Other Experiments

I enjoy tinkering with things, and I have tried several different ways to improve fuel economy on my Escort ZX2. Unfortunately most of them have not been particularly successful in the end:

(1) Driving slower - This is clearly an effective way to improve fuel economy. The EPA web site (see "Web Links") says that each 5mph above 60mph hurts fuel economy by about 7-8%. This is consistent with my simulations, but my actual data shows a smaller benefit. I did a careful back-to-back experiment during my regular commute, which is about 90% highway driving. I normally drive 70mph on the highway, but I limited myself to 60mph for about 500 miles of driving. I measured 35.8mpg during this time, which is about 5% to 8% better than the 70mph driving immediately before and after the experiment. This is 3% to 4% per 5mph.

(2) Injector cut-out cylinder deactivation - Several production engines use cylinder deactivation, including some versions of the Chrysler/Dodge "Hemi" V8, GM/Chevrolet V8, and Honda V6. These engines can deactivate some of the cylinders at light loads, when full torque is not needed.  Cylinder deactivation is accomplished by turning off the fuel injectors AND disabling the intake and exhaust valves (for more details, see SAE paper 2001-01-3591). This prevents both fuel and air from flowing through the deactivated cylinders. It is not feasible to retrofit valve deactivation on an engine which was not designed for it. But in theory, most of the fuel economy benefit can be attained by only turning off the fuel injectors.

I installed a switch which allowed me to turn off two of the fuel injectors on the 2.0L DOHC 4-cylinder in my Escort ZX2 (cylinders 2 and 3). The car did not idle or accelerate well with two cylinders disabled, but it ran surprisingly smoothly above about 50mph. I drove for 1,300 miles on my normal commute (about 90% highway), with two fuel injectors switched off at highway speeds, and switched back on for city driving.  Surprisingly, my data showed a clear fuel economy penalty for injector cut-out! My fuel economy was about 2% to 4% worse than the normal driving before and after the experiment. I suspect that the engine control computer was running non-optimal spark timing and/or air/fuel ratio; the "Check Engine" light was on, and there were OBD-II error codes for misfire and for oxygen sensor.

(3) Low rolling resistance tires - According to a couple of published reports (see "Web Links"), low rolling resistance tires should be worth about 4% to 8% fuel economy compared to typical tires. The latest data I've seen (from 2002) shows that the lowest rolling resistance tires were Bridgestone B381 (originally developed for the Honda Civic Hybrid).  I bought a set of B381 tires in April 2007, and I carefully measured fuel economy just before and after the switch. I measured no change in fuel economy! This is partly because the old tires were almost completely worn out, which also improves rolling resistance. But I still expected to measure some benefit...

(4) Higher tire pressure - The same reports mentioned above also say that tire inflation pressure is important for fuel economy. One "rule of thumb" indicates that 5psi higher pressure should be worth about 10% lower rolling resistance, which should be worth about 1% to 2% better fuel economy. I tried increasing inflation pressure from 32psi front, 36psi rear (recommended by the vehicle manufacturer) to 44psi front and rear (maximum allowed according to the tire sidewall). I did not measure any improvement in fuel economy! I suspect that lower than normal tire pressure hurts fuel economy, but higher than normal pressure gives little (if any) benefit.

(5) Accelerate/coast driving - A common "HyperMiler" strategy for fuel economy improvement is "pulse and glide" (see "Web Links"). Clearly this is effective at low vehicle speeds, and it is commonly used in fuel economy competitions. It is an effective strategy because engines become less efficient at light loads (low vehicle speeds), where friction and parasitic losses become a larger percentage of total work. But at higher vehicle speeds, the engine is already at fairly high load (high efficiency), and aerodynamic drag force increases as the square of vehicle speed (aero power increases as the third power of speed).

Therefore, I doubted that "pulse and glide" would work well at typical highway speeds of about 70mph, and I took some data to find out. Instead of cruising at a steady 70mph on the highway, I accelerated to 75mph, then took my foot off the gas and coasted down to 65mph, then accelerated back up to 75mph, etc. I used an OBD-II connection to my laptop (see "Web Links") to confirm that the engine was not running rich during acceleration, and that the fuel injectors were cut off during deceleration. I drove this way for 575 miles during my normal commute (~90% highway driving), and I took careful fuel economy data. I measured about 3% worse fuel economy than my normal steady 70mph driving before and after the experiment! I think that two factors could explain this result: the non-linear increase in aerodynamic drag at higher speeds, and possibly non-optimum spark timing during acceleration (the engine may be knock-limited at high loads).

Last modified: January 6, 2009  (moved from PeoplePC to Google Sites November 2009)