Add spanwise shape variables #75

joanibal · 2021-03-16T21:42:03Z

Purpose

These new design variables control all the FDD nodes along a line in a chosen spanwise direction

This means that for 2D shape optimization with ADflow, you can use spanwise local design variables to parametrize the 2D shape without the need for added linear constraints.

I think this is a lot easier to explain with an examples

Consider the following rectangular FFD with it i,j, and k axes as shown in the figure

If a airfoil surface mesh is embedded in the FFD as shown

Then adding spanwise shape variables like this

DVGeo.addGeoDVSpanwiseLocal("shape", 'k', lower=-0.5, upper=0.5, axis="z", scale=1.0)

would add the 6 design variables that correspond to the six colors in this figure. Each design variables would control the z location of two FFD points.

The test case uses the existing cylinder FFD

Type of change

What types of change is it?
Select the appropriate type(s) that describe this PR

Bugfix (non-breaking change which fixes an issue)
New feature (non-breaking change which adds functionality)
Breaking change (non-backwards-compatible fix or feature)
Code style update (formatting, renaming)
Refactoring (no functional changes, no API changes)
Documentation update
Maintenance update
Other (please describe)

Testing

Explain the steps needed to test the new code to verify that it does indeed address the issue and produce the expected behavior.

Checklist

Put an x in the boxes that apply.

[no] I have run flake8 and black to make sure the code adheres to PEP-8 and is consistently formatted
^formatting introduced a large amount of change not relevant to the feature
I have run unit and regression tests which pass locally with my changes
I have added new tests that prove my fix is effective or that my feature works
I have added necessary documentation
^ I will after feedback

joanibal · 2021-03-16T21:43:52Z

I following the same process that Nick used for sectionLocal, but we may want to consider refactoring DVGeometry at some point.

anilyil · 2021-03-17T14:35:26Z

I don't understand DVGeometry's inner working as well as @marcomangano so I will pass this down to him. The figures in the PR does a great job at explaining what it does, is there a way we can include them in the docs as well?

marcomangano · 2021-03-17T19:05:05Z

@joanibal I will carefully review this tomorrow! Just a few points for now:

The tests are failing, is that because you trained your .ref files locally?
Is this for 2D optimization only, or do you envision a broader use?
2.b I need to check the code/examples, but can you select any of x,y,z and/or "single rows" of control points? Just to get a grasp of the flexibility of this implementation.

I see the point of having this extra feature but I am trying to figure out how usable it can be for general cases.
Also agree that the figures are a great added value and we should add them to the docs somehow :)

joanibal · 2021-03-18T02:33:05Z

@marcomangano

guilty as charged. I did train them locally. Why is that a problem again?
I envision broader use. It think it is use for 3D wings too. If you start from a rectangular wing then you can change the shape while keeping a constant airfoil along the span. But I think the best use is for 2D cases.

can you select any of x,y,z and/or "single rows" of control points?

You can do both. Selecting a "single rows" is a bit tricky, but you would do the same way you would for any set of local shape variables (use pointSelect and/or volList)

marcomangano · 2021-03-18T14:14:28Z

I am not sure if this is an issue for pure-python packages too, but when I added a test (and a locally trained .ref file) for ADflow the tests on the docker configuration were not matching the prescribed tolerances on the docker builds. I guess our local machines always have some different settings and configurations that can alter the results, even in the I/O phase. @nwu63 can you explain this more in detail? I have a pretty qualitative overview of the issue. Anyway, not sure if it applies here, will check the tests soon.
I see your point, this feature does not actually add new capabilities but rather simplify and make our usual parametrization much leaner. Just the benefit for 2D cases is already enough imho
I see, makes sense, will check the details in my review.

Just one last thought before I dive into the code: if not in place already, I would add an assertAllClose check that compares FFD locations and sensitivities of the new DV vs the usual multi-DV approach. This way we can catch any discrepancy in future developments.

ewu63 · 2021-03-18T18:13:26Z

I am not sure if this is an issue for pure-python packages too, but when I added a test (and a locally trained .ref file) for ADflow the tests on the docker configuration were not matching the prescribed tolerances on the docker builds. I guess our local machines always have some different settings and configurations that can alter the results, even in the I/O phase. @nwu63 can you explain this more in detail? I have a pretty qualitative overview of the issue. Anyway, not sure if it applies here, will check the tests soon.

I am not sure where this discrepancy comes from. Sometimes I've had no problems training stuff locally, other times things behave in really strange ways. In this case though, it appears that the ref file was not added to the PR branch?

marcomangano

A lot of new code - but luckily pretty close to the previous GeoDVSectionLocal implementation. As this is an entirely new feature, I don't see risks of breaking anything (unless the user mixes vars in a weird way?) so I think this is pretty much ready - see the comments on testing though.

