In 2005, after local gas prices went over $2.00 per gallon for the first time (on their way near $4.00/gallon a couple years later), I thought about what's involved in fuel efficiency in my car. Since then, I've seen a lot of advice about ways to use less gas, but not much explanation for why it works; and in some cases, the advice doesn't help enough to bother.
Since I teach physics, I decided to apply the ideas I teach to this problem of fuel efficiency. I came up with several strategies; and by loosely applying them, I managed to increase my fuel efficiency by about 6%. I'm sharing these techniques to help you benefit from what I tried, and to build your appreciation and understanding of physics.
With one of my previous cars, I thought its fuel efficiency suddenly got worse not long before the engine died. So, I decided to start tracking my fuel efficiency. The way I do it: every time I buy gas, I always fill the tank completely, always get a receipt (which has the number of gallons on it), and always write my odometer reading on it. Take the number of miles (using the difference in miles from two receipts), divide by the number of gallons, and you get the miles per gallon. Not a perfect method, but it gets quite close.
In the years since, I've learned this is a pretty common method to know what your real fuel efficiency is - much more reliable than manufacturer estimates, or what it shows on your dashboard.
Since doing that, I've read about these and other techniques in other places. With even slight adjustments (like I did), you should be able to notice the difference. You can improve even more if you stick to them pretty strictly.
This article uses ideas that should be familiar to those who have taken high school physics or a first-semester college physics course. I'll mainly use metric units for mass (kilograms) and energy (Joules), since those are the most common units used in the sciences and engineering. But I'll mainly use English units for volume (gallons) and speeds (miles/hour), since my current target audience is in the USA, and that combination seems likely to make the most sense to American students of science and engineering.[*]I usually don't like to mix metric units with English units, but I want this page to make the most sense to the most people. If there's enough interest, I'll write a version of this page using metric units only.
If you don't care about the "why" and just want the tips, jump down here; hopefully the science will still pique your interest.
IMPORTANT: Safety First
Don't risk an accident to save a little fuel. You'd use less fuel if you never had to stop at a traffic light or stop sign. But far and away the most inefficient thing you can do, even if you're focused on the narrow issue of energy, is get into a traffic accident. Injuries, stress, insurance bills, police reports, and lost time are all more important than energy. But even if no one is injured and it's not a big deal to the people involved, consider this: it takes a lot of energy to tow a car, repair it, and return it to you. Fuel often leaks as the result of a crash. So, a single accident will probably cost more energy than you'd save in a lifetime of high-efficiency driving. The human cost is even greater.
It's All About Energy
People have been using math to analyze physical systems for centuries. Once "energy" was defined mathematically, people found that it's never created or destroyed, it just moves from one system to another (or changes form within the system). It's easy to fool yourself, since you have to make sure all of the energy really is tracked down - which can be very hard to do.
For over two hundred years, people didn't need to worry about anything other than a few forms of potential energy, kinetic energy, heat transfer, and energy transfer by forces (called "work"). Then came radioactivity. Some say "Energy is conserved, except during nuclear reactions," for which you need to add in some special rules. But a better picture is: matter happens to be a very compact and stable form of energy. So stable, in fact, that you usually can treat it as something different from energy. If you do need to treat it as energy, just use E=mc2 (where m is the mass of the thing, and c is the speed of light in vacuum (roughly three hundred million meters per second, or 186,000 miles per second)).
Energy has to come from somewhere, and it has to go somewhere. That's the main idea behind conservation of energy, a concept used throughout physics and many of the sciences. It's a very important concept, one worth learning.
When physicists say "system", we mean "a well-defined group of objects." For the purposes of this page, the "system" is "the car" (and all components of it, including the gas tank and the gas inside).
Sometimes, it makes sense to keep the system simple ("system = one ball"); other times, it actually makes calculations easier if the system has more to it ("system = three cars that collide with each other"; or "system = the earth, sun, and a spaceship").
In terms of an equation, you may state it as: the energy put into any system, plus the energy it had to start with, is equal to the energy that it has at the end, plus the energy taken away. Or, more compactly:
Einto system + Eof system initially = Eof system final + Eout of system
The Energy you Have to Spend
Most cars have just a few types of energy worth considering (in the Eof system initially term):
Chemical Potential Energy, stored in the gasoline. It varies a lot by the blend of fuel, and somewhat by the temperature, but several sources quote that burning one gallon of gasoline releases around 120 million Joules of energy.
