Forum

Unisciti allacomunità

Accurate battery status instead of single number for range

Proposal

Rather than display a single range estimate, display an accurate indication of the state of charge remaining and the ranges that result based on a variety of possible conditions. For example, (using U.S. units)

42.0 kWh remaining.

Consumption over last 10 minutes (average speed Z mi/h): x mi/kWh -> range X mi
Expected consumption at 55 mph: y mi/kWh -> range Y1 mi with climate control on, Y2 mi with climate control off
Expected consumption at 65 mph: z mi/kWh -> range Z1 mi with climate control on, Z2 mi with climate control off

etc. The question is how best to present this information in an easily digested form. If the destination is entered, the car can calculate the route and highlight in red any of the range numbers that are lower than the required range, thus showing you that you really can't get there if the destination is 1000 feet up a mountain and you average 75 mph with the climate control on, say.

What do you think? Does anyone want to work up a suggested display? Do you agree that the single range number estimate is a problem?

Motivation

A lot of anti-EV press is based on the premise that you can't trust the car's range indicator. Some writers take this simple observation to ridiculous conclusions such as EV manufacturers lying about range, that EVs really don't work, and so on.

Think about the problem for a moment. Users are expecting a single number to give them accurate information about the future. This is simply impossible. Your range depends on the energy you have available in the battery and your power consumption. The latter depends on your speed, on your use of heating or cooling, on elevation change, on the weight of the car, on headwind vs tailwind, and a bunch of smaller factors. There is no way to accurately predict all of this, though you hypothetically could come close if you assume the driver goes the speed limit and you have extensive geographical data.

When you get in an ICE car, do you see a range number anywhere? No, you see a fuel gauge that tells you roughly how much energy you have left.

Now, what happens in the world of aviation? Pilots are required by law to compute fuel requirements prior to takeoff, taking into consideration the expected cruising altitude and winds at that altitude. They also have to allow for the possibility of a missed approach at destination and a flight from there to their alternate. And even with all that, the number is padded (there is a reserve) to allow for conditions not being as predicted. Further, if anything goes wrong during the flight, the pilot may need to make a decision, instantly, to alter the flight plan and land short of the destination to refuel.

That is reality. The exact same thing applies to ICE cars, it's just that "landing short" is usually not a problem because of the ubiquity of gas stations, and the speed at which a "recharge" is possible.

People are lazy - they want so much range that they don't ever have to think about it. Catering to that laziness by providing one guesstimate of range is a strategy which I think is backfiring. How can people trust a number that changes to be completely unrealistic when going up or down a hill, or as soon as you take the car out of the garage into a really cold, snowy environment? They can't. And those who, for whatever reason, want EVs to fail, take advantage of this issue.

I really like the proposal and would put it near the top of my list of enhancements. Since the solution is essentially software, it should not be too difficult to implement. What's the process for getting it on TM's enhancement plan?

I understand your reasoning, but I respectfully believe that this is unnecessary (for me, at least).

Everyone drives differently, and every owner learns quickly what the range number means to them. If someone unfamiliar with the car (i.e., an idiot reporter) decides that the range number is somehow a contract with TM and proceeds to drive right through to zero, then no amount of data displayed will prevent idiots from being themselves.

TM is going to sell a million of these cars, and for a myriad of reasons. The 'range anxiety' fear experienced by fence-sitters will be overcome by the thundering herd of rabid fan/owners (myself included) who view this as a non-issue.

Tesla should certainly add smart assistance to the car. It would be great if the car informed users when their navigation destination was out of mileage range.

Also, Tesla should add locations of all known charging locations in your area to assist novice drivers. It's seriously naive to expect every driver to understand how to survive in an EV. Blame the driver all you want but situations like with NY Times will become the norm more than the exception as EVs become more mainstream.

What would work best is to tie in the weather forcast (including wind speed and direction along with expected temps) and gps of trip with elevation info.
Not sure when or if the elevation info will be accessable to google though.
If it is then with this info and the driving habbits obtained from past driving along with user selectable fast normal and slow, the computer on the car could calculate quite accurately the range.
However one would be required to log in the destination to achieve the most accurate estimate.
Just a bit of software! An app coming soon? Work for someone here.

What I would really like to see; on the map the location of the charging stations, speed, price, and busy/free status on my dash for stations in the direction I'm traveling, or perhaps in any direction around the car. Since the car knows it's potential range, it should be able to draw a circle around the present location, size depending on the available range. At any given time, I really only need to be aware of stations inside that circle since it's within my range. As I drive, the stations come into my range, and as I drive, my circle gets smaller as my range decreaces.

Saves all the hard work about using campground maps, chargepoint search, blink search, and the supercharger map, then doing the math to calculate range. If i could SEE my range, and where the charging stations are, how could I be anxious about my range?

@Hogfighter

Let me guess... You live in California?

