[lightarrow]

The amount of fuel in the tank has no effect on the efficiency of the engine. The mass of the fuel in the tank does have an effect on the fuel consumption of the vehicle, but not very much.

Ok, but I think you forgot to say that if go along a (long) road which on average is a slight descent, without having to accelerate, having more mass *decreases* fuel consumption  []

[Geezer]

Here you go Wolfekeeper. I did it for you -

Engine Loss 62%
Standby 17%
Accessories 2%
Driveline Losses 6%
Aero 3%
Rolling 4%
Braking 6%

By an amazing coincidence, when you add those numbers together, the result is 100.

On the urban driving diagram, 19% is the energy output from the engine. It's what is left after you subtract 62, 17 and 2 from 100.

The 13% on the diagram is what's left of the engine's energy output after you take off another 6% to account for friction losses in the transmssion and driveshafts. Remarkably, 3+4+6 = 13

[Geezer]

The amount of fuel in the tank has no effect on the efficiency of the engine. The mass of the fuel in the tank does have an effect on the fuel consumption of the vehicle, but not very much.

Ok, but I think you forgot to say that if go along a (long) road which on average is a slight descent, without having to accelerate, having more mass *decreases* fuel consumption  []

Well, it might, but, as Wolfekeeper points out, the extra weight would also increase the rolling friction (just not as much as he seems to think  []) so it would sort of depend on how steep the hill was.

[lightarrow]

Well, it might, but, as Wolfekeeper points out, the extra weight would also increase the rolling friction (just not as much as he seems to think  []) so it would sort of depend on how steep the hill was.

Certainly  []

[wolfekeeper]

Geezer, you've conclusively proved to everyone, apparently except you, that you don't have the slightest clue what those numbers mean; they don't mean that only 4% of the fuel is used to propel the vehicle against rolling friction; they mean that only 4% of the fuel is usefully used to propel the vehicle; there's a huge tail of wasted fuel energy that is used to produce any useful energy; by a factor of 7.7, the extra 6.7 ends up as wasted heat, as hot air.

So the energy you need in the gallon for that 4% is the 4 x 7.7 = 31% percent of each gallon, because you can't generate propulsive energy without wasting 87% of the initial energy. Cars really are that inefficient, overall, and the losses are listed in that table.

It's not like you have a choice, you can't decide to not burn the energy to make drive train losses today to save energy, they're present whatever or however you drive your car, for any propulsive force your car generates, and for whatever it's being used for. You can't even cut the idle losses, your car is stuck in traffic for a certain proportion of the journey, these are supposed to be a typical journey.

[CZARCAR]

With an emptier tank, the gas sloshes,stirs, & aerates itself?

[CliffordK]

See FIGURE 3-1. http://onlinepubs.trb.org/onlinepubs/sr/sr286.pdf

Following the URL on the page:
http://www.fueleconomy.gov/feg/atv.shtml



With slightly different numbers for city, hwy, or combined.

Oddly they don't seem to separate acceleration and deceleration, perhaps the mode of acceleration doesn't play as big of a factor as one might think as it would be a difference between a short, intense burst of energy vs a long, gradual energy use.

Anyway, if you increase energy requirements at any one point of the system, then that will cause increases throughout the system.

So, consider it this way.

Say you have 100KW energy input.
70% (70KW) is "lost" by the engine.
5% (5KW) is consumed by the water pump and alternator.
5% (5KW) is lost in friction in the transmission and drive train.
20% (20KW) gets to the wheels and is divided by 10KW wind, 5KW rolling, 5KW braking.

Now, if you increase the "rolling resistance" by a modest amount from 5KW to 6KW (about a 20% change).

This would propagate through the entire system.
Rolling, up by 20%
Wind - Same
Braking - Also up by 20% if weight was the cause of the increased power needs.
Engine - Also up by 20%, as it is working harder to supply energy to the system.
Water Pump, Alternator, etc.  May increase slightly.  More cooling requirements?  Is it electrically or mechanically cooled??
Idling, - About the same.  Traffic and congestion doesn't care about how full or empty your tank is.