Kinetic Energy, the energy of motion. Every moving object has energy associated with its motion. This depends a lot on how heavy the car is and how fast it's moving, since the equation for it is K= ½mv2 (m is the mass and v is the speed). One example: a typicalsedan has a mass of around 1500 kg. If it's moving at 60 miles per hour, it has around 500,000 Joules (half a million Joules) of kinetic energy. (So if there were no other types of energy and perfect efficiency, it would only take 1/240 of a gallon to go from zero to 60 mi/hr (miles per hour[*]I prefer mi/hr instead of mph because it makes it much easier to see and understand that you're dividing by hours - which helps in many calculations.). Obviously, there's a lot more going on.)
Gravitational Potential Energy. If you're at the top of a hill, you have a lot of energy stored up, which you can use to speed up. The equation for it is Ug= mgy (g=9.8 m/s2 is the gravitational field near the earth, and y is the height above any reference point you decide to pick). So, the same typical sedan 200 ft above the bottom of a hill can get around 900,000 J by going down the hill. If all of that energy could be converted into kinetic energy, the car would go over 75 mi/hr. Again, there's a lot more going on.)
Electric Potential Energy, stored in the batteries. For most gasoline-powered cars, this is tiny in comparison with the others, and will typically be about the same at the start of the trip as at the end of the trip. (So, it would show up as the same value in both Eof system initially and Eof system final terms, and cancel out.) But for hybrid and full-electric vehicles, this can be extremely important. Since most people concerned with fuel efficiency run entirely on gas, I'll ignore this on this page. Still, many of the tips apply just as well for any moving vehicle, including hybrid and electric.
Very often when people think about energy, they only think about the energy in the fuel. Certainly, you've spent good money to get that fuel. But you should also consider the other sources, and use them wisely: you've spent good fuel to get that kinetic energy, and you shouldn't just throw it away.
How You Spend That Energy
Most cars also have a pretty small list of ways you spend that energy (in the Eof system final and Eout of system terms) under normal driving conditions[*]e.g. ignoring the energy to break the chemical bonds of a car tire when you "burn rubber".:
When discussing car engines, it's common to discuss the concept of engine efficiency. According to the laws of thermodynamics, no engine operating in the real world[*]Since we don't have access to anything at absolute zero temperature can be 100% efficient. In fact, gasoline-powered cars typically waste around 2/3 of of the energy available in the gasoline. But the energy doesn't just vanish: it mostly goes into Wasted Chemical Potential Energy, Waste Heat, and Fighting Friction.
So, this list describes causes and effects of efficiency - there's no need to account for it separately.
Kinetic Energy. If you're going faster at the "end" of some process than at the beginning, then you've gained kinetic energy. And even if you slow down, you still have some kinetic energy until you come to a stop.
Gravitational Potential Energy. It also takes energy to go up a hill. Since trucks can have trouble converting the energy in their fuel into gravitational potential energy fast enough, you may find them building up a lot of kinetic energy reserves as they approach a steep hill. This makes sense, but can make driving stressful.
Wasted Chemical Potential Energy. Almost every engine wastes a lot of the available energy. If gasoline didn't have any impurities (and were just a pure hydrocarbon), the only things in car exhaust would be carbon dioxide and water. Carbon monoxide, ozone, and other pollutants are common, however. It would be possible to get more energy out of them, but your engine doesn't, for practical reasons.
Waste Heat. Your exhaust is hotter than the air that was brought into the engine; energy had to be spent heating it up. Also, every time you hit the brakes, some of the kinetic energy that you had heats up the brake pads. And, each section of your tires heats up as it gets repeatedly squashed and relaxed as the tire rotates.[*]Physicists sometimes call the resistance due to squashing tires "rolling friction."
Doing Work Against Air Resistance. In physics, work done is equal to force times distance, and is one of the most important sources of Einto system and Eout of system: it takes energy to do work. The stronger the force and the larger the distance over which you apply that force, the more work you need to do.
The force of air resistance (or "drag force" FD) is calculated using
FD= ½CρAv2
where C is the "drag coefficient" (how easily the shape pushes air out of the way), ρ is the density of the air you're moving through, A is the area of the car that's face-on to the wind, and v is the speed of the car relative to the air.
