Sponsored

NewsCoulomb

Member
Joined
Jun 24, 2021
Threads
0
Messages
21
Reaction score
5
Location
California
Vehicles
Bolt EV, Ranger Electric
I do understand. You think ABRP's model is incorrect and that, therefore, the program can't be used.

The fact that you even expect it to be "accurate" means that you aren't quite up to speed on how it works. You cannot expect it to be accurate. If the model is good it will produce a plan representative of what might be expected under nominal conditions.


Again your saying that conditions exactly match the base line is, forgive me, naive, because they can never do so. Do you drive at constant speed? Is that the speed you put into ABRP? Is it dead calm throughout the entire trip?


This says you are really missing the concept so let's see if maybe a bit of examination of an ABRP session will help you grasp it. Before I launch into that I will point out that I do have a fair amount of experience with optimization and estimation and I find it useful as do thousands of others so either you are missing something or the rest of us are wrong.


Let's get this out of the way first as it probably a good place to start. The reason the final SoC is 10% is because YOU told the program you want to arrive at the destination. It will, therefore, optimize over the route parameters in order to get you there as fast as possible subject to the constraint that you arrive with 10%. As I gather the route between these two places is Veterans/Steens highway or back roads it clearly has to pick that highway for best time. What does that allow it to adjust? The charge time at Winemucca. It will tell you to charge for as short a period as you can and still be pretty sure (its an estimate) of 10% charge at the destination.

So I loaded this route into ABRP but used Rivian parameters rather than Bolt. The Bolt model is probably better than the Rivian model because there are real cars out there. ABRPs model represents guesses. It tells me that I will drive 2:18 to arrive at Winemucca with 39%, charge there for 13 minutes to 65%m drive another 3:24 to the destination arriving with 10% left in my battery. Some assumptions went into this:
a)400 Wh/mi consumption at 65 mph
b)Leave with 85% SoC on the battery
c)Drive at the speed limit
d)No wind.
e)I want to get there as quickly as possible under these constraints.

Now it is true that if someone in Reno says "Hey, I want you to take the R1T up to Burns tomorrow" and I say "How far is that" and he says about 400 miles I know immediately that I'm going to use about .4*400 =160 kWh and that with a 170 kWh battery I am going to need to charge once on the way. But when I put the trip into ABRP I know that I will be using a bit more than 400 Wh/mi because of terrain and or speeds above 65 mph and that, given the speed limits, it will take me about 5:44 of driving and 13 min of charging and that Winemucca is a good place to charge because it has big chargers. I also see right away that I have a choice of Roberto's only 801 ft. from the chargers, Dos Amigos 1127 feet away and 4 other choices. I also see right away that if I am feeling lucky and have a couple of hours to kill I can run up to the Say When Bar and Casino for a bit while I charge at the 40 kW station there.

I now know lots more about the road from Reno to Burns than I did 10 minutes ago. And I am guessing that all this information is accurate.

Now as for learning something about my vehicle and the trip let's play some games. If I have to charge 13 minutes leaving it with 85% on board could I not save some time if I left with 95%? Indeed I could. If I enter that into ABRP then I'd only have to charge 11 minutes. Suppose I decide to take advantage of that by driving faster say 110% of the speed limit then I'll use more energy (471 Wh/mi) and I'll have to charge 15 minutes but I'll be 5:15 on the road for a total of 5:30 for the trip as opposed to 5:51 at the slower speed and departure at 95%.

I've now got a pretty good picture of what the morrow will bring. Will I arrive at Winemucca after 2:18 with 39% on the battery? Very probably not. A head or tail wind will probably get me there earlier or later and the consumption will be more or less than 400 Wh/mi. But the number is going to be close to 39%. Will I leave with 65%? No, probably not. I'll dawdle over my MicyD's and/or decide to put in 70 or 75% for some margin at the destination. Will I arrive at the destination with 10% SoC? No, it will be more than that. Has ABRP been a big help to me in planning this trip? Yes. Could someone using ABRP learn something from an exercise like this one? We hope so.
Thank you for illustrating my point. If ABRP was a teaching tool, it would not require years of experience in "optimization and estimation" and a lengthy explanation in order to use effectively, would it? And again, that is one of my key issues. It's set up to be easy to use as though their data is accurate and up to date. For some EVs and charging networks, it appears to be. For others, not so much.

Obviously, no trip is going to perfectly match ABRP's baseline numbers for typical driving (that would be impossible); however, if the conditions are reasonably similar, then one can expect the real-world results to be reasonably similar. But they aren't. Sure, maybe I stated that I wanted to arrive at ~10%, but how (exactly) do I leave Reno with 100% and somehow magically use exactly 90% to arrive at the specific charger I need when the next nearest charger is 50 miles away? Did I just magically plug in a route where the first leg required exactly 90%? Something's wrong with ABRP's calculation.

I can tell from experience with the 2020 Bolt EV that, under typical driving (what should be ABRP's baseline), I would arrive at Winnemucca with ~25% (approximately 15% more than ABRP's projection). That's too large of a margin of error for a tool that is supposed to be providing estimates on typical conditions and driving.

So over to you. What's wrong with the picture below?

