A complete report on the aerodynamic and strength design of an ul aircraft

HomeBuiltAirplanes.com

Help Support HomeBuiltAirplanes.com:

christos

Well-Known Member
Joined
Nov 9, 2018
Messages
52
Location
Greece
[FONT=&quot]My name is Christos and i am a civil engineer and aviation enthusiast from greece. First of all i would like to thank you all for your interesting and useful threads. It helped me throughtout the design. Now i am going to build my own design 2 seat ultra light composite aircraft. For the design and development of the aircraft all tools available to the modern engineer have been properly used.
This is the technical report of the design.

Best regards!

[/FONT]
https://www.lisafea.com/citations.html
https://lisafea.com/pdf/Design_of_Ultralight_Aircraft.pdf
 

proppastie

Well-Known Member
Log Member
Joined
Feb 19, 2012
Messages
4,568
Location
NJ
I did not see any structural details so I am somewhat confused as to how you were able to do an FEA...are the elements assigned values?
 

christos

Well-Known Member
Joined
Nov 9, 2018
Messages
52
Location
Greece
Good evening sir! I am sorry about late response, but i am construction engineer and now i am working on pinios' river dam restore and i worked today.
I want to thank you for your interest!
According to your question, i will try to explain you (i think i understand you).
Lisa is a powerful finite element analysis program like ansys, code aster, etc.
If you want to solve a model in one of above softwares you have to:
1) make a 3d mesh model (i prefer stl or step but it is a personal choice)
2) choose materials and put on model
3) make a good 3d elements' grid.
4) choose your support and loads
5)Run it and that is all!

I publish only a technical report and i you can see only the results (of finite element analysis).
But i don't have any problem to share some picture of mesh and grid.
pic1: model
1.jpg

pic2: grid
2.jpg

pic3:material
3.jpg

pic4: material
4.jpg

pic 5: analysis results (visual)
5.jpg

pic 6: analysis results (values)
6.jpg

And one openfoam's analysis animation ( Pressures around the aircraft at speeds just before stall with fully extended flaps)

Webp.net-gifmaker (1).jpg
It will be a really fast and low stall speed aircraft!
Feel free to contact me if need something else!

Best regards!
 
Last edited:

proppastie

Well-Known Member
Log Member
Joined
Feb 19, 2012
Messages
4,568
Location
NJ
Lisa is a powerful finite element analysis program like ansys, code aster, etc.
I have a copy of Lisa but found it too difficult to use...the learning curve is high. I use Grape FEA....but it probably is not as powerful.
 

christos

Well-Known Member
Joined
Nov 9, 2018
Messages
52
Location
Greece
Dear sir,

I suggest you make an stl drawing (in a cad software) and put it in lisa. Also i am able to help you!

Best regards!
 

christos

Well-Known Member
Joined
Nov 9, 2018
Messages
52
Location
Greece
Finally i made a video


I hope you will like it :)
 
Last edited by a moderator:

christos

Well-Known Member
Joined
Nov 9, 2018
Messages
52
Location
Greece
And one non linear Code Aster's analysis (hard landing)

 
Last edited by a moderator:

Retiree

Well-Known Member
Joined
Nov 29, 2010
Messages
81
Location
USA
Hi christos,
How many points are you using in your CFD grid and how many points do you use to resolve the boundary layer?
Thanks for sharing your results.
Doug
 

rv6ejguy

Well-Known Member
Joined
Jun 26, 2012
Messages
3,749
Location
Calgary, Alberta, Canada
Cool project. The total empty weight, landing gear weight and engine/ redrive weight seem pretty low at first glance compared to other aircraft in this size range. The last time I weighed an EA81 engine, it was about 66kg bare- no exhaust, intake, carb, redrive etc.
 

christos

Well-Known Member
Joined
Nov 9, 2018
Messages
52
Location
Greece
Hi christos,
How many points are you using in your CFD grid and how many points do you use to resolve the boundary layer?
Thanks for sharing your results.
Doug
Dear sir Doug,

Depended on analysis type. For low speeds (every speed below 0.3 mach is low for fluid dynamics) a k-omega sst model or k-e is a good choice. About boundary layer, is (maybe) the most important subject of turbulence's flow theory. There are a lot of different solutions out there, but i know that the theory is one and you have to calculate y+ correctly for your application and after that design a good mesh with a correct amount of boundary layers.