If there's no wind, v is the same as the speed of the car; if you're driving into the wind, it's the speed of the car plus the speed of the wind; if the wind is at your back, it's the speed of the car minus the speed of the wind (and if that ever goes negative, the wind is helping to push you rather than slowing you down). If the wind is at an angle, things are a little more complicated, but that's not too important for this discussion.
Air resistance is an interesting kind of force: if you're not moving relative to the wind, then there's no air resistance. The faster you go, the stronger the force. In fact, the strength of the air resistance force goes up roughly with the velocity squared. So, going a distance of one mile at a speed of 60 mi/hr would require four times the energy to fight air resistance as it would if you went that same mile at 30 mi/hr.
Air resistance also depends strongly on the size and shape of your vehicle (and its effective shape can change a lot by strapping things to the roof, having the windows open, etc.).
Fighting Friction. Vehicles contain many moving parts (whether or not the vehicle itself is moving). Every time something rubs against something else, you're spending energy. This heats things up (so you may prefer to think of this in the "Waste Heat" category), but often that heat is dissipated to the surrounding air quickly enough that you might not notice unless you put it in its own category.
Powering Electric Devices. This is usually a small factor in comparison to all of the others listed, except under a few circumstances. If you are running your air conditioner, or if you're powering things usually used in a home, you will notice a need for more fuel.[*]"If it shows up as the same value in both Eof system initially and Eof system final, why does it matter?" One way to think of it: when there are more electric devices, it's takes more force to spin the alternator (since there's more current, so the magnetic forces are stronger). More force means more work to generate that electricity. Another way: that electric energy can't go back in the fuel or the batteries, so it has to go out to the surroundings somehow (heat, light, sound, etc.).
On one trip, I travelled mostly downhill (a difference of elevation around 600 ft.). Traffic conditions forced a steady moderate speed of around 50 mi/hr (without needing to hit the brakes), for a two hour ride without stops. It was a cool day, so my windows were closed; but not cold enough to need the heater. The car was running well. The conditions weren't perfect since there was some drizzle, so my headlights and wipers were on. Still, I got about 5 miles per gallon better than normal.
So, in order to have the ideal fuel efficiency, you should travel downhill, at a constant moderate speed (without hitting the brakes), downwind, with the windows closed, lights and air conditioner off, with a very efficient engine. Some of these things you can control, some you can't. Some will be a big effect, some small. So let's do some rough calculations to see how to deal with it.
It's not quite all about energy.
If you care most about cutting costs rather than focusing on how to drive more efficiently, there are some things that you can do that have nothing to do with energy. Probably the biggest effect is timing your purchases for days when gas is less expensive, or looking for stations that charge less at the current time. That's just common sense, with no physics knowledge needed. Perhaps the most significant effect for which physics is helpful is buying fuel when it's the coldest possible.
Fluids expand according to
ΔV = βV0ΔT
where β is the coefficient of volume of expansion of the liquid V0 is its initial volume, ΔV is its change in volume, and ΔT is its change in temperature.
To get a change in volume of 1% (i.e. ΔV/V0=0.01), the gasoline must change temperature by around 19 F°, or around 10 C°.
This calculation isn't perfect: no material expands in a perfectly linear way, but it's close enough at these temperatures, for this simple back-of-the-envelope calculation.
Why does this matter? Almost everything, including gasoline, expands as it gets warmer. But gas stations dispense fuel by volume, not by mass (which is also referred to as the "quantity of matter"). So: buy fuel when the fuel in the tanks is at its coldest, but use it when the engine is warmest (which usually happens anyway, as the car runs).
This can be very tricky to establish: the temperature of the gasoline in the tank is a lot different from the temperature of the surrounding air. It will be somewhat closer[*]Only "somewhat" because the trucks hold a lot of fuel (so it takes a lot of energy to change the temperature), and very well-insulated (so, like a thermos, it's harder to get the energy to the fuel). just after a delivery truck arrives. (So, during the summer, you want to avoid filling up for very many hours after the tank is filled; during the winter, it's probably the opposite.) Likewise, short-term changes tend to get averaged out: the ground is a pretty good "heat sink" keeping temperatures pretty steady; and underground gas tanks are double-walled, insulating the contents (similar to a thermos). But especially in spring and fall, there can be swings where the average temperature changes by 15F°[*]Why is the degree sign after the F? Because being at a temperature of 30°F is so much different from changing by 30 F° (for example, from 60°F to 90°F). So, scientists always used to use °F to mean "the temperature you're at" and F° to mean "how much the temperature changed." People don't follow that rule as closely as they used to, but I find it useful. over a few days. I'd be surprised if there weren't a significant change in tank temperatures, even though somesources suggest that consumers shouldn't consider this (though commercial users andsuppliers do).
