www.recumbents.com
www.recumbents.com
www.recumbents.com
Home | Profile | Register | Active Topics | Members | Search | FAQ
 All Forums
 www.recumbents.com
 Technical / Bike Building
 Streamliner Fairing Construction and Design

Note: You must be registered in order to post a reply.
To register, click here. Registration is FREE!

Screensize:
UserName:
Password:
Format Mode:
Format: BoldItalicizedUnderlineStrikethrough Align LeftCenteredAlign Right Horizontal Rule Insert HyperlinkInsert EmailInsert Image Insert CodeInsert QuoteInsert List
Videos: Google videoYoutubeFlash movie Quicktime movieWindows Media videoReal Video
   
Message:

* HTML is OFF
* Forum Code is ON
Smilies
Smile [:)] Big Smile [:D] Cool [8D] Blush [:I]
Tongue [:P] Evil [):] Wink [;)] Clown [:o)]
Black Eye [B)] Eight Ball [8] Frown [:(] Shy [8)]
Shocked [:0] Angry [:(!] Dead [xx(] Sleepy [|)]
Kisses [:X] Approve [^] Disapprove [V] Question [?]

 
   

T O P I C    R E V I E W
Matthew Martin Posted - 09/19/2017 : 20:42:30
Hi, I am a high school student who has been interested in the concept of streamliners and their construction. The question that has been bothering me is not knowing how to attach a windscreen (canopy;plastic view port) to a fiberglass fairing almost seamlessly without the use of tape. I have been working tirelessly on a cad model of a bike that I hope to build. The bike is a somewhat practical streamliner that is high enough to be moderately visible in traffic. I need to be able to attach a large vacuumed formed windscreen to go with my design.As I don't know how to insert my image ( not online so no url) my windscreen needs to be similar to the velotilt
25   L A T E S T    R E P L I E S    (Newest First)
Matthew Martin Posted - 11/17/2018 : 08:38:06
Wow cool ideas, but I am not sure if they are practical in the real world. These were some of the ideas that I believe Matt Weaver tried to use. Ultimately I think there needs to be a balance between theoretical ideas and overall practicality of it. I believe if Matt Weaver kept the bike simpler, it would have gone faster. But at the same time, I think its people like Matt Weaver who keep trying less conventional ideas, that will eventually find something that works. I think the next generation of stream liners may incorporate some of Matt's aero devices.

But for my bike, all I am going to do is keep it simple and efficient. Once I have the molds, If I want, I could experiment, but I will see what my budget looks like then. I am going to stick to what has worked in the past the best, mostly because I don't have the cash to invest in a untested idea.

That brings up the question from aveland, how will I build the thing? Well, I would love to CNC a plug, and I currently looking for places that offer this service. However I am not seeing anything near me. So I may have to go the old fashioned way and do it by hand. Though less efficient, there is a lot more to learn when doing it by hand. However I still want precision, so I will use many bulkheads (3mm ply wood) taken directly from CAD with a stack 2in foam in between them. Then once I get is roughly to shape using the bulkheads and templates as guide, I am planning on taking a 3D scan of the plug, and comparing it to the CAD model. then I will use bondo, or another automotive body filler to find the low spots for me. then surface primer, and polish to a mirror finish. I plan to use a method very similar to what Mike Arnolds used to build a world record plane. I am so thankful they decided to post these videos after he past away.
https://www.youtube.com/watch?v=qStFDcZs5Og
Balor Posted - 11/16/2018 : 08:03:07
An other interesting book on active laminar layer control:

https://www.nasa.gov/centers/dryden/pdf/88792main_Laminar.pdf

Well, after reading some papers on the stuff, I find this to be way over my head. There is a certain 'goldilocks' zone where you must have porosity, suction/speed coefficient and distribution of suction surfaces/slots/perforations *just right* for it to have intended effect. Too little is ineffective and once transition took place suction cannot reverse it. Too much can trip the layer by itself, not to mention that power to pump it is not free.
That fact that it was avoided by university teams suggest that it is too complex even for them, so for a lone builder to try it is a lost cause, unfortunately.

Boundary layer *blowing* into separation bubbles is a more interesting concept for a *practical* vehicle, because you can have the system to be passive (inlets and air hoses), I've already thought something along this lines, but a true liner should avoid separation bubbles/wide wake in the first place...

P.S.
Something alone this lines:

Drag can significantly be reduced by channeling the high pressure from the stagnation point at the leading edge and into the wake by means of a bypass channel. Drag reductions in the order of 50% have been measured in a wind tunnel on a sphere with a hole through the middle.
Balor Posted - 11/14/2018 : 16:05:02
Oh, so it was implemented on a streamliner! (I know it was done on aeroplanes it worked with limited success, but incurred too much maintenance costs).
Is there any data available about this bike?