ABRP.png
As for what's specifically wrong with that picture, I'd say pretty much everything. If I were "planning" that route for myself, first, I'd complain because nobody has chargers in Susanville or Alturas (Highway 395 is my preferred route as I'd typically be coming from west of Reno). Then I'd open PlugShare. Based on what I saw, I'd plan a trip in a 2020 Chevrolet Bolt EV accordingly:
  1. Leave Reno with a full battery.
  2. Arrive at Winnemucca Electrify America with 20% to 30% battery (possibly having to adjust along the way for conditions). That's what can be expected at baseline, freeway efficiency in average conditions.
  3. I'd charge up as long as the stop made sense. There's not much to do at that location (nor are there places I'd like to eat), so other than walking over to LUCID Cannabis and being disappointed it wasn't an EV showroom, I probably wouldn't want to spend more than 30 minutes charging.
  4. As a result, I'd expect to leave with between 60% and 70% battery, which is not enough to make the remaining 222 miles at posted speeds.
  5. I refuse to drive less than the posted speed unless it's an emergency or weather dictates, so I'd jam it up Highway 95 at 70+ mph, leaving me 70 miles short of making it to Burns.
  6. After about an hour, I'd stop a Quinn River Market in McDermitt, NV (with what I'd estimate to be about 30% battery) and charge up on one of the two Greenlots 50 kW BTC Power Slim units that ABRP never even presented to me as an option when attempting to use it to plan the route.
  7. I'd probably charge just a little bit extra in case, and after 30 to 40 minutes, I'd expect to leave with close to 80%.
  8. I'd drive the speed limit to Burns and arrive with 10% to 15% battery.
Mind you, my cheeky response took about five times longer to type out than this "plan" took to formulate. It's entirely possible that, while I was driving, I might decide to stop for a quick top up at Orovada along the way (another Greenlots charging site that ABRP didn't even bother mention as an option). Of course, it's also entirely possible that I really liked the LUCID store and decided to charge to near full while hanging out in Winnemucca, at which point, because I checked the route ahead of time, I'd still drive with the flow of traffic and stop at McDermitt (though I'd spend substantially less time charging).

So again, bringing it back to Rivian... The only thing that would change with this route plan is that I'd have an additional option for charging at a RAN station in Winnemucca. I don't know what amenities would be nearby, so I don't know how long I'd want to charge there. I also haven't seen data on how long the 180 kWh R1T takes to charge to full, but 80% still might not be enough for me to comfortably drive 220 miles at 70+ mph. Because RAN doesn't appear to have plans for Highway 95 north of Winnemucca, I'd hope that (unlike ABRP) Rivian includes those Greenlots chargers on their navigation and route planning.

And again, I can now illustrate my other issue with ABRP. I'd be new to the R1T (not experienced like I am with the Bolt EV), so I'd be relying on ABRP's baseline numbers to be somewhat accurate. A 15% margin of error could easily leave me stranded in the desert, and caveat emptor or "you should have slowed down" on one of the fastest driven highways in the country won't cut it.
Sponsored

 

ajdelange

Well-Known Member
First Name
A. J.
Joined
Aug 1, 2019
Threads
9
Messages
2,883
Reaction score
2,317
Location
Virginia/Quebec
Vehicles
Tesla XLR+2019, Lexus, Landcruiser, R1T
Occupation
EE Retired
Thank you for illustrating my point. If ABRP was a teaching tool, it would not require years of experience in "optimization and estimation" and a lengthy explanation in order to use effectively, would it?
No, and ordinarily it doesn't. Engineers, mathematicians, physicists and others in the sciences have what the medical profession calls "the knack" This is a "... a rare condition characterized by inherent understanding of all things electrical and mechanical and utter social ineptitude." At least that's how the medics in Dilbert's world describe it. There is another end to the spectrum. It's sort of like full color blindness. Folks at that end, while otherwise being fully intelligent and capable people, have no intuition whatsoever for anything quantitative. It's not that rare. I live with such a person. You are apparently such a person and it, therefore, takes a lengthy explanation to try to get you to understand what the program does, is intended to do, and its limitations.

And again, that is one of my key issues. It's set up to be easy to use as though their data is accurate and up to date. For some EVs and charging networks, it appears to be. For others, not so much.
It's entirely possible, in fact it is certain, that it is more "up to date" on Tesla vehicles (with the exception of the CT) than it is on Rivian or Lucid or Fisker as Telsa's are out in the real world. Tesla also would be at an advantage as the people who developed the program originally developed it for Tesla and the Tesla model is more refined.


Obviously, no trip is going to perfectly match ABRP's baseline numbers for typical driving (that would be impossible); however, if the conditions are reasonably similar, then one can expect the real-world results to be reasonably similar. But they aren't.
But they are. Thousands of users know that based on their experiences. You think they aren't because you haven't the perspective to interpret what the program is telling you.



Sure, maybe I stated that I wanted to arrive at ~10%, but how (exactly) do I leave Reno with 100% and somehow magically use exactly 90% to arrive at the specific charger I need when the next nearest charger is 50 miles away?
You don't. You told the program you want to arrive at chargers with a minimum of 10%. It will adjust the trip parameters to get you there with about 10%.
Did I just magically plug in a route where the first leg required exactly 90%? Something's wrong with ABRP's calculation.
No. That's just about how much energy it takes to cover that terrain at the speed limit given the estimated consumption of the car.

I can tell from experience with the 2020 Bolt EV that, under typical driving (what should be ABRP's baseline), I would arrive at Winnemucca with ~25% (approximately 15% more than ABRP's projection). That's too large of a margin of error for a tool that is supposed to be providing estimates on typical conditions and driving.
You make the common error of assuming you represent the population or that the population should conform to your way of doing things. In fact ABRP's baseline should represent the population. So what it does is use consumption numbers representative of the population (i.e. a mean or median). ABRP evidently puts some margin into the consumption figure to protect neophytes from running out of gas. If you know, based on experience, that your consumption is less than the ABRP default, use the number that applies to your personal driving habits. I do that as would any prudent user. You would probably argue that you shouldn't have to do that. In fact you did. You clearly stated that ABRP should reflect your experience.




As for what's specifically wrong with that picture, I'd say pretty much everything. If I were "planning" that route for myself, first, I'd complain ....
So there is, in fact, nothing wrong with that picture. You list a whole lot of things you would, or might do differently based on your experience with that route. In my last post I showed how proper use of ABRP reveals that many of those options are possible. That is the value of it to me, or anyone else, who has no familiarity with that route.