Careful timing on weeks with large temperature swings may approach (but are unlikely to reach) a savings of one percent. So, the cost savings per gallon in cents would be the less than the cost of the gas itself in dollars. If you're willing to drive out of your way to save a cent or two per gallon, you may be willing to adjust your timing when you fill up your tank.
There are other considerations, but have smaller effects, so I haven't mentioned them here. If you know another big effect, send me an email and I may update this page.
How to Spend Your Energy Well
-or- "Shut up and tell me what to do"
Minimize braking. Hitting the brakes is a complete waste of energy (and worse): you're spending energy (mostly kinetic) to heat up your brakes - and wearing out your brakes faster.[*]Electric and hybrid vehicles are more efficient largely because they recover some of this energy. With "regenerative braking," pressing the brakes makes the car's electric motor act as a generator to recharge the batteries. It's not perfectly efficient, but it's a lot better than losing it all. (When you need to stop quickly, standard friction brakes also engage.)
There are other methods to recover some of the energy from braking: flywheelsystems, and "hybrid hydraulic"systems that store braking energy in compressing a fluid (liquid or gas).
You can reduce braking by looking ahead and anticipating. If you'll need to stop/slow down anyway (at a light, stop sign, or traffic congestion), keeping your foot on the accelerator costs a lot of energy in exchange for zero or minimal time. Instead, you should coast to the slow/stop.[*]Think about what you naturally tend to do when riding a bike.
But remember: safety first! Give yourself plenty of reaction time - don't get too close to car in front of you. And in some cities, driving like this can inspire road rage, aggressive driving, and accidents. Don't let your desire to save some fuel override your common sense.
For a smaller car in one moderate traffic jam, you'd save around 0.03 gallons (or 3 cents × the cost of the gas in dollars) by driving somewhat conservatively vs. driving aggressively. In a bigger vehicle and/or driving more conservatively can save even more.
Suppose in a mild traffic jam, you could average the same speed by either going between zero and 40 mi/hr ten times, or between 10 and 30 mi/hr ten times.
Using the equation for kinetic energy, you find[*]Kinetic energy K= ½mv2, and 1 Joule of energy = 1 kilogram·meter2/second2 Kat 40 mi/hr = ½(1500 kg)[(40 mi/hr)(1 hr/3600 s)(1609 m/mi)]2=2.40×105 J Kat 0 mi/hr = ½(1500 kg)[(0 mi/hr)(1 hr/3600 s)(1609 m/mi)]2= 0 J
⇒ Klost each stop = 2.40×105 J
⇒ Klost in 10 stops = (10)(2.40×105 J)(1 gal/120×106 J)(3) ≈ 0.06 gal
(The ×3 is from the engine only making good use of 1/3 of the available energy.) Kat 30 mi/hr = ½(1500 kg)[(30 mi/hr)(1 hr/3600 s)(1609 m/mi)]2=1.35×105 J Kat 10 mi/hr = ½(1500 kg)[(10 mi/hr)(1 hr/3600 s)(1609 m/mi)]2=1.50×104 J
⇒ Klost each slowdown = 1.20×105 J
⇒ Klost in 10 slowdowns = (10)(1.20×105 J)(1 gal/120×106 J)(3) ≈ 0.03 gal that you lose around 0.06 gallons' worth of kinetic energy going between zero and 40 mi/hr ten times, while you only use around 0.03 gallons going between 10 and 30 mi/hr.
This ignores differences in air resistance, energy spent idling, and high rpm's when accelerating (which would favor going between 30 and 10 mi/hr), and differences in engine efficiency at various speeds (which would favor going between zero and 40 mi/hr), but this should be in the right ballpark.
Of course, if you drive a heavier vehicle, or if conservative driving leads to a steadier speed, then the savings are even greater.
