PK Journal by mayork · Pull Request #93 · convexengineering/SPaircraft

mayork · 2017-07-07T19:41:36Z

@pgkirsch @1ozturkbe I'm starting my attempt at pulling apart the full model to an updated PK journal model. Will keep you posted on progress.

pgkirsch · 2017-07-07T19:44:31Z

Awesome, thank you.

1ozturkbe · 2017-07-07T19:51:00Z

sweet, looking forward to seeing how this goes!

mayork · 2017-07-07T20:06:52Z

alright already got climb removed. was easier than I anticipated.

mayork · 2017-07-07T20:09:33Z

@pgkirsch looks like one enhancement is going to be discretizing the breguet range (it's going to require extra work to remove that feature, i dont see any point in doing that)

pgkirsch · 2017-07-07T20:11:37Z

Fine by me to have a discretized (more accurate) breguet range.

mayork · 2017-07-07T22:00:50Z

subs_optimal_737.py

            'TSFC': 0.5,
            'F_TO': 30000*units('lbf'),
+            'A_{2}': 3*units('m^2'),
    }


@pgkirsch you should update these four engine subs to what you want. as of now they're 100% arbitrary.

mayork · 2017-07-07T22:07:34Z

@pgkirsch @1ozturkbe I think this is done. What we have now is Philippe's paper plus the all the updates that we discussed (since I just removed from our current code base). The updates I believe this includes are:

Breguet range discretization (works for any number of segments)
HT and VT drag models [Martin]
VT sizing [Martin]
HT sizing alteration (Philippe see below) [Martin]
Fuselage bending [Berk]
Moment of inertia [Berk]
CG (plays into the HT sizing) [Berk]
Wing structural improvements [Berk]
Wing drag improvements [Martin]

Philippe note that I believe this changes a few of the HT sizing constraints (you should check this...I made the change because of some complications in flight profile integration, namely needing to include a CG)...it's described in the attached write up. Let me know your thoughts about this.
Basic_HT_Sizing_Model (3).pdf

In the bulleted list above I noted in brackets who should update the manuscript to reflect each new model. @1ozturkbe and @pgkirsch thoughts on that? Philippe you tell us how you want us making manuscript updates.

Berk - if you could VSP the airplane this outputs that would be good for debugging anything I messed up.

Philippe feel free to posts questions once you look over the code and see how the solution compares to what is currently in the paper (update engine subs first, noted in a commit comment)

@whoburg and the other who advocated for working from current code I think that was the right call. This only took me ~3 hours (could be some bugs still, needs VSP and solution comparison).

mayork · 2017-07-07T22:08:25Z

@pgkirsch if you update the engine subs and post a solution comparison to what you had in the paper manuscript I can take a look at any issues tomorrow.

mayork · 2017-07-07T22:22:48Z

@pgkirsch forgot to mention that to run the model import the optimal 737 class from the pk_SP_aircraft.py file. All the aircraft level constraints are again still in D8.py

pgkirsch · 2017-07-08T05:12:08Z

@mayork You're a machine, thanks for doing all of this. I'm going to spend some time getting my bearings and figuring out what's what.

pgkirsch · 2017-07-08T05:29:15Z

Does it usually take a ridiculously long time to build the model? If so, why is that?

@pgkirsch you probably need to pull master. It is a much larger problem then your baseline models due to the vectorization but should solve in a few seconds (3.7 seconds for me on verbosity=4)

pgkirsch · 2017-07-08T05:58:48Z

For manuscript editing, I think we should PR to the pkjournal branch of the pubs repo. Give me a chance to understand better how much has changed from my original stuff though before we do that.
Btw @mayork what happened in the last few commits to the pkjournal branch? You seem to have made some commits, reverted them, and also reverted one of my commits?

@pgkirsch I removed statelinking which I thought we didn't need but i'm pretty sure we do so I ended up reverting back to that. I lost some changes I made to to the substitution file in the process of some local experimental reverts so I ended up just repushing it all back to the branch.

Which one of your commits did I revert? All I see is commits I made.

I think PR on the pkjournal branch of pubs is a good idea. @1ozturkbe is that good?

pgkirsch · 2017-07-08T06:04:11Z

One general question I have about the current code structure, is how should I interpret the module naming? I think we previously had aircraft.py, fuselage.py, vtail.py, wing.py etc. and now we have D8.py, D8_Fuselage.py, D8_VT_yaw_rate_and_EO_simple_profile.py, D8_Wing_simple_profile.py etc. Do these names tell me something I need to know? Why did they become so long and descriptive? If everything is for a D8, why are they all called D8_something? If everything can actually be applied to a conventional-config aircraft, why are they all called D8_something?

