Structural stress penalty function for optimization, fix some CConfig problems #1176

pcarruscag · 2021-01-22T17:37:56Z

Proposed Changes

A differentiable maximum VM stress function for optimization.
Simplifies the "DoubleArray" option to avoid one array of default values and another of actual values.

PR Checklist

I am submitting my contribution to the develop branch.
My contribution generates no new compiler warnings (try with the '-Wall -Wextra -Wno-unused-parameter -Wno-empty-body' compiler flags, or simply --warnlevel=2 when using meson).
My contribution is commented and consistent with SU2 style.
I have added a test case that demonstrates my contribution, if necessary.
I have updated appropriate documentation (Tutorials, Docs Page, config_template.cpp) , if necessary.

…ts it

pcarruscag · 2021-01-22T17:41:53Z

Common/src/grid_movement/CSurfaceMovement.cpp

-    Omega  = (config->GetRotation_Rate(3)/config->GetOmega_Ref());
+    Omega = config->GetRotation_Rate(2)/config->GetOmega_Ref();


Using static types lets the compiler find bugs for us at compile time :)
Using bare-pointer does not let the compiler do anything :(

I am curious how the compiler found that now that the underlying Rotation_Rate is static because is has 3 elements initialized to 0. So how did the compiler know but I guess I miss sth here

Because this is not matlab or fortran xD

Oh man am I blind... insert palmface here

As blind an any programmer, we should have XYZ enums

pcarruscag · 2021-01-22T19:21:55Z

SU2_CFD/src/output/CFlowOutput.cpp

-  SetHistoryOutputValue("AVG_NORMALVEL", Tot_Surface_Enthalpy);
+  SetHistoryOutputValue("AVG_NORMALVEL", Tot_Surface_NormalVelocity);


Thanks @WallyMaier , I literally changed the line below 2 days ago and missed this.

WallyMaier

This looks good to me. I appreciate the clean-up done. Especially getting rid of SU2_MSH!

WallyMaier · 2021-01-26T02:43:38Z

Common/src/CConfig.cpp

@@ -1262,8 +1239,8 @@ void CConfig::SetConfig_Options() {
  /*!\brief INC_DENSITY_INIT \n DESCRIPTION: Initial density for incompressible flows (1.2886 kg/m^3 by default) \ingroup Config*/
  addDoubleOption("INC_DENSITY_INIT", Inc_Density_Init, 1.2886);
  /*!\brief INC_VELOCITY_INIT \n DESCRIPTION: Initial velocity for incompressible flows (1.0,0,0 m/s by default) \ingroup Config*/
-  default_vel_inf[0] = 1.0; default_vel_inf[1] = 0.0; default_vel_inf[2] = 0.0;


I like the removal of all the default_

TobiKattmann

Can't say much to the FEA stuff but thanks 💐 for the Config cleanup. Yet another net-code-reduction :)

TobiKattmann · 2021-01-28T07:21:23Z

Common/include/CConfig.hpp

@@ -117,6 +109,7 @@ class CConfig {
  su2double Opt_RelaxFactor;              /*!< \brief Scale factor for the line search. */
  su2double Opt_LineSearch_Bound;         /*!< \brief Bounds for the line search. */
  su2double StartTime;
+  unsigned short SmoothNumGrid;           /*!< \brief Smooth the numerical grid. */


Out of curiosity, what is the numerical grid (or is it just the mesh) and when is it smoothed?
greping through all cpp,hpp,inl of the code base didn't really enlighten me, just found some evidence in Cpoint

Left over from SU2_MSH I believe, there is some geometry routine that uses those variables of CPoint, but it is never called.

TobiKattmann · 2021-01-28T09:05:17Z

TestCases/radiation/p1model/configp1.cfg

@@ -23,6 +23,7 @@ INC_DENSITY_MODEL= VARIABLE
 INC_ENERGY_EQUATION = YES
 INC_DENSITY_INIT= 0.00597782417156
 INC_TEMPERATURE_INIT= 288.15
+INC_VELOCITY_INIT= (0, 0, 0)


Why do you have to add this option here? I might have missed that in the code but if the option is not set I understood your changes that the default remains the same, but with out explicitly having containers for the default value? (default here would be 1,0,0)

TobiKattmann · 2021-01-28T09:12:53Z

Common/src/grid_movement/CSurfaceMovement.cpp

-    Omega  = (config->GetRotation_Rate(3)/config->GetOmega_Ref());
+    Omega = config->GetRotation_Rate(2)/config->GetOmega_Ref();


I am curious how the compiler found that now that the underlying Rotation_Rate is static because is has 3 elements initialized to 0. So how did the compiler know but I guess I miss sth here

pcarruscag · 2021-01-28T10:01:15Z

Common/src/CConfig.cpp

-  default_vel_inf[0] = 0.0; default_vel_inf[1] = 0.0; default_vel_inf[2] = 0.0;
  /* DESCRIPTION: Coordinates of the rigid motion origin */
-  addDoubleArrayOption("MOTION_ORIGIN", 3, Motion_Origin, default_vel_inf);
+  addDoubleArrayOption("MOTION_ORIGIN", 3, Motion_Origin);


@TobiKattmann the default velocity was not 1,0,0 because it was being re-used here as the default value for something else.