I kind of suspect that for this system to work with limited human power, there should be *no* tire openings (completely sealed wheel wells) - otherwise tires will pump much more air *in* than out due to magnus effect, hence with holes in your fairing you will get boundary layer 'inflation', not suction. Dumping air should be only at the trailing edge...
Speedy Posted - 11/14/2018 : 14:54:25
Matt Weaver built a bike with boundary layer suction.
Suction was created by a custom design centrifugal air pump driven by the rear tire.
Suction flow being exhausted out the rear tire opening.
The bike publicly ran at Battle Mountain 2000 / 2001 and Casa Grande 2004
alevand Posted - 11/14/2018 : 13:33:22
How are you going to make the fairing?

C:
Tony Levand
Balor Posted - 11/14/2018 : 10:04:46
Interesting "to read" material concerning boundary layer suction:
http://sci-hub.tw/10.1007/s11012-015-0100-9
Balor Posted - 11/14/2018 : 06:28:36
By the way, you may want to take a look at this for some "expectation vs reality" pictures :)

http://sci-hub.tw/10.1016/j.proeng.2012.04.054
Balor Posted - 11/14/2018 : 04:16:54
By the way, an interesting tidbit from Bicycling Science:

By 'sucking in' increasingly turbulent air at and beyond transition point, you can (highly theoretically, of course, I don't think any real experiments were made) delay boundary layer detachment and 'reattach it', and extend laminar flow from 60% to 95%, which should give you much greater savings (hundred(s) of watts) in aero drag at cost of about 20 watts of 'suction power' (even given inefficiencies of generating it using some sort of human-powered fan).

Section on bicycle aerodynamics, pages 193-196
I don't recall it mentioned anywhere here, great book btw.
I wonder how this 'suction inlets' should look like... a series of Naca ducts? A simple mesh grille?

This section also stresses that minor undulations from pedalling (and balancing corections) and road vibration will wreak havoc on laminar flow. Plus, a perfectly 'laminar' airfoil will likely stall at a hint of yaw in wind direction, be it 'undulations' or side wind.

I wonder if you can have your shell not just suspended, but gyroscopically stabilized as well?

Edit:
Totally not NACA ducts, they were designed *exactly* to deflect boundary air away and dip into freestream for better airflow with minimum extra drag.
Balor Posted - 11/07/2018 : 01:30:35
That's comfy, I think even I can fit that :). Make sure that you your width does not taper too fast compared to your shoes, or you'll be scraping your toes when ankling (which is a good idea to maximise power).
Matthew Martin Posted - 11/06/2018 : 17:43:18
It all makes sense now...

I was still thinking about the cd value that I was getting from the simulation, and just knew it had to be wrong, if Eta had a Cd of .038, and my bike, which I believe should be less aerodynamic than Eta, had a lower Cd value.

So I re-started my investigation. I already knew that my simulations had no errors and the conditions were set up correctly, however I could not see what I was missing. I almost began to doubt the Open Foam solver's reliability, but was reassured that the system was capable of producing accurate results, after seeing that Mercedes F1 uses it to design there cars. So after getting frustrated I decided take a break for a few hours, when it came to me. I had forgotten to take into account viscus drag! I had previously only been taking into account pressure drag. I then solved for both viscus drag and pressure drag and came to a much more reasonable result.

The model shown had a Cda around .016, so with further refinement I managed to get this number down.
I eventually made some changes to the front wheel fairing, and tail to achieve a Cd value of .04249 and a frontal area of .3409m^2, giving a Cda of .01448.. Now I am satisfied with this result, however on the road it will more than likely be higher due to uncontrollable variables. I am still tring different devices such as different tail shapes and front wheel fairings, however I think I am set on the main body, which has nice 46cm(18.1in) at the shoulders and gives the rider at least 12.5cm(5in) width for the feet when using a q-factor of 150mm


Now I redid the calculation for Eta and,a lower rolling resistance was all it took for the numbers to make sense.
Balor Posted - 11/04/2018 : 13:48:13
By the way:

"I also did some calulations for the Cda of based off a statistic that I found in an article which stated that Eta would only need 198 watts to go 90 kph. This would require (using a Crr of .005 and a bike weight of 6o lb and a air density of 1.22kg/m^3) a Cda value of about .007 @ 90 kph which is far lower than the cda of my design. of course this value would decrease a higher speeds."