Actually, except for the kwhs remaining, a lot of this information is available on the trip page. My ice car won't tell me how many gallons remain, but it estimates remaining range. I just drove 40 miles on 15 kwhrs. Not great, but not bad either. I have my "gas gauge" as well. I'm not a fan of wh per mile. That's a little like cups of gas per mile. I would rather see miles per kwhr displayed. Until then i can divide in my head or maybe there will be an app...

Electric Avenue, this is exactly the kind of proposal that would help the entire situation. In fact, I made almost exactly the same proposal a few minutes ago on the NY Times web site and have suggested the same in these forums.

The airplane analogy is a spot-on. The fact is that an EV faces a bigger penalty for running low on charge than does an ICE (though certainly not as severe as an airplane).

To the extent we can use smart software to to avoid problems in these early years Tesla owners will do better.

I have been shocked by owners who have rejected any proposal to help alleviate range anxiety.

@DanD I agree! It's arrogance, plain and simple. Smart range software for EVs is what's gonna help EV driving break into mainstream. It's purely a software problem so I have no doubt Tesla will address it eventually.

The problem with the number is that it says "miles" but it doesn't mean miles. I would prefer to see usable kWh as the measure, and then I can calculate the range. I would also prefer that we use the convention of miles per kWh rather than watt-hours per mile, since that is what I am accustomed to. I typically get about 2.5 miles/kWh (400 watt-hours/mile), but I know how to adjust that based on various conditions. If my energy gauge says I have 20 kWh remaining, and conditions are typical, it's easy to calculate the range as 2.5 x 20 =50 miles.

A possibly stupid question: But on Energy screen to right of graph, isn't the 'Projected' miles exactly what you are talking about here? If you toggle on Average, then set option for last 5, 15 or 30 miles of usage to project remaining miles based off remianing charge.

Of course, the current glitch is that the car doesn't remember to use Average mode, and reverts to Instant each time you restart. Also, it would be nice to put that (remembered mode) up in main display instead of 'rated' miles. Both relatively simple software fixes, and directly comparable (and probably much more accurate) to the Distance-to-Empty displays most modern ICE cars have.

@mfelisky That's a godo point. i think they should put more emphasis on projected miles instead of rated miles. The default UI should focus more on how I am driving rather than the standard, hard to achieve scenario that rated miles are based on.

Hi all. Thanks for the feedback so far.

I wasn't thinking when I wrote kWh/mi, it should have read mi/kWh. Sorry about that. I have edited the original text, as I think that's probably less confusing than just noting the mistake here.

mfelisky: The screen you describe is close to what I'm talking about, but what I'm trying to get across is the importance of not highlighting a single mileage number, or at least putting a big bloody asterisk beside it to indicate which calculation was used to arrive at that number. That "miles remaining" display right in front of the driver needs to be based on a calculation that the driver selects. Or maybe it shouldn't be there at all (point of discussion). The driver knows better what the future is likely to hold, and it is the driver's responsibility to make that call. Rather than the car saying "in the past you've been consuming X average power, therefore you have Y range remaining (assuming the future is the same as the past)", the car would say "your current energy remaining is A, you have told me you expect from this point on to be consuming B average power, therefore you can go C distance."

To go back to the aviation analogy: "That's time in your tanks." When flying an airplane, you know your fuel flow either from a gauge or from a table that maps power setting to fuel flow. You also know, again either from a gauge or from calculation, how much fuel you have. This gives you the time you have remaining until you run out of fuel. If you're encountering a headwind instead of an expected tailwind, for example, your range will be lower (possibly much lower) than expected, but you know this by simply multiplying your ground speed by the time you have left. If you let the computer do the calculation based on what happened in the past, it might be optimistic as it's based on winds that applied in the past.

I am suggesting giving the driver the responsibility for estimating their power consumption, based on what they know, rather than trying to have software just guess what it might be. It really is the driver's responsibility to do this. Pilots don't get to say "well, I would have had enough fuel if the winds hadn't changed, it's the plane's fault" after their plane crashes in a field, so why do EV drivers get to blame the car when they are stranded because they are relying on an inaccurate range calculation?

Just found out that google maps does give elevations for any piont on earth.
This means that all the data is available to do an accurate estimate of power usage. It is just a good piece of programing away. The cars computer has considerable processing power so I dont see problems for number crunching in doing this. See my last posting for what this is about.

On trips with our ICEs, when the gas level gets down to 1/4 full, I start looking for a place to fill up. I would prefer a simple equivalent, say KW remaining, or 1/4 of KW capacity remaining. I don't need all of the arcane miles/time predictions to know that it is time to look for a charging station.

