[Nicles43]

Hi

I am trying for many days to make a simulation for a re-entry bluff body, an Apollo like capsule, at hypersonic velocity.

My intial goal was to study it at h = 38 km, Mach = 5, with air-5 model for air, p and T found in internet for 38 km altitude.

However, simulation was not converging. So i decided to consider at first a subsonic mach, Mach = 0.5, at the same altitude of 38 km. HOwvwer, still the simulation is not converging.

I am wondering many questions:
- why the simulation starts only with AUSM scheme? WHen i try to use ASUM+UP or AUSM+M , the simulation does not make any iteration, displaying "SU2 has diverged"


-what is my possible error? I have done so many tentatives. There is an error on the mesh? My mesh is a fully unstrcutred one, with smaller elements near the capsule. Myabe, i am putting something worng inside the config?


Here, below, i will copy the config and the mesh.geo texts

I will attach some images about: mesh and results of the simulation.

This is the config: -------------------------------------------------------------------------------
SOLVER= NEMO_EULER
MATH_PROBLEM= DIRECT
RESTART_SOL= NO
READ_BINARY_RESTART= NO
TIME_DOMAIN= NO
%
FROZEN_MIXTURE=NO
TRANSPORT_COEFF_MODEL = CHAPMANN-ENSKOG
KIND_TURB_MODEL= NONE
%
FLUID_MODEL= MUTATIONPP
GAS_MODEL= air_5
GAS_COMPOSITION= (0.0, 0.0, 0.00, 0.79, 0.21)
%
% Reference origin for moment computation : CG coordinates
REF_ORIGIN_MOMENT_X = 1.1455
REF_ORIGIN_MOMENT_Y = 0.1600
REF_ORIGIN_MOMENT_Z = 0.00
%
% Reference length for pitching, rolling, and yawing non-dimensional moment
REF_LENGTH= 3.9116
%
% Reference area for force coefficients (0 implies automatic calculation)
REF_AREA= 12.020
%
INIT_OPTION= TD_CONDITIONS
MACH_NUMBER= 0.5
AOA= 0.0
FREESTREAM_PRESSURE= 377.1
FREESTREAM_TEMPERATURE= 244.8
FREESTREAM_TEMPERATURE_VE= 244.8
%
MARKER_FAR = ( FARFIELD )
MARKER_EULER = ( CAPSULE )
%
MARKER_PLOTTING= ( CAPSULE )
MARKER_MONITORING= ( CAPSULE )
%
ITER= 1000000

NUM_METHOD_GRAD= WEIGHTED_LEAST_SQUARES

CFL_NUMBER= 0.05

%
CFL_ADAPT= YES
%
CFL_ADAPT_PARAM= (0.0005, 1.0005, 0, 1)

RK_ALPHA_COEFF= ( 0.66667, 0.66667, 1.000000 )

LINEAR_SOLVER= BCGSTAB

LINEAR_SOLVER_ERROR= 1E-6

LINEAR_SOLVER_ITER= 5

%
CONV_NUM_METHOD_FLOW= AUSM

MUSCL_FLOW= NO
SLOPE_LIMITER_FLOW= VENKATAKRISHNAN_WANG
VENKAT_LIMITER_COEFF= 0.01

TIME_DISCRE_FLOW= EULER_EXPLICIT
%

% Min value of the residual (log10 of the residual)
CONV_RESIDUAL_MINVAL= -50
%
MESH_FILENAME= apoll.su2
MESH_FORMAT= SU2
TABULAR_FORMAT= CSV
SOLUTION_FILENAME= restart_flow5_nemo.dat
%OUTPUT_FILES= (RESTART_ASCII, PARAVIEW, SURFACE_PARAVIEW)
CONV_FILENAME= history5_nemo
RESTART_FILENAME= restart_flow5_nemo
VOLUME_FILENAME= soln_volume5_nemo
SURFACE_FILENAME= soln_surface5_nemo

OUTPUT_WRT_FREQ= 1000
SCREEN_WRT_FREQ_TIME= 100
SCREEN_WRT_FREQ_OUTER= 100
SCREEN_OUTPUT= (INNER_ITER, RMS_DENSITY_4, RMS_DENSITY_5, RMS_ENERGY, RMS_ENERGY_VE, DRAG, CAUCHY_DRAG)

VOLUME_OUTPUT= (COORDINATES, SOLUTION, PRIMITIVE, AUXILIARY, RESIDUAL)

HISTORY_OUTPUT= (ITER, RMS_RES, CFL_NUMBER)
-----------------------------------------------------------------------------------------------



This is my geo file: ---------------------------------------



H = 0.75;
h = 0.0075;
R = 80;

