Big Project | Project | Project Manager |
|---|---|---|
Fixed Wing | Wings |
Table of Contents
Task Description
We need to determine what shapes of airfoils we want to try for the
This may include trying 2 general shapes, and several versions of these shapes with dimensions tripled or halved to allow us to understand what these changes do for the flight characteristics of the drone.
Check out NACA airfoil generator to generate plots of different airfoil shapes: NACA 4 digit airfoil generator (NACA 2412 AIRFOIL) (airfoiltools.com)
Other resources:
How To Read NACA airfoils (4 digit, 5 digit, 6 digit) (youtube.com)
Constraints
This is written pre-scoping and architecture meeting, so for now there are no concrete constraints, this will be mostly research to start and we will get a better idea of what we’re looking for later.
Constraints | Written By | Append Date |
|---|---|---|
Supports an aircraft of approximately 5kg (target) |
| |
Wingspan of approximately 1.2m (target) |
| |
Cruise speed of 30m/s (approx. 100kph) |
|
Constraints marked as “target” are rough estimates and will most likely change as they are not finalized. More constraints can be added as targets to help guide the airfoil design process in a general direction.
Assignees
Assignee | Asana Task | Date |
|---|---|---|
Nathaniel Li Evan Janakievski Thushanth Parameswaran Conall Kingshott | link to asana task assigned | Date assigned |
Key Terms
Nathaniel Li put this here for clarity.
Chord: imaginary straight line joining the leading edge (usually left) and trailing edge (usually right) of airfoil
LE: leading edge
TE: trailing edge
CL: lift coefficient, determined from type of airfoil and angle of attack
CD: drag coefficient
AoA: angle of attack, angle at which chord meets relative wind
CM: pitching moment coefficient
Clockwise (pitch up) positive, counterclockwise (pitch down) negative
Author: Nathaniel Li Updating Date:
Summary of How To Read NACA airfoils (4 digit, 5 digit, 6 digit)
4 digit - Ex. NACA 2412
First digit 2: max camber of airfoil
2% of chord (or 0.02c)
Second digit 4: location of max camber
40% of chord (or 0.4c)
Third and fourth digit 12: thickness of chord
12% of chord (or 0.12c)
Chord is straight line from leading edge (LE) to trailing edge (TE) going from left to right
In this case 0.4c is the location of max camber
t = 0.12c is the thickness
5 digit - Ex. NACA 46015
First digit 4: design lift coefficient (CL)
Multiply by 3/2 and divide by 10 to get CL
Ex. 4 x 3/2 = 6 → 6 / 10 = 0.6 → CL = 0.6
Second and third digit 60: location of max camber
Multiply by 1/2 and divide by 10
Ex. 60 / 2 = 30 → 30 / 10 = 0.3 → 30% of chord from LE (or 0.3c)
Fourth and fifth digit 15: thickness in %
Ex. 15% of chord length (or 0.15c)
6 series - Ex. NACA 64-320
First digit 6: series #
Second digit 4: location of min pressure
40% of chord (or 0.4c from LE)
When airfoil is at zero lift condition
Third digit 3: CL
No factor multiplication → CL = 0.3
Fourth and fifth digit 20: thickness in %
20% of chord length (or 0.2c)Author: Nathaniel Li Updating Date:
Research of Various NACA Airfoils
Other NACA airfoils include: 16-series, 7-series, 8-series
Quick summary based upon course content from Stanfords AA200 Applied Aerodynamics course
Family | Advantages | Disadvantages | Applications |
|---|---|---|---|
4-digit |
|
|
|
5-digit |
|
|
|
16-series |
|
|
|
6-series |
|
|
|
7-series |
|
| Seldom used |
Author Thushanth Parameswaran Updating Date: Jun 12th 2024
Based on the research so far any of the NACA 4 digit airfoil that meets the requirement below will be a good starting point for us to investigate further.
Camber % 4 to 6
chord from 9" to 12"
thickness % 10 to 14
One of the important constraints is the ease of manufacturing.
