700759 vertices including ghost points.

---------------------------------------------------
The cfg file is based on the oneraM6 Turbulent Test Case.

I ran the case on SU2 4.1.0 on 48 processors for 8h. It ended before reaching cauchy convergence at 12911 iterations. A figure on the residuals evolution is added as a joint file.

After analysing the residuals I came to the conclusion that

(1) The convergence criteria was not the best (should probably use residual instead of cauchy).
(2) Maybe I shouldn't adimensionalize the pb (so that I work with actual values intead of normalized ones).
(3) The mesh is probably not the best, but I need to know where it fails.
(4) I'm still wondering why the residuals (rho, rho*energy) tend to "explode" after the "minimum residual" has ben reached (around iteration 9500).

Thanks.

Oliver

----------------------------------------------------

% File Version 4.1.0 "Cardinal" %
% %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% ------------- DIRECT, ADJOINT, AND LINEARIZED PROBLEM DEFINITION ------------%
%
% Physical governing equations (EULER, NAVIER_STOKES,
% TNE2_EULER, TNE2_NAVIER_STOKES,
% WAVE_EQUATION, HEAT_EQUATION, LINEAR_ELASTICITY,
% POISSON_EQUATION)
PHYSICAL_PROBLEM= NAVIER_STOKES
%
% Specify turbulence model (NONE, SA, SA_NEG, SST)
KIND_TURB_MODEL= SA
%
% Mathematical problem (DIRECT, CONTINUOUS_ADJOINT)
MATH_PROBLEM= DIRECT
%
% Restart solution (NO, YES)
RESTART_SOL= NO

% -------------------- COMPRESSIBLE FREE-STREAM DEFINITION --------------------%
%
% Mach number (non-dimensional, based on the free-stream values)
MACH_NUMBER= 0.6
%
% Angle of attack (degrees, only for compressible flows)
AoA= 0.0
%
% Side-slip angle (degrees, only for compressible flows)
SIDESLIP_ANGLE= 0.0
%
% Free-stream temperature (288.15 K by default)
FREESTREAM_TEMPERATURE= 216.65
%
% Reynolds number (non-dimensional, based on the free-stream values)
REYNOLDS_NUMBER= 93.598E6
%
% Reynolds length (1 m by default)
REYNOLDS_LENGTH= 25.0

% ---- IDEAL GAS, POLYTROPIC, VAN DER WAALS AND PENG ROBINSON CONSTANTS -------%
%
% Different gas model (STANDARD_AIR, IDEAL_GAS, VW_GAS, PR_GAS)
FLUID_MODEL= STANDARD_AIR
%
% Ratio of specific heats (1.4 default and the value is hardcoded
% for the model STANDARD_AIR)
GAMMA_VALUE= 1.4
%
% Specific gas constant (287.058 J/kg*K default and this value is hardcoded
% for the model STANDARD_AIR)
GAS_CONSTANT= 287.058
%
% Critical Temperature (131.00 K by default)
CRITICAL_TEMPERATURE= 131.00
%
% Critical Pressure (3588550.0 N/m^2 by default)
CRITICAL_PRESSURE= 3588550.0
%
% Critical Density (263.0 Kg/m3 by default)
CRITICAL_DENSITY= 263.0
%
% Acentric factor (0.035 (air))
ACENTRIC_FACTOR= 0.035

% --------------------------- VISCOSITY MODEL ---------------------------------%
%
% Viscosity model (SUTHERLAND, CONSTANT_VISCOSITY).
VISCOSITY_MODEL= SUTHERLAND
%
% Molecular Viscosity that would be constant (1.716E-5 by default)
MU_CONSTANT= 1.716E-5
%
% Sutherland Viscosity Ref (1.716E-5 default value for AIR SI)
MU_REF= 1.716E-5
%
% Sutherland Temperature Ref (273.15 K default value for AIR SI)
MU_T_REF= 273.15
%
% Sutherland constant (110.4 default value for AIR SI)
SUTHERLAND_CONSTANT= 110.4


% ---------------------- REFERENCE VALUE DEFINITION ---------------------------%
%
% Reference origin for moment computation
REF_ORIGIN_MOMENT_X = 0.25
REF_ORIGIN_MOMENT_Y = 0.00
REF_ORIGIN_MOMENT_Z = 0.00
%
% Reference length for pitching, rolling, and yawing non-dimensional moment
REF_LENGTH_MOMENT= 1.0
%
% Reference area for force coefficients (0 implies automatic calculation)
REF_AREA= 0
%
% Reference element length for computing the slope limiter epsilon
REF_ELEM_LENGTH= 0.1

% -------------------- BOUNDARY CONDITION DEFINITION --------------------------%
%
% Navier-Stokes wall boundary marker(s) (NONE = no marker)
MARKER_HEATFLUX= ( BWB, 0.0 )
%
% Far-field boundary marker(s) (NONE = no marker)
MARKER_FAR= ( FARFIELD )
%
% Symmetry boundary marker(s) (NONE = no marker)
MARKER_SYM= ( SYMMETRY )
%
% Marker(s) of the surface to be plotted or designed
MARKER_PLOTTING= ( BWB )
%
% Marker(s) of the surface where the functional (Cd, Cl, etc.) will be evaluated
MARKER_MONITORING= ( BWB )

% ------------- COMMON PARAMETERS DEFINING THE NUMERICAL METHOD ---------------%
%
% Numerical method for spatial gradients (GREEN_GAUSS, WEIGHTED_LEAST_SQUARES)
NUM_METHOD_GRAD= GREEN_GAUSS
%
% Courant-Friedrichs-Lewy condition of the finest grid
CFL_NUMBER= 4.0
%
% Adaptive CFL number (NO, YES)
CFL_ADAPT= NO
%
% Parameters of the adaptive CFL number (factor down, factor up, CFL min value,
% CFL max value )
CFL_ADAPT_PARAM= ( 1.5, 0.5, 1.0, 100.0 )
%
% Runge-Kutta alpha coefficients
RK_ALPHA_COEFF= ( 0.66667, 0.66667, 1.000000 )
%
% Number of total iterations
EXT_ITER= 50000

% ------------------------ LINEAR SOLVER DEFINITION ---------------------------%
%
% Linear solver for the implicit (or discrete adjoint) formulation (BCGSTAB, FGMRES)
LINEAR_SOLVER= FGMRES
%
% Preconditioner of the Krylov linear solver (NONE, JACOBI, LINELET)
LINEAR_SOLVER_PREC= LU_SGS
%
% Min error of the linear solver for the implicit formulation
LINEAR_SOLVER_ERROR= 1E-6
%
% Max number of iterations of the linear solver for the implicit formulation
LINEAR_SOLVER_ITER= 5

% -------------------------- MULTIGRID PARAMETERS -----------------------------%
%
% Multi-Grid Levels (0 = no multi-grid)
MGLEVEL= 0
%
% Multi-grid cycle (V_CYCLE, W_CYCLE, FULLMG_CYCLE)
MGCYCLE= V_CYCLE
%
% Multi-grid pre-smoothing level
MG_PRE_SMOOTH= ( 1, 2, 2, 2 )
%
% Multi-grid post-smoothing level
MG_POST_SMOOTH= ( 2, 2, 2, 2 )
%
% Jacobi implicit smoothing of the correction
MG_CORRECTION_SMOOTH= ( 0, 0, 0, 0 )
%
% Damping factor for the residual restriction
MG_DAMP_RESTRICTION= 0.7
%
% Damping factor for the correction prolongation
MG_DAMP_PROLONGATION= 0.7

% -------------------- FLOW NUMERICAL METHOD DEFINITION -----------------------%
%
% Convective numerical method (JST, LAX-FRIEDRICH, CUSP, ROE, AUSM, HLLC,
% TURKEL_PREC, MSW)
CONV_NUM_METHOD_FLOW= ROE
%
% Spatial numerical order integration (1ST_ORDER, 2ND_ORDER, 2ND_ORDER_LIMITER)
SPATIAL_ORDER_FLOW= 2ND_ORDER_LIMITER
%
% Slope limiter (VENKATAKRISHNAN, MINMOD)
SLOPE_LIMITER_FLOW= VENKATAKRISHNAN
%
% Coefficient for the limiter (smooth regions)
LIMITER_COEFF= 0.3
%
% 1st, 2nd and 4th order artificial dissipation coefficients
AD_COEFF_FLOW= ( 0.15, 0.5, 0.02 )
%
% Time discretization (RUNGE-KUTTA_EXPLICIT, EULER_IMPLICIT, EULER_EXPLICIT)
TIME_DISCRE_FLOW= EULER_IMPLICIT