Carry speed through slowdowns. There may be times when you need to slow down, but not stop entirely. Going as fast as possible once you're free to do so allows you to keep that kinetic energy that you paid for.
If you're used to the timing of traffic lights, you can judge when a red light is going to turn green. So, by doing some extra braking early, you may be coasting at a faster speed when you get to the intersection with the light green. But as always, BE CAREFUL![*]Don't enter the intersection early, or even right when the light turns green: cars often enter just as their light turns red (even though they shouldn't). And don't brake too early if there's a car close behind you: they may run into you (since they're not expecting you to slow down), or you may inspire anger and aggressive driving (even though you're actually saving them time and fuel, as long as they don't hit you).
At ramps from a highway to another major road, you know you'll be slowing down, so it makes sense to coast as you approach the exit, then stay at that same speed through the curve (as long as it's not rainy or icy) to be close to speed you want for next road. Most people go slower than necessary on ramps because going around a curve gives a sideways acceleration[*]Described in physics class as "centripetal acceleration" ar=v2/r, where v is your speed, and r is the radius of the curve you're on. that feels uncomfortable. But as long as your tires are good and it's not rainy or icy, it's less stressful on your tires than a sudden stop.
In heavy traffic, look as far ahead as you can. If you're moving much faster than cars up ahead, start coasting, or maybe even apply the brakes lightly. Give yourself a little victory whenever you have some speed when traffic starts moving again, and never come to a complete stop. But you need to keep very aware of everything around you: not only do you need to look far ahead, you need to watch out for someone cutting in from another lane, and for people behind you getting angry and aggressive.
Drive at the optimum speed for your car. This depends a lot on the car you're driving. Some examples:
Source: WikiMedia Commons[*]This uses data from from Table 4.24: Steady Speed Fuel Economy of this pdf. That document got the data from "B.H. West, R.N. McGill, J.W. Hodgson, S.S. Sluder, D.E. Smith, Development and Verification of Light-Duty Modal Emissions and Fuel Consumption Values for Traffic Models, Washington, DC, April 1997, and additional project data, April 1998" - which I've been unable to locate online. Please let me know if you find it. Source: blog.automatic.com[*]This seems to be the original source, but the site is now offline, and I haven't found the methods they used.
There are a few obvious features for all of vehicles:
At high speeds, drag dominates, so cars get less efficient above a certain speed. (Remember: the drag force depends on your speed squared.) That usually happens to larger cars sooner than for smaller cars. (Remember the face-on area A in the drag equation.) Do your best to avoid those high speeds - you save surprisingly little time for that energy loss, except on very long trips.
At low speeds, fighting friction dominates. At zero mi/hr, you get zero miles per gallon since you're spending energy not going anywhere.
This is pretty easy, assuming I can trust the graphs above, and that I can get reasonable estimates just by looking at the graphs.
Suppose you have a 50 mile trip that you're able to drive on a freely-flowing highway at whatever constant speed you want. Assuming a Honda Civic, driving anywhere between 45 to 55 mi/hr or around 70 mi/hr seems to give around 42 mi/gallon. So, (50 mi)×(1 gallon/42 miles)≈1.2 gallons used. Driving at 80 mi/hr seems to be around 34 mi/gallon, so you use (50 mi)×(1 gallon/34 miles)≈1.5 gallons - around 0.3 gallons worse, or (0.3/1.2)=25% extra fuel burned.
Is that time worth it to you? Let's check:
Driving 50 miles at 80 mi/hr takes (50 mi)(1 hr/80 mi) ≈ 0.625 hr ≈ 38 min.
Driving 50 miles at 70 mi/hr takes (50 mi)(1 hr/70 mi) ≈ 0.714 hr ≈ 43 min.
Driving 50 miles at 55 mi/hr takes (50 mi)(1 hr/55 mi) ≈ 0.909 hr ≈ 55 min.
But "slower is better" is clearly not true: that Honda Civic at around 30 mi/hr seems to give around 32 mi/gallon. So, (50 mi)×(1 gallon/32 miles)≈1.6 gallons used.
The sources for those plots don't give much information - it will depend on model year, maintenance, season of driving, etc. Still, this gives a sense of how to figure what will happen for you.