Given their custom tires, assuming crr of 0.002 is not unreasonable, given that you can get very close to that with Corsas (0.0023 CRR)
Balor Posted - 11/04/2018 : 09:32:28
Well, theoretically, using fat(ish) low-pressure tires takes care of that as well, but practically it would likely result in too much rolling resistance/aero hit at speeds you are willing to attain. At 60+ mph speeds and given extra fairing weight rolling resistance losses are huge. Of course, fatter tires actually roll a bit BETTER everything being equal, but the best tire you can get without going custom (huge $$$) is tubeless Vittoria Corsa speed, 23mm as far as BRR site is concerned, and I tend to trust their data. Pumped up to "near" max pressure it should provide outstanding rolling resistance (nearly half that of most racing tires) and still quite adequate vibration absorption.
So far as I know, Eta used some kind of custom tires, 20mm, likely tubulars and pumped up to enormous pressures. If you play around with your tire pressure, you *might* get away without any 'shell suspension' with a relatively small rolling resistance hit.
Matthew Martin Posted - 11/04/2018 : 07:04:33
Balor I completely agree, however the first fairing I will make, will probably be made with fiberglass and a thin foam core, and have a steel frame (no shell suspension)

My point is that once I have the molds I can begin to work towards eliminating sources of drag and to maintain laminar flow using internal devices. I will build another frame to accomplish this. My current thinking is, if I design a fairing that has enough clearance for the rider (what I am doing now), later on I can design a frame with various internal devices to make the bike as efficient as possible. That's what I meant by later, but you are right, I need to be thinking about these things now.
Balor Posted - 11/04/2018 : 00:30:08
Point is, a typical unibody liner and something like ETA with entire 'laminar' portion of the shell mounted externally and suspended are very different designs - not something you can add on later.
I think this is why ETA owns speed record despite not being the slipperiest shape 'on paper'. I don't think this is much harder to do (I might be very wrong though) - just *different* and a design choice you should make from the start.
jjackstone Posted - 11/04/2018 : 00:02:47
Matthew, do what you want. It's great that you are learning. By the way. A 70 mph hat as a junior is a world record. The other thing is. Not one of the junior world record holders have built their own bikes. Good luck in your endeavors.

JJ
Balor Posted - 11/03/2018 : 23:25:47
Well, it might not be a *small* source of inefficiency, in fact it might be a difference between laminar flow over 2/3 of your fairing or less than 1/3 for instance - basically, a difference between a genuine chance of setting a record and at most (not that it is a small achevement by itself) a '70 mph hat'. Notice that at high altitude you'll likely be able to generate only half the power you think you'll be able to, unless you can afford a hypobaric chamber training :) (and bent already your own personal hypobaric chamber as far as perfusion in your legs is concerned).
What's the point of carefully manipulating the shape of a fairing to achieve maximum laminar flow if you are not going to have it anyway?
Matthew Martin Posted - 11/03/2018 : 17:11:11
Thank you Terry for the offer, however the main point of this whole project is to learn how to work with composites. So I am going to do all of the work myself.

Yes Dreamer, I will make a ergonomics jig first, and check clearances.

Balor, I will worry about the small sources of inefficiencies later,when I build the molds I can build different iterations, with various internal devices, however that is later.
Balor Posted - 11/01/2018 : 14:58:35
Just a thought I've had - there is a lot of talk how road vibrations kill off laminar flow, which actually makes sense.
I think bents like Eta mount their shells with vibration dampers for that reason. It might be prudent to make the bent as light as possible, yet make the shell as *heavy* as possible, massively stiff and/or introduce vibration-damping layers in layup (think automotive soundproofing).
Terry Posted - 11/01/2018 : 14:41:03
Matthew,
That is a very fast looking shape you are developing .
Are you planning to build the shell yourself? What shell weight are you aiming for?
Would you like some help?

If you are interested, I would offer to build the shell for you. I would use wood strip Fiberglass composite construction.

Let me know your thoughts, I would love to be a part of a record breaking attempt!!
Terry
Dreamer Posted - 10/31/2018 : 22:03:20
quote:
... soon I will build a simple steel frame .. check clearances.

I recommend doing this before even starting to build the frame. I always start my final build with a mock up of the internal cockpit with the actual cranks and other moving parts along with the seat temporarily installed in their design position. Using a strip construction method inside a simple frame mockup to form the inside cockpit shape takes a couple hours to build but ensures that knees, heels and toes, etc fit when cranking at full speed and also that road visibility, steering and braking ergonomics will have room to perform and/or function as intended.
jjackstone Posted - 10/30/2018 : 20:16:00
Thanks for the link. Trying to post pix. Not working.







JJ
Matthew Martin Posted - 10/30/2018 : 19:17:22
Yes here is the link for Taurus :http://www.policumbent.it/blog/category/cfd
Its towards the bottom of the page.

In the end I wont really know how fast my bike is until its done. All I can do is create the best shape relative to my designs.