@mayork @1ozturkbe @whoburg

@pgkirsch treat D8.py as aircraft.py, D8_Wing_simple_profile.py as wing.py, etc. The D8 tags on everything is a hold over from a previous naming convention that we need to change. Another change I made is that I integrated all of the sub components like wing and HT etc. with a simple aircraft and flight profile model for testing purposes, hence the tag simple_profile on the files we are importing from. We used to have a two VT files hence the additional VT_yaw_rate_and_EO_ there, but it's nothing important. Just an identifier Berk and I used that needs to be changed.

mayork · 2017-07-08T12:49:46Z

@pgkirsch responses are above beneath your comments.

mayork · 2017-07-08T17:21:23Z

@pgkirsch @1ozturkbe I updated the sub values and I think the solution is looking pretty decent. It's a bit more accurate to the reference aircraft than what was in your original manuscript.

mayork · 2017-07-08T17:25:57Z

@1ozturkbe you should generate a VSP of the aircraft...could be good for the paper

@mayork super time + internet limited atm, but will do my best.

pgkirsch · 2017-07-08T20:21:51Z

@mayork Thanks for updating the sub values and checking the solution.

Solving the model is indeed relatively quick, but actually just building it (i.e. importing the method) takes > 20 seconds. Is that typical?

Sorry I was referring to the pkjournal branch of the pubs repo (you pushed some commits to that branch it seems).

Ok, understood on the naming. It seems like from a reducing-code-duplication point of view, it would make sense to import the simple profile part for each of those models. I am happy to go through and change names if that is useful forward work? Although I guess it isn't a priority at the moment.

A VSP would be great! That's such a cool new feature.

mayork · 2017-07-08T20:29:53Z

@pgkirsch yeah i somehow was on that branch and not master and pushed my engine paper to it then reverted and pushed the engine paper to master...i didn't think that i overwrote any of your commits. Sorry if I did!

It takes under a second to perform the import on my computer.

Berk and I can handle the naming. Parts of the simple profile should be imported but it is difficult to import a lot because all the simple subsystem models change depending on which detailed model is used due to how the coupling between the detailed subsystem models works.

pgkirsch · 2017-07-08T20:37:06Z

Gotcha, makes sense.

pgkirsch · 2017-07-08T20:46:19Z

And as for pubs, it's commits 449291d and b56838 that I'm most confused about. The former is a merge that I'm not sure I understand, and the latter is reverting a commit I made.

# Conflicts: # horizontal_tail.py

mayork · 2017-08-21T14:26:31Z

aircraft.py

            # HT/VT moment arm constraints
            aircraft.HT['l_{ht}'] <= aircraft.HT['x_{CG_{ht}}'] - xCG,
-            aircraft.VT['l_{vt}'] <= aircraft.VT['x_{CG_{vt}}'] - xCG,
+            aircraft.VT['l_{vt}'] <= aircraft.VT['x_{CG_{vt}}'] - 0.25*aircraft.VT['c_{root_{vt}}'] - xCG,


@pgkirsch @1ozturkbe

mayork · 2017-08-21T14:26:36Z

aircraft.py

            # HT/VT moment arm constraints
            aircraft.HT['l_{ht}'] <= aircraft.HT['x_{CG_{ht}}'] - xCG,
-            aircraft.VT['l_{vt}'] <= aircraft.VT['x_{CG_{vt}}'] - xCG,
+            aircraft.VT['l_{vt}'] <= aircraft.VT['x_{CG_{vt}}'] - 0.25*aircraft.VT['c_{root_{vt}}'] - xCG,


@pgkirsch @1ozturkbe

# Conflicts: # aircraft.py

…s problematic.

mayork · 2017-11-05T18:52:01Z

alright @1ozturkbe I've finally got a few hours so I want to start integrating this with master. Might need your support.

I'm going to start by add flags to remove the engine and integrate this engine model. And then I think we need to remove the flight profile and that'll be it right?

@pgkirsch

mayork · 2017-11-05T19:11:08Z

#110

mayork · 2017-11-17T19:25:52Z

spajoa.py

-    sol = m_relax.localsolve(verbosity=0, iteration_limit=50, reltol=0.01,
-                             modifylastgp=True)
+    sol = m_relax.localsolve(verbosity=0, iteration_limit=50, reltol=0.01)
    post_process(sol)


@1ozturkbe email any udpated plots to @pgkirsch , he has to give them individual files

mayork · 2017-11-23T02:40:18Z

fuselage.py

                Wfuse >= Cfuse*(Wshell + Wfloor + Winsul + \
-                    Wapu + Wfix + Wwindow + Wpadd + Wseat + Whbend + Wvbend),
+                    Wapu + Wfix + Wwindow + Wpadd + Wseat + Whbend + Wvbend + Wcone),
            ])