Manufacturing technique
-Skins for Airfoil
Oracover heat film
-Ribs
balsa wood
aluminum
Author: Nathaniel Li Updating Date:
Preliminary Selection of NACA Airfoils
Based on searching through forums, NACA 2412 and 4412 are popular choices for RC planes
Relatively thinner designs provide lower drag
The slight under cambered design provides higher lift
Although a flat bottom wing could provide more speed (due to lower drag), there are lift penalties
To help narrowing down the options, I’m opting to prioritize stall characteristics initially before optimizing to make the initial flying/testing easier
This would favour a design like NACA 4412 since the under cambered design produces more lift at slower speeds and stalls at a higher angle of attack compared to a flat bottom airfoil https://forum.flitetest.com/index.php?threads/airfoil-type-for-slower-flights.69366/
Higher lift at lower speeds can also be reproduced by thickening wings but of course, that increases drag exponentially
To get a sense of CD and CL, I decided to run through a quick CFD simulation of a NACA 4412 to get some preliminary data and also learn to use Ansys following Smile’s guidehttps://uwarg-docs.atlassian.net/wiki/x/CoCqiQ
Can be found in Fixed Wing → Sims → Fixed Wing CFD → NACA 4412 - 2m Wingspan
The test wasn’t very formal and more for learning so I’m summarizing it shortly here
Size: 2m wingspan by ~1m chord (CAD taken from online)
I realized afterwards that we had decided on a 4ft wingspan which is my bad, will use in future https://uwarg-docs.atlassian.net/wiki/x/E4Jjm
Material: Aluminum
Speed: 80km/h (an approximation of a cruising speed)
AoA: 0 degrees (cruising again)
CL: 1.8331316
CD: 0.16140058
The bounding box was ~ 5 times bigger than the airfoil
200 iterations didn’t take too long (~30 mins) so can probably increase that in the future
Don’t feel like much can really be taken from the data
From here, I think some good next steps are:
Determining if there is a specific constraint or priority in mind (eg. low speed takeoff/landing which requires lower stall speeds, high speed flights which requires lower drag)
Choosing some other designs to play around with or even test with some popular designs for RC planes being:
NACA 2412
Another under cambered design good for lots of lift but high drag
Semi-symmetrical
Similar to under cambered designs but thicker
Lift-drag ratio is typically good enough for most applications
Clark-Y is heavily used in general purpose aircraft design
Author Thushanth Parameswaran Updating Data: 25th June 2024
Reynold’s Number and Some Airfoil Choices (Selig High Lift)
Background
Some of the important factors in choosing an airfoil are
Coefficient of Lift (CL)
Coefficient of Drag (CD)
Reynold Number
 | = | reynolds number |
 | = | |
 | = | |
 | = | characteristic linear dimension (In our case Chord length) |
 | = | dynamic viscosity of the fluid |
The Reynolds number is the ratio of inertial forces to viscous forces. The Reynolds number is a dimensionless number used to categorize the fluids systems in which the effect of viscosity is important in controlling the velocities or the flow pattern of a fluid.
AoA: angle of attack
Airfoil choices
Before choosing an airfoils, we need to know the plane's velocity. Since we do not have any fixed velocity constraint. I am assuming the velocity to be from 20kph to 40kph (low Reynold number). I will also be choosing the chord length to be 9 in. We know the Wing span is constrained to 4 ft (48 in). Using this we can calculate the aspect ratio which is Wingspan/Chord length. 48/9 is 5.333 which our aspect ratio
40 km/h to 80 km/h is low speed which means the Camber of the airfoil needs to be high to produce high lift. But having a High camber comes with a side effect which is drag. So we have to strike the right balance between these factors.
Initially, I asked my friend from UBC Aero Design who recently competed in the SAE west Micro division. He told me the Airfoil they use is S1223. The specification of the this airfoil is listed below. I started to reference this research paper (https://m-selig.ae.illinois.edu/uiuc_lsat/Low-Speed-Airfoil-Data-V1.pdf). which has much more wind tunnel-tested airfoil data.
The link below gives much more technical data regarding the S1223 airfoil.
http://airfoiltools.com/airfoil/details?airfoil=s1223-il
Final thoughts
It is very difficult to conclude what is the right airfoil. I feel air foil similar to SP1223 are very good starting point for us to investigate flight behaviour and make necessary changes to the airfoil as needed. The conclusion is based on the assumption that velocity will be from 20kph to 40kph. If our flight operating window is not within range then reducing the camber of the Airfoil will be one solution. since this decrease drag and lift might have minimal changes.