So why are these graphs bumpy? That depends on the transmission, and what gear you're in. A traditional technique for "hypermiling" is to use the highest gear that's practical for the conditions. In lower gears, the components of the engine are spinning faster, leading to more energy lost fighting friction. Even with an automatic transmission, you can look or listen for shifting to lower rpm's (revolutions per minute). As long as your engine is well-maintained, you'll tend to save fuel by keeping it at lower rpm's. (That's not a perfect indicator [cars often get a little more efficient at slightly higher rpm's than the minimum for each gear], but it's easy to track, and gets you fairly close to peak efficiency.)
Minimize drag. You don't have much choice about the density ρ of the air you're moving through[*]It depends on things like altitude, temperature, and humidity., but you can control some things. Only put things on the roof or back of your car if you really need them for that trip (otherwise, you're increasing both the face-on area A and the drag coefficient C). And at high speeds, using air conditioning is more efficient than having open windows.[*]Air conditioning requires you to spend energy powering electric devices; that doesn't depend on speed. So, at slow speeds, open windows are better than air conditioning. But open windows increase the drag coefficient of a vehicle. When you're at highway speeds, the v2 in the equation for air resistance makes any change in C a big deal.
Minimize mass. Remember the equation for kinetic energy: K= ½mv2. If you're using your car as a storage unit, you're spending useless energy to speed up that extra mass every time you drive. If you never had to speed up or slow down, that wouldn't be much of an issue. But since you do, you're just wasting energy when you speed up, and heating up your brakes more (and wearing them out faster) when you slow down.
Minimize idling. When you're not moving, you're spending energy without going anywhere, and getting zero miles per gallon. While it's true that starting the engine does cost some fuel, somestudies showed that "Idling is only fuel-efficient for approximately seven seconds—after that, it’s better to shut off your engine." But keep safety first: unless your car is designed to do that,[*]This is another big benefit of most hybrid cars: they're designed to shut off when stopped (since they can use electric motors until the gasoline engine can get going again). don't shut off the ignition at every opportunity, or you may cause an accident. (Other drivers anticipate that you'll start moving as soon as possible). Probably the best opportunity to shut off the engine is in drive-through lanes.
When possible: choose routes and times with fewer lights, fewer stop signs, shorter distance, less traffic. Flex time can be wonderful. Driving on a road packed with stores at 8:30 AM can be easy. This helps you minimize braking, stick to a reasonable speed, reduce idling, and hopefully reduce the number of miles to spend those gallons on.
Maybe: As long as it doesn't make your vehicle less safe, slightly over-inflating the tires will reduce how much they squash as they roll. Since squashing of a tire is what we mean by "rolling friction", reducing that makes the car more efficient. (But it's less comfortable, especially if the road is bumpy.) The biggest issues: overly-inflated tires may have less grip on the road; and very overly-inflated tires may burst.
Seeing is believing. It's much easier to see the effect of air resistance on bicyclists than cars, since you don't need to consider the cars' engines. Here's a low-resolution version of one of my favorite clips about this[*]For higher resolutions, just search around for video clips of Michael Guerra and/or superman cycling. (Apparently, he's not the first to try it, but his videos are easier to find.):
Some other shots of the same race show the other riders passing Michael Guerra during the uphill portion, before he goes into his "superman" position.
Even though he's not pedaling (converting chemical potential energy into kinetic energy), his posture on the bike reduces his face-on area A and drag coefficient C enough that he's able to maintain a higher speed v than the other riders.
A scholarly research article has been published about the importance of aerodynamic bike race positions. It includes several wonderful figures showing many different racing positions. Here's just one:
FD= ½CρAv2. It's a powerful effect, once you know what to look for.
You may find advice to accelerate slowly from a stop. There is some wisdom to that, but it's easy to overdo it. The best thing to do: stay in a reasonable gear (so the rpm's don't get huge) while getting to your most efficient speed reasonably quickly.
The following graph helps, but it's easy to misinterpret it:
Source: antranik.org[*]This clearly wasn't the original source, but I haven't been able to locate an earlier one, and I haven't found who made it, nor what methods they used.
Accelerating faster causes you to be in a lower gear at higher rpm than at lower accelerations. That is less efficient (at each moment) than being in higher gear. That's a large part of the reason why the peak for maximum acceleration is taller than the peak for slow acceleration.
