45psi matter? CANbus data turning me into a drag-queen..er..

[tbleakne]

More on my modeling approach:

I plan to use the method described in my previous post to collect pairs of data (Fd, Veq),
where Fd is the total drag and Veq is the equilibrium zero-power speed for descending the given grade.
I will fit this data to the expression:
Fd = Crr + A*v + {Cd * A * 1/2 * rho * V^2}
I will use least-squares find the best fit of parameter A and Crr. I have confidence that the aerodynamic drag as computed by Sparky is accurate, so it has no parameter in this fit.

Of course it is important to control for tire pressure, Sparky's original interest. My tires are at 39 lbs/in^2.

Getting rho right for each data point is important as well. Rho is a function not just of altitude and temperature, but also humidity. The molecular weight of dry air (80% N2 + 20% O2) is:
(.8*14*2) + (.2*16*2) = 28.8.
H2O has a molecular weight of 16+2 = 18.

Humid air is lighter. Pilots correct for this fact when computing how much runway they need to take off. We should be able to drive the Leaf further at faster speeds on warm, humid days at high altitude.

Alternatively, if we have the barometric pressure, we can get rho from the ideal-gas law:
rho = p/(RT)

 

[TomT]

Actually, we don't. We compute density altitude based on actual field altitude combined with field air temperature. This gives the effective attitude at which the plane "thinks" it is operating.
Some of the newest, fanciest jets with top-of-the-line Flight Management Systems and Air Data Computers might also factor in humidity but for most pilots, it is not a factor that is considered.

Tom
Comm/Inst

Pilots correct for this fact when computing how much runway they need to take off.

 

[TonyWilliams]

Getting rho right for each data point is important as well. Rho is a function not just of altitude and temperature, but also humidity........
Humid air is lighter. Pilots correct for this fact when computing how much runway they need to take off. We should be able to drive the Leaf further at faster speeds on warm, humid days at high altitude.

Pilots, quite simply, call that "high, hot, and humid", since those are the ingredients of "density altitude". Every part of the airplane is affected; wings, propellers, engine power, etc. Humidity has a bigger affect on engine power, which doesn't apply to the LEAF.

You rightly say that temperature and elevation will affect the LEAF, but I wouldn't spend much time on humidity. I've NEVER seen airline data include humidity for performance calculations. Runway calculations are more dependent on wind, aircraft weight, how much power derate on the engines for pressurization, anti/de-icing, braking performance (for an abort), and of course, density altitude (air pressure/temperature).

Tony
ATP

 

[TonyWilliams]

Some of the newest, fanciest jets with top-of-the-line Flight Management Systems and Air Data Computers might also factor in humidity but for most pilots, it is not a factor that is considered.

I've flown a fancy jet or two, and I'm not aware of any that have that capability.

 

[sparky]

More on my modeling approach:

I plan to use the method described in my previous post to collect pairs of data (Fd, Veq),
where Fd is the total drag and Veq is the equilibrium zero-power speed for descending the given grade.
I will fit this data to the expression:
Fd = Crr + A*v + {Cd * A * 1/2 * rho * V^2}
I will use least-squares find the best fit of parameter A and Crr. I have confidence that the aerodynamic drag as computed by Sparky is accurate, so it has no parameter in this fit.

Interesting approach. I'll be curious to see what numbers you get.
A couple of thoughts:

While I agree that derivatives can be noisy, I'd be concerned about relatively large errors coming from the uncertainty of the grade.
It seems that error, (a 10% error on a 5km 2% grade can come from a starting ending altitude error of only 10 m) could easily
exceed my dv/dt noise. Plus, your least squares fitting; like my slope regression, will add noise. Plus, tail winds one way or the other.

I'm not following your equation: Fd = Crr + A*v + {Cd * A * 1/2 * rho * V^2}

I assume Fd is kg*m/s^2, Crr is unit-less, and I don't know what A*v is; m^2 * m/s ? Those don't add up to kg/m^2.

Perhaps SOC is not precise enough either. Here's a time series plot of the ( 0x5BC ) field we use for SOC vs an Amp-hours series I derived
from CANbus amps x time. You can see the SOC is kind of coarse.
This came from a recent drive of several miles down and up a 2-3% grade.

 

[TomT]

I believe that the A380 ADC can use humidity. I was reading the ops manual some time back and it was mentioned as an input...

Some of the newest, fanciest jets with top-of-the-line Flight Management Systems and Air Data Computers might also factor in humidity but for most pilots, it is not a factor that is considered.