Reference Book:- From The Ground Up by “Sandy” AF MacDonald (I have a physical book and could not find Online pdf version)
Author: Sohee Yoon Updating Date:
Summary of Everything so Far
Based on all the information above and some further research, I do agree with Nathaniel Li where I found NACA 2412 and 4412 to be the most popular airfoils for RC planes (both under-cambered). Also mentioned by Nathaniel Li is to look into semi-symmetrical airfoils. Below is just the difference:
I found this journal article that compares the foils Nathaniel Li and Thushanth Parameswaran looked into and it states, “They compared lift and drag coefficients for these two airfoils [NACA 4412 and S1223] and suggested that, NACA 4412 is suitable for use in Sports Plane, whereas, S1223 is suitable for Heavy lift Cargo Plane due to its high lift.” ~ https://www.ijert.org/research/an-aerodynamic-comparative-analysis-of-airfoils-for-low-speed-aircrafts-IJERTV5IS110361.pdf (of course the conditions these airfoils were placed is different in our case).
The airfoils we should possibly test based on the research above:
NACA 4412
NACA 2412
S1223
Very similar to S1233 in terms of Camber.
Author: Nathaniel Li Updating Date:
Decisions and Next Plans From 2024-06-26 Mechanical Meeting Minutes
Narrowed down to airfoil choices to NACA 4412 and a similar airfoil with ~8% camber
Reasoning: simpler design with lots of data readily available online, easier to manufacture
NACA 8412 was a choice for 8%, however it is not a standardize airfoil so it’s much hard to find data for it (any standardized NACA airfoils have plenty of data online)
To choose airfoils, the plan is to try 2 airfoils with low and high camber (4-8%) and optimize for something between
Camber determines a majority of flight characteristics so its better to start by varying only 1 factor
AT350 + 14 x 7” props being used → this may be useful further in airfoil development
4ft wingspan gives us a starting point for various chord lengths
Next steps:
Choose another (hopefully standardized or with plenty of data available) airfoil with 8% camber and compare to NACA 4412
Choose some conditions that are important to determine like Reynolds number and aspect ratio (high aspect ratio for low speed) to guide the optimization process
Discuss control surfaces like ailerons and rudder
Most likely stay away from dual rudder and instead use simpler single rudder
Author: Nathaniel Li Updating Date:
Factors when Selecting/Optimizing Airfoils
We’ve decided to use metric units going forward for airfoil design based on https://uwarg-docs.atlassian.net/wiki/x/AYBEmw It is also easier to stick with basic units (eg. m/s instead of km/h) to prevent unit errors.
CL and CD → CL/CD Ratio
Mass
This determines how much lift we need but only becomes more useable when doing sims
Flight speed
100km/h (or 28m/s) was discussed in discord as a target for cruising speed
Chord length and wingspan
These go hand-in-hand; an aspect ratio (wingspan : chord length) is a measure of the wing’s slenderness and 4:1 was mentioned in the mech meeting
Wingspan of ~1.2m to ~1.4m from https://uwarg-docs.atlassian.net/wiki/x/P4BLmw but is up for discussion
Wing loading is another parameter → Directly affected by surface area of wing (ie. chord length and wingspan)
With a 5ft wingspan (1.5m) and 8” (0.2m) chord length, it seems that the wing loading may be too heavy so the chord length and wing span may need to increase to reduce this (ie. increase upper surface area)
Reynolds number
Flying altitude
Generally, at higher altitudes the air is less dense so less lift and drag is produced
These are some of the major dictating factors that we’ll attempt to change and optimize through our design. By constraining these factors more (especially flight speed, chord length, and wingspan) it can make design of the wings easier.
High vs. Low Camber Wings
As mentioned in https://uwarg-docs.atlassian.net/wiki/x/AYBEmw an easy way to determine airfoil design is to choose one simple design and only optimize for one factor. This may be a bit limiting but for now it should be sufficient, especially given that there are several designs used in RC planes already that can be good starting points. In the meeting, it was decided to specifically research 2 wings with low and high under camber. More camber produces more lift with a drag penalty, and vice versa is also true. The camber profile of an airfoil also heavily influences its flight characteristics (eg. stall characteristics, manoeuvrability).