Though I may not go to battle mountain next year I would like to go in 2020 to attempt to break the Junior land speed record
jjackstone Posted - 10/30/2018 : 18:56:36
Sorry, guess I should have listed my sources. Some of them may no longer be accessible.
Haven't seen the Taurus estimates. Do you have a link to that one. Actually because your shape is very similar to the top bikes I could see it being close to those numbers. Good job on the design. Hope you get the physical build done. Please bring it to Battle Mountain next year.

Aerovelo Eta 0.012 2016 Aerovelo video http://www.aerovelo.com/mission-log/2015/7/24/questions-and-long-answers
U of Toronto Vortex 0.023 2011 Vortex design report https://www.rose-hulman.edu/hpv/design-reports/2011/Vortex-Design-Report-2011-u-of-toronto.pdf
U of Toronto Ace 0.054 2010 Vortex design report https://www.rose-hulman.edu/hpv/design-reports/2011/Vortex-Design-Report-2011-u-of-toronto.pdf
Delft Velox 0.032 2011 Velox design report http://ac.els-cdn.com/S1877705812016670/1-s2.0-S1877705812016670-main.pdf?_tid=f1c0902a-1251-11e7-a25f-00000aacb35d&acdnat=1490553061_819bd7c645aed16c48543c55689cb5c4
Policumbent Pulsar 0.0225 2015 Published article by Paolo http://aec-analisiecalcolo.it/static/media/riviste/Num71_web.pdf
Damjan Eivie 1 0.02 2002 Eivie website http://www.eivie4.com/
Damjan Eivie 2 0.0155 2004 Eivie website IIRC all of Damjans cda's are based on roll down tests from Battle Mountain so may not be particularly accurate. Or maybe he did other testing also.
Damjan Eivie 3 0.0125 2010 Eivie website
Damjan Eivie 4 0.0083 2013 Eivie website


JJ
Matthew Martin Posted - 10/30/2018 : 17:36:33
JJackstone, you are right to be skeptical about the numbers, and I was too when I saw the result of the shape and tried to dismantle it and some how it made sense at the time, thanks to you I have gone back and double checked, doing some calculations and even wasting some core time just to simply go in a circle. Ironically enough I was trying to see what I could do to try to increase the result to make it more reasonable, but my simulations were set up correctly, and nothing could be changed.

I do not see any reason this bike couldn't be one of the fastest on paper, however I know it should be slower than the top BM bikes as my frontal area is larger.

So I began investigating why the computer was giving me such low numbers. Unsuccessful in trying to debunk the numbers, I tried to justify them mathematically. So I started by changing the wind speed (15m/s) which did effect the Cd values however the results were still quite low (Cda .0087). So after doing a lot of nonsense, I still could not find a reason that my number could not work. Then I began to look at the references.

I believe the Cda you have for Eta, maybe wrong. I first get this idea because one the Team Policumbent website they state "Considering a speed of 145 kph, air density of 1.18 kg/m3 and a frontal area of 0.284 m2, we get a Cd of 0.0218".

All of a sudden my numbers don't seem as unreasonable.

I also did some calulations for the Cda of based off a statistic that I found in an article which stated that Eta would only need 198 watts to go 90 kph. This would require (using a Crr of .005 and a bike weight of 6o lb and a air density of 1.22kg/m^3) a Cda value of about .007 @ 90 kph which is far lower than the cda of my design. of course this value would decrease a higher speeds.

Of course simulations do not provide a 100% accurate portrayal of the real world but can certainly help for analyzing trends from iteration to iteration.

So to sum up, I believe the Cda value of Eta is a lot lower that what you have listed. This not only would make sense based off my findings, but the results of Team Policumbents Taurus speed bike which give a Cda of .0061912, and as we know, they don't hold the world record at this time, so it may be a safe assumption to say that Eta is more aerodynamic.

I hope this provides some clarity, and thanks for challenging the results, I ended up having a nice brain workout.

jjackstone Posted - 10/30/2018 : 09:54:26
"At a wind speed of about 80 mph it only generates 6.25 Newtons of drag, which comes out to a Cd of .023. The bike has a frontal area of .34435 m^2."

Are you sure of your numbers? Just asking because if these are correct, they would be in line with the some of the fastest streamliners in the world.

The following are cda's estimated by their designers.
Aerovelo Eta 0.012
U of Toronto Vortex 0.023
U of Toronto Ace 0.054
Delft Velox 0.032
Policumbent Pulsar 0.0225
Damjan Eivie 1 0.02
Damjan Eivie 2 0.0155
Damjan Eivie 3 0.0125
Damjan Eivie 4 0.0083

Your cdA comes out to .0079 from your listed values.



JJ

www.recumbents.com © 2017 www.recumbents.com Go To Top Of Page
Snitz Forums 2000