I've flown a fancy jet or two, and I'm not aware of any that have that capability.

 

[surfingslovak]

Perhaps SOC is not precise enough either. Here's a time series plot of the ( 0x5BC ) field we use for SOC vs an Amp-hours series I derived
from CANbus amps x time. You can see the SOC is kind of coarse.

Interesting topic. What caught my attention was your note that you interpret each SOC tick as 250mAh. I assumed that each SOC point was 75Wh, and from a few trips reports on the forum I was able to confirm that it's in that ballpark (70 to 72.5Wh). Would you have any more data that would support this assumption?

On a related note, I speculated that each battery gauge bar consists of 20 SOC points, and started to use that for range predictions. Several times a day I would catch myself calculating my range while the guessometer was showing the same or very similar number. You can probably imagine how that surprised me.

 

[sparky]

Interesting topic. What caught my attention was your note that you interpret each SOC tick as 250mAh. I assumed that each SOC point was 75 Wh, and based on the a few trips reports confirmed that it's in that ballpark (70 to 72.5 Wh). Would you have any more data that would support this assumption?

Not really. It's a guess, nothing more. Seems just as plausible as kWh. Battery cells are typically rated in Ah, not kWh. Amp-hours will be proportional to kWh since the pack voltage stays pretty constant. To measure kWh requires combining three sensors (V, A, time) whereas Ah, just two. But, my estimate of 250Ah per SOC count is just a play on the binary numerics. 0.25 decimal is 0.01 binary. 75 doesn't come from a nice binary integer but of course neither does 280.

 

[surfingslovak]

Not really. It's a guess, nothing more. Seems just as plausible as kWh. Battery cells are typically rated in Ah, not kWh. Amp-hours will be proportional to kWh since the pack voltage stays pretty constant. To measure kWh requires combining three sensors (V, A, time) whereas Ah, just two. But, my estimate of 250Ah per SOC count is just a play on the binary numerics. 0.25 decimal is 0.01 binary. 75 doesn't come from a nice binary integer but of course neither does 280.

Interesting, thanks for responding. The 75Wh number originated from a similar train of thought. I figured that if the available pack capacity is represented by 280 SOC points, then the rated capacity must add up to something round as well. I took a few other shreds of data into account and thought that 320 would be a good number.

Given that the nominal voltage per module is listed as 7.5V, does this look right?

48 x 7.5V x 0.25Ah = 90Wh

 

[Herm]

Batteries work in Ah but motors work in Wh, it will be close since the voltage is relatively flat (due to the large size of the pack) and constant for lithium batteries.. but it would be better to use Wh.

 

[palmermd]

sorry if this has been answered in the thread somewhere, I did not read the whole thing, just the first post and some other random ones. Why not just to an A-B-A test with tire pressure. Run the 35psi and do 5 coast down tests from 60mph logging the rpm as you did earlier vs time, then change to 45psi and do the same thing, and then back to 35psi to verify it goes back to the original values. The difference in the coast-down should show if you have made any difference. I prefer empirical data to "calculated" results. I suspect the end results will be similar, but it would be a more definitive test.

 

[jumpjack]

Interesting but unusual coastdown plot, which probably highlights what discussed in subsequent posts: drivetrain friction is not negligible, but similar to air drag and friction roll in strength; else, the chart would look different.

Here are some literature data (not on Leaf):
V Vehicle 1 Vehicle 1b Vehicle 2 Vehicle 2b Vehicle 3 Vehicle 4 Vehicle 5 Vehicle 6 Vehicle 6
120 5,88 6,28 5,64 4,75 4,88 7,06 5,91 6,08 6,47 6,68
100 7,55 8,03 7,25 6,21 6,66 8,88 7,73 7,6 8,15 8,38
80 9,4 10,34 9,46 8,07 8,98 10,91 10,02 9,5 10,31 10,2
60 12,1 13,52 13,28 11,01 10,93 13,6 12,97 12,25 12,75 12,93
40 15,19 15,08 18,03 14,6 15,92 15,07 15,48 15,76 16,38 16,24
20 18,88 19,42 24,43 19,92 21,58 18,68 20,18 20,86 20,34 20,57
Mass kg 1492 1566 1316 1277 1124 1962 1652 1523 1634 1724

My ongoing study on Cd Crr extrapolation:

Experimental values:
0 19.44
10 16.76
20 14.31
30 12.08
40 10.35
50 8.92
60 7.55

(seconds, meters/seconds)