Since the 8% camber wing previously found is not standardized, there is little data to support research, at least without running a bunch of sims. For now, I’ve found that NACA 2412 (2% camber) and NACA 6412 (6% camber) are good choices. The both are popular in the RC world, have same parameters except for the camber and are relatively simple and easy to manufacture.
Wings with above 6% camber (like the S1223 with 12%) can become difficult to manufacture accurately.
Given these 2 wings, it is possible to compare the data using resources like http://www.airfoiltools.com/airfoil/naca4digit (specifically has the ability to compare airfoils in its database). https://www1.grc.nasa.gov/wp-content/plugins/cheerpj-integration/lib/applets/FoilSimStudent/FoilSimStudent.html can also be used for faster references, though it is more limited in data availability.
Author: Nathaniel Li Updating Date:
NACA 2412 vs. NACA 6412 (High vs. Low Camber) at Cruising Speed
Main resource being used: http://www.airfoiltools.com/airfoil/naca4digit
Has NACA 2412 and 6412 data available
Context/Background
Have to determine a Reynold’s numbers based on parameters below:
Velocity: 30m/s (108 km/h)
This makes a big change but first we will target cruising speed
Chord width: 0.3m (approx. 1ft)
Also makes big changes but 0.3m is realistic for size
Kinematic viscosity: 1.5111E-5m^2/s
For air at 20C and 1 atm
This results in a Reynold’s number of 595,593
The nearest approximation we can use from the data is 500,000
A critical N-Factor (Ncrit) must also be chosen
Parameter in XFOIL (software used to generate data in airfoil tools) that controls the transition of laminar to turbulent flow over the airfoil
Higher Ncrit represents a less sensitive flow (ie. More likely to remain laminar) and lower Ncrit represents a higher sensitive flow (ie. More likely to transition to turbulent)
There isn’t an exact way of choosing but based on online forums, 9 is the standard for most applications
If we encounter a lot of turbulence in the future, lower Ncrit values of 4-8 can be used to simulate “dirtier air flow”; the table below was found from the air foil tools website
Note: Airfoil tools only supports Ncrit of 5 or 9
Situation | Ncrit |
|---|---|
sailplane | 12 to 14 |
motorglider | 11 to 13 |
clean wind tunnel | 10 to 12 |
average wind tunnel | 9 |
dirty wind tunnel | 4 to 8 |
Generated Data
Based on the above parameters, I used the airfoil comparison feature to plot and compare
It then generates a bunch of plots comparing CL, CD, CM (pitching moment coefficient, causes airfoil to pitch downwards) and AoA (Alpha)
NACA 2412: max CL/CD of 87.3 at α = 5°
NACA 6412: max CL/CD of 114.2 at α = 6.25°
Analysis
Based on the max CL/CD, we can tell that 6412 will give us a better peak CL/CD
Can be beneficial in takeoffs/landings with shorter runways; takeoff and approach speed can be slower
Based on the CL vs CD plot:
6412 consistently has a higher CL/CD
6412 initially reaches peak CL/CD faster as CD increase
CL/CD tapers off slowly for both
Based on the CL vs Alpha plot:
6412 consistently has a higher CL for a given AoA
Both have similar CL vs AoA profiles so as AoA increases, the lift characteristics remain relatively the same
Based on the CL/CD vs Alpha plot:
Overall, 6412 maintains a higher CL/CD as AoA increases
Specifically starting around ~ 5° the CL/CD increases much faster compared to the 2412
Past the peak, CL/CD for 6412 drops severely
Possible efficiency concerns when going past 5°
Possible stall concerns as CL/CD drops very quickly
Based on the CD vs Alpha plot:
As expected, CD increases exponentially as AoA increases but both are very close
For brief moments between 5-10°, 6412 has lower CD
This is the same zone as on the CL/CD vs Alpha plot where 6412 has a much higher CL/CD but drops rapidly
Beyond 10° and the drag penalties kick in hard, most likely need to avoid this as much as possible
Based on the CM vs Alpha plot:
I’m not entirely sure about this as I’m still researching into CM but it is important for stability
Don’t be thrown off by the negative value, it just means it’s pitching down
For both, they have a positive slope meaning as AoA increases, CM also increases (ie. less pitching down)
This positive feedback loop can lead to instability as the wing pitches up uncontrollably