So again, bringing it back to Rivian... The only thing that would change with this route plan is that I'd have an additional option for charging at a RAN station in Winnemucca. I don't know what amenities would be nearby, so I don't know how long I'd want to charge there. I also haven't seen data on how long the 180 kWh R1T takes to charge to full, but 80% still might not be enough for me to comfortably drive 220 miles at 70+ mph.
This is an excellent example of the power of ABRP. In terms of planning it doesn't make much difference if you use the Winnemucca EA or RAN charger which is only a couple of miles away. As soon as the RAN go on line ABRP will pick them up of course. ABRP does give one the option to prefer or abhor a particular brand of charger so users will have the ability to favor one or the other if they wish. More significant here is how the tool helps one to make decisions about charge levels. We don't like to charge our batteries over 80% if we don't have to. We can leave Reno with 80% and arrive at Winnemucca with 14%. That's comfortable as there are 2 350 EA chargers and 2 150 and soon, we hope, an RAN station. Now should we charge above 80% eliminating a stop at McDermitt where there is only 1 40 kW CCS or plan to stop there risking that it might be off line or occupied. If we charge only to 80% then we have to charge 54 minutes at McDermitt bringing total trip time to 6:45 (I am assuming we want to drive at 110% of the speed limit). Charging to 98% takes 58 minutes for a total trip time of 6:09. If total time is the criterion then clearly we want to go to the higher charge level but we know we don't want to do that on a regular basis.


Because RAN doesn't appear to have plans for Highway 95 north of Winnemucca, I'd hope that (unlike ABRP) Rivian includes those Greenlots chargers on their navigation and route planning.
Which are the ones that ABRP doesn't show. It shows the one at McDermitt which, curiosly enough, isn't shown on Plugshare.



And again, I can now illustrate my other issue with ABRP. I'd be new to the R1T (not experienced like I am with the Bolt EV), so I'd be relying on ABRP's baseline numbers to be somewhat accurate. A 15% margin of error could easily leave me stranded in the desert, and caveat emptor or "you should have slowed down" on one of the fastest driven highways in the country won't cut it.
Evidently ABRP is quite conservative with consumption so you shouldn't get stranded. You have been complaining about 15% margin on the safe side. This should prevent you from getting stranded. Not to mention that you should be monitoring your battery condition as you go using either the tools in the car or ABRP. If you see that battery status is below the projection at any point on a leg you had better understand why and what to do about it. This has nothing to do with any route planner. A good driver just knows this.

So in summary: If you hand me a Stradivarius and ask me to play one of the Partitas I can't. If I hand you ABRP and ask you to use it to plan a trip you can't. It's a waste to put a Strad in my hands. So also is it a waste to offer ABRP to you. Forget about it.

This is a bad analogy in the sense that you don't have to be a Perlman to use ABRP. Most can use it to profit. Some can't.
 

NewsCoulomb

Member
Joined
Jun 24, 2021
Threads
0
Messages
21
Reaction score
5
Location
California
Vehicles
Bolt EV, Ranger Electric
No, and ordinarily it doesn't. Engineers, mathematicians, physicists and others in the sciences have what the medical profession calls "the knack" This is a "... a rare condition characterized by inherent understanding of all things electrical and mechanical and utter social ineptitude." At least that's how the medics in Dilbert's world describe it. There is another end to the spectrum. It's sort of like full color blindness. Folks at that end, while otherwise being fully intelligent and capable people, have no intuition whatsoever for anything quantitative. It's not that rare. I live with such a person. You are apparently such a person and it, therefore, takes a lengthy explanation to try to get you to understand what the program does, is intended to do, and its limitations.
You just can't help yourself from making it personal, can you?

It's entirely possible, in fact it is certain, that it is more "up to date" on Tesla vehicles (with the exception of the CT) than it is on Rivian or Lucid or Fisker as Telsa's are out in the real world. Tesla also would be at an advantage as the people who developed the program originally developed it for Tesla and the Tesla model is more refined.
Maybe so, but again, my issue is that I submitted significant amounts of data that, apparently, ABRP decided modify or ignore. Yes, I was specifically told that their model for the Bolt EV over 65 mph was entirely modeled on my data logs (at least initially).

But they are. Thousands of users know that based on their experiences. You think they aren't because you haven't the perspective to interpret what the program is telling you.
Not quite. I know full well how to use the program and interpret what it's telling me. That's not my issue with the program; I actually find it fairly intuitive and simple to use. My issue is that, in order to use ABRP effectively, you need to already know how your car operates on long trips, how the charging infrastructure works, and how to plan a trip in an EV. Essentially, by the time an EV owner knows how to use ABRP effectively and set it up to provide reliable outputs, that EV owner no longer needs it (even on routes they are not familiar with). Ergo, it doesn't serve its intended purpose.

You don't. You told the program you want to arrive at chargers with a minimum of 10%. It will adjust the trip parameters to get you there with about 10%. No. That's just about how much energy it takes to cover that terrain at the speed limit given the estimated consumption of the car.
And that is a flaw in the program's design and interface. When someone checks a box saying they want to arrive with at least 10% SOC, the program shouldn't then clandestinely adjust baseline efficiency or driving speeds to reduce efficiency to the point that the 10% SOC on arrival is the predicted outcome. That's just poor modeling.

