For the past 6 months or so I've suspected my mileage has been varying much more than expected so I started performing some tests.

First I started resetting the trip computer and measuring the mileage the first 5 minutes of my morning commute to work and then repeating for the afternoon commute home. Both trips are around 15 minutes +/- 1 minute. I almost always stop at the same lights and for close to the same amount of time. Average speed one way versus the other doesn't vary by much more than a couple of mph. The route is flat and there appears to be no appreciable difference in elevation between the two locations.

The first few weeks I did this test the morning temperature in my garage was between 45 and 55. Afternoon temperates for the trip home were usually in the same range but a couple of times a little lower.

What I noticed is most of the morning trips to work averaged 12-13 mpg while the afternoon trips home averaged 22 mpg. However, 2 morning trips were at 20 and 21 mpg when I got to work. On one of these occasions the temperature was actually more than 10 degrees lower than the prior tests. Measuring only the first 3 minutes resulted in many times as low as 8 mpg (this includes an average of 30 seconds waiting at the first light and then 2-3 minutes at 40-45 with no stops). The two times it was higher it was reading 16-18 by the end of the first 3 minutes.

When the car is fully warmed up and the CC is on 45 it will average 32-34 mpg. Doing this as soon as possible after first start of the morning will result in values between 14 and 24 mpg.

I contacted the local dealer and Acura USA to try to find out just how much richer the mixture would be during the warmup to see if this is out of line with design. No one has been able to find the real data but conversations with other car dealers service departments seems to hint at 15-20% richer during warming. What I'm seeing is 100%-200% increase during warmup - but not all the time. The local dealer has looked at the car 4 times but they finally admitted there is no good way to test the car when it is cold (open loop mode) because all the sensors are disabled while the engine warms up.

Has anyone else seen their mileage vary this widely, especially during cold morning warmups? Does anyone know what the fuel/air mixture increase is during warming, i.e., the ECU's warmup lookup table values?

 

when the motor is not up to operating temperatures, the drive by wire system will initiate a process to get it up to operating temperatures as quickly as possible. This consists of a few procedures:
The car will idle at 1.1K RPM's to warm up quickly, this is only in PARK.
In DRIVE, the car will retain the gears for a longer period to keep RPM's up to increase heat.
The car will not engage into VTEC and will simply act as a regular motor as during non-vtec RPM's.

The TL has an average idle of 700-800 RPM's and time to time will regulate fuel during start-up for the ULEV 2 System (emissions)

Also you must account the rolling resistance created by the tires, cold tires mean flat-spots, and it will take more fuel to spin a tire that is out of round and has no grip than a hot, round tire with grip.

As well, you must account that colder air means more fuel added to keep the mixture in check.

 

For the last two tanks I've been letting my TL idle to warm up for 2-3 minutes instead of only 30 seconds or so and my overall average mileage has gone up 10-15%. So much for the "modern cars do not need any warm up period".