Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Can exhaust heat be used to reduce automotive drag?

  1. Mar 22, 2015 #1
    I was reading a BBC article about the automotive challenges of building a 300 MPH capable, street-legal vehicle and had a the following thought/question: Could one dissipate enough of the considerable exhaust heat generated by these supercars into the front bumper/air-splitter, so that the air passing over these surfaces would be instantly heated enough to reduce overall drag on the vehicle?

    Ignoring the obvious risk of BBQing a few stray pedestrians, could this technique work?

    I guess the same could be asked for an airplane's nose cone... if you intentionally increased the temperature of the cone well beyond the temperatures that would naturally occur due to air resistance, would this reduce overall drag?
     
  2. jcsd
  3. Mar 22, 2015 #2

    mfb

    User Avatar
    2016 Award

    Staff: Mentor

    Why do you expect lower drag?
     
  4. Mar 22, 2015 #3
    I'm assuming that hotter air would be less dense (or if nothing else, slightly more chaotic) and therefore easier to push a car through. I guess I'm seeing the whole "hot knife through butter" cliche in my head, with a hot car starring as the butter knife, and the air playing the role of butter.
     
  5. Mar 22, 2015 #4

    Doug Huffman

    User Avatar
    Gold Member

    Knife through butter and the Shkval both change the phase of the medium, from solid to liquid and liquid to vapor respectively
     
  6. Mar 22, 2015 #5
    Fair enough, air will be air in this case, no phase changes... however, if we instead look at it as an energy transfer problem, would injecting wasted heat energy into the air in front of the car supplement the kinetic energy of the vehicle itself in moving air particles out of the way?
     
  7. Mar 22, 2015 #6

    mfb

    User Avatar
    2016 Award

    Staff: Mentor

    Well, the air does not vanish suddenly, you still have to push the same mass away.
    And turbulence often increases drag.

    It could make cooling the engine a bit more effective.
     
  8. Mar 22, 2015 #7

    boneh3ad

    User Avatar
    Science Advisor
    Gold Member

    Further, in air, viscosity increases with temperature, so by causing the boundary layer to be much warmer, you will increase viscous drag. Injecting the hot exhaust air is also likely to trip the boundary layer to turbulence if it isn't already turbulent by that point, so that would increase viscous drag further in those regions rather dramatically (as in an order of magnitude increase).
     
  9. Mar 22, 2015 #8

    SteamKing

    User Avatar
    Staff Emeritus
    Science Advisor
    Homework Helper

    Here's one challenge I bet the article in the Beeb didn't mention:
    What happens if Jeremy Clarkson is behind the wheel of one of these 300 mph supercars and he's feeling a bit peckish? :nb)
     
  10. Mar 22, 2015 #9

    russ_watters

    User Avatar

    Staff: Mentor

    If you use the heat to expand air behind the car, it will push the car forward. That's how a jet engine works.
     
  11. Mar 23, 2015 #10
    Thanks for the replies everyone, it's definitely helping me better understand the error of my ways.

    So looking at this from a mass-energy equivalence standpoint: injecting more energy into the air in front of the car would essentially be the same as making the air more massive. The basic F=MA principle would then indicate that the air is applying increased (drag) force against the vehicle, thus decreasing its forward momentum instead of supplementing it.
     
  12. Mar 23, 2015 #11

    boneh3ad

    User Avatar
    Science Advisor
    Gold Member

    Wait, is this a relativistic car?
     
  13. Mar 23, 2015 #12

    mfb

    User Avatar
    2016 Award

    Staff: Mentor

    We can certainly neglect relativistic effects for street-legal vehicles.

    If red traffic lights appear green to you, then you are going too fast!
     
  14. Mar 23, 2015 #13
    I wouldn't simplify it like that. It might lead you to think about the problem wrong later on. If I had to simplify it down to something as easy as F=ma, it's more about not being able to accelerate the mass as much as you would normally with the same force because you're changing the fluid flow.
     
  15. Mar 23, 2015 #14

    Q_Goest

    User Avatar
    Science Advisor
    Homework Helper
    Gold Member

    Hi David. I guess I'll have to disagree with the responces so far. I don’t have a calculator to do real cars, only for cylinders and spheres so I did a calculation for a cylinder driving down the road. Something like this:
    fred_flintstone_car.jpg

    Drag is linearly dependent on density so if you heat the air by 1%, density decreases by 1% and drag also drops by 1%. See for example:
    http://exploration.grc.nasa.gov/education/rocket/drageq.html

    As Boneh3ad points out though, viscosity increases with an increase in temperature. This gets into the calculation through the coefficient of drag which changes as a function of Reynolds number. See this graph for example:

    dragsphere.jpg

    As temperature increases, viscosity increases, Reynolds number decreases and coefficient of drag increases so drag force increases. But the affect is very small. The affect of gas density is much more significant. Doing a calculation for Fred’s car, I find the drag is linearly proportional to absolute temperature.
     
  16. Mar 23, 2015 #15

    boneh3ad

    User Avatar
    Science Advisor
    Gold Member

    Here's the problem with this assessment: the drag equation you are using to make this conclusion depends on the free-stream density, not the density in the boundary layer. The exhaust gas will only affect the very thin layer of air by the car, so drag is not going to vary linearly (or even hardly at all) with that density change in the boundary layer.

