Acceleration Efficiency Tests (Subaru AWD)
 

I know someone already made a post about an Automatic toyota, but here is mine for a mildly modified Subaru impreza 2.5i AWD. I got almost the opposite result as the other member with matrix.

The Test Strip:
565 meters of smooth asphalt track with sunny conditions, 45% relative humidty, 16'c temprature and calm winds.

The car:
2009 Subaru Impreza 2.5i AWD with 89 octane gas, stock internals.

test runs:
multiple runs were done in multiple engine loads. The goal was to achieve 50mph and maintain, then let off at a check point and call mileage at second check point. The runs were timed.

consumption---LOD------RPM
12.7L/100km....95% ....~3500rpm
12.3L/100km....100% ....~4000rpm
11.2L/100km....70% ....~2500rpm
11.0L/100km....95% ....~2500rpm
11/0L/100km....80% ....~2500rpm
10.7L/100km....80% ....~3000rpm
10.7L/100km....70%....~2500rpm
10L L/100km....60% ....~2000rpm

Run....Load(%).....Shifts(RPM)...Consumption(L/100km)...Acceleration Time(s)...
A.........WOT.............5500RPM...........15.8L. 100km
B.......+90%.............3000rpm............13.3L/100km
C.........80%.............2300rpm............12.3L/100km
D.........60%.............1800rpm.............12.1 L/100km
E.........80%.............2300rpm.............12.0
F.........80%.............2500rpm............12L/100km
G.........90%.............3000RPM............11.9L/100km
H.........90%.............3000rpm.............11.9
I..........60%.............1800RPM............11.5 L/100km
J..........60%.............1800rpm.............11. 5

Result:
the above data was used to create this chart:

http://imageshack.us/a/img33/6985/7tqa.png

Conclusion:
Based on the raw data above it seems to me that lower engine RPM along with moderate loads (60%) is essential for optimum fuel efficiency in this car. The lower RPM is understandable due to less fuel enrichment and internal frictions. The reasoning for lower load benefits could be three folds. One it could be due to higher ignition timing at low loads. Higher spark advance improves volumetric efficiency which in turn improves torque and pumping losses. It could also be that rate of acceleration doesn't increase as much when engine load increases pass a certain point. From above chart you can see how steep the engine load incline gets while acceleration rate continues to be steady. So pass a certain engine load the engine is breathing more air yes, it might still be running stoich yes, but it is not producing anymore power to make up for the higher fuel flow. There are also sometimes anomalies such in some low-moderate loads such as temporary lean conditions since most cars are Not equipped with a high-resolution wideband sensor.

http://imageshack.us/a/img545/5759/ater.jpg

^read:
http://download.springer.com/static/...a2357&ext=.pdf

I'd like to see time to accelerate to 50 MPH. Also keep testing lower accelerations until you find the optimal acceleration for gas mileage.

A better way to show your results would a plot with LOD on the horizontal axis and L/100km on the vertical axis.

Nice testing. You didn't mention one key variable though, automatic or manual trans?

SGII shows best acceleration load. We need to know distance traveled and time to accelerate to speed to make a useful comparison between your different acceleration loads and fuel consumption data.

time was taken into account. One thing noted was past ~2/3 manifold pressure (%66) the amount of increase in acceleration was insignificant assuming shifts were at same RPM. i will post some time logs later when i get home. car is automatic but the triptronic system was used to shift at desired RPM.

UPDATE ---

I can't find the numbers for the load testing under 2500rpm, but here are the numbers for when the car is left in fully automatic D (drive).

RPM--MPG--Load--Acceleration
1800--23.5-60--- 38s
2500--22---70---36
3000--22---80---34
4000--19.1--95-- 32

I have also been advised by several subaru members that variable valve lift system could play a role in improving swirl (BSFC) at low RPM and moderate loads. I'm not sure how this works but it's apparently very similar to honda's 3-cylinder vtec engine, where it has multiple sweet spots along the powerband (1500-4000rpm).

MORE DATA ---
to 60kmh and coast to end of 350m runway:
8.1lhk ~2200rpm 39s 65% Drive
9.4lhk ~2200rpm 40s 65% tiptronic
9.1lhk ~2500rpm 43s 70% tiptronic
11.8hjk ~3600rpm 36s 90% tiptronic
10.1lhk ~3600rpm 40s 80% tiptronic

0 to 70km/h
60% 18.1lhk 20s ~2000rpm
70% 23.2lhk 16s ~2200rpm -> 28% more fuel vs. 60%load but only 20% faster
85% 35.7lhk 11s ~2500rpm -> 53% more fuel vs 70%load but only 40% faster

acceleration?
 

Taking 40 seconds for 0 to 60 kph (37 mph) is almost like real acceleration. In normal traffic you would be a moving road hazard. For a 3000 lb vehicle the acceleration power is only about 6 hp--hard to believe any ICE could operate very efficiently at such low power levels.

Putting a coast down phase into an acceleration measurement seems to be confusing and creates difficulty in understanding the data. For example your runs G and H indicate very nearly the same fuel mileage as I and J--what is the meaning of that--e.g. how much is contributed due to acceleration and how much due to coasting?

Sentra's testing showed that 90% load acceleration uses less fuel than lesser loads for both acceleration and P&G.

sorry where did u get 40s to 60kmh from? I never posted this. and my car hits open loop at 90% load and dumps enough fuel to power a prius for a week. the coasting starts after car hits desired speeds. so the faster you accelerate the more you coast. in this case the extra coasting did not make up for yhr loss of fuel due to higher load acceleration. and maybe sentera drives a different car? this is a bsfc for an older gen prius:

http://expha.com/articles/images/10_graph_2.gif

as you can see the bsfc sweetspot is not at the highest load in this case. I have also seen the bsfc for a geo metro and for that car too its at moderate load and engine rpms (3k rpm and 60% load). every car is different and factors like ignition timing, AFR, EGR, EGT, vvt, valve lift and transmissions play a key role.

for example honda and subaru are using a variable valve lift system that eliminates pumping losses due to throttling by fully opening the throttle at low loads and switching to a lowlift cam that retards the intake valve closure time. this way you can accelerate slowly but without the pumping losses of a traditional throttle system.

honda's i-VTEC (watch at 2:25 for low load tunning):
https://www.youtube.com/watch?v=aFf_sw69Gf4

The difference in the results IMHO suggests my points in the previous thread stand - this is based on instruments and not a scientific test, and no the two are not the same.

Just my 2p.