In my opinion (for k-omega sst model, and low speeds) if you want to run a model and you don't want high accuracy, take value 1 for y+, 20 boundary layers and 15-20 million elements. But it isn't good enough if you want high accuracy (98-99% of experimental one). And it works well in low speeds. Also if you choose k-e, you need a different value of y+.

Also i designed a motorglider and this is a pic of his mesh.
1.jpg
 

christos

Well-Known Member
Joined
Nov 9, 2018
Messages
52
Location
Greece
The fuel consumption is very optimistic. What engine did you design the plane for?
What is the cruise L/D and the L/D max?
It is a flaperon aircraft. And flaperons operate as cruise flaps also as a result max L/D 21 (25 with retractable landing gear).

I would like to notice that fuel consumption is liters per 100km and not liters per hour! :) yes it still remains economy aircraft. But it is a full laminar flow design and it is well know that this type of aircrafts are efficiently.
 

christos

Well-Known Member
Joined
Nov 9, 2018
Messages
52
Location
Greece
Cool project. The total empty weight, landing gear weight and engine/ redrive weight seem pretty low at first glance compared to other aircraft in this size range. The last time I weighed an EA81 engine, it was about 66kg bare- no exhaust, intake, carb, redrive etc.
EA81 is an excellent also engine. But i will use an A71 one, which is 7-8 kg lighter. I own it and it is (87ps or ~85 hp)version so it is an easy solution for me. Carb'sweight included in fuel and ignition in electrical equipment.

I am planing to use this type of engine because i trust subaru and this engine is getting close to rotax 912 ul so i will be able to compare it with other aircrafts
 

Retiree

Well-Known Member
Joined
Nov 29, 2010
Messages
81
Location
USA
Dear sir Doug,


In my opinion (for k-omega sst model, and low speeds) if you want to run a model and you don't want high accuracy, take value 1 for y+, 20 boundary layers and 15-20 million elements. But it isn't good enough if you want high accuracy (98-99% of experimental one). And it works well in low speeds. Also if you choose k-e, you need a different value of y+.
View attachment 75870
Thanks for the info. My next question is what kind of a computer are you using to make the calculations and how long does it take?
 

christos

Well-Known Member
Joined
Nov 9, 2018
Messages
52
Location
Greece
Thanks for the info. My next question is what kind of a computer are you using to make the calculations and how long does it take?
Depends on the solver's settings. The bigger the tolerance the lower the time. But if you choose too big tolerance, could not convergence your solver . I bought core hours and it was simple to me run these models. I don't know exactly how many hours need a pc to run these models, but you need a strong enough pc, at least i5 or i7 and 64 gb ram and (i suppose) something like 30-40 hours for each run. Or lower if you choose a little bigger tolerance. But you will need a lot of hours when you run flutter analysis because it is time depended analysis.

Also a good choice is to buy a used workstation (with one or two good processors and big ram) or build a good amd pc. I don't believe that intel's processors are better than amd's for this type of work. You can go with intel but it will be more expensive solution.
 

Retiree

Well-Known Member
Joined
Nov 29, 2010
Messages
81
Location
USA
Technology has progressed a lot in 20 years. Back then a coworker of mine got a 16 node (16 pc's connected in parallel) Linux cluster. He was able to do similar solutions to what you are doing. I was doing similar calculations on remote Linux clusters requesting 48 processors. Time marches on...
 

dino

Well-Known Member
HBA Supporter
Joined
Sep 18, 2007
Messages
677
Location
florida
The EA81 Subaru comes in at different weights depending on model. Single carb version which I have converted is about 124lbs bare. EA71 is a good choice.
 
Top