Quantcast

Trace based simplified composite analysis. Opinions?

HomeBuiltAirplanes.com

Help Support HomeBuiltAirplanes.com:

Hot Wings

Grumpy Cynic
HBA Supporter
Log Member
Joined
Nov 14, 2009
Messages
7,267
Location
Rocky Mountains
Ran across this recently and it looks like it might help those of us that are weak in the math needed for traditional composite analysis. Being one of those souls afflicted with apprehension at the mention of tensors and such this obviously interested me.

Would any of those HBA members that have real world experience with composite design care to look over this subject and comment on it's validity or usefulness?

CompoSIDE and Prof. Stephen Tsai collaborate to demonstrate new composites theory | JEC Group

www.sciencedirect.com/science/article/pii/S0266353814002188/pdfft?md5=b9f45ef859874bec2abbcab227dd6c00&pid=1-s2.0-S0266353814002188-main.pdf

 
Last edited by a moderator:

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,126
Location
Saline Michigan
I am one of the guys who recommends Tsai and Hahn for composite analysis. Same guy in the video... Failure criteria are complicated to get and apply. A transformation of strength space could work, and would simplify the whole process. And if Tsai is involved, I bet that it does. The hard part is still getting good failure data in each direction needed and then convincing yourself that your work is good enough...

Bill
 

Hot Wings

Grumpy Cynic
HBA Supporter
Log Member
Joined
Nov 14, 2009
Messages
7,267
Location
Rocky Mountains
Ran across an upcoming Webex by Tsai regarding this method via my ASTM membership. Don't know if it's intended strictly for ASTM members but I suspect not. You may have to register with Webex but it's pretty painless.

Session date: Monday, October 17, 2016
Starting time: 1:00 pm, Eastern Daylight Time (New York, GMT-04:00)
Duration: 1 hour
Description:
Agenda:
Session number: 214 186 763

I have a PDF overview of the webinar, but it's too big to post here.

Dropbox link
 
Last edited:

Aerowerx

Well-Known Member
Joined
Dec 1, 2011
Messages
5,610
Location
Marion, Ohio
The hard part is still getting good failure data in each direction needed and then convincing yourself that your work is good enough...

Bill
If I understand what this is about, you only need data for one direction---along the direction of the uni fiber?
 

Vigilant1

Well-Known Member
Lifetime Supporter
Joined
Jan 24, 2011
Messages
5,145
Location
US
I might be more interested in purchasing the book if I could see some sample pages online. No luck at Amazon. I'm not capable of using a high-level book on composite theory, but I could use a hands-on book with plenty of real examples using real units presented in an understandable way. The video doesn't make me optimistic that this book is for me.
 

stanislavz

Well-Known Member
HBA Supporter
Joined
Aug 21, 2016
Messages
632
Location
Lt
Two cup of tea later an watching fancy tutorial video. 4 metre wing, with strut at the middle, 1.1 m chord, without a nose and tail. it will break at this point (red x)

1597442153067.png
Displacement:
1597442183989.png
Loaded :

1597442376288.png
wing itself :
1597442427443.png
Section :
1597442463825.png
Stack for skins is 300 gsm woven + 300 gsm ud, for shear web only woven, also 300 gsm..
 

stanislavz

Well-Known Member
HBA Supporter
Joined
Aug 21, 2016
Messages
632
Location
Lt
And by moving strut from 50 to 60% of wing span we got better results.. 1.1 vs 1.7 reserve

1597443101764.png
 

stanislavz

Well-Known Member
HBA Supporter
Joined
Aug 21, 2016
Messages
632
Location
Lt
Wing left for later time, now tried to rebuild ch-701 rear fuselage part in "black wood"

Skin only, from 300 gsm x 2 ud + 300 gsm biax, weight ~ 16 kg. load - 600 kg from horizontal stab.

1597786822819.png

removed one biax , 11 kg

1597786906727.png

wood thinking :

1597786964943.png
At dark section 3 x 300 biax + 2 300 ud

added load from vertical taill to, at 4 kn

on light side side got this :
1597787338189.png

Some more toying and + 4 kg, most by changing light skin from one plier to two pliers:

1597787391977.png

And again light skin to two plier, but one is ud with vertical orientation

1597787737475.png

Better on deformation this is from one ud with proper orientation:

1597787772591.png

And this is from two biax :

1597787828406.png
 

stanislavz

Well-Known Member
HBA Supporter
Joined
Aug 21, 2016
Messages
632
Location
Lt
Register for trial and play with it. It is quite straight-forward. And it allow to see big difference from small changes.
 

stanislavz

Well-Known Member
HBA Supporter
Joined
Aug 21, 2016
Messages
632
Location
Lt
One more sesion, modeling D cell of wing.