//CAPSULE

Point(1) = {0.3743, 1.8368, 0, h};
//+
Point(2) = {0.3743, -1.8368, 0, h};
//+
Point(3) = {0.5543, 1.9558, 0, h};
//+
Point(4) = {0.5543, -1.9558, 0, h};
//+
Point(5) = {3.3254, 0.1938, 0, h};
//+
Point(6) = {3.3254, -0.1938, 0, h};
//+
Point(7) = {3.4306, 0, 0, h};
//+
Point(8) = {0.554227, 1.760289, 0, h};
//+
Point(9) = {0.554227, -1.760289, 0, h};
//+
Point(10) = {3.1995, 0, 0, h};
//+
Point(11) = {4.6939, 0, 0, h};
//+
Point(12) = {0.6608, 1.9242, 0, h};
//+
Point(13) = {0.6608, -1.9242, 0, h};
//+
//+
Circle(1) = {1, 8, 12};
//+
Circle(2) = {2, 9, 13};
//+
Circle(3) = {5, 10, 6};
//+
Circle(4) = {1, 11, 2};
//+
Line(5) = {12, 5};
//+
Line(6) = {6, 13};


//FARFIELD

Point(15) = {R+30, 0, 0, H};
Point(16) = {30, R, 0, H};
Point(17) = {-R+30, 0, 0, H};
Point(18) = {30, -R, 0, H};

// farfield center
Point(19) = {30,0,0,H};

Circle(7) = {15, 19, 16};
Circle(8) = {16, 19, 17};
Circle(9) = {17, 19, 18};
Circle(10) = {18, 19, 15};
Physical Curve("CAPSULE", 1) = {5, 1, 4, 2, 6, 3};
//+
Physical Curve("FARFIELD", 2) = {7, 8, 9, 10};
Curve Loop(1) = {8, 9, 10, 7};
//+
Curve Loop(2) = {1, 5, 3, 6, -2, -4};
//+
Plane Surface(1) = {1, 2};
//+
Physical Surface("VOLUME", 1) = {1};
---------------------------------------------------------------------

[giovanni.medici]

Possibly I'm getting it wrong (and there may be other details to correct) but in your solution and in your cfg it is simulating Mach 0.5 and not 5.0.

I'd revise the cfg and re-try.

[Nicles43]

Hi


Thanks for your interest, Giovanni. Sorry for the late repsonse.

My initially intention was to make convergence for a subsonic case, like M = 0.5. Anyway, i managed to reach convergence for M = 5.



1) the mesh is an unstrctured one, refined gradually near the capsule. I made 2 concentric circles around the capsule in order to impose in these 'construction' lines a desired mesh dimension so that from the farfield to the capsule the sizes are less and less. Moreover, i reduced the extranl domain radius from 80 to 50 since the computational effort with 80 was to high, so that i could not see convergence even attending 4/5 hours.



2) in the config file i put M = 5 and these are the main options:
SOLVER= NEMO_EULER
MATH_PROBLEM= DIRECT
RESTART_SOL= NO
READ_BINARY_RESTART= NO
TIME_DOMAIN= NO
FROZEN_MIXTURE=NO
TRANSPORT_COEFF_MODEL = CHAPMANN-ENSKOG
KIND_TURB_MODEL= NONE
FLUID_MODEL= MUTATIONPP
GAS_MODEL= air_5 %N O NO N2 O2
GAS_COMPOSITION= (0.0, 0.0, 0.00, 0.79, 0.21)
REF_ORIGIN_MOMENT_X = 1.1455
REF_ORIGIN_MOMENT_Y = 0.1600
REF_ORIGIN_MOMENT_Z = 0.00
REF_LENGTH= 3.9116
REF_AREA= 12.020
INIT_OPTION= TD_CONDITIONS
MACH_NUMBER= 5
AOA= 0.0
FREESTREAM_PRESSURE= 377.1
FREESTREAM_TEMPERATURE= 244.8
FREESTREAM_TEMPERATURE_VE= 244.8
MARKER_FAR = ( FARFIELD )
MARKER_EULER = ( CAPSULE ) %%%
MARKER_PLOTTING= ( CAPSULE )
MARKER_MONITORING= ( CAPSULE )
ITER= 120000
NUM_METHOD_GRAD= WEIGHTED_LEAST_SQUARES
CFL_NUMBER= 0.25
CFL_ADAPT= YES
CFL_ADAPT_PARAM= (0.005, 1.005, 0.1, 1)
RK_ALPHA_COEFF= ( 0.66667, 0.66667, 1.000000 )
LINEAR_SOLVER= BCGSTAB
LINEAR_SOLVER_ERROR= 1E-6
LINEAR_SOLVER_ITER= 5
CONV_NUM_METHOD_FLOW= AUSM
MUSCL_FLOW= NO
SLOPE_LIMITER_FLOW= VENKATAKRISHNAN_WANG
VENKAT_LIMITER_COEFF= 0.01
TIME_DISCRE_FLOW= EULER_EXPLICIT
CONV_RESIDUAL_MINVAL= -50


Now my new next challenge will be to reach convergence for a second order scheme, with MUSCL active.
I tried just restarting from the first order, putting MUSCL = YES, limiter = VENKATAKRISHNAN_WANG and AUSM scheme. However, residuals oscillate for 2/3 hours around +3

I don't know what i could do, since when for example i try to change my numerical scheme in AUSM+M, the simulation does not even starts.


Maybe i have to refine mesh so that AUSM+M could start? Any other suggestion?