% -------------------- TURBULENT NUMERICAL METHOD DEFINITION ------------------%
%
% Convective numerical method (SCALAR_UPWIND)
CONV_NUM_METHOD_TURB= SCALAR_UPWIND
%
% Spatial numerical order integration (1ST_ORDER, 2ND_ORDER, 2ND_ORDER_LIMITER)
SPATIAL_ORDER_TURB= 1ST_ORDER
%
% Slope limiter (VENKATAKRISHNAN, MINMOD)
SLOPE_LIMITER_TURB= VENKATAKRISHNAN
%
% Time discretization (EULER_IMPLICIT)
TIME_DISCRE_TURB= EULER_IMPLICIT

% --------------------------- CONVERGENCE PARAMETERS --------------------------%
%
% Convergence criteria (CAUCHY, RESIDUAL)
%
CONV_CRITERIA= CAUCHY
%
% Residual reduction (order of magnitude with respect to the initial value)
RESIDUAL_REDUCTION= 6
%
% Min value of the residual (log10 of the residual)
RESIDUAL_MINVAL= -10
%
% Start convergence criteria at iteration number
STARTCONV_ITER= 10
%
% Number of elements to apply the criteria
CAUCHY_ELEMS= 100
%
% Epsilon to control the series convergence
CAUCHY_EPS= 1E-6
%
% Function to apply the criteria (LIFT, DRAG, NEARFIELD_PRESS, SENS_GEOMETRY,
% SENS_MACH, DELTA_LIFT, DELTA_DRAG)
CAUCHY_FUNC_FLOW= DRAG
CAUCHY_FUNC_ADJFLOW= SENS_GEOMETRY

% ------------------------- INPUT/OUTPUT INFORMATION --------------------------%
%
% Mesh input file
MESH_FILENAME= mesh-00.cgns
%
% Mesh input file format (SU2, CGNS, NETCDF_ASCII)
MESH_FORMAT= CGNS
%
% Mesh output file
MESH_OUT_FILENAME= mesh_out.su2
%
% Restart flow input file
SOLUTION_FLOW_FILENAME= solution_flow.dat
%
% Restart linear flow input file
%SOLUTION_LIN_FILENAME= solution_lin.dat
%
% Restart adjoint input file
SOLUTION_ADJ_FILENAME= solution_adj.dat
%
% Output file format (PARAVIEW, TECPLOT, STL)
OUTPUT_FORMAT= PARAVIEW
%
% Output file convergence history (w/o extension)
CONV_FILENAME= history
%
% Output file restart flow
RESTART_FLOW_FILENAME= restart_flow.dat
%
% Output file restart adjoint
RESTART_ADJ_FILENAME= restart_adj.dat
%
% Output file flow (w/o extension) variables
VOLUME_FLOW_FILENAME= flow
%
% Output file adjoint (w/o extension) variables
VOLUME_ADJ_FILENAME= adjoint
%
% Output objective function gradient (using continuous adjoint)
GRAD_OBJFUNC_FILENAME= of_grad.dat
%
% Output file surface flow coefficient (w/o extension)
SURFACE_FLOW_FILENAME= surface_flow
%
% Output file surface adjoint coefficient (w/o extension)
SURFACE_ADJ_FILENAME= surface_adjoint
%
% Writing solution file frequency
WRT_SOL_FREQ= 250
%
% Writing convergence history frequency
WRT_CON_FREQ= 1
%

% --------------------- OPTIMAL SHAPE DESIGN DEFINITION -----------------------%
%
% List of design variables (Design variables are separated by semicolons)
% From 1 to 99, Geometrycal design variables.
% - HICKS_HENNE ( 1, Scale | Mark. List | Lower(0)/Upper(1) side, x_Loc )
% - NACA_4DIGITS ( 4, Scale | Mark. List | 1st digit, 2nd digit, 3rd and 4th digit )
% - DISPLACEMENT ( 5, Scale | Mark. List | x_Disp, y_Disp, z_Disp )
% - ROTATION ( 6, Scale | Mark. List | x_Axis, y_Axis, z_Axis, x_Turn, y_Turn, z_Turn )
% - FFD_CONTROL_POINT ( 7, Scale | Mark. List | FFD_BoxTag, i_Ind, j_Ind, k_Ind, x_Mov, y_Mov, z_Mov )
% - FFD_DIHEDRAL_ANGLE ( 8, Scale | Mark. List | FFD_BoxTag, x_Orig, y_Orig, z_Orig, x_End, y_End, z_End )
% - FFD_TWIST_ANGLE ( 9, Scale | Mark. List | FFD_BoxTag, x_Orig, y_Orig, z_Orig, x_End, y_End, z_End )
% - FFD_ROTATION ( 10, Scale | Mark. List | FFD_BoxTag, x_Orig, y_Orig, z_Orig, x_End, y_End, z_End )
% - FFD_CAMBER ( 11, Scale | Mark. List | FFD_BoxTag, i_Ind, j_Ind )
% - FFD_THICKNESS ( 12, Scale | Mark. List | FFD_BoxTag, i_Ind, j_Ind )
% - FFD_VOLUME ( 13, Scale | Mark. List | FFD_BoxTag, i_Ind, j_Ind )
% From 100 to 199, Flow solver design variables.
% - MACH_NUMBER ( 101, Scale | Markers List )
% - AOA ( 102, Scale | Markers List )
DEFINITION_DV= ( 1, 0.001 | airfoil | 0, 0.1 ); ( 1, 0.001 | airfoil | 0, 0.2 )