You make the common error of assuming you represent the population or that the population should conform to your way of doing things. In fact ABRP's baseline should represent the population. So what it does is use consumption numbers representative of the population (i.e. a mean or median). ABRP evidently puts some margin into the consumption figure to protect neophytes from running out of gas. If you know, based on experience, that your consumption is less than the ABRP default, use the number that applies to your personal driving habits. I do that as would any prudent user. You would probably argue that you shouldn't have to do that. In fact you did. You clearly stated that ABRP should reflect your experience.
Again, in case you missed or glossed over it: ABRP's freeway speed efficiency for the Bolt EV *was based on my personal >65 mph driving logs.* So someone made an arbitrary decision to adjust those numbers after the fact. A 5% adjustment I can understand. Even a 10% adjustment to add a cushion might be understandable. A 15% to 20% adjustment with no data to support it? I don't know how that can be justified.

So there is, in fact, nothing wrong with that picture. You list a whole lot of things you would, or might do differently based on your experience with that route. In my last post I showed how proper use of ABRP reveals that many of those options are possible. That is the value of it to me, or anyone else, who has no familiarity with that route.

This is an excellent example of the power of ABRP. In terms of planning it doesn't make much difference if you use the Winnemucca EA or RAN charger which is only a couple of miles away. As soon as the RAN go on line ABRP will pick them up of course. ABRP does give one the option to prefer or abhor a particular brand of charger so users will have the ability to favor one or the other if they wish. More significant here is how the tool helps one to make decisions about charge levels. We don't like to charge our batteries over 80% if we don't have to. We can leave Reno with 80% and arrive at Winnemucca with 14%. That's comfortable as there are 2 350 EA chargers and 2 150 and soon, we hope, an RAN station. Now should we charge above 80% eliminating a stop at McDermitt where there is only 1 40 kW CCS or plan to stop there risking that it might be off line or occupied. If we charge only to 80% then we have to charge 54 minutes at McDermitt bringing total trip time to 6:45 (I am assuming we want to drive at 110% of the speed limit). Charging to 98% takes 58 minutes for a total trip time of 6:09. If total time is the criterion then clearly we want to go to the higher charge level but we know we don't want to do that on a regular basis.

Which are the ones that ABRP doesn't show. It shows the one at McDermitt which, curiosly enough, isn't shown on Plugshare.
No, there's still a lot wrong with that picture, and I actually have no previous experience with that route at all. I only mentioned it because someone mentioned Burns, OR earlier in the thread, and Rivian appears to have no RAN support for Highway 95 north of Winnemucca. I've only taken the Highway 395 route to Burns years ago, which is still impassable in any EV without lengthy L2 AC stops. I do, however, have a lot of experience with the Bolt EV and the disconnect with ABRP's models, which is what I was demonstrating.

I do have to apologize to ABRP, though, because I apparently did have a filter in place (I haven't opened ABRP in over a year). As I recall, I set Electrify America preferred exclusively to illustrate how ABRP had been natively (no filters or preferences set) prioritizing 40 kW chargers over Electrify America chargers along the same route. That being said you also might have a filter set on PlugShare that you don't realize because both the McDermitt and Orovada charging sites appear on my PlugShare.

I unfiltered the charging providers, and ABRP kicked back this gem. Apparently, now, it requires 3% battery to drive 7 miles across town in order to charge from 97% to 98%. :rolleyes: Then, despite leaving with 2% additional battery capacity than I had under the previous plan, I still arrive in Winnemucca with 11% battery instead of the 10% from before. So ABRP's results aren't even consistent from route plan to route plan. :facepalm:

Rivian R1T R1S RAN Rivian Charging Stations Locations Map via Google Maps ABRP 02


Evidently ABRP is quite conservative with consumption so you shouldn't get stranded. You have been complaining about 15% margin on the safe side. This should prevent you from getting stranded. Not to mention that you should be monitoring your battery condition as you go using either the tools in the car or ABRP. If you see that battery status is below the projection at any point on a leg you had better understand why and what to do about it. This has nothing to do with any route planner. A good driver just knows this.
It's interesting because we've had the same discussion in Bolt EV owners groups regarding the inaccuracy. Most seem agree that the overly conservative margins of error for the Bolt EV are fine when ABRP is being used as a route planner. Bolt EV owners were typically just happy to spend half the time and money charging that ABRP predicted.

However, two problems arise if a 15% to 20% margin of error is considered acceptable. First is, the pendulum swings both ways. I'm reasonably sure you won't see anything close to 400 Wh/mi in a Rivian R1T at freeway speeds (oddly, my ABRP lists the consumption as 516 Wh/mi at 65 mph, which is probably more accurate). If ABRP is 15% off the wrong direction for the R1T, then that's going to create a lot of problems for people who were used to using ABRP when it was either more accurate or 15% to the conservative side.

The second problem is, I most often didn't see ABRP used for route planning. At least not in the Bolt EV groups. As I said, it was most often used as a tool to denigrate or dismiss the Bolt EV (i.e., a marketing and sales tool). "You can't make this trip because ABRP says...." or "See how much slower or more expensive it is to travel in a Bolt EV because ABRP says...." At that point, when ABRP is no longer being used as a route planning tool, a 15% to 20% conservative margin of error is misleading and harmful.

So in summary: If you hand me a Stradivarius and ask me to play one of the Partitas I can't. If I hand you ABRP and ask you to use it to plan a trip you can't. It's a waste to put a Strad in my hands. So also is it a waste to offer ABRP to you. Forget about it.

This is a bad analogy in the sense that you don't have to be a Perlman to use ABRP. Most can use it to profit. Some can't.
Making things personal is often an indication that you don't have a sound argument. If it doesn't add to the discussion, it's probably not worth including.
 

ajdelange

Well-Known Member
First Name
A. J.
Joined
Aug 1, 2019
Threads
9
Messages
2,883
Reaction score
2,317
Location
Virginia/Quebec
Vehicles
Tesla XLR+2019, Lexus, Landcruiser, R1T
Occupation
EE Retired
Maybe so, but again, my issue is that I submitted significant amounts of data that, apparently, ABRP decided modify or ignore. Yes, I was specifically told that their model for the Bolt EV over 65 mph was entirely modeled on my data logs (at least initially).
So which was it. They ignored it or they made it the basis for their model? I hope not the latter as they should know better than to take data from a single user as the basis for their model.