Using this chord-vise load graph with spanwise load of 300kg/m, on wing section 1 meter long and 1.2 chord, modeling only front D nose:

1597952632121.png
Arbitrary profile from two circle and curve:

1597953164623.png

Loaded at top skin only, loaded using surface load at 0.0068 N/mm2, section is 320mm wide and 1000mm long.

1597953232390.png

Best so far with no foam, but nose ribs each 250 mm from single layer of 300gsm biax without any stiffner up here:

1597953297080.png

And deformations :

1597953332453.png

Ribs with foam inside and two layers of 300 gsm

1597953379766.png

Middle rib removed (RF reserve factr still > 2):
1597953417208.png

And with foam (and it did fail, but very small displacement) :

1597953498073.png

And yes, it is spar-less. Just for nose
 

stanislavz

Well-Known Member
HBA Supporter
Joined
Aug 21, 2016
Messages
632
Location
Lt
And for PS.

Got piece of fabric from fw 44 biplane - this piece 0.46mx 0.47m - weighted 88grams !!! Or 420 gsm. And yes, it is nor load bearing at all.

1597954142166.png
 

stanislavz

Well-Known Member
HBA Supporter
Joined
Aug 21, 2016
Messages
632
Location
Lt
Full wing - from two D cells, section 1200 mm long, ribs spaced each 600mm, false ribs each 300mm. Fron d Cell skin from three pliers of 300 gsm cf. Rear from two.. All else in drawings.. Loaded same as previous D cell - 0.0068 N/mm2 + 0.002 after spar.

Later, will compare to ribless wing with cored sandwitch.
 

Attachments

stanislavz

Well-Known Member
HBA Supporter
Joined
Aug 21, 2016
Messages
632
Location
Lt
vWww styled wing done at last. Two pieces were modeled 1.25 and 2.5 m long, fixed at both side. Nose is omitted, due to giving wrong result in meshing. Kind of overkill, but mass is ok i think. Any comments ?

numbers from best result for 2.5m section

1598132515661.png

1598132538324.png
 

Attachments

WonderousMountain

Well-Known Member
Joined
Apr 10, 2010
Messages
2,100
Location
Clatsop, Or
It looks good, did you optimize weight, or load distribution?
Are the results with high lift flaps roughly the same? Is the
bond strip double thickness, or counted as one? Can we
assume +-45 degree webs weaves? Or is this extruded...

I might like to use this if it can give accurate E-glass results.
 

stanislavz

Well-Known Member
HBA Supporter
Joined
Aug 21, 2016
Messages
632
Location
Lt
1. I do load is per chord as in my first post - 2.28 x max pressure on first 25% of profile, 1.14 to 0 for rest 75%.. Found it in old glider design book.

+ 25% of same load on bottom skin, but different direction - "pushing"

And i optimize weight (minimum layouts/different orientation ) trying to get similar reserve factor on all shells. But i am not optimising it span wise, due to first design for wing with spar, and to be easier. Sometimes one ud placed in correct place makes a miracles.

2. I do not know. Now - i am just toying. Theory behind this is good. Its do have E glass too.
 

stanislavz

Well-Known Member
HBA Supporter
Joined
Aug 21, 2016
Messages
632
Location
Lt
Some more screens, re-read Jim Marske book, its have some good references on D cell skin and rib spacing from fiberglass. Redone it in carbon, again no-foam :

1598474788987.png

1598474830746.png

And if it is deformation are kept in reasonable level - we got huge rf factor.. It was skins only with ribs and some shear web/ minimum spar, but still - 4.1 kg per 1 metre of span of wing having 1.1m chord is ok. And comparable with aluminium.

Tried with foam - its need thinner cf, in the area of 120-150 gsm, it is some lighter, and needs less ribs. But have reserve factor of 2-3 max.

This here is at 0.006N/mm2 (~10psf/45kg/m2 at 6g) at nose of wing, going to zero at the end of the wing.

Lighter version 3.3 kg per metre of span , all skin done from 300gsm biax x 2, addition rf is 4.6 on 6g...

1598475601910.png
 

stanislavz

Well-Known Member
HBA Supporter
Joined
Aug 21, 2016
Messages
632
Location
Lt
Could someone clarify me on huw much epoxy is used extra in three-layered sandwitch for foam glueing ? I have in my note number of 1 kg extra, but it is unreal i think..
 
Top