As pitch increases, there’s a chance that the wing pitches up even more eventually leading to stall
A negative slope is desired to combat the instability issue, but a positive slope doesn’t necessarily mean there will be instability issues, it should just be something we look out for
This is something that will become more relevant when looking at horizontal stabilizers and elevators; may need to design them with this in mind
Main Takeaways
Seems like CL/CD is overall much better with 6412
Harsh drag penalties may occur but only in specific scenarios past 5° of pitch
Lower takeoff and landing speeds possible with 6412
CL/CD tapers off slowly for both
Possible concerns of dramatic stall characteristics with 6412 based on exponential drop off in CL/CD vs AoA
Under 10°, CD similar for both
2412 lower CD from ~ 0-5°
6412 lower CD from ~ 5-10°
CM is a concern for both 2412 and 6412 and there’s a chance of instability (wings pitching up a lot randomly)
This is more of risk and control thing, I think it’ll be more understandable with real life testing
Generally, I think 6412 is a good choice based on how much extra lift it can provide us
Extra lift should make initial flying easier plus more room for error in takeoff/landing
If we realize it’s too much drag, we can reduce the camber and try something like 4412
It is possible to do more testing at different speeds and with different chord widths but everything is mostly arbitrary at the moment
Author: Nathaniel Li Updating Date:
Updates from 2024-07-31 Mechanical Meeting Minutes and Designs
Plan to use 1/2” aluminum box tube for spars (also being used for frame)
Sohee Yoon working on a design for balsa ribs
RPC laser cutters cuts up to 9.5mm thick pieces
If too thin, can glue two pieces together
Should keep in mind control surfaces that also need to be added (ie. ailerons and possibly flaps)
My current WIP design:
0.23m chord length, 1.5m wingspan, NACA 4412 based on fixed wing calculator
Wingspan is end-to-end length but will most likely break into 2 sections so that it can connect to fuselage
5mm rib thickness although 2-3mm (1/8”) seems common for rc planes with wingspan around 1-2m
Will look into what kind of Balsa sheets we can get, anything within 2-5mm range will do but should be able to get away with thinner side
Also can save lot of weight by cutting out material
Rib spacing is kinda unknown but ~ 2-3” and up to 5” is common
5-10cm seems like a good range
Main issues to tackle:
Supporting LE and TE to prevent concaving issues
Thinking of doing a 3d print that covers them to give rigidity
Positioning of main and rear spar as discussed with Smile Khatri → seems like 25-30% of chord length is ideal for main spar
For simplicity, usually a spar at LE and TE would suffice and also give the desired profile (and rigidity) but cutting the profile is very hard for Balsa tubes
Author: Sohee Yoon Updating Date:
Balsa Ribs Update
Designed a cut-out for the ribs. Based on this research article, elliptical cut-out shapes result in lower stresses in the aircraft ribs (studied using ANSYS) since sharp corners was what leads to higher stresses. Hence, I used elliptical cut-out shapes for our ribs. I also used a similar design in their studies (e.g., 3 ellipses) and chose the size in respect to their ratio between ellipses.
But…I did read on reddit that this doesn’t matter if your wingspan is greater than 6 inches…but that’s also from a random redditor 🤣
The placement of the ellipses solely depended on the spar cut-out and the thickness of the outline. There is at least 2-3 mm distance of clearance and 12.5 mm distance from the outer edge of an ellipse to another ellipse.
The ellipses and spar cut-our follow (or on) the camber line.
I wasn’t sure if the main spar and rear spar will be the same so I just made a cut-out with the one Evan Janakievski added in the Wings file. This is welcome to change
I’m curious if we should fillet the corners to reduce stress but idk if the spar will fit
Thickness is 5 mm but we can def slim it down to 2-3 mm like Nathaniel Li suggested
Wings Update
8 ribs are spaced 80 mm apart and the wing skin is 600 mm (all are estimates rn)
LE and TE Concave Issue
3D prints that cover them is one solution but I was also considering a long wooden rod like in the images below and we can cut-out a circle in the front to insert it. But this will only for the the LE issues.