----------------------------------

SU2 log:


-------------------------------------------------------------------------
| ___ _ _ ___ |
| / __| | | |_ ) Release 4.1.0 "Cardinal" |
| \__ \ |_| |/ / |
| |___/\___//___| Suite (Computational Fluid Dynamics Code) |
| |
...
------------------------ Physical Case Definition -----------------------
Compressible RANS equations.
Turbulence model: Spalart Allmaras
Mach number: 0.6.
Angle of attack (AoA): 0 deg, and angle of sideslip (AoS): 0 deg.
Reynolds number: 9.3598e+07.
No restart solution, use the values at infinity (freestream).
Dimensional simulation.
The reference length/area will be computed using y(2D) or z(3D) projection.
The reference length (moment computation) is 1.
Reference origin (moment computation) is (0.25, 0, 0).
Surface(s) where the force coefficients are evaluated: BWB.
Surface(s) plotted in the output file: BWB.
Surface(s) belonging to the Fluid-Structure Interaction problem:
Input mesh file name: mesh-00.cgns

---------------------- Space Numerical Integration ----------------------
Roe (with entropy fix) solver for the flow inviscid terms.
Second order integration with slope limiter.
Venkatakrishnan slope-limiting method, with constant: 0.3.
The reference element size is: 0.1.
Scalar upwind solver (first order) for the turbulence model.
First order integration.
Average of gradients with correction (viscous flow terms).
Average of gradients with correction (viscous turbulence terms).
Gradient computation using Green-Gauss theorem.

---------------------- Time Numerical Integration -----------------------
Local time stepping (steady state simulation).
Euler implicit method for the flow equations.
CFL adaptation. Factor down: 1.5, factor up: 0.5,
lower limit: 1, upper limit: 100.
Courant-Friedrichs-Lewy number: 4
Euler implicit time integration for the turbulence model.

------------------------- Convergence Criteria --------------------------
Maximum number of iterations: 49999.
Cauchy criteria for Drag using 100 elements and epsilon 1e-06.
Start convergence criteria at iteration 10.

-------------------------- Output Information ---------------------------
Writing a flow solution every 250 iterations.
Writing the convergence history every 1 iterations.
The output file format is Paraview ASCII (.vtk).
Convergence history file name: history.
Forces breakdown file name: forces_breakdown.dat.
Surface flow coefficients file name: surface_flow.
Flow variables file name: flow.
Restart flow file name: restart_flow.dat.

------------------- Config File Boundary Information --------------------
Far-field boundary marker(s): FARFIELD.
Symmetry plane boundary marker(s): SYMMETRY.
Constant heat flux wall boundary marker(s): BWB.

---------------------- Read Grid File Information -----------------------
Reading the CGNS file: mesh-00.cgns.
CGNS file contains 1 database(s).
Database 1, BASE#1: 1 zone(s), cell dimension of 3, physical dimension of 3.
Zone 1, mesh-00.uns: 1473852 vertices, 4220332 cells, 0 boundary vertices.
Reading grid coordinates.
Number of coordinate dimensions is 3.
Loading CoordinateX values into linear partitions.
Loading CoordinateY values into linear partitions.
Loading CoordinateZ values into linear partitions.
Distributing connectivity across all ranks.
Number of connectivity sections is 4.
Loading section FLUID of element type Mixed.
Loading section BWB of element type Triangle.
Loading section SYMMETRY of element type Mixed.
Loading section FARFIELD of element type Triangle.
Successfully closed the CGNS file.

Loading CGNS data into SU2 data structures.
Three dimensional problem.
4220332 interior elements before linear partitioning.
Building the graph adjacency structure.
1473852 grid points before linear partitioning.
3 surface markers.
78751 boundary elements in index 0 (Marker = BWB).
21268 boundary elements in index 1 (Marker = SYMMETRY).
517 boundary elements in index 2 (Marker = FARFIELD).
Calling ParMETIS...
Finished partitioning using ParMETIS (528211 edge cuts).
Communicating partition data and creating halo layers.
4518314 interior elements including halo cells.
2152016 tetrahedra.
2359192 prisms.
7106 pyramids.
1700759 vertices including ghost points.
Establishing MPI communication patterns.