@1ozturkbe or @pgkirsch please check this to be sure I didn't miss anything

pushing initial file to run PK journal aircraft

6e65117

removed climb

9905a55

mayork added 7 commits July 7, 2017 16:31

removing cruise climb, not straight breguet range

eace44a

percent diff works with this branch

e0577dc

runs with simple enigne model

104ba3c

removed state linking...still runs

213896a

adding statelinking back in, aircraft flies at constant altitude

a99e829

updated simple engine

c740300

fixing A2, fixing M (required due to simple engine)

bae2c4b

mayork commented Jul 7, 2017

View reviewed changes

updating subs

480fe7a

pgkirsch and others added 13 commits August 20, 2017 17:13

HT model cleanup, standalone stuff moved down

53429b2

Merge branch 'pkjournal' of https://github.com/hoburg/d8 into pkjournal

bc6c842

Removing duplicated variable xCGlg and constraint. Changing some units.

e027e4f

Merge remote-tracking branch 'origin/pkjournal' into pkjournal

0bc0ec2

# Conflicts: # horizontal_tail.py

V minor ht cleanup

e4bcbf1

Removing duplicated constraint

34868b3

N->lbf in ac, HT, VT.

db108b9

changing drag fits

c4cd69c

Merge remote-tracking branch 'origin/pkjournal' into pkjournal

c1f5cf1

# Conflicts: # horizontal_tail.py

Finalizing merges...

cec768f

Merge branch 'pkjournal' of https://github.com/hoburg/d8 into pkjournal

71d121e

TSFC to TASOPT 737-800 initial cruise TSFC

e9366d1

possible fix to VT moment arm

aa21376

mayork commented Aug 21, 2017

View reviewed changes

1ozturkbe added 10 commits August 21, 2017 12:29

Changing modifylastgp for speed.

8b296b8

Added dxtrail and dxlead for HT.

2922321

Merge remote-tracking branch 'origin/pkjournal' into pkjournal

7f4fd1d

# Conflicts: # aircraft.py

Reintegrated LE/TE constraints for both VT and HT. Untight constraint…

02ce4c9

…s problematic.

Making CG compatible with TE/LE constraints.

827e85f

Changing m.a.c. for VT/HT to SigEqs.

19dc6a8

Thorough variable name clean-up.

a579df6

Changing ReqRng to Rreq.

00bd4fa

Changing WHT to Wht.

c4dccab

Edited plot ranges.

05c7e9e

Minor plot modifications for sweeps.

9bb4fc8

mayork commented Nov 17, 2017

View reviewed changes

Wcone part of Wfuse

c9411f9

mayork commented Nov 23, 2017

View reviewed changes

Conversation

mayork commented Jul 7, 2017

Uh oh!

pgkirsch commented Jul 7, 2017

Uh oh!

1ozturkbe commented Jul 7, 2017

Uh oh!

mayork commented Jul 7, 2017

Uh oh!

mayork commented Jul 7, 2017

Uh oh!

pgkirsch commented Jul 7, 2017

Uh oh!

mayork Jul 7, 2017

Choose a reason for hiding this comment

Uh oh!

mayork commented Jul 7, 2017

Uh oh!

mayork commented Jul 7, 2017

Uh oh!

mayork commented Jul 7, 2017

Uh oh!

pgkirsch commented Jul 8, 2017

Uh oh!

pgkirsch commented Jul 8, 2017 • edited by mayork Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pgkirsch commented Jul 8, 2017 • edited by mayork Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pgkirsch commented Jul 8, 2017 • edited by mayork Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mayork commented Jul 8, 2017

Uh oh!

mayork commented Jul 8, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mayork commented Jul 8, 2017 • edited by 1ozturkbe Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pgkirsch commented Jul 8, 2017

Uh oh!

mayork commented Jul 8, 2017

Uh oh!

pgkirsch commented Jul 8, 2017

Uh oh!

pgkirsch commented Jul 8, 2017

Uh oh!

mayork Aug 21, 2017

Choose a reason for hiding this comment

Uh oh!

mayork Aug 21, 2017

Choose a reason for hiding this comment

Uh oh!

mayork commented Nov 5, 2017

Uh oh!

mayork commented Nov 5, 2017

Uh oh!

mayork Nov 17, 2017

Choose a reason for hiding this comment

Uh oh!

mayork Nov 23, 2017

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

pgkirsch commented Jul 8, 2017 •

edited by mayork

Loading

pgkirsch commented Jul 8, 2017 •

edited by mayork

Loading

pgkirsch commented Jul 8, 2017 •

edited by mayork

Loading

mayork commented Jul 8, 2017 •

edited

Loading

mayork commented Jul 8, 2017 •

edited by 1ozturkbe

Loading