My issue is that, in order to use ABRP effectively, you need to already know how your car operates on long trips, how the charging infrastructure works, and how to plan a trip in an EV.
Really? And to use a Strad you have to know how to play a violin. Perspective, perspective



And that is a flaw in the program's design and interface. When someone checks a box saying they want to arrive with at least 10% SOC, the program shouldn't then clandestinely adjust baseline efficiency or driving speeds to reduce efficiency to the point that the 10% SOC on arrival is the predicted outcome. That's just poor modeling.
THAT IS EXACTLY WHAT OPTIMIZATION ALGORITHMS ARE DESIGNED TO DO. Of course it shouldn't be clandestine and it isn't. The results are presented to the user.



Again, in case you missed or glossed over it: ABRP's freeway speed efficiency for the Bolt EV *was based on my personal >65 mph driving logs.* So someone made an arbitrary decision to adjust those numbers after the fact. A 5% adjustment I can understand. Even a 10% adjustment to add a cushion might be understandable. A 15% to 20% adjustment with no data to support it? I don't know how that can be justified.
As it would be foolish to use.... this is a waste of time. You don't understand largely because you don't want to understand. Put it on the shelf. Forget about it. Leave it to those who do understand.






That being said you also might have a filter set on PlugShare that you don't realize because both the McDermitt and Orovada charging sites appear on my PlugShare.
Yes, I had the minimum power slider too high. They now both appear in PlugShare and ABRP.



I unfiltered the charging providers, and ABRP kicked back this gem. Apparently, now, it requires 3% battery to drive 7 miles across town in order to charge from 97% to 98%. :rolleyes: Then, despite leaving with 2% additional battery capacity than I had under the previous plan, I still arrive in Winnemucca with 11% battery instead of the 10% from before. So ABRP's results aren't even consistent from route plan to route plan.
There are rounding errors but ABRP does do some quirky things from time to time. That's why I keep emphasizing that you have to know how to play your instrument in order to produce music. You also need to understand that ABRP gives you a rather flexible but still limited set of optimality criteria and constraints. Sometimes you will want to impose additional ones. This is fairly easily done but would probably be in the "advanced techniques" part of the manual (if there were a manual - that's a complaint I have).


However, two problems arise if a 15% to 20% margin of error is considered acceptable. First is, the pendulum swings both ways.
It does swing both ways. That's why they apparently boost the consumption so that when it swings negative the neophyte is still protected.

I'm reasonably sure you won't see anything close to 400 Wh/mi in a Rivian R1T at freeway speeds (oddly, my ABRP lists the consumption as 516 Wh/mi at 65 mph, which is probably more accurate).
I'm pretty sure you won't either. I picked 400 to make the math easier but, of course, 500 is easier still and is closer to what ABRP lists as default. I did all the runs in the last post using the program default (516). And also 110% of posted speed just to stress the truck a bit and because people out there like to drive that fast.



If ABRP is 15% off the wrong direction for the R1T, then that's going to create a lot of problems for people who were used to using ABRP when it was either more accurate or 15% to the conservative side.
We don't have Rivians. We have their promise of 400 EPA miles on a 170 kW battery. That implies 420 Wh/mi rated consumption. Going to 65 means an increase on consumption of about 10% bringing us to 467.5. The default of 516 is 10% higher than that. So it is definitely not off "in the wrong direction". It is, as are all ABRPs numbers, fairly conservative.

At this point we use ABRP as a tool to determine whether the Rivian would be suitable for this particular trip. Using the conservative default, driving faster than the charging limit and even limiting charge levels to 20 - 80% we see that the answer is that we can make these trips with adequate margins and that a max R1T will do it quite nicely.
 

NewsCoulomb

Member
Joined
Jun 24, 2021
Threads
0
Messages
21
Reaction score
5
Location
California
Vehicles
Bolt EV, Ranger Electric
So which was it. They ignored it or they made it the basis for their model? I hope not the latter as they should know better than to take data from a single user as the basis for their model.
They used it as the basis of their model, but they arbitrarily adjusted it. I thought that part was clear.

Really? And to use a Strad you have to know how to play a violin. Perspective, perspective
You assume that I don't know how to use the tool, but I clearly do. What your perspective is preventing you from seeing is that, despite knowing how to use ABRP, I don't find it to be good or useful. Yet here you are, comparing it to a Stradivarius violin. ABRP isn't the worst EV route planner, but it certainly isn't the best.

THAT IS EXACTLY WHAT OPTIMIZATION ALGORITHMS ARE DESIGNED TO DO. Of course it shouldn't be clandestine and it isn't. The results are presented to the user.
Only if arriving with exactly 10% SOC is the goal. It's not. It's a parameter. The goal is to complete the trip as quickly as possible, and if driving 110% of the posted speed limit results in arriving at the first stop with 20% SOC, so be it. ABRP shouldn't be adjusting efficiency and driving speeds beyond 110% in order to manipulate the 10% SOC on arrival outcome.

As it would be foolish to use.... this is a waste of time. You don't understand largely because you don't want to understand. Put it on the shelf. Forget about it. Leave it to those who do understand.

***

There are rounding errors but ABRP does do some quirky things from time to time. That's why I keep emphasizing that you have to know how to play your instrument in order to produce music. You also need to understand that ABRP gives you a rather flexible but still limited set of optimality criteria and constraints. Sometimes you will want to impose additional ones. This is fairly easily done but would probably be in the "advanced techniques" part of the manual (if there were a manual - that's a complaint I have).

