[sciencegeek]

Testing regen efficiency in real world situations is difficult because there are so many parameters that matter for consumption and regeneration that it's tough to set up the right experiment. In addition, to make sure it's not a fluke, you have to do the same thing over and over to get reliable data (i.e., statistical averages and estimates of variance, etc.)

So by accident I came upon an idea that would test regen efficiency in a controlled way; at first I was just interested in how much energy it takes to lift the I-Pace uphill, with me and my bike and not much else. Which inspired this thread:

How much battery do you use uphill?

This thread is inspired by Steve's post ONE discharge to 0% causes >5% battery capacity loss Each 1000ft of climbing (roughly 300 meters) uses 1.89 kWh if it's just you. Each additional passenger will be about 1/25 more. You should get back 90% of that on the downhill if you drive at the same...

www.i-paceforum.com

Then it occurred to me that one can measure the regen efficiency by comparing the energy spent going uphill with the energy regenerated while going the same downhill.

I drive to a trailhead several times a week. It's exactly 690 ft above my house, 10.6 miles one way. I go back and forth the same way except for a quarter mile at the end where the way into our neighborhood is different from the way out, but the overall distance is the same and it's flat there anyway.

Distance-wise and regarding all other parameters, the way there and back are identical. On average it takes me 23 minutes to get there and 24 minutes back. Consequently, the average speed going there is slightly higher than coming back (28 mph vs 27 mph, respectively). Not much different, probably reflecting time spent at stoplights which are slightly more forgiving on the way there (and I tend to drive slightly faster on the way downhill.) Anyway, point being that there and back are identical except: the elevation.

Soooo.

I know it takes 1.89 kWh to lift my I-Pace with me and my bike (2270 kg) 304.8 meters (1000 ft). Therefore, it takes 0.69*1.89 = 1.3 kWh to lift my I-Pace 690 feet.

Going there I use 4.7 kWh on average. 1.3 kWh of that is the potential energy required for lifting. Therefore, 4.7 - 1.3 = 3.4 kWh are for the driving.
Going back I use 2.3 kWh on average. Assuming the driving consumption (see above) is the same, I therefore regenerate 3.4 - 2.3 = 1.1 kWh of potential energy. That's 85% (1.1 / 1.3 = 0.85) of what I spent lifting the car. Therefore, by this calculation, the regen efficiency is 85%.

Independently of that calculation, the car also provides the amount of regenerated energy (last column in the spreadsheet). Going there (uphill) I regenerate 1.2 kWh on average. Going back I regenerate 2.4 kWh. The difference should be what I regenerate due to recuperating potential energy going the net downhill of 690 feet: 2.4 - 1.2 = 1.2 kWh. Using this figure, I get back 1.2 / 1.3 = 0.92, implying a regen efficiency of 92%.

Bottom line, there's of course some measurement variance and some minor confounders but it's pretty clear that regen efficiency is very high, somewhere around 80 to 90 percent. 🍺

 

[NewJag in Bama]

Help me with one thing, how do the energy consumed and the regen numbers relate? For instance, going uphill, you used an average of 4.7 and regenerated 1.2. Did you actually use 5.9 with 1.2 off set, or did you consume 4.7 and regen 1.2 for a net of 3.5?

 

[sciencegeek]

The former. Net consumption is 4.7 kWh in this case. Regen value is provided for entertainment/information only.

 

[NewJag in Bama]

Interesting data. Did you have it set for High Regen or Low on the way down? Wonder what difference that would make?

 

[sciencegeek]

I always drive in Dynamic with High Regen, no creep.

 

[NewJag in Bama]

Wonder what the numbers would look like in Normal mode and Low Regen? There was another post in a different section where this was discussed. The theory there was that in low regen you would coast more and consume much less energy even though you don't regen as much. This would be a perfect way to answer that question.

 

[sciencegeