[SGVLUG] OT: Hybrid efficiency (was:New Linux Lug)

[David Lawyer]

On Mon, Feb 20, 2006 at 10:08:47AM -0800, Dustin Laurence wrote:
> On Sun, Feb 19, 2006 at 07:23:48PM -0800, David Lawyer wrote:
> 
> > On Thu, Feb 16, 2006 at 04:35:09PM -0800, Dustin Laurence wrote:
> > > No.  It cannot *ever* do so for engines tuned for their respective jobs
> > > and a hybrid that is meant to be the real thing (and not the "hybrid
> > > assist" stuff some are apparently designed for).
> 
> > ...First, the hybrid engine must be
> > able to achieve high speed on climbing mountains, in which case there
> > will not be nearly enough energy stored in the battery for this.  So
> > the engine size of a comparable non-hybrid only needs to be a little
> > larger.
> 
> No, you just don't understand the tradeoff.  In simple terms, you can
> maintain the same peak power in a smaller engine if you narrow the power
> curve.  The downside is that you have to run the smaller engine in a
> narrower RPM range to get that peak power--something that a hybrid can
> do but a direct drive engine cannot.  The effect can be quite
> dramatic--I have seen the curves in engine design textbooks when I was
> an undergrad.
Not so.  Look again at how efficiency varies with RPM.  It doesn't
vary all that much.  For example, fig. 2.2 of an internal Lockheed
report shows a variation from .458 to .520 in brake specific fuel
consumption as one goes from 1800 to 3600 rpm in a mid-sized v-8
engine.
> 
> The truth is that a reciprocating engine is not the best choice for an
> advanced hybrid.  Turbines reach better thermodynamic efficiency, and
> the reason we don't use them yet is because if very high initial cost
> and very high operating temperatures (in fact the latter is where a lot
> of the efficiency comes from, so you can't reduce it).  The downside of
> a turbine is that it wants to be run nearly flat-out and just won't
> accept a wide range of RPMs, so they aren't very good for direct-drive.
> But the power storage of a hybrid means you can run it at it's most
> efficient point to charge the batteries and not otherwise.
> 
Why haven't they been successful in railroad locomotives?  The
"turbo-liners" had terrible fuel efficiency.  Charging batteries is
only 36% efficient in the Prius II.  
> I seem to recall there are a number of other problems with turbines, but
> I think the chances are good that a fully evolved hybrid would use some
> form of one.  The problems really relate to their use in a new
> application (passenger vehicles), since in other areas turbines are a
> very mature technology.
> 
> Until we have them, the most efficient practical choice I am aware of is
> a highly tuned hybrid which runs it's engine at it's optimum RPM.
> 
> > ...The weight of the batteries and motor-generator tend to make
> > the hybrid heavier.  Hence more rolling resistance and a little more
> > aerodynamic drag since the body must be slightly larger.
> 
> Most of the drag at highway speeds is wind resistance, and the
> difference there is essentially zero: notice that some cars can be
> purchased in either configuration.  It is the same body shape and size,
> thus the same drag.  In fact, probably the only effect on drag would be
> underbody airflow (smoother is better), and there is a good chance that
> a hybrid is better there (I suspect less ventilation is needed).  I
> think I'll just look under a Prius the next time I see one.  Better
> would be to go to a Honda dealer and compare a Civic hybrid and a
> regular Civic side-by-side.
The aero drag become equal to the rolling resistance at speed of from
40-50 mi/hr.  See S. F. Hoerner: Fluid Dynamic Drag, published by the
author, 1965, p.12-7.  It shows curves for various drag coefficients
so that it doesn't become obsolete over time.  It also agrees with my
own calculations based on SAE papers.
> 
> At city speeds rolling resistance is more of a factor but the additional
> rolling resistance due to weight is again negligible.  A random web
> search turns up this:
"More"?  It's the predominant factor.  Also the energy used for
acceleration and climbing hills is directly proportional to weight.
> 
> http://reviews.cnet.com/5208-7810-0.html?forumID=77&threadID=104113&messageID=1194686
> 
> Which lists the additional weight for a Prius (a fuzzy concept since
> there is no direct-drive Prius) as 200 lb.  A better test is for the
> same car, and Ars Technica lists the 2003 Civic hybrid as 200 lbs
> heavier than it's direct-drive twin.
What about the payload?  Is it the same for both vehicles?
> 
> http://arstechnica.com/reviews/other/2003-civic-hy.ars/1
> 
> In other words, the drive train weighs about as much as an extra
> passenger.  Just not a big factor, as anyone who has recorded the
> mileage of his car for a while knows.  And low-speed driving is mostly
> stop-and-go in actual practice, where the additional weight is requiring
> additional braking, so the difference will be even smaller for the
> hybrid.
It will be worse since the energy used for acceleration in stop-and-go
will be higher.
> 
> > But worst of all is the poor efficiency of regenerative braking.
> 
> No, it is not "worst of all," because the direct-drive engine has no
> counterpart.  If you will, the efficiency of regenerative braking for a
> direct-drive car is zero, and it costs very little to add it to a hybrid
> because the drive train is already there and electric motors are easy to
> run as generators.  In fact it's nearly free.
Not at all free.  If you preserve it as KE, then it may be possible to
recover 100% of it by coasting.
> 
> As for it being "low", this is far from the case.
> http://www.greencarcongress.com/2004/09/under_the_hood_.html
> 
> Lists the Accord hybrid as achieving 95% of available kinetic energy
> into electricity, an improvement over the Civic's efficiency of 93.5%.
> I challenge you to make a coherent case that this is "low."  It then
> has to be converted back into kinetic energy, but electric motors are
> good at this and it's no worse than the situation with newly generated
> electricity.
> 
> > Thus the efficiency of recovery of kinetic energy of a vehicle via
> > regenerative braking is poor.  Perhaps only 50%.
> 
Sorry, I made a serious over-estimate of efficiency (and I won't bore
you with how I did it).  It's actually only 36%.

> Reference, please, or it shall be disregarded by the court.

"http://www.hybridcars.com/flywheels.html" Hybrid Cars - Flywheel
Power to Increase Hybrid Efficiency says"
   ...
   The four energy transformations undermine the overall level of
   efficiency. For example, if the motor/generator operates at 80%
   efficiency under peak load, in and out, and the battery charges and
   discharges at 75% efficiency at high power, the overall efficiency over
   a full regenerative cycle is only 36%, almost the same as the figure
   Toyota quotes for the Prius II.
> 
> In any case, 50% is by no means "low"; it's better than the efficiency
> of the engine in the first place.

No, the energy was created by a 25% efficient engine.  With a 36%
conversion efficiency the overall efficiency is only 9% (.36 x .25).

> 
> > with a non-hybrid if one drove optimally to conserve energy?  How
> > does on recover kinetic energy for a non-hybrid?  By simply coasting
> > with the engine off.
> 
> This is a false comparison on many grounds.  First, you are positing a
> significant change in driving habits for the direct-drive car but not
> for the hybrid.  One can also save energy in a hybrid by driving more
> sensibly, especially at highway speeds where so much energy goes into
> pushing air.  In fact, one can do the same trick with a hybrid.  In
> fact, one does; the car essentially does it automatically because the
> engine need not kick on to recharge power that is not being used.  A
> hybrid is better at this by design than your direct-drive car.  Also it
> only does it when it helps (when you want to stop already, or going down
> a hill), unlike what you apparently intend (which seems to be on level
> ground).
> 
I implied that there would be a change in driving habits.  Coasting is
for both level, downhill, and in some cases uphill.

> Second, your suggestion is not a realistic method.  It is dangerous,
> because you lose power steering and (after maybe a couple of pedal
> presses) power brakes.  If you accidentally turn the key all the way to
> off, you may also lock the steering column, which would be disastrous.
> And finally, if you do this often enough on, say, long hills, you will
> probably lose lubrication in the drive train.  There is a reason that
> you are not supposed to tow a rear-drive car any distance without
> disconnecting the drive shaft.  I'm not sure of the situation for a
> front-wheel drive car, however; towed the front wheels are generally off
> the ground whereas in your scenario they are freewheeling, so for
> front-wheel drive we can't use the simple towing directions.

The old cars didn't need power brakes because they used self
energizing brakes which utilized KE of the car to help apply the
brakes..  Cars need to be redesigned to make them suitable for
coasting with the engine off.
> 
> Consider the extra emissions and the greater number of cars you will
> consume over your life and the emissions involved in constructing them
> before you pat yourself on the back.

