Page 30 of 37

Re: Nissan Engineering Team Visit Dec. 3rd: Recap

Posted: Sun Dec 04, 2011 4:27 pm
by garygid
Wanting a "better algorithm" for calculating DTE is asking for a better prediction of the FUTURE.

If you can do that, do not waste it on a GOM, but use it in the stock market, horse races, etc.

If anybody was suggesting that the "gid" could be an estimation of the Battery's Open Circuit Voltage (OCV), then ...
I suggest that PERHAPS the GID/2 (0 to 140.5) is an estimate of the remaining USABLE pack voltage (about 394v full and 254v "exhausted").
I will have to do some graphing of some logs to see if that scaling and "meaning" works.

However, I believe that the TRUE low limit for the Pack really should be based upon the voltage of the lowest cell-pair, NOT on the Pack Voltage (to avoid damage to that lowest cell-pair).

Re: Nissan Engineering Team Visit Dec. 3rd: Recap

Posted: Sun Dec 04, 2011 4:38 pm
by astrorob
garygid wrote:Wanting a "better algorithm" for calculating DTE is asking for a better prediction of the FUTURE.
yes, but if you have a route programmed into the nav computer and the nav computer knows the elevation change on the route, then it has some idea of the future.

that was one of the proposed ideas.

Re: Nissan Engineering Team Visit Dec. 3rd: Recap

Posted: Sun Dec 04, 2011 4:42 pm
by evnow
TonyWilliams wrote:Chief engineer: they monitor the 96 cell pairs: DTE = SOC x km/kWh
But, what kind of km/kWh does it use ? Last 5 miles ... last several miles with varying weightage ... ?

Re: Nissan Eng. Team Visit: Reservations Closed

Posted: Sun Dec 04, 2011 5:20 pm
by lemketron
astrorob wrote:perhaps rather than trying to nail an exact number they could give a best-case and worst-case estimate. i think they are always battling between confusing the ordinary user with a load of information and simplifying the information too much. they seem to be painfully aware that they have to present the information in such a way that it can be understood by the average person, and not an engineer EV enthusiast. my personal opinion is that they are going to try very hard to come in on the conservative side of any estimate. the worst thing that could happen is that they tell you 20 miles are left and there's only 5 in real life. their lawyers probably worry about what could come of consistently overestimating the range...
Yes, this (best/worst case range) is something I would prefer over a SOC meter. To me, a SOC (in gids or %) means just another number I have to mentally convert into "how far can I go" and as was pointed out at the meeting, it's also somewhat of an estimate.

The currently displayed DTE is always some number that lies within a range of "best case" and "worst case" values. I have to believe that they can also calculate (with reasonable accuracy) those best/worst case values within which they are currently displaying the DTE guess. The problem is that we can't tell whether the currently-displayed DTE is optimistic or pessimistic (cf: earlier "bipolar" discussion).

Perhaps some clever bit of UI/UX engineering could result in a display that shows both of the best/worst case values (as calculated by the car), with some visual indicator (font size, or some other emphasis) as to which one the car thinks is closer to the truth. Or it could just display them both evenly and let the driver decide, since (especially for routes that the driver is already intimately familiar with) we likely know better than the car, which end of the range is more appropriate.

The problem with the current DTE estimate, as well as the Leaf's 12-bar SOC, is that even if you know how far you are from home (in SOC or miles), the car really doesn't give you any way to tell whether or not you can make it since you don't know how the DTE was calculated (as was pointed out with a concrete example at the meeting by Arnold).

I believe the current gid/SOC-meter could be programmed to display these two numbers -- as "High.Low DTE" instead of the gid or percentage. I'd probably choose (from the range chart) the 7/bar and 4.5/bar columns for best/worst, so 100% would show "97.62", 80% would show "80.51", 51% would show "50.32", and 26% would show "19.11", etc. (of course, the high/low data columns could even be user-configurable).

Re: Nissan Engineering Team Visit Dec. 3rd: Recap

Posted: Sun Dec 04, 2011 5:27 pm
by lemketron
I don't know if it's been mentioned in other threads here on MNL, but someone at the BayLeaf meeting yesterday suggested a pretty interesting in-dash SOC alternative that we already have, but none of us had ever thought of...

The 120V "time to charge" display in the center of the dash. It counts up by half-hours needed to re-charge, meaning that it counts roughly in 2% units of SOC. Unfortunately it counts up rather than down, and if you have an 80% timer set, it only displays time to re-charge to 80%, and it's not clear exactly what the "max" time is. But once those values are known, you can watch (and do your mental math) to know roughly how much SOC remains. It's just too bad the existing 12-bar SOC isn't divided into more segments to give us this same level of granularity (without the math).

Anyway, several of us thought this was a very clever approach to using information that is already available.

Re: Nissan Engineering Team Visit Dec. 3rd: Recap

Posted: Sun Dec 04, 2011 5:39 pm
by astrorob
lemketron wrote:I don't know if it's been mentioned in other threads here on MNL, but someone at the BayLeaf meeting yesterday suggested a pretty interesting in-dash SOC alternative that we already have, but none of us had ever thought of...
huh, i have been using the 120V time to charge indicator as a rough "kw used" since i got the car. didn't realize that it was not a common idea. but i think it counts off in 10 minute increments - in fact just as i got home now i noticed that it was asking for 10 minutes at 120V to get back to 80%.

i wonder if the car is showing you true wall time with the charger efficiency factored in. i've just been using 1 hour = 1kw for the mental math. 1200w = 60 minutes might be just as easy though. somewhere between 200W and 166W used per 10 minutes at 120V.

it's all the wondering that bothers me. i wish nissan could just tell us crazies all about how these various things work and then we would just know.

Re: Nissan Engineering Team Visit Dec. 3rd: Recap

Posted: Sun Dec 04, 2011 5:59 pm
by garygid
IF
1. the Nav data included elevation (it does not even have all the STREETS).
2. the car measured the wind and rain drag, AND knew what they would be in the future.
3. the car could know about the traffic conditions AND how fast you would be driving IN THE FUTURE.
4. the "SOC" was reliable, especially at the low end.
5. your FUTURE usage of lights, accessories, Heater, etc. was known ...

THEN, it could do a better job estimating range ...

But, neither YOU nor IT can SEE the FUTURE.

Yes, an ACCURATE GOM, even any meaningful High and Low ranges (unless you think 150 and 0.1 are reasonable), are just not possible.

If you pick a mi/kWh (the way YOU expect to use energy), or the car uses the instantaneous mi/kWh (the way YOU are using energy right NOW), the DTE estimate can be meaningful.

Re: Nissan Engineering Team Visit Dec. 3rd: Recap

Posted: Sun Dec 04, 2011 6:07 pm
by mitch672
Speaking of SOC of the battery, what about "Coulomb Counting"?

http://www.mpoweruk.com/soc.htm" onclick="window.open(this.href);return false;

http://liionbms.com/php/wp_soc_estimate.php" onclick="window.open(this.href);return false;

MAXIM had an IC for this in 2000: http://www.maxim-ic.com/app-notes/index.mvp/id/485" onclick="window.open(this.href);return false;

This would probably be the most accurate..

Re: Nissan Engineering Team Visit Dec. 3rd: Recap

Posted: Sun Dec 04, 2011 6:25 pm
by evnow
garygid wrote:But, neither YOU nor IT can SEE the FUTURE.

Yes, an ACCURATE GOM, even any meaningful High and Low ranges (unless you think 150 and 0.1 are reasonable), are just not possible.
As I said perfect shouldn't be made the enemy of good. Just because DTE can't be perfect doesn't mean it can't get better.

If Leaf just took account of (having given a destination)
- elevation changes along the route
- Max speed along the route
- Current temperature

It can still do a better job than what it does now. More over, when the speed of the car or elevation changes in the future, the car won't go into some bi-polar extremes in DTE.

For widespread adoption of EVs, we need
- Better DTE or
- Much higher range at same price

I think better DTE is easier to get than much higher range at same price.

Afterall, there are less variables than something like future climate forecast.

Re: Nissan Engineering Team Visit Dec. 3rd: Recap

Posted: Sun Dec 04, 2011 6:30 pm
by garygid
What DTE does a typical ICE require to be "mainstream"?