It does swing both ways. That's why they apparently boost the consumption so that when it swings negative the neophyte is still protected.
Again, you keep getting caught up on this "not knowing how to use the tool." I know perfectly well how to use the tool, but my three points of concern are:

1. If a "neophyte" uses the tool, they don't know enough to "play their instrument well," so their route plans are going to be of limited use and result in questionable trip experiences.

2. ABRP requires EV and trip planning experience to use effectively, at which point, why go through the work to use a tool that is going to produce a lower quality trip plan than you could generate for yourself with less initial effort?

3. ABRP is primarily being used as a resource by experienced EVers to explain to "neophytes" what to expect on their trips; however, the experienced EVers that are prescribing these route plans often don't have experience with the EV in question and are not (as a result) "playing their instrument well." The result is, a misleading trip plan for new or prospective EV owners that either discourages them from taking the trip or buying the EV in question, or on the other other hand, instilling a false sense of confidence that could result in buying the wrong EV (e.g., battery size) for their needs or in catastrophic trips that end in requiring roadside assistance.

I'm pretty sure you won't either. I picked 400 to make the math easier but, of course, 500 is easier still and is closer to what ABRP lists as default. I did all the runs in the last post using the program default (516). And also 110% of posted speed just to stress the truck a bit and because people out there like to drive that fast.

We don't have Rivians. We have their promise of 400 EPA miles on a 170 kW battery. That implies 420 Wh/mi rated consumption. Going to 65 means an increase on consumption of about 10% bringing us to 467.5. The default of 516 is 10% higher than that. So it is definitely not off "in the wrong direction". It is, as are all ABRPs numbers, fairly conservative.
The R1T has an estimated range of 400 miles on a 180 kWh battery (2.2 mi/kWh or 455 Wh/mi). EVs interact with their EPA rated ranges differently than ICE vehicles, and even small, aerodynamic EVs will see a 20% increase in consumption at freeway speeds. Large EVs see closer to 30%. EV trucks are likely to be even worse, but the R1T is a relatively small truck, so 30% is probably a good initial estimate. As a result, I'd anticipate ~1.6 mi/kWh or 600 Wh/mi as a baseline freeway consumption for the R1T.

As a result, ABRP might very well be off in the wrong direction.

At this point we use ABRP as a tool to determine whether the Rivian would be suitable for this particular trip. Using the conservative default, driving faster than the charging limit and even limiting charge levels to 20 - 80% we see that the answer is that we can make these trips with adequate margins and that a max R1T will do it quite nicely.
And again, this is possibly the most common use I see for ABRP, and for this purpose, I don't think it's particularly helpful. Optimization algorithms are only as effective as the quality of their input criteria and data. ABRP has already shown itself to be near useless for certain EVs where copious amounts of data have been provided (such as the Bolt EV). I seriously question its value here as a predictive tool for the R1T when data inputs are being chosen arbitrarily.

I can say with certainty based on years of experience with both EVs and trucks that the Highway 95 route north of Winnemucca should be worrisome for R1T owners, even with the 400+ mile range option.
 

Sponsored

skyote

Well-Known Member
Joined
Mar 12, 2019
Threads
55
Messages
2,725
Reaction score
5,647
Location
Austin, TX
Vehicles
Jeeps, 2500HD Duramax, R1S Preorder (Dec 2018)
AJ & News, y'all should definitely keep going back & forth on this. I feel like you're both so close to convincing the other!
 

Atlrivian

Well-Known Member
Joined
Apr 14, 2021
Threads
12
Messages
580
Reaction score
961
Location
Atlanta
Vehicles
Grand Cherokee, XC90
Clubs
 
AJ & News, y'all should definitely keep going back & forth on this. I feel like you're both so close to convincing the other!
They should just go ahead and block each other and be done with it
 

azbill

Well-Known Member
First Name
Bill
Joined
Jun 8, 2020
Threads
14
Messages
1,261
Reaction score
1,558
Location
Arizona
Vehicles
R1T, Mach E, Hummer EV SUT
Occupation
Engineer
Living out west where chargers are few and far between, I like to use SRP, Simple Route Planner. Look at my route, in this example Phoenix to Las Vegas, and see what chargers are on the route (Plugshare works for that):

Rivian R1T R1S RAN Rivian Charging Stations Locations Map via Google Maps 1625254497071


Kingman is 168 miles, then 111 more to Vegas, so I will stop in Kingman and charge until I have 111 miles plus margin. Simple Route Planner using my gray matter.
 
OP
OP

Autolycus

Well-Known Member
Joined
May 2, 2021
Threads
14
Messages
2,027
Reaction score
3,111
Location
ATL
Vehicles
ICE only :(
AJ & News, y'all should definitely keep going back & forth on this. I feel like you're both so close to convincing the other!
Yeah, and especially about the accuracy and usefulness of ABRP while in a thread about a map of approximate locations showing really just the overall strategy for the Rivian Adventure Network…
 

ajdelange

Well-Known Member
First Name
A. J.
Joined
Aug 1, 2019
Threads
9
Messages
2,883
Reaction score
2,317
Location
Virginia/Quebec
Vehicles
Tesla XLR+2019, Lexus, Landcruiser, R1T
Occupation
EE Retired
They used it as the basis of their model, but they arbitrarily adjusted it. I thought that part was clear.
As well they should.



A)You assume that I don't know how to use the tool, but I clearly do.
B)I don't find it to be good or useful.
B proves A to be false


ABRP isn't the worst EV route planner, but it certainly isn't the best.
The B stands for "better", not best. Out of curiosity I'll ask which is the best?

Only if arriving with exactly 10% SOC is the goal. It's not. It's a parameter. The goal is to complete the trip as quickly as possible,
Minimum trip time is one criterion for optimality which is merged with others depending on how you set the optimality slider. The goal is sought by varying the independent variables subject to constraints. Mininum and maxium SoC levels are at charger and destination are constraints.

