I think there is a misconception of how much battery capacity is actually used by the algorithms during normal operation, what the battery symbol really is, and trying to save EV for later will results in a significant boost to FE. Based on my observations, what other's have observed, and common practices used in hybrid operations to ensure battery longevity, the PCM algorithms limit the usable capacity of the high voltage battery to between about 30% and 70% of the maximum battery capacity of 1.4 kWh. But in normal operation on fairly flat roads at all speeds, I have observed / recorded that generally the state of charge (SOC) of the HVB stays between about 40% and 55% of the capacity. Very seldom do I see the SOC dropping below 38% and above 65%. There really isn't a lot of usable energy to save for a rainy day. So, 15% of the capacity is only 0.21 kWh (0.15 x 1.4 kWh). Yet, looking at the battery symbol one would think that the usable energy would be a significantly higher portion of the 1.4 kWh.
The battery symbol display does not represent the SOC of the battery but appears to represent the % of the usable capacity. I see about the same relationship that hannawcu saw in this post between usable capacity and HVB SOC. Also, the battery symbol requires a "calibrated eyeball" to read with any degree of accuracy and does one count the tip of the battery symbol or not. I can say that when the battery symbol appears to show 50% (to me), the SOC is 48+%. When the battery symbol appears to show about 1/3, the SOC is about 45% and when the battery symbol shows about 2/3, the SOC is about 52%. When the battery level is about a lines width from touching the top of the battery (not the tip), the SOC reads about 64%.
Also, one may think a road is flat but it usually isn't. I use my Garmin to display elevation while traveling and use Google Earth and topo maps after the fact to see how elevation changes may have affected my recorded data. Small changes in elevation, wind and so forth affect load on the vehicle and how the powertrain algorithm reacts to maintain speed - run ICE, run EV, coast, run ICE and EV.
Bottom line is that trying to second guess why the algorithms choose to run ICE instead of EV, EV instead of ICE, or both and to think that operating the powertrain differently would have been better with such limited data is a fruitless exercise. However, this is not saying that anticipating driving conditions and forcing certain powertrain operations can't improve FE. Hypermilers anticipate driving conditions all the time to maximize FE. But I'll say this again, hypermilers typically sacrifice time for FE. Their average speed is lower than they think as they P&G, coast, and so forth.
Below I've posted a trip I recorded recently to show how the SOC changes over 42 miles. Also note that ECO cruise was set at 55 mph just before the 800 second mark for about 5 miles. SOC climbed from low 40s to about 58% where it leveled off. This is the point many refer to as high ICE. There is virtually no change in SOC and RPM remains fairly constant varying for the slight changes in load. FE looks to average about 47 mpg into nearly 10 mph sustained wind. But I wondered why the FE dropped during this period from a high of about 48 to 46 mpg. Even though the road looked flat during the entire 5 miles, the first part was downhill and the last 1 1/4 miles was flat.
Edited by Plus 3 Golfer, 03 April 2014 - 09:08 AM.