How does this affect the fuel spent for the whole trip? Since this is grams of fuel vs. time, the total fuel spent over the entire time span is equal to the area under each curve. By eye, there doesn't seem to be too big a difference between these areas.
But notice that this graph is fuel spent vs. time. When considering energy spent (and the fact that we want to travel somewhere), what matters more is fuel vs. distance.[*]That's why we use "miles per gallon" not "hours per gallon." If you just wanted the fuel to last the longest time, you'd never go anywhere. If this were a graph of fuel vs. distance, the total area under each curve would give you an apples-to-apples comparison.
Once you recognize that you cover more distance in the same time when accelerating faster at the start, you can see there is a difference in the efficiency once you're at your desired speed and are in high gear - but it's still not huge. So, you should try your best not to jump to too low a gear, but it's also possible to waste fuel by accelerating too slowly (spending more time and distance in the least efficient mode for your car).
And if you're interested in saving money more than energy, buy fuel at days and times when the gas underground is cool.
Many other tips can be found online, such as those at ecomodder.com
How it Worked Out for Me
When I changed my driving habits, I mainly worked to minimize braking and listened for low-rpm driving. I made this change around March 2005 while driving a 1995 Geo Prizm mostly for 20-mile commutes, with about 10 of those miles on the highway. I still had to get to work, go shopping, etc., so this is much more of a real-world application than I've found in other places online.
This graph shows a moving average of five fill-ups[*]averaging the mi/gal for the fill-up from that day with those for the two prior fill-ups and the two following fill-ups.; otherwise, the scatter in the points distracts from the trends.[*]I think most of the scatter was due to variations in how full the tank was when the pump stopped. The trends are still clear in the raw data, but the eye is attracted to high and low outliers. For the 48 months prior to the change in driving habits, I averaged 31.0 miles/gallon. For the 48 months afterwards, I averaged 32.9 miles/gallon - an improvement of around 6%.[*] Now, since I didn't drive exactly the same amount each year (and each season within each year), and since road conditions couldn't be exactly the same, I can't trust this percentage too exactly. But since these both average over two full years and the effect is fairly large, I'm confident that these changes had a real effect at least around 5%.
There's also a clear seasonal effect to fuel efficiency. I have a few thoughts about reasons for that:
Spending energy heating up the engine. At lower temperatures, lubricants have a higher viscosity, so you spend more energy fighting friction. Also, while the engine is cold, igniting the fuel doesn't just push the piston - some energy goes into heating the engine. Another way to think of it: the combustion gasses get cooled by cold engine walls, so the gas pressure doesn't increase as much, so the push on the pistons is weaker.
Different fuel blends. In winter, gasoline is allowed to include more volatile fluids (i.e. stuff that evaporates relatively easily). The more expensive additives used in summer blends contributearounda 2% increase[*]These linked pages describe this effect, but don't cite any reliable sources. If you know of one, I would appreciate your input. in energy density.
Other effects of the cold. At lower temperatures, air is denser, leading to more air resistance (all other things being equal). At lower temperatures, gases contract; so unless you inflate your tires, their air pressure is lower, leading to more rolling friction. In winter, there's a chance of snow; pushing that out of the way requires a lot more force (and energy expense) than smooth rolling. In winter, there are fewer hours of daylight (that's why it's cold to begin with). So, you're more likely to use your headlights, spending more energy on electric devices. It's cold, so you use the heater more, again spending more energy on electric devices.
Air conditioning. I don't like the quality of the air coming out of most car air conditioners, so I rarely use them. But when it's very hot, I do. That's the main cause for the dips in the summer. There may also be a little more loss due to the gasoline in the tank evaporating, or by buying warmer gas, but I doubt those would be noticed at this scale.
Since I saw a more than 20% difference between summer and winter, it must have been a combination of the above. I'm hardly alone: "Fuel economy tests show that, in city driving, a conventional gasoline car's gas mileage is roughly 15% lower at 20°F than it would be at 77°F. It can drop as much as 24% for short (3- to 4-mile) trips." -- fueleconomy.gov
Full disclosure: some years later, in part because of these thoughts, I didn't drive as conservatively as I should have, and got into an accident. A lifetime of increased efficiency won't make up for that. Since then, I've still applied these techniques, but been more defensive to give myself time to react to other drivers' odd driving. If that prevents an accident in the future (for me or any of you), that will help everyone involved.