Anyway, once you sum it all up.
You started at 100KW giving the equivalent of 5KW "rolling resistance".
But, you are up to between 115 and 120KW total energy input to give you that additional 1KW at the wheels (6KW rolling resistance).

Fuel tanks are generally sized to give a typical vehicle 300 to 400 miles range. 

Subcompact, perhaps 10 gallon tank.
Mini-Pickup, 15 gallon tank.
Full sized pickup, 20-30 gallon tank.
Semi Truck, perhaps twin 150 gallon tanks.

At 8 lbs to the gallon (density of fuel is actually slightly less than for water), that gives about 80 lbs for the subcompact, 240 lbs for the big pickup, and 2,400 lbs for the semi.

As mentioned by techmind, the weight of the full tank of fuel is probably somewhere around 2% to 5% of the total vehicle weight. 

If 100% of that additional weight goes towards increased rolling resistance, acceleration, and braking energy, then it will also mean increased engine requirements and a 2% to 5% increase in fuel consumption.

80 lbs of fuel in a subcompact may not seem like a lot.  However, certainly the loaded vs unloaded fuel consumption of a pickup is noticeable.  And 240 lbs would be about ¼ the rated capacity of a half-ton pickup.

Keep in mind that extra trips to the gas station may not be efficient.  Do you go out of the way to get gas?  Do you end up idling the car while waiting in line to fill up, or does your car konk out 1/2 mile from the gas station due to being out of gas.  You coast half that distance, and end up getting out and pushing it the last ¼ mile, saving additional gas. []

 

[Geezer]

they don't mean that only 4% of the fuel is used to propel the vehicle against rolling friction;

When did I ever say that only 4% of the fuel is used to propel the vehicle? Of course that's ridiculous. What I said was that only 4% of the thermal energy released by burning the fuel goes into overcoming rolling friction.
And, if you had even bothered to read the link I posted, you would have found this on page 39;

"For urban driving, only 10 to 15 percent of the fuel energy is ultimately transmitted as power to the wheels."

and

"For both urban and highway driving, the mechanical energy that does make its way through the driveline to turn the wheels is consumed by three (energy) sinks: aerodynamic drag, rolling resistance, and braking."

Your loud assertion that "Just about 4/13 of the energy in the gasoline is spent overcoming rolling friction!" is just plain wrong. In the urban model it is simply 4% of the total thermal energy in the gasoline.

If you do the math you will discover that 4% of the thermal energy in a quantity of the fuel put into the engine will be very similar to the energy consumed in overcoming rolling friction (which, BTW, is mainly dissipated as heat too.) You'll have to assume a value for the coefficient of friction of course, but 0.01 might be about right.

The other mistake you are making is that you are getting all hung up on the term "useful". Why is the energy that is dissipated as rolling friction, aerodynamic drag (friction) and braking (friction) any more useful than the heat produced by the engine while it is doing work? The only really "useful" bit is the work done by the engine, but ultimately that is all converted into "useless" heat in propelling the vehicle.

It's really much more satisfactory just to analyse where the energy released from the gasoline ends up, and the diagrams on page 40 do a very good job of explaining that.

[Geezer]

70% (70KW) is "lost" by the engine.
5% (5KW) is consumed by the water pump and alternator.
5% (5KW) is lost in friction in the transmission and drive train.
20% (20KW) gets to the wheels and is divided by 10KW wind, 5KW rolling, 5KW braking.

Now, if you increase the "rolling resistance" by a modest amount from 5KW to 6KW (about a 20% change).

This would propagate through the entire system.
Rolling, up by 20%
Wind - Same
Braking - Also up by 20% if weight was the cause of the increased power needs.
Engine - Also up by 20%, as it is working harder to supply energy to the system.
Water Pump, Alternator, etc.  May increase slightly.  More cooling requirements?  Is it electrically or mechanically cooled??