Why would more cars be needed?
> 
> > ...This way one recovers all the the kinetic
> > energy (=KE), not just 50%.
> 
You recover it for the purpose of using it to overcome vehicle
resistance.  In other words you recover it to help more the auto.
> This is absolutely untrue.  You still lose it as heat, just at a slower
> rate and not into the brake pads and rotors.  Then you have to re-start
> the engine (the most polluting thing you can do to a normal car, though
> not so bad now that the engine is warm) and accelerate back to speed
> (with attendant fuel cost).  
Restarting an engine by putting it in gear shouldn't be very
polluting.  You don't need a rich mixture since it's already warm.
Car fuel systems need to be redesigned to avoid enriching the mixture
during this type of a start up.  Yes, it takes energy to accelerate
but you get it back when you coast again.

> Further, wind resistance will be higher than driving steady at the
> same average speed, because of the v^2 nature of fluid drag in this
> regime, and thus you need still more fuel.

Not if you are willing to go at a slower average speed.

> 
> I'd be interested to watch you do an actual test of this idea, but you
> won't be abusing my car to do it.

You don't abuse cars much by doing this.  I've done it and have had to
do it at times to keep from running out of gas in areas with no gas
stations open.  Cars need to have electric oil pumps that would turn
on just before starting to avoid the extra wear that happens when one
starts the engine.
> 
> > ...Even if one coasts (on level) only until half
> > the cruising speed is reached, KE recover is 75% since KE is
> > proportional to v^2.
> 
> I am not sure what you even mean by this, but to the extent it makes
> sense you have it backwards.  At v/2 you only have a fourth of the
Typo.  You mean V^2.
> original KE left.  This is *NOT* "recovery," but if it were the
> efficiency would be 25% not 75%.  If you mean that you "recovered" the
> three-quarters of the KE you wasted into heating the roadway, drive
> train, and (mostly) air, then you are very confused.
No I'm not.  The energy of vehicle resistance must be provided.  Thus
you recover it to use it to propel the auto by overcoming vehicle
resistance (aero drag and rolling resistance in the broad sense).  Of
course almost all of this energy is turned into heat but a tiny part goes
into deformation and breaking of minerals in the soil and pavement.

> 
> > > Non-hybrids must have a broad power curve, which greatly lowers peak
> > > efficiency.
> > Not at all, since the driver only powers the car at the points near max
> > efficiency and most of the time coasts with the engine off.
> 
> This jackrabbiting is going to really kill your performance, as I said.
> But beyond that your argument is truly fallacious.  Drivers cannot
> reliably do this.  
They can do it, but it would be much easier if the cars were designed
for it.