    The primary effects will be viscous in nature, in which case the likely result is greater viscosity and a tripped boundary layer if it isn't already turbulent at the injection point, both of which increase drag. Also, while a turbulent boundary layer can reduce drag if it's effect on drag due to separation is larger than its effect on viscous drag, on a car the boundary layer is likely turbulent before reaching any separation point anyway. So there's no benefit there.

    On the other hand, if you injected the air parallel to the wall just upstream of the separation point at the rear of the car, you can probably delay separation a bit if the separation point is in a location amenable to this approach. In that case it may be possible to reduce drag. It probably wouldn't work if the separation point is already so far back that it occurs at the sharp corner near the trunk, though.
     
  17. Mar 23, 2015 #16

    Q_Goest

    User Avatar
    Science Advisor
    Homework Helper
    Gold Member

    Hi Boneh3ad. You may be right about difficulties in creating a 'bubble' of hot air around the car. Regardless, the analysis isn't incorrect, the question is whether or not it is applicable. I don't think we can make that determination definitively and without a lot more assumptions about the system used to create this hot air bubble. But at the very least, my responce should be informative and provides some basis to learn a bit about drag forces.
     
  18. Mar 23, 2015 #17

    boneh3ad

    User Avatar
    Science Advisor
    Gold Member

    But ultimately, the free stream is still the atmosphere in this case. There isn't nearly enough exhaust from a car to create a large enough temperature "bubble" to change the free stream density used with the typical drag equation. The effect will be confined to the boundary layer.
     
  19. Mar 23, 2015 #18

    Q_Goest

    User Avatar
    Science Advisor
    Homework Helper
    Gold Member

    A hot air bubble around the car would seem implausible, so it must be easy to show mathematically if that's the case, right?

    I think the OP has an interesting suggestion that may have some value for learning about drag on objects.
     
  20. Mar 24, 2015 #19

    jack action

    User Avatar
    Science Advisor
    Gold Member

    Even with the biggest bubble you can imagine, that bubble (moving with the car) will still have to displace the «outside» cold air from the atmosphere. If the bubble is big enough, the frontal area it creates will probably make things worst.

    But let's throw some numbers around for fun to see if such a bubble is plausible:

    Imagine an engine that consumes 1000 cfm. That is a very large engine (http://www.summitracing.com/expertadviceandnews/calcsandtools/cfm-calculator [Broken]), running at WOT. What comes in must come out ... only hotter. Since the exhaust temp will be about 3-4 times larger than atm temp (absolute), the density will be about 3-4 smaller; so say 4000 cfm out.

    If we take a very small car (Honda Insight), you could have a 5 ft² frontal area. So how much exhaust gas will there be in front of the car (just in front)? 4000 ft³/min / 5 ft² = 800 ft/min. So, after 1 min, the engine will produce an 800 ft «thickness» of exhaust gas in front of the car.

    The problem is that if the car goes at 9 mph (15 km/h), after 1 min, it will have traveled that 800 ft also, thus reaching the end of that "bubble".

    And this was assuming we installed a 5L engine in a Honda Insight, and run it at 10000 rpm, WOT!

    Unless I made some wrong assumptions or calculations, the hot air bubble around the car would seem mathematically implausible.
     
    Last edited by a moderator: May 7, 2017
  21. Mar 24, 2015 #20

    Q_Goest

    User Avatar
    Science Advisor
    Homework Helper
    Gold Member

    I like it Jack! Thanks... I think making those types of estimations come in very handy as an engineer. We often don't need to do a super accurate analysis of something if a conservative guess will get us an answer that proves something is good or not. I'll use your idea and add a few things...

    Consider the gas coming out of the exhaust and mixing 100 to 1 with the air. Instead of some very high temperature, the temperature will be just slightly above atmospheric. For the sake of argument, let's have the gas coming out at 1500 R and assume atmosphere is 500 R for a dT of 1000 F. Mix 100 to 1 and the temperature is down to 10 F above atmospheric, so 510 R. Round numbers. But we now have 100 times the amount of air the engine puts out. That's another issue we could play with, but let's just assume what comes out of the engine is a gasseous mixture of air (mostly unburned nitrogen?) with plenty of water and carbon dioxide. To make it easy, let's just assume it's air and see what happens...

    Now assume a more modest engine, around 2.8 L (0.1 cubic foot) and 2000 RPM. Assuming a 4 stroke (only 1 exhaust per 2 revolutions) that's 100 cubic feet times 100 because of the mixture, so around 10,000 cubic feet of 510 F air. For a frontal area of 5 square feet, that means we can sweep a volume 10,000 ft3/5 ft2 = 2,000 ft long in one minute. And all the air in that bubble will be 10 degrees warmer than atmosphere. That's a rather slow 23 mph, but we could cut the temperature in half and go 46 mph.

    Given the drag is roughly proportional to density (proportional to absolute temperature), then for a 10 F rise, that's a 2% reduction in drag. For a 5 F rise, that's a 1% reduction in drag. Nothing to brag about but we're getting closer to a realistic answer. Did I do the math right? What else? What about the method of exhausting the gas in front of the vehicle (instead of behind it)? And does the 5 square feet frontal area make sense? Doesn't that mean we have 5 square feet of hot air ALL AROUND the car? It's an interesting question.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: Can exhaust heat be used to reduce automotive drag?
Loading...