Control Surface Idea
I thought this idea was pretty cool, where the servo to control the flaps are inside the wing (attached to the ribs): https://timtuxworth.me/2021/07/18/aileron-servos-on-my-guillows-sopwith-camel/. I’ll prob look into it more tho
Author: Nathaniel Li Updating Date:
Balsa Selection
Many hobbyists prefer 1/16” or 1/8” (~1.5-3mm) for wingspans of around 90” (over 2m) https://maestronet.com/forum/index.php?/topic/116884-rib-thickness/ https://www.rcuniverse.com/forum/scratch-building-aircraft-design-3d-cad-174/1524641-rib-thickness-spacing.html
1/16” is thinner and saves weight
Ribs are approx 30mm x 230mm
https://www.allrockets.ca/Balsa many sheet thicknesses (1/32”, 1/16”, 3/32”, 1/8”) and sizes
https://www.allrockets.ca/Build/Balsa/4/1-16 1.5875mm thick: 101.6mm x 914.4mm
LE and TE Concave Solution
Working on making a 3d print that would cover both LE and TE to preserve profile
Author: Nathaniel Li Updating Date:
3D Print Cover for LE and TE
200mm length (print bed is 210mm x 210mm x 250mm for prusa)
LE follows curve profile and is 1mm thick into the rib
Not sure if 1mm is too thin/will fail easily but can always change to be thicker
TE is a triangular piece as following the curve profile would be too thin
Modelled in assembly
Weight Savings - 1/16” vs 1/8” ribs
Currently have 1/16” modelled
1/16” rib weight: 0.795g
0.795g x 8 ribs/wing x 2 wings = 12.72g
1/8” rib weight: 1.591g
1.591g x 8 ribs/wing x 2 wings = 25.456g
Total weight savings of 12.736g if using 1/16” instead of 1/8”
Tbh, this isn’t much of a weight savings so to be safe it might be better to go with 1/8” for more rigidity
Monokote and Ultracote
https://brodak.com/finishing-products/monokote-covering.html This appears to be the monokote that other teams were talking about at comp
https://rickyshobbycorner.ca/products/ultracote-white?variant=40243315310667¤cy=CAD&utm_medium=product_sync&utm_source=google&utm_content=sag_organic&utm_campaign=sag_organic&gad_source=1&gbraid=0AAAAADfUTBvo4sIm38bDJqPspsC39Y-TN&gclid=Cj0KCQjw2ou2BhCCARIsANAwM2E1sL2Q8GOIF4Rf5u66TLLfwIu-VuL_QfwHyNeufalCFkkv4_ME-bYaAnOMEALw_wcB Ultracote is another material that kept coming up in my searches, it’s essentially the same product as monokote, it is the american name for Oracover (previously mentioned in our research)
Any hobbyist store would sell it, I found that https://rickyshobbycorner.ca has a lot of selection and is in Canada, they sell a variety of UltraCote
https://www.rcgroups.com/forums/showthread.php?2881437-Monokote-vs-Ultrakote
Seems like most people find Monokote harder to apply compare to Ultracote
https://www.rcuniverse.com/forum/kit-building-121/5595054-monokote-vs-ultracote.html However, it seems like Monokote holds up better long-term
I think it’s worth using Ultracote for development as it’s easier to get started and will be faster for making a bunch of wings
In the future, we can look at using Monokote for a final product to make it more durable and look nicer (more colour selection apparently)
Author: Alison Thompson Date:
WING SYNC
Alison’s thoughts so far:
These are looking great!
Disagree on 12g being a small weight savings! 16th ribs > 1/8th ribs imo → this is 24g overall for the drone!
Don’t design around prusa dims, we should have bambu labs printer back
Unless you have a reason not to, I would make the cutouts larger but keep more vertical supports (more material in the vertical direction will help reduce bend more than more material in the horizonal direction, think like beam bending)
How are we thinking of adding ailerons to these?
How are we spacing out the ribs?
maybe LW PLA spacers?
TE print could probably be bigger, LE slightly thicker
How are the LE and TE prints attached? → glue
End cap + print that bolts to spar to keep wood on aluminum (shoutout Sohee)
You may want to make a test rib to check tolerance of the square cut-out after laser-cutting, talk to the person who does the laser cutting to find out what tolerances you can expect → exact sizing almost never just works for mating parts
tbh you addressed most of my concerns in design updates (no more second spar)