[/QUOTE]and if driving 110% of the posted speed limit results in arriving at the first stop with 20% SOC, so be it. ABRP shouldn't be adjusting efficiency and driving speeds beyond 110% in order to manipulate the 10% SOC on arrival outcome.[/QUOTE] No and it doesn't. ABRP does not adjust consumption except by speed, terrain, and weather. It does not adjust speed. That's a constraint. If you tell it to do the run at 110% of the speed limit it will do it at 110 of speed limit. The speed constraint can be softened if you allow the algorithm to adjust speed in order to satisfy one of the other constraints. For example if you select 20% minimum arrival SoC and you can't do that driving 110% of 80, the segment lights up orange and the notification "maximum speed 75" lights up.

Again, you keep getting caught up on this "not knowing how to use the tool." I know perfectly well how to use the tool,
Again, in your comments I just addressed you make it quite clear that you don't.



but my three points of concern are:
1. If a "neophyte" uses the tool, they don't know enough to "play their instrument well," so their route plans are going to be of limited use and result in questionable trip experiences.
Certainly they will produce better plans as they gain experience with the tool. That's why I encourage people in waiting here and on the CT site to get familiar with it now (if they are not already). They can learn a great deal about where they can and cannot take an R1T, for example, and what the effects of a 30 kt headwind are likely to be.


2. ABRP requires EV and trip planning experience to use effectively,
You are alone or in a very small group in holding that opinion. It ignores the value of it as a teaching tool.

...at which point, why go through the work to use a tool that is going to produce a lower quality trip plan than you could generate for yourself with less initial effort?
You think a neophyte knows how to calculate wind load or gravitational load or inertial load or the effects of rain on rolling resistance or has sufficient experience from driving BEV to formulate a plan? No. Of course not. But ABRP (and other programs too) have models for all these things and from them a neophyte can learn about these phenomena before he ever gets into his new car. That's why.

3. ABRP is primarily being used as a resource by experienced EVers to explain to "neophytes" what to expect on their trips;
ABRP is generating about 4080 plans per hour as I write this. How would you know what any of them are being used for or by whom?


...even small, aerodynamic EVs will see a 20% increase in consumption at freeway speeds. Large EVs see closer to 30%.
And we wonder where you got this idea from. I drive an X. It's a pretty hefty chunk of metal. Consumption goes up by less than 10% to 310 Wh/mi at 65 mph. ABRP in its conservatism defaults this car at 332, another 7%, which comes to 118% of the rated consumption. ABRP defaults the R1T max to 516 which is 121% of the presumed rated consumption. That's clearly a much more realistic estimate than



EV trucks are likely to be even worse, but the R1T is a relatively small truck, so 30% is probably a good initial estimate. As a result, I'd anticipate ~1.6 mi/kWh or 600 Wh/mi as a baseline freeway consumption for the R1T.
But you'd probably be way off. 516 is more realistic than 600 for sure. And it better be. 600 implies something like 700 at 75 (assuming drag is half the load at 65) which in turn implies a working range of a little under 200 miles at 75 mph cruise. There's going to be a lot of screaming from the west if it turns out to be that bad.


As a result, ABRP might very well be off in the wrong direction.
As an, I hope, decent analyst I recognize that the probability that it will be off in the optimistic direction but the evidence points the other way.



I seriously question its value here as a predictive tool for the R1T when data inputs are being chosen arbitrarily.
Well that's why we don't chose the constraints and consumption values or optimality criteria arbitrarily.



I can say with certainty based on years of experience with both EVs and trucks that the Highway 95 route north of Winnemucca should be worrisome for R1T owners, even with the 400+ mile range option.
A good analyst would know better than to say "with certainty" but it's pretty easy to debunk this statement without any trip planners.
1)R1T rated consumption is going to come in close to 445 Wh/mi. At that rate it would take 99 kWh to cover the 222 miles. The elevation difference isn't much so with conservative driving - and that's well below the posted speed limits on this road, one would use 58.1% of the battery leaving 41.9% margin.
2)Take off 20% for the end keep out zones and you still have 21.9% margin
3)Drive conservatively (i.e. at the speed limit) and charge conservatively (i.e. not above 80%) and you will arrive with 10% margin
4)Drive conservatively but take on extra charge and you will arrive with extra margin (above 10%).
5)Live dangerously. Drive at 110% of the speed limit. You'll have to charge to 88% (which isn't bad to do occasionally) and you will still get there with 10% margin.
6)Charge to say 85% and hit the road monitoring the consumption graph (I assume there will be one) watching the predicted arrival SoC. Modulate your driving speed to keep the arrival SoC where you want it to be. You are now doing ABRPs job - adjusting one of the independent variables to minimize time en route while not violating the restrictions.
 
Last edited:

Sponsored

ajdelange

Well-Known Member
First Name
A. J.
Joined
Aug 1, 2019
Threads
9
Messages
2,883
Reaction score
2,317
Location
Virginia/Quebec
Vehicles
Tesla XLR+2019, Lexus, Landcruiser, R1T
Occupation
EE Retired
AJ & News, y'all should definitely keep going back & forth on this. I feel like you're both so close to convincing the other!
Yes, yes I know I'm wasting my time.

But I do want people to understand that ABRP and the other programs can be of great benefit to those trying to learn about BEV.

I promise to waste no more time or bandwidth on this.
 

skyote

Well-Known Member
Joined
Mar 12, 2019
Threads
55
Messages
2,725
Reaction score
5,647
Location
Austin, TX
Vehicles
Jeeps, 2500HD Duramax, R1S Preorder (Dec 2018)
Yes, yes I know I'm wasting my time.

But I do want people to understand that ABRP and the other programs can be of great benefit to those trying to learn about BEV.

