MFD vs. hand calc. MPG Effect

The common wisdom here appears to be that when averaged over time the MFD mileage will differ from the hand calculated fill mileage by about 1-2%. Some believe this is the work of gas stations, others tire wear, etc. In a previous thread I pondered whether or not there would be some sort of physical explanation for this in the fuel system--perhaps in addition to the above.

One thing that occurred to me was that while the MFD would use the injection pulse and count to calculate fuel consumption and that this would likely not be temperature compensated. Therefore, colder fuel in winter would actually be slightly undercounted on the order of 1% (VERY approximate) because the fuel density changes more rapidly with temperature than the injector orifice dimensions.

Anyway, what I noticed is that as the ambient daily temp averages dropped below freezing, the MFD consumption vs. calculated fill error went from bouncing around zero, to consistently overestimating mileage/underestimating consumption by the MFD. The shift in overall consumption average was about -1% (an undercount by the MFD.)

The wierd part is that if the bladder shrank with cold weather, then one would expect to underfill and therefore get higher hand calculated mileage rather than lower for a few fills. And when spring arrived one would expect to get a lower calculated mileage for a few fills as the bladder re-expanded. Instead, I'm seeing the opposite overall. (I may have had a bladder shrinkage issue in late Fall fill or two, but that was soon replaced by a trend in the other direction, and Spring has not shown an obvious re-expansion event.)

Since I have very few points and only a single year to go on, I can't make any real conclusion yet from my own values. (I might just be reading in non-existent trends from noise in a few samples.) Is anyone else seeing the same thing: winter MFD calcs being more optimistic (more error compare to hand calcs) than summer? I'm particularly interested to see whether this holds up by geography... Over a year I would expect hotter climates to be closer to zero than colder ones if my reasoning is correct.

 

Could be related, but fuel shipped in winter tends to be pretty cold by the time it leaves the tanker. I've unloaded some tankers of other material (heavy solvents and very light ones) and they tend to be much like the ambient. I mostly assume that the fuel temp will be similar to the underground temp. During summer there is likely to be substantial flashing/evaporative cooling during the unload so that the fuel is not quite so hot. This can be recovered with vent condensers (typical in the process industry, but I'm not sure how the fuel terminals and average station is configured.) Then there is the ground acting as a heatsink as well.

Definitely fuel in the tank of the car is going to be closer to the surrounding temps. Garaged cars' fuel should be warmer on winter trips out, but less so if they sit outside in a parking lot all day before the return leg.