Nice link Clifford, however, I see a slight problem.

Rolling resistance is simply the resistance to rolling! It's only a function of the tires and the weight they are carrying, and it is also influenced by the nature of the road surface. Wind resistance (drag) has nothing to do with rolling resistance, so it does not follow that if one changes the other will also change.

For example, you can increase the rolling resistance by letting some air out of the tires  [] 

[Geezer]

Clifford,

Perhaps you are saying that if one quantity changes, the other will also have to change too. I'd agree with that.

If you upped the rolling resistance from 5 to 6kW, it's no longer 20% of the power that is getting to the wheels. 1 kW is 1% of the power, so, it's now 21% that is going to the wheels, but if we simply do that, we will end up with a total power of 101%, which is obviously wrong.

To fix this, we will need to adjust all the values so that the total power still sums to 100%, if you see what I mean.

[CliffordK]

Wind - Same

Nice link Clifford, however, I see a slight problem.

Wind resistance (drag) has nothing to do with rolling resistance, so it does not follow that if one changes the other will also change.

Precisely what I said...
Wind resistance would be essentially unchanged by an increase in internal mass.

So, in fact, by changing the internal mass, you will change the actual energy consumption of certain aspects of the car (rolling resistance, acceleration, braking, engine, etc).  Other aspects will remain largely unchanged for energy consumption (wind resistance, car stereo, AC, idling, etc).

If you increase your rolling resistance and braking by about 20%, you will increase your overall energy consumption by about 20%.

I.E. in my example, changing your energy consumption from about 100KW to just under 120KW (perhaps 118KW).

It will also shift the relative percentages slightly.  The power lost to wind resistance will stay the same, so the relative percentage would actually drop slightly.

However,
We've concluded that a full fuel tank is only between 2% and 5% of the total vehicle weight.  So, the savings of an empty tank vs full tank would likely be somewhat less than 5%.

If you regularly top off your tank, then you might realize the cost of a continuously full tank.  However, if you drive from full to empty, then refill, your average mass would be just over half a tank (1 to 2.5% of the cars weight). 

If, on the other hand, your fuel miser drives the car in the range ½ to empty, then the average would be about ¼ of a tank, or about ½% to 1.25% of the total vehicle weight.  And, thus the actual average weight & fuel savings of the fuel miser would be on the order of ½% to 1.25%.

So, from my notes above, the subcompact with a 10 gallon tank would have a maximum capacity of about 80 lbs fuel.  But, the person driving from full to empty would average 40 lbs of fuel, and the person driving half to empty would average 20 lbs fuel.  The savings is thus, only about an average of about 20 lbs. 

For many of us, we might be better off just dieting  [xx(], and filling the tank to the max  [^]

[Geezer]

Oh! When you said "same", I thought you meant it would also increase by 20%, but what you meant was it is unchanged! I see what you mean.

Yes, a 20% increase in rolling resistance and braking friction will have a subsantial impact on fuel consumption, although precisely how much will depend on how fast the vehicle is travelling because the energy that goes into overcoming drag is unchanged (assuming the same set of conditions), so it might be a bit less than 20%.

Rather than adjusting percentages, I suspect it's safer to use a model and apply real values, then derive the percentage values, otherwise things tend to go pearshaped very quickly.

[damocles]

Time for a bit of science: the original question was Does the amount of fuel in the tank affect petrol consumption?

The discussion has moved on to How much does the amount of fuel in the tank affect petrol consumption?

There is a consensus in the discussion, and good reason to believe that (apart from dirty petrol and condensation as I earlier pointed out) the only significant factor is the mass of fuel involved. If that is the case then the mass does not have to be petrol. There is a simple experiment that can be done.

Run your family car for a month, filling the tank quite full at the beginning and the end. Get an accurate check on fuel consumption.
Then, put a 20 kg sandbag in the boot, and leave it there for the next month. Take a little trouble to see that you mostly do a similar amount and style of running around in the car, including issues like how much you use the air conditioner, and how much driving is done on wet or dry roads. Compare the fuel consumption for the second month with the first, and then think carefully about the amount of experimental error involved.

[Geezer]

Time for a bit of science:

Oi! We must have missed the bit where you came riding in on your gleaming white stallion.  []

[Geezer]

Wolfekeeper, old pal [],

I think I understand your problem - it looks as if you are somehow equating power and energy. The diagrams in that paper are all about energy. What you are concerned about is power. It takes power to overcome rolling friction and air friction, and power requires the conversion of energy in time.

As you implied, the mechanical power produced by the engine is only a fraction of the "thermal power" being produced by the fuel, so any increase in the mechanical power consumed by rolling friction has a corresponding multiplication effect on fuel consumption to generate the required amount of mechanical power, and the multiplication factor will depend, to some extent, on the thermal efficiency of the engine.

I say "to some extent" because it's not that simple. In urban driving, the vehicle's speed is anything but constant, so a lot of the rolling resistance occurs while the vehicle is slowing down and the engine is developing hardly any power. During acceleration, the power produced by the engine to accelerate the vehicle "swamps" the power required for rolling friction, and the engine's thermal efficiency during acceleration is also very different. The energy values in the paper average all that out for us.

Ah-ha!, you might say, so why don't we keep the conditions constant. Well, yes, we can. That's when we might use the highway driving model, in which case the rolling resistance is much less than the air resistance, so that throws everything askew as well.

The "bottom line" is that you can't simply equate a percentage increase in vehicle weight with a percentage increase in fuel consumption without specifing the conditions in which the vehicle is operating.

EDIT: BTW, I realized that your 4/13ths estimate might actually be a bit conservative. You shouldn't really count the braking component because it's unlikely to occur while the engine is producing power  []. On the other hand, a lot of the rolling resistance happens while the engine is in standby, so that takes things in the other direction. It's very complicated.

The energy analysis in the paper is useful as a means to determine, averaged over a long interval, where all the energy goes, but it's not much use as a real-time model.

[CliffordK]

Then, put a 20 kg sandbag in the boot, and leave it there for the next month. Take a little trouble to see that you mostly do a similar amount and style of running around in the car, including issues like how much you use the air conditioner, and how much driving is done on wet or dry roads. Compare the fuel consumption for the second month with the first, and then think carefully about the amount of experimental error involved.

Yes,
A 10 or 20 kg weight would be representative of the different driving styles.

If you put 20 of the in the trunk (400 kg), the differences should be more obvious, but not as representative of the actual fuel.

Experimental Error?
Good point.
And, it is likely low enough that you could only tease the difference out (for the 10 or 20 kg weight) with an average over many tankfuls. 

Also, consider a "blinded" trial.  Have someone else fill (or empty) the trunk.  Then seal it so you won't know whether it is full or empty.

In fact, I think the last time I was watching my fuel consumption, I saw a significant difference between summer and winter.  So, as we are approaching winter, the fuel efficiency might naturally drop.

[Geezer]

Then, put a 20 kg sandbag in the boot, and leave it there for the next month. Take a little trouble to see that you mostly do a similar amount and style of running around in the car, including issues like how much you use the air conditioner, and how much driving is done on wet or dry roads. Compare the fuel consumption for the second month with the first, and then think carefully about the amount of experimental error involved.

Yes,
A 10 or 20 kg weight would be representative of the different driving styles.

If you put 20 of the in the trunk (400 kg), the differences should be more obvious, but not as representative of the actual fuel.

Experimental Error?
Good point.
And, it is likely low enough that you could only tease the difference out (for the 10 or 20 kg weight) with an average over many tankfuls. 