I promise to waste no more time or bandwidth on this.
FWIW, I'm with you on this matter...
 

Gshenderson

Well-Known Member
First Name
Greg
Joined
Sep 28, 2019
Threads
13
Messages
1,229
Reaction score
2,767
Location
Park City, UT / Kemmerer, WY
Vehicles
2015 Tesla S 85D, 2019 4Runner TRD Offroad, R1T
AJ & News, y'all should definitely keep going back & forth on this. I feel like you're both so close to convincing the other!
Does anyone whose read all of it care to give the rest of us who just scroll (and scroll, and scroll and scroll) past it the Cliff Notes recap?
 

DuckTruck

Well-Known Member
First Name
Duck
Joined
Dec 23, 2020
Threads
33
Messages
2,343
Reaction score
6,228
Location
PNW
Vehicles
Corvair, BMW325, Acura Legend, XC60, '16 Caddy ELR
Clubs
 
The no services for XXX miles is where it starts to get interesting (but those signs are generally found only on well travelled routes). We rode one stretch for 80+ miles last week where we saw only 3 other vehicles on the road - 1 brand new Mercedes with temp tags and two pickups that were obviously worked hard every day (as they were intended).

We stopped for lunch and borrowed the shade on the porch of the General Store/Gas Station/Bar/Restaurant/Post Office (all closed save the Post Office function)
1625159561898.png


Got a chance to chat with several locals as we hung out. One worked a 30K acre ranch with 1,200 "pairs" (cow and calf). Another didn't have a ranch, just 150 acres and a few head. They really want someone to buy and reopen the store (we declined).

This "town" is in the RAN hole in Oregon coverage as discussed above. John Day and Burns could both use inclusion in the RAN. There are relatively new ChargePoint 62.5 kW stations in both with mixed reliability ratings.
1625161693045.png
DucRider,

I love your travel story. Having more than a single DCFC station in both John Day and Burns would allow much more Adventure away from US 26 and US 20, rather than simply allowing us to drive through these areas. Per ChargePoint, ChargeHub, and ABRP, each city reportedly has one Level 3 CCS DCFC unit available, of either 50 or 62.5 kW. The John Day area, with its namesake Fossil Beds (Google that, folks) would be a nice way to show those old dinosaurs some respect that ICE-mobiles just can't offer. After all, those gas-powered beasts running up and down US 395 are burning up Dino's family and friends.

To complete the patchwork of multiple bay chargers necessary to traverse and really explore Eastern Oregon, Western Idaho, and Northwest Nevada, also putting stations in Jordan Valley and/or Rome, Oregon, would be a huge help. Both are on what's called the "ION Highway", the central stretch of US 95. "ION" stands for Idaho, Oregon, Nevada, although US 95 goes all the way from Mexico to Canada.

As you can imagine, that's quite a stretch of impressive land along the entire length of US 95. Once you get across Interstate 84, North of Boise, you skirt Hells Canyon on your way North towards Lake Coeur d'Alene and Lake Pend Oreille. Both are beautiful bodies of water with mountains all around providing great places to explore. As Lil'O Annie and I have discussed here, some of the geological forces that impacted and sculpted the Northwest corner of the nation (including the Great Missoula Floods) are on full display along the Northern stretches of US 95. The only problem is the absence of large commercial DCFC charging between I-80 in Winnemucca, NV, (Where there's an ElectrifyAmerica) until you hit the intersection with I-90 in Coeur d'Alene Idaho. Even at that, you have to go East or West some distance along I-90 to find another EA station. North of Winnemucca to the Canadien border, there is no real commercial DCFC charging available. The only DCFC access near Boise is quite a bit East of US 95. That won't really suffice for opening this North/South scenic route, as you'd burn too much time and energy getting there and back to US 95.

Just as teasers, I'm leaving some pics of the places along US 20 and US 95 in Eastern Oregon, as well as US 95 in Western Idaho. These places all scream Adventure, and I hope someone with the ability to influence DCFC station placement is listening.

For those people who think this area is desolate, you are correct. But that's part of its allure. It is beautiful and full of wonderful adventures, if people would stop, look, and explore.

Steens Mountain:
Rivian R1T R1S RAN Rivian Charging Stations Locations Map via Google Maps 4c099556

Rivian R1T R1S RAN Rivian Charging Stations Locations Map via Google Maps download (2)


Kiger Gorge on Steens Mountain:
Rivian R1T R1S RAN Rivian Charging Stations Locations Map via Google Maps steens-kiger


The Kiger Mustangs, a wild horse family on Steens Mountain:
















Alvord Desert:
Rivian R1T R1S RAN Rivian Charging Stations Locations Map via Google Maps Desert-Flame-Alvord-Desert-2

Rivian R1T R1S RAN Rivian Charging Stations Locations Map via Google Maps kevbrc52yjf31


Where the Alvord Desert meets Steens Mountain:
Rivian R1T R1S RAN Rivian Charging Stations Locations Map via Google Maps download (3)


The Pillars of Rome
Rivian R1T R1S RAN Rivian Charging Stations Locations Map via Google Maps PillarsOfRomePano1


The Upper Owyhee River:
Rivian R1T R1S RAN Rivian Charging Stations Locations Map via Google Maps CReeser_Owyhee

Rivian R1T R1S RAN Rivian Charging Stations Locations Map via Google Maps place_image-image-56717895-5c20-4a1e-8d2c-ffe53d06629d

Rivian R1T R1S RAN Rivian Charging Stations Locations Map via Google Maps owyhee-rafting-lambert-gorge


Hells Canyon:
Rivian R1T R1S RAN Rivian Charging Stations Locations Map via Google Maps HellsCanyon7_VisitIdaho-700x1049
Sponsored

 
 




Top