Quantcast

Flying Wing for Search and Rescue

HomeBuiltAirplanes.com

Help Support HomeBuiltAirplanes.com:

karan

New Member
Joined
Oct 27, 2014
Messages
2
Location
India
Hey Guys,
I've been working on designing a flying wing UAV since 2 months now. Human search and rescue using an autonomous drone is the prime objective of this project.

First Prototype.
One of the major aim of our project is to show how a visual human detection system that uses images from a normal camera can be implemented in software. The system would fit into the larger context of disaster management and more specifically how it can benefit search and rescue operations.
So the first prototype was made but it couldn't fly all that well. The tips stalled while rolling(probably because of no twist and less sweep) and led to a disastrous crash. As soon as our first prototype failed we got down to designing the second version.
I need some advice with the design of the second air-frame.

1.jpg
Currently I'm using E334 for the root section and MH 49 for the remaining plane. I'm using a twist of around -4 degrees and I've a non linear variation throughout the wing. First one third has 1/3 of the twist. Next 1/3rd has the same amount and the last 1/3rd has -4
Also I'm using a symmetrical profile for the tip because it performs well for -ve aoa as well.
Although the problem is when I applied the pankin twist formula to this design it churned out -9 which reduced the performance drastically.

2.jpg

The one with the points is the current prototype.
The problem is I'm not really confident about the design and I don't want to take risks after my first crash. So I'd like your general opinion on this design and also any suggestions that would help me improve the design.

I've been reading a lot of this forum lately and this finally decided to register and ask a few questions from you guys.
Also is anyone willing to sell the book Tailless Aircrafts in Theory and Practise, I'd be interested in buying.

Also some updates on the image processing I've done to detect humans.

HOG detector successfully tested on a sample camera feed.

Thank you for your time and consideration.

Cheers,
Karan
 

karoliina.t.salminen

Well-Known Member
Joined
Oct 7, 2008
Messages
407
Location
Finland
Your first prototype has higher aspect ratio, and probably higher L/D than the simulation in the XFLR5(?). I would like to know what made you change from that wing geometry to the lower aspect ratio, what was the design parameter that made you do that call?
 

karan

New Member
Joined
Oct 27, 2014
Messages
2
Location
India
Hey
The model name that you see as Discovery 1 in those polars is just a version of the current design. It has a smaller root and tip chord the former being 250 mm and the latter being 200 mm. That obviously increased the performance but I had to increase the root chord since I needed to make enough space for all the processors I'm loading the plane with. So that's a compromise I had to make.

Currently this model has CLmax of .78 at 11.4 degrees. It stalls after that.
 

Riggerrob

Well-Known Member
Joined
Sep 9, 2014
Messages
1,621
Location
Canada
Increasing sweep will definitely improve pitch control authority, because it moves the control surfaces farther from the center of gravity.
The new constant chord is also an improvement. First, constant chord simplifies construction. Secondly, constant chord increases pitch stability. Thirdly, a larger tip chord allows larger manufacturing tolerances at tiny Reynolds numbers IOW a large tip chord tolerates bug splat better.
To look at extremely large tip chords, research Withold Kasper's flying wing gliders from the 1950s and the Kasperwing powered ultralight from the 1970s. Kasper used triangular elevons which had tip chords longer than the elevon root chord.
IOW Kasper's flying wings had straight, swept leading edges with the same sweep angle across most of the trailing edge but sweep angle increased along the elevon trailing edge.
 

oriol

Well-Known Member
Joined
Dec 31, 2009
Messages
770
Location
Barcelona, Spain.
Karan,


I think that the best UAVS for working on reduced spaces in case of natural disasters, eartquakes and the like, are hands down small multicopters. This a video recorded by one of those small unmanned aircrafts inside the ruined and abandoned concrete buildings of "battleship island" in Japan.




If comparing drones with traditional aircrafts for surveillance tasks, the latter ones are still by far more economical to operate. I do not think that with a conventional RC sized airplane, like the one you are working with, you will have enough range to cover large areas: it is still comparable to an electric multicopter.

On the other hand even If you are working on that interesting project as a student: just to learn and experiment things. I would recommend to focus your efforts only on one aspect of the multiple practical aspects, for example artificial vision versus aerodynamics. Instead of spending your valuous time trying modifications with a CFD program, and risking many crashes before you work everything out, just use a conventional RC airplane.

There are plenty of RC clubs all over the world and of forums devoted to that subject with free information on the net.



Oriol
 
Last edited by a moderator:

bmcj

Well-Known Member
HBA Supporter
Joined
Apr 10, 2007
Messages
13,541
Location
Fresno, California
Karan,


I think that the best UAVS for working on reduced spaces in case of natural disasters, eartquakes and the like, are hands down small multicopters. This a video recorded by one of those small unmanned aircrafts inside the ruined and abandoned concrete buildings of "battleship island" in Japan.


[video=youtube;73B5Dv0aNJM]https://www.youtube.com/watch?v=73B5Dv0aNJM[/video]
Did it take a swim, then pull itself out of the water at the 1:20 mark?
 
Last edited by a moderator:
Top