------------------------- 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.
Volume of the computational grid: 3.23e+10.
Searching for the closest normal neighbors to the surfaces.
Compute the surface curvature.
Max K: 1.67e+06. Mean K: 346. Standard deviation K: 1.77e+04.
Computing wall distances.
Area projection in the z-plane = 158.

------------------------- Driver Preprocessing --------------------------
Instantiating a single zone driver for the problem.

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

------------------------- Solver Preprocessing --------------------------
Viscous flow: Computing pressure using the ideal gas law
based on the free-stream temperature and a density computed
from the Reynolds number.
Force coefficients computed using free-stream values.
-- Input conditions:
Fluid Model: STANDARD_AIR
Specific gas constant: 287.058 N.m/kg.K.
Specific gas constant (non-dim): 287.058
Specific Heat Ratio: 1.4
Viscosity Model: SUTHERLAND
Ref. Laminar Viscosity: 1.716e-05 N.s/m^2.
Ref. Temperature: 273.15 K.
Sutherland Constant: 110.4 K.
Laminar Viscosity (non-dim): 1.716e-05
Ref. Temperature (non-dim): 273.15
Sutherland constant (non-dim): 110.4
Conductivity Model: CONSTANT_PRANDTL
Prandtl: 0.72
Free-stream static pressure: 18695.5 Pa.
Free-stream total pressure: 23846.2 Pa.
Free-stream temperature: 216.65 K.
Free-stream density: 0.300613 kg/m^3.
Free-stream velocity: (177.043, 0, 0) m/s. Magnitude: 177.043 m/s.
Free-stream total energy per unit mass: 171150 m^2/s^2.
Free-stream viscosity: 1.42155e-05 N.s/m^2.
Free-stream turb. kinetic energy per unit mass: 117.541 m^2/s^2.
Free-stream specific dissipation: 248563 1/s.
-- Reference values:
Reference specific gas constant: 1 N.m/kg.K.
Reference pressure: 1 Pa.
Reference temperature: 1 K.
Reference density: 1 kg/m^3.
Reference velocity: 1 m/s.
Reference energy per unit mass: 1 m^2/s^2.
Reference viscosity: 1 N.s/m^2.
Reference conductivity: 1 W/m^2.K.
-- Resulting non-dimensional state:
Mach number (non-dim): 0.6
Reynolds number (non-dim): 9.3598e+07. Re length: 25 m.
Specific gas constant (non-dim): 287.058
Free-stream temperature (non-dim): 216.65
Free-stream pressure (non-dim): 18695.5
Free-stream density (non-dim): 0.300613
Free-stream velocity (non-dim): (177.043, 0, 0). Magnitude: 177.043
Free-stream total energy per unit mass (non-dim): 171150
Free-stream viscosity (non-dim): 1.42155e-05
Free-stream turb. kinetic energy (non-dim): 117.541
Free-stream specific dissipation (non-dim): 248563

Initialize Jacobian structure (Navier-Stokes). MG level: 0.
Initialize Jacobian structure (SA model).

----------------- Integration and Numerics Preprocessing ----------------
Integration Preprocessing.
Numerics Preprocessing.

------------------------------ Begin Solver ----------------------------

[asonda]

Saving intermediate solutions without overwriting them is such a useful functionality (Monitoring the flowfield evolution tells a much clearer story than residuals etc.) Is there any easy way to hack the code to implement it? Thanks a lot,

Alberto

[harikaushik]

Originally Posted by OVS

Hello,

I'm running a RANS-SA calculation on a full aircraft but it doesn't seem to converge to acceptable residuals... the residuals go to a minimum before exploding into a neverending enveloppe (oscillatory solution).

I usually cannot see this until I finished the calculations and then the only solution file I have is the one corresponding to the last iteration.

How can I save the files every n iterations? I know I can specify the frequency of solution files writting, but what if I want to keep these files instead of overwritting them?

Doing it "by hand" is not an option since I'm running on a cluster. Is there an option in the configuration file to do this?

Thanks
Oliver

Hey did you find a method to keep the intermediate files without them being over written?

[bigfootedrockmidget]

To all: a feature request has been created:


https://github.com/su2code/SU2/issues/1463