"so why do EV drivers get to blame the car"

... by "EV drivers" I am including anyone in the media who are predisposed to slag off electric cars as unreliable and not ready for prime time. :-)

I realize that probably most Model S owners know not to trust the range calculation in certain circumstances. Wouldn't it be better to be able to trust the range calculation because you know it should be relatively accurate, rather than messing around in your head with "well, usually the range is off by about 10% when ..."

mallynb: the difference is the time to refuel vs the time to recharge. I agree with what you say in principle, and I think there should be a display of kWh remaining in big numbers right in front of the driver, but unless we're talking about a world where Superchargers are within range everywhere, stopping to charge involves a major time delay. It is more important in an EV to be confident of reaching your destination, and it is likely always going to remain thus. There is no way around making the calculation, unless you know that you have so much energy compared to how much you'll need that it can't possibly be an issue.

I'm a big fan of EVs, but the reality of longer charging times can't be waived away.

To @mallynb's question: doesn't the "battery bar" turn yellow at some low SoC, and then red at a lower SoC?
Of course, on a road trip, due to current charging infrastructure, when you get to yellow SoC, you might not have enough juice to get somewhere with a sufficiently powerful EVSE. That's why we early adopters need to plan ahead.

I drove a LEAF for 14 months, and I found I completely ignored the projected miles, because it was based on the past and not on the future (which it obviously can't know) and changed constantly. Instead, I looked at the 12 charge bars, and knew how many mile I could get per bar for different ways I drive (surface, slow on the interstate, fast on the interstate).

I much prefer the rated miles display in the Model S. It still isn't accurate for the future, but it is close enough to typical that it means something rather than a number that jumps all over the place. For example, on the way home tonight I was in a traffic jam for an accident, and the 5mi-average projected range was showing me 26mi, while the rated miles was 196 -- a big difference.

After you drive an EV for a while, you just don't worry about it except in the rare cases when you need to, and you instinctively know when those are. I don't think showing a lot of different numbers for different scenarios will help, and in fact I think it will confuse people more who are prone to range anxiety -- the man with one watch always knows what time it is, but the man with two is never sure.

@mallynb - then you should be happy, you already have that in the charge gauge just above the rated miles.

@ElectricAvenue - great minds think alike. I just posted a similar suggestion here: http://www.teslamotors.com/forum/forums/43-scheduled-charging-does-anyon...

Sightly off topic, but related. I remember someone posting a link to a web site with an EV charge usage estimator. The site allowed you to pick a start, destination, and type of car (Yes Model S was in the list). Assuming you went the speed limit and followed the NAV route it would calculate % of charge required for the trip, taking into account terrain. I tried and can't find the post. Did someone save the link?

If I could build an "app" for the Model S it would be this.

You can get elevations (and to some extent traffic) from Google as has been mentioned. That combined with either a weather forecast or an input weather assumption from the user, an input assumption on driving style (i.e. cruise control @ 60mph) or default to past style, a weight input (or does the car have a sensor) and the battery level (Tesla needs to expose the API) AND presto, problem solved.

Add integration with Chargepoint, the Supercharger Locations maybe others and you've got a car that will keep you driving.

Adding Superchargers to the system would be gravy.

Of course, I'd need access to the formulas that Tesla has to predict mileage based on temp, incline, weight, speed, etc. What a fun programming challenge.

To make it even better, the software learns over time to improve on the engineering formulas.

nickjhowe: Oooo, I like it. :-)

I had an idea regarding what could be displayed in the small display in front of the driver: imagine the current graphic indicating proportion of battery charge, but it says "42.0 kWh" in big letters and below it, somehow

Destination 122.2 miles
Need to beat 2.9 mi/kWh
Last 10 min: 4.2 mi/kWh

Or make the efficiency figures be needles on a circular dial with the difference shown either in red or green.

On the big screen: the energy usage graph would have a horizontal line showing the required efficiency to reach the destination or next charging point (that you select), and a horizontal line showing last 10 minutes or whatever average or expected usage you want to show, with an indication in green if what you're getting is above what you need, and in red if what you're getting is below what you need.

I have to say it is a difficult engineering problem to determine the SoC of a battery that has been sitting in the cold. Power is measured by multiplying current with voltage. The problem with the battery chemistry is that voltage fluctuates with temperature. So you can easily tell how much energy you put in and how much energy you took out but can't tell how much energy is stored. That has to be calculated after pulling a bit off the battery. I drive with ideal range displayed and treat it like a gas tank knowing the behavior of the car. There is a big difference from the owners who have considerable mileage on their vehicles and low mileage owners. After just 1000 miles, I am very comfortable with my vehicle for my transportation needs.

@dtesla wrote:
Sightly off topic, but related. I remember someone posting a link to a web site with an EV charge usage estimator. The site allowed you to pick a start, destination, and type of car (Yes Model S was in the list). Assuming you went the speed limit and followed the NAV route it would calculate % of charge required for the trip, taking into account terrain. I tried and can't find the post. Did someone save the link?

http://www.jurassictest.ch/GR/

rgds
Norbert