> If they could, it wouldn't matter, because as I said
> above the engine has poor peak efficiency to maintain drivability.  No
> production car is made any other way.  If you are positing a new car
> designed for this sort of driving, then you should get rid of the
> driver's judgement and let the computer start and stop the engine
> (basically what Cadillac years ago tried to do with individual
> cylinders, which was a disaster, and what Honda at least does now with
> greater success).  But if you can do that, I'll design my own hybrid for
> the average speed you make and greatly exceed your performance because
> I'm not jack-rabbiting and wasting a lot of extra fuel in wind
> resistance and starting.
As I previously explained, wind resistance would be less.  Starting
doesn't take much energy if it's supplied by the car's momentum.
> 
> > > A true hybrid, by comparison (one where the engine generates power
> > > only and is not directly connected to the wheels), can have an
> > > engine tuned with a very narrow--and thus very highly peaked--power
> > > curve and run at the most efficient point regardless of vehicle
> > > speed.
> > True, but what about the poor efficiency of KE recovery?
> 
> First, it doesn't exist, and second, it's irrelevant to the tuning of
> the engine.  Everything I said about the engine is true if you
> programmed the computer not to do regenerative braking.
Not so.  See above.
> 
> > > This is why, in fact, there has been interest in continuously
> > > variable transmissions--you can get a similar benefit without the
> > > electric drive train.  
> > 
> > You really don't need CVT if you shift at the proper points and coast
> > at the optimal points.
> 
> Nonsense.  A trivial disproof: you cannot coast up to a higher speed in
> general, so you must use the engine to speed up.  In between optimal
> points you are working the engine both hard and suboptimally.
Not really.  It's a little suboptimal but much less than the 64% loss
in the regenerative braking cycle.
> 
> > ...Car's aren't designed for this, especially
> > ones with automatic transmission.  They need to be redesigned.
> 
> And will never be as good as a car designed properly, for the physical
> reasons I mentioned above.
What reasons?
> 
> > ...When I
> > try to get max efficiency by pushing the gas pedal to the floor, the
> > AT shifts into a lower gear, decreasing efficiency.
> 
> This is ridiculous, sorry.  If you push the gas pedal to the floor you
> are using a lot of extra gas for a small improvement in torque.
Look at the curves.  The engine efficiency at max torque is only a
little degraded.  And I think a reason for this degradation may be
the supplying of an enriched mixture for maximum power.  So if cars
were designed only to provide this enrichment when told to do so by
the driver, then the best efficiency would be a maximum torque (I
think).
> 
> > > However, you still don't get the ability to run a smaller engine
> > > that only needs to satisfy the average power demand (letting the
> > > batteries take up the slack).  Non-hybrids, even with a CVT, must
> > > size the engine to the *peak* demand, and this hurts average
> > > efficiency.
> > They can also just not supply what you call "peak demand".
> 
> Then the corresponding hybrid also can have the same peak demand and a
> correspondingly smaller engine.  You can't win that race because every
> improvement you make can also be made by a hybrid, for the same reasons.
> 
> > At any rate I think that *if* one had a non-hybrid car that could be
> > fully controlled (so as to operate at full throttle in high gear)
> 
> You make think so, but it isn't so.
Oh yes it is.
> 
> > that traffic, signals, and laws were such as to promote coasting,
> 
> Ever heard of light sequencing?
It needs to be a lot more advance than this.  A car approaching a
signal need to know when it's going to change and what rights it has
to change that.  The coasting period may need to be shortened so that
the car can make it thru on the green.  If the car is bound to stop,
it needs to know this well in advance so that it can go into the
coasting mode well before the signal.  Thus even if the light is
green, you should be coasting if you are going to have to stop.
> 
> > *if( non-hybrids were designed to have bodies lighter than hybrids,
> > then such non-hybrids would be more energy efficiency *if* driven
> > optimally.
> 
> And if they drove at a quarter the speed of the hybrid, they would also
> be more efficient.  Both are apples-and-oranges comparisons.
> 
> > ...Some of the energy saving in hybrids is that they use more
> > lighter weight aluminum etc.  So in any comparison, one needs to
> > design a non-hybrid this way too.
It's true that with a non-hybrid and costing one would go slower.  But
one can show mathematically, than no matter what your time is worth,
it always pays to coast some.  It might only be for 1 second if your
time is extremely high priced.
> 
> First you argued that the hybrid had to be heavier.  Have you changed
> your argument?
No, it does need to be heavier but I read that they were intentionally
made lighter at great expense.

> You can by an Accord and a Civic in either configuration, for starters,
> and the hybrid is slightly heavier (according to the web reference
> above) and still gets better fuel efficiency.  So this is not going to
> help you enough.
> 
> > I think auto manufacturers should be required to provide efficiency
> > curves for their engines so that drivers can select efficient
> > operating points.
> 
> Most drivers can neither select efficient operating points nor
> understand the efficiency curve.  However, I've driven cars that tried
> by urging you to shift early to keep RPMs low and efficiency high rather
> than late to keep up torque.
What counts for efficiency is engine torque, not torque at the wheels.

> > ...Why not a display on the dash that would show
> > efficiency?
> 
> I believe a Prius has this (Tom can tell us for sure).  Still sure you
> don't want a hybrid? :-)
> 
> That said, the most efficient forms of transportation for a single
> person have two wheels because of weight and wind resistance.  They
> brake much better, too.
Well, my old bicycle with rim brakes worn smooth wouldn't stop in the
rain on a steep hill.
> Pity that it's so very, very hard to walk away from an accident.  I
> wouldn't mind riding a motorcycle but I do intend to live to see
> Eric grow up and I also don't intend to encourage him to do
> something that would also tend to lower the likelihood of that
> happening.

The drag coefficient for a bicycle is extremely high, almost 1.  The
result is that aero drag for a bicycle is significant, even though the
speed is much lower than an auto.  More efficient is the pedal
railroad train (my invention) with less rolling resistance per person
than a bicycle and far less aero drag.  It's not practical since it
would be expensive and not go fast if there weren't enough people to
pedal it.

> Hey, maybe we should buy dirt bikes or something.  At least there if you
> hit a tree it's pretty much your doing, not some joker who ran a red
> light.

Such bike riding can be pretty dangerous.  But I do it on an old
3-speed road bike.  It's not very dangerous if you go slow.  I don't
really do this by choice since I've got damage to the backs of my
heels that is made worse if I were to hike instead of bike.
	
> 
> Dustin, contemplating the interesting idea of a hybrid motorcycle (but
>         where to store the electricity on such a small vehicle)?
> 
> 
			David Lawyer