I added a bunch of comments, mostly for me to try and better understand how pyGeo works. Hope to turn this PR into a chance to improve the code current documentation, any concise comment or explanation is welcome.

Also, I completely agree that at some point pyGeo will need some major refactoring

examples/childffd_warning/warning_childFFD.py

tests/reg_tests/warning_childFFD.py

pygeo/DVConstraints.py

marcomangano · 2021-03-18T14:58:58Z

pygeo/DVGeometry.py

@@ -144,9 +145,11 @@ def __init__(self, fileName, complex=False, child=False, faceFreeze=None, *args,
        self.nDVG_T = None
        self.nDVL_T = None
        self.nDVSL_T = None
+        self.nDVSW_T = None


I see you had to come up with a different name because SL was taken, right?
Maybe we could add some line comments here to explain every acronym, just to make things a tiny easier

See the new comments in code to address this

pygeo/DVGeometry.py

marcomangano · 2021-03-18T17:55:00Z

pygeo/DVGeometry.py

@@ -1955,7 +2122,7 @@ def computeTotalJacobianCS(self, ptSetName, config=None):
        return

    def addVariablesPyOpt(self, optProb, globalVars=True, localVars=True,
-                          sectionlocalVars=True, ignoreVars=None, freezeVars=None):
+                          sectionlocalVars=True, spanwiselocalVars=True, ignoreVars=None, freezeVars=None):


Add documentation for sectionlocalVars and spanwiselocalVars below.

marcomangano · 2021-03-18T18:09:18Z

pygeo/DVGeometry.py

+                N = self.FFD.embededVolumes['child%d_coef'%(iChild)].N
+                self.children[iChild].dCcdXdvl = numpy.zeros((N*3, self.nDV_T))
+
+            iDVSpanwiseLocal = self.nDVSW_count


I see most of this comes from _sectionlocalDVJacobian but it would be great to have some more inline comments, I am struggling a bit to follow the flow.

see new added comments

pygeo/DVGeometry.py

marcomangano · 2021-03-18T18:23:33Z

tests/reg_tests/test_Cylinder.py

+            leList = [[0, 0, 0 ], [-radius/2, 0, height]]
+            xAxis = [-1, 0, 0]
+            yAxis = [0, 1, 0]
+            DVCon.addLERadiusConstraints(leList, nSpan=5, axis=yAxis,


I see this comes from the previous test, but I am not familiar with this script. Why are we using it?

marcomangano · 2021-03-18T18:25:50Z

tests/reg_tests/test_Cylinder.py

+                                         chordDir=xAxis, scaled=False)
+            # DVCon.writeTecplot('cylinder_constraints.dat')
+
+            funcs = {}


It looks we are only testing the DVCon sens here... which I guess it is pretty much the same than testing the DV sens directly? See my comment before the review about an additional test - if needed

marcomangano · 2021-03-18T18:32:20Z

I am not sure where this discrepancy comes from. Sometimes I've had no problems training stuff locally, other times things behave in really strange ways. In this case though, it appears that the ref file was not added to the PR branch?

L M A O, good catch

bbrelje

I apologize for the delay in review. I am going to focus on the testing because I don't totally understand the implementation code and I trust you did it generally correctly.

Is there any way of verifying with the cylinder test as written that the feature is working as intended, and not just that it's producing the answer you expect. I am not sure a LERadiusConstraint alone is enough for that. a 2d grid of thickness constraints might be (would want to verify that the numbers are the same in the spanwise direction

There also ought to be a verification that this DV isn't used on a non-2D problem; or the test coverage needs to be extended to cover 3D cases.

joanibal · 2021-03-26T21:48:10Z

I added another test using a 2d grid of thickness constraints as Ben suggested.
I also addressed the comments Marco had in the code.

It is ready for review again.

ewu63 · 2021-03-26T22:57:59Z

Not sure why but there are some merge conflicts, could you resolve those so the newly added test gets run on Azure?

marcomangano

Looks good, thanks for the comments! Let's wait for the tests to run

marcomangano

I think we are good to go! (see comment below). @bbrelje we need your approval of the changes before we merge this.

pygeo/DVConstraints.py

bbrelje

I still think it would be best to test the situation where someone tries to use this functionality for a 3D problem (in my view it should throw an error and the test should assert that it throws an error). One minor cleanup item (remove the extra debug statement from DVConstraints).

pygeo/DVConstraints.py

joanibal · 2021-03-30T13:39:23Z

@bbrelje maybe I'm not getting what you mean by 3D problem. In the newest test, these design variables are used for a 3D problem. I agree the main application is 2D, but it isn't wrong to use these design variables for 3D.

ewu63 · 2021-03-30T17:33:26Z

I still think it would be best to test the situation where someone tries to use this functionality for a 3D problem (in my view it should throw an error and the test should assert that it throws an error). One minor cleanup item (remove the extra debug statement from DVConstraints).