This trips a lot of people (myself included) but the actual parsing of the options does not take place in this method, it creates a list of option parsing objects but it does not read immediately. Hence the need for the config postprocessing method to check for conflicting things and so on.

TobiKattmann · 2021-01-28T10:28:39Z

Common/src/CConfig.cpp

+  vel_init[0] = 1.0; vel_init[1] = 0.0; vel_init[2] = 0.0;
+  addDoubleArrayOption("INC_VELOCITY_INIT", 3, vel_init);


@pcarruscag Ok I see. This really seems so straight forward with what it replaces but it isn't

rafapalacios · 2021-01-28T16:39:48Z

Got my first glasses half a year into Fortran… and never looked back. From: Pedro Gomes <notifications@github.com> Sent: 28 January 2021 11:13 To: su2code/SU2 <SU2@noreply.github.com> Cc: Subscribed <subscribed@noreply.github.com> Subject: Re: [su2code/SU2] Structural stress penalty function for optimization, fix some CConfig problems (#1176) @pcarruscag commented on this pull request.

________________________________ In Common/src/grid_movement/CSurfaceMovement.cpp<#1176 (comment)>:

- Omega = (config->GetRotation_Rate(3)/config->GetOmega_Ref());

+ Omega = config->GetRotation_Rate(2)/config->GetOmega_Ref(); As blind an any programmer, we should have XYZ enums — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub<#1176 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AATH4JKAAEAXKNHHCUGOQT3S4FBB7ANCNFSM4WOZ6DBA>.

pcarruscag added 4 commits January 19, 2021 15:47

initial implementation, not working with AD

ec4eb9f

fix the AD problem and avoid stopping the preacc outside of what star…

162265a

…ts it

Merge remote-tracking branch 'upstream/develop' into stress_penalty

cb3d924

fix COptionDoubleArray, const correctness, and one bug

a5e8a74

pcarruscag added the changelog:feature label Jan 22, 2021

pcarruscag added this to In progress in General Maintenance via automation Jan 22, 2021

pr-triage bot added the PR: draft label Jan 22, 2021

pcarruscag commented Jan 22, 2021

View reviewed changes

pcarruscag added 2 commits January 22, 2021 18:29

remove more SU2_MSH

aa3fa3f

gauss average instead of nodal (same, but different)

b7b731a

pcarruscag commented Jan 22, 2021

View reviewed changes

pcarruscag marked this pull request as ready for review January 22, 2021 19:40

pr-triage bot added PR: unreviewed and removed PR: draft labels Jan 22, 2021

pcarruscag and others added 4 commits January 22, 2021 20:28

fix p1 case

9d077df

remove deprecated options from testcases

d39ec8e

Merge branch 'develop' into stress_penalty

286e3c2

option to consider only the surface of a reference geometry

813ff06

WallyMaier approved these changes Jan 26, 2021

View reviewed changes

pr-triage bot added PR: reviewed-approved and removed PR: unreviewed labels Jan 26, 2021

simplify a few things

30b2554

pr-triage bot added PR: unreviewed and removed PR: reviewed-approved labels Jan 26, 2021

pcarruscag merged commit c008fd4 into develop Jan 27, 2021

General Maintenance automation moved this from In progress to Done Jan 27, 2021

pcarruscag deleted the stress_penalty branch January 27, 2021 15:50

pr-triage bot added PR: merged and removed PR: unreviewed labels Jan 27, 2021

TobiKattmann reviewed Jan 28, 2021

View reviewed changes

pr-triage bot added PR: unreviewed and removed PR: merged labels Jan 28, 2021

pcarruscag commented Jan 28, 2021

View reviewed changes

TobiKattmann reviewed Jan 28, 2021

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Structural stress penalty function for optimization, fix some CConfig problems #1176

Structural stress penalty function for optimization, fix some CConfig problems #1176

pcarruscag commented Jan 22, 2021

pcarruscag Jan 22, 2021

TobiKattmann Jan 28, 2021

pcarruscag Jan 28, 2021

TobiKattmann Jan 28, 2021

pcarruscag Jan 28, 2021

pcarruscag Jan 22, 2021

WallyMaier left a comment

WallyMaier Jan 26, 2021

TobiKattmann left a comment

TobiKattmann Jan 28, 2021

pcarruscag Jan 28, 2021

TobiKattmann Jan 28, 2021

TobiKattmann Jan 28, 2021

pcarruscag Jan 28, 2021

TobiKattmann Jan 28, 2021

rafapalacios commented Jan 28, 2021 via email

		Omega = (config->GetRotation_Rate(3)/config->GetOmega_Ref());
		Omega = config->GetRotation_Rate(2)/config->GetOmega_Ref();

		SetHistoryOutputValue("AVG_NORMALVEL", Tot_Surface_Enthalpy);
		SetHistoryOutputValue("AVG_NORMALVEL", Tot_Surface_NormalVelocity);

		vel_init[0] = 1.0; vel_init[1] = 0.0; vel_init[2] = 0.0;
		addDoubleArrayOption("INC_VELOCITY_INIT", 3, vel_init);

Structural stress penalty function for optimization, fix some CConfig problems #1176

Structural stress penalty function for optimization, fix some CConfig problems #1176

Conversation

pcarruscag commented Jan 22, 2021

Proposed Changes

PR Checklist

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

WallyMaier left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

TobiKattmann left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rafapalacios commented Jan 28, 2021 via email