Show messages:
1-4
5-24
25-31
From: Hamish Mead (HAIRYKIWI)
'fraid not Pilou - just one person and all the glorious blue sky they can eat ...maybe with cake and coffee at a neighbouring field for desert. ;)
- Hamish
From: Hamish Mead (HAIRYKIWI)
As people may well start using the NACAAirfoilGenerator script (NACAAG) to make airfoils for real wings intended to take litigious cargo safely into the air - and return it just as safely to terra firma, but mostly out of personal pride, I thought I ought to continue making an effort to check the accuracy and quality of its output. That said, I make no claim as to the suitability of the script for any purpose so please DOUBLE-CROSS-CHECK everything it generates meets requirements before relying on it. The full legal disclaimer is included the JavaScript file. Right, enough quasi-legalese waffle... Here's what I've done and discovered so far:
I generated a NACA 23012 airfoil of '100 stations' and 100mm chord length with NACAAG, and compared it firstly with the ordinates and secondly the curve made from the ordinates generated by
naca456.exe* - using the input file '23012_dencode3.nml' (below) to generate the greatest number of output points possible; '98 stations'.
Comparison 1
At a number of randomly chosen stations, I measured the perpendicular distance from points generated by naca456.exe to the NACAAG generated NACA 23012 airfoil curve.
Differences of between 0.00001mm to 0.000049mm were observed. Those differences are well within manufacturing tolerances, but why any difference at all? I suspect it may be due to (cumulative) rounding errors - or possibly truncation error, due to the naca456 airfoil ordinates (to 6 DP) being scaled from unit 1 to 100 in CAD, rather than in script. Any other ideas?
Comparison 2
Using Rhino's
Curvature Graph tool, I compared relative smoothness of the NACAAG generated NACA 23012 and the through point curve created using the MoI ImportPointFile script and the file 'NACA 23012 points from naca456_dencode3.xyz' - which is simply the actual naca456 ordinate output file, '23012_dencode3.out' reformated to xyz and rearranged to enable one continuous through point curve be created. Considering the relatively small differences measured in Comparison 1, I find the contrast between the two curvature graphs in the following screen capture quite surprising.
It's reassuring to know that the smoother of the two airfoil curves was generated by the NACAAG script, but I have my doubts a wing made using either method would look much different in real life after coming out of the paint-shop. Any thoughts from anyone who uses curvature graphs as part of a pre-manufacturing inspection process? What for example, is a reasonable rule of thumb for setting curvature graph display scale to avoid unnecessary rework of a curve or surface?
As an indication of just how much cross-checking needs to be carried out to verify the results of a new script, the airfoil ordinate generator basic code I previously mentioned -
http://m-selig.ae.illinois.edu/ads/naca980820.bas appears be generating unreliable output. Now it could be that the three DEF functions I needed to removed from that program so as to make it acceptable to the QB64 basic compiler on my Win7-64 machine were in fact being used, but I couldn't see where. I'll try to find out why the results are not as expected directly from the author. For the meantime, and especially because I'm not certain I haven't created the problem myself, I won't publish any comparison results - just be aware.
* naca456.exe is a public domain airfoil generator program written in Fortran by renowned aeronautical engineer Ralph L. Carmichael. As he notes on his site pdas.com, naca456 is a further development of NASA's work, first documented in the 1970's, to create a computer program to obtain ordinates for NACA airfoils.
- Hamish
Attachments:
23012_dencode3.nml
23012_dencode3.out
Comapison of NACA 23012 by NACAAirfoilGenerator MoI script and naca456_exe.3dm
NACA 23012 points from naca456_dencode3.xyz
Image Attachments:
Comapison of NACA 23012 by NACAAirfoilGenerator MoI script and naca456_exe - Curvature Analysis .png
From: Hamish Mead (HAIRYKIWI)
Latest (minor) release:
Point access syntax bug discovered under MoI V4 beta - now compatible with MoI V3 and V4.
Bug squashed by r2d3 - many thanks again Ralf :)
Attachments:
NACAAirfoilGenerator_2018-03-12.zip
From: r2d3
:-)
From: Frenchy Pilou (PILOU)
Seems i must update my French version! :)
http://moiscript.weebly.com/aile-naca.html
From: Hans
Hello Moi user.
If somebody is calculating a new homebuildt aircraft and think about using NACA profiles please read the booklet from
Harry Riblett.
GA Airfoils
A CATALOG OF AIRFOILS FOR GENERAL AVIATION USE
To buy at Aircraft Spruce and Spezialitys for about 20$.
It is an eye opener and may safe your live.
With best regards
Hans
From: Hamish Mead (HAIRYKIWI)
Hans, my apologies for this late reply and thanks for adding your sound advice.
I am in no way qualified, however from what I've learned about NACA's early studies - and read about Harry Riblett's work - I couldn't agree with you more.
There's a super article about Harry Riblett here, for anyone interested:
https://www.kitplanes.com/the-airfoil-adventures-of-harry-riblett/
Show messages:
1-4
5-24
25-31