My understanding here is that these DVs can be used in 3D cases, and there are tests/docs covering usage. I'm going to merge this but if you think this PR can be improved, please make an issue @bbrelje.

joanibal added 10 commits February 21, 2020 12:00

added child warning example

6be0c7f

merged in master and moved test

9ec85d9

formatting changes

c84ef1b

added unittest.main

af48a98

tweaked unittest.main

4c8f49a

tweaked unittest.main again

94065c9

add issue ref in code

cfc6f28

inital implementation

405915f

Merge branch 'master' into spanwise_local

2b72469

added test

9f51fe4

joanibal requested a review from a team as a code owner March 16, 2021 21:42

joanibal requested review from bbrelje and anilyil March 16, 2021 21:42

anilyil requested review from marcomangano and removed request for anilyil March 17, 2021 14:35

marcomangano reviewed Mar 18, 2021

View reviewed changes

added missing ref

fe68af3

bbrelje requested changes Mar 24, 2021

View reviewed changes

joanibal added 2 commits March 24, 2021 18:08

fixed mapping for cons

8e1cf76

added another test

dfad88f

joanibal requested review from bbrelje and marcomangano March 26, 2021 21:48

marcomangano reviewed Mar 27, 2021

View reviewed changes

Merge branch 'master' into spanwise_local

98395fe

ewu63 requested a review from marcomangano March 29, 2021 14:13

marcomangano reviewed Mar 29, 2021

View reviewed changes

pygeo/DVConstraints.py Outdated Show resolved Hide resolved

Update DVConstraints.py

1c7164b

joanibal requested a review from marcomangano March 29, 2021 17:57

marcomangano previously approved these changes Mar 30, 2021

View reviewed changes

bbrelje requested changes Mar 30, 2021

View reviewed changes

pygeo/DVConstraints.py Outdated Show resolved Hide resolved

Update DVConstraints.py

8529930

joanibal dismissed marcomangano’s stale review via 8529930 March 30, 2021 13:40

joanibal requested review from bbrelje, ewu63 and marcomangano March 30, 2021 13:41

ewu63 approved these changes Mar 30, 2021

View reviewed changes

ewu63 merged commit 974b087 into mdolab:master Mar 30, 2021

sseraj mentioned this pull request Mar 31, 2021

Fixed wrong variable name #78

Merged

12 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add spanwise shape variables #75

Add spanwise shape variables #75

joanibal commented Mar 16, 2021

joanibal commented Mar 16, 2021

anilyil commented Mar 17, 2021

marcomangano commented Mar 17, 2021

joanibal commented Mar 18, 2021 •

edited

Loading

marcomangano commented Mar 18, 2021

ewu63 commented Mar 18, 2021

marcomangano left a comment

marcomangano Mar 18, 2021

joanibal Mar 26, 2021

marcomangano Mar 18, 2021

joanibal Mar 26, 2021

marcomangano Mar 18, 2021

joanibal Mar 26, 2021

marcomangano Mar 18, 2021

marcomangano Mar 18, 2021

marcomangano commented Mar 18, 2021

bbrelje left a comment

joanibal commented Mar 26, 2021

ewu63 commented Mar 26, 2021

marcomangano left a comment

marcomangano left a comment

bbrelje left a comment

joanibal commented Mar 30, 2021

ewu63 commented Mar 30, 2021

Add spanwise shape variables #75

Add spanwise shape variables #75

Conversation

joanibal commented Mar 16, 2021

Purpose

Type of change

Testing

Checklist

joanibal commented Mar 16, 2021

anilyil commented Mar 17, 2021

marcomangano commented Mar 17, 2021

joanibal commented Mar 18, 2021 • edited Loading

marcomangano commented Mar 18, 2021

ewu63 commented Mar 18, 2021

marcomangano 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

Choose a reason for hiding this comment

Choose a reason for hiding this comment

marcomangano commented Mar 18, 2021

bbrelje left a comment

Choose a reason for hiding this comment

joanibal commented Mar 26, 2021

ewu63 commented Mar 26, 2021

marcomangano left a comment

Choose a reason for hiding this comment

marcomangano left a comment

Choose a reason for hiding this comment

bbrelje left a comment

Choose a reason for hiding this comment

joanibal commented Mar 30, 2021

ewu63 commented Mar 30, 2021

joanibal commented Mar 18, 2021 •

edited

Loading