Also, consider a "blinded" trial.  Have someone else fill (or empty) the trunk.  Then seal it so you won't know whether it is full or empty.

In fact, I think the last time I was watching my fuel consumption, I saw a significant difference between summer and winter.  So, as we are approaching winter, the fuel efficiency might naturally drop.

It's a lot more than experimental error. That would only apply if all conditions were carefully controlled, but in this case there are so many uncontrolled variables that the results could be very misleading. I would think a simulation with accurate input data would yield much more meaningful results.

Clifford, did you see more or less MPG with lower temperatures? I think you might expect slightly more MPG with lower temperatures, but there are a lot of other factors that could take it in the opposite direction.

[CliffordK]

Clifford, did you see more or less MPG with lower temperatures? I think you might expect slightly more MPG with lower temperatures, but there are a lot of other factors that could take it in the opposite direction.

It has been a few years since I've watched it, but I thought I was getting lower fuel mileage in the winter than in the summer.  For my Ranger, it was something like 20-22mpg in the winter, and 22 to 25 mpg in the summer.  But, I could have got that reversed. 

NO AC, of course.   [xx(]

I'm not quite sure what seemed to be causing the difference.  Fuel mixes may have been different, with municipalities often adding more ethanol in the winter.  Or, perhaps the colder fuel/air mix didn't burn as efficiently.  Obviously there is a warm-up period in the winter, but I think I was also noticing the difference with freeway driving.  The Ranger would have had a mechanical water pump and fan, so likely not much change with hot vs cold.  Heaters are run off of water, so not much difference there either.

[Geezer]

Clifford, did you see more or less MPG with lower temperatures? I think you might expect slightly more MPG with lower temperatures, but there are a lot of other factors that could take it in the opposite direction.

It has been a few years since I've watched it, but I thought I was getting lower fuel mileage in the winter than in the summer.  For my Ranger, it was something like 20-22mpg in the winter, and 22 to 25 mpg in the summer.  But, I could have got that reversed. 

NO AC, of course.   [xx(]

I'm not quite sure what seemed to be causing the difference.  Fuel mixes may have been different, with municipalities often adding more ethanol in the winter.  Or, perhaps the colder fuel/air mix didn't burn as efficiently.  Obviously there is a warm-up period in the winter, but I think I was also noticing the difference with freeway driving.  The Ranger would have had a mechanical water pump and fan, so likely not much change with hot vs cold.  Heaters are run off of water, so not much difference there either.

Well, at least we know it's not the AC!

I would have thought that you might get a slight thermodynamic advantage because of the lower air temps, but, come to think of it, the bastards probably screw that up by adulterating your gas in the "interests" of air-pollution.

(Don't get me started on expletive ethanol.)

[damocles]

From Geezer:

It's a lot more than experimental error. That would only apply if all conditions were carefully controlled, but in this case there are so many uncontrolled variables that the results could be very misleading. I would think a simulation with accurate input data would yield much more meaningful results.

Geezer, in our local dialect of the English language "experimental error" means "the uncertainties arising from design, execution, and measurement in an experiment". It is hard to see how the effects that you refer to, which are undoubtedly real ones, are "...much more than..." this.

I put together the outline of a very crude experiment that should probably undergo considerable refinement of experimental design in an attempt to snag most of these "uncontrolled variables" before it would show significant and trustworthy results at a low error level.

But the question has always been about real operation of real vehicles. I would submit that testing of some sort on real vehicles should always be preferred to simulations or models, which are always subject to missing or mis-translation of real operational effects.

So you should run the tests on two days with identical weather conditions with a professional driver on a vehicle testing track with a sealed fuel tank and petrol from the same batch ... etc. The real danger is that the in going down this track, you may produce a poorer simulation of operation of an ordinary family vehicle in ordinary conditions. But there is little doubt in my mind that a computer type model or simulation would be even more artificial, and much more fraught with the danger of overlooking important but unforeseen factors.