GOM range algorithm

[Blue20]

It is suggested that the GOM uses the last 18 miles to calculate the range. This certainly doesn't work for me. Does the 18-mile algorithm work for anyone?

As an example, I travelled 32 km (19.9 miles) since the last full charge. It is now at 78%. If the calculation is based on last 18 miles, my remaining should be 113 km (78 miles). The GOM and phone app are showing, I have 78 km (48 miles) remaining. My "All trips consumption" is 4.5 mi/kWh (13.8 kWh/100 km)

 

[Jeffj]

It doesn't just use the average from the last 18 miles, it uses samples from that period. Also, the outside temp, the inside temp, the heating/cooling load and the terrain (if you have a destination set) are all factored in.

In my own experience, I can generally explain whether its high or low when I think back on what the last 18 miles were vs. what I'm currently doing. So, yes, it sort of works for me. But realistically, I use the blue bars and/or the SOC display to figure out how much further I can go.

 

[jadnashuanh]

Well, maybe more accurate would be to say that the car uses your driving style for the last 18-miles to figure out the power use, and extrapolates that to how long before (given the same conditions) you'd discharge the battery. This could be way off if you started out on the highway, and are now on low-speed roads, or vice-versa. Or, if you were sitting in traffic with lots of stop and go, and now you're out in the country. The car isn't omniscient! It only knows how it was being driven, not how it is going to be, or where (well, if you were following a route in the nav system, it takes some of that into account since in many areas, that has grade information along with speed limits).

 

[Jeffj]

If you have a destination set in the Nav, it can make an enormous difference in what hte GOM displays. Last weekend I drove up into the mountains in Colorado and while I had 100+ miles on the GOM when I started (I had been driving pretty slow running errands the night before), as soon as I put in the destination (which included > 10,000 ft of climbing and about 4,000 ft. of descending), the GOM dropped to 52 miles, as it was taking the first part of the climb into account. Clearing the destination brought the GOM back up to 100.

 

[I33t]

I think it needs improvement...

In my case, I live up a 900ft hill, so when I head off in the morning it always tells me I have low range. On the way back home it tells me I have amazing range because a substantial part of the trip has been coasting downhill on regen.

Maybe there should be some selectable options for typical routine situations that would help the GOM do better, like living on a hill or living in a valley, or living on the flat.

 

[epirali]

If you have a destination set in the Nav, it can make an enormous difference in what hte GOM displays. Last weekend I drove up into the mountains in Colorado and while I had 100+ miles on the GOM when I started (I had been driving pretty slow running errands the night before), as soon as I put in the destination (which included > 10,000 ft of climbing and about 4,000 ft. of descending), the GOM dropped to 52 miles, as it was taking the first part of the climb into account. Clearing the destination brought the GOM back up to 100.

That is interesting and makes perfect sense! I didn't even think of it. I'll definitely have to try it. I know on the i8 it actually uses terrain and nav route to make drivetrain decisions. I wonder if i3 does something similar with regen. It would be great if they used the info for the Rex version to engage Rex depending on route to maximize range.

 

[jadnashuanh]

That is interesting and makes perfect sense! I didn't even think of it. I'll definitely have to try it. I know on the i8 it actually uses terrain and nav route to make drivetrain decisions. I wonder if i3 does something similar with regen. It would be great if they used the info for the Rex version to engage Rex depending on route to maximize range.

The i3 DOES use terrain information when you put in a route. And, in an upcoming software update, when you have the REx, it will turn it on earlier if it knows you are going to experience elevation changes big enough to potentially impact your operation. They had hoped it would be ready this spring, but those testing it made additional requests and comments, and it went back for more work.

Unless you tell the car where you're going, it has no way to know, and can only base future range estimates on what you have done in the past. So, if you want it to be more accurate, put in a route, even if you know exactly where you're going and how to get there. Save them, and they are really quick to engage. You might also find an alternate way to go that will increase your range by playing with some of the options.

 

[Blue20]

I tied setting destination when I went home from work yesterday. The distance is about 15 km (10 miles). The range did change, but only by 1 km.

It seems GOM should be used as a tool to check if you can get to your destination, rather than using it to project how far you can go.

 

[jadnashuanh]

1/15=6.67%, in reality, that's a lot!

 

[NHi3REx]

Has anyone else noticed that the GOM also appears to account for past driving habits for a given day (ie weekends versus weekdays)?

I bring this up because, I have noticed that for five days a week (M-F) when I take the car to work (mostly highway), the GOM predicts a much lower expected range (70-80mi). However, on Saturday mornings when I take my daughter to dance class (only a handful of miles in town), the GOM consistently reads a substantially longer range (90-100+ mi).

If the algorithm only took into account the last ~18mi driven and other current/recent conditions, I would expect to see the same expected range on Saturday mornings that I see on T-F based on the previous day's commute. It's almost as if the algorithm is also trying to predict what my driving pattern is going to be for that given day.

BTW, I never program my route, so the algorithm has no knowledge of where I'm ever going.

 

[epirali]

Has anyone else noticed that the GOM also appears to account for past driving habits for a given day (ie weekends versus weekdays)?

I bring this up because, I have noticed that for five days a week (M-F) when I take the car to work (mostly highway), the GOM predicts a much lower expected range (70-80mi). However, on Saturday mornings when I take my daughter to dance class (only a handful of miles in town), the GOM consistently reads a substantially longer range (90-100+ mi).

If the algorithm only took into account the last ~18mi driven and other current/recent conditions, I would expect to see the same expected range on Saturday mornings that I see on T-F based on the previous day's commute. It's almost as if the algorithm is also trying to predict what my driving pattern is going to be for that given day.

BTW, I never program my route, so the algorithm has no knowledge of where I'm ever going.

That is a very interesting observation. I have noticed that the reported range will JUMP just by calendar time changing (but the car hasn't been driven). If your observation is correct it would explain it, as the day changes it adjusts to the learned pattern?

 

[jadnashuanh]

Are you leaving later in the morning and it's warmer out? Battery temp has a part in max range.

 

[epirali]

Are you leaving later in the morning and it's warmer out? Battery temp has a part in max range.

Not me, I am pretty data detailed and the jumps are large and not temperature related. And way too much considering the thermal mass of the battery. My best guess is simple calender impact, just time going by. So either data is aged out by time or as poster mentioned the car may be doing a smarter average by day to capture typical patterns.

 

[NHi3REx]

My departure times are pretty much the same (give or take 30min), and I agree with epirali, the jumps are too big to be temperature related (outside of the winter months with and without battery preconditioning active).

 

[RJSATLBA]

If you have set a destination in the satnav the range prediction is significantly lower when motorway is involved.

 

[jadnashuanh]

When you have a route activated, the terrain and speed limit information is taken into account. Otherwise, it uses averages, which may not be realistic for your actual intended trip. FWIW, the car assumes you'll drive it like you did the last 18-miles or so and has no way to know otherwise unless you tell it where you are going.

 

[NHi3REx]

I never program my route or really even use the sat nav at all. So the car should have no knowledge that I would not be simply driving back to work on a Saturday morning unless it recognizes, learned, and remembers that on Saturdays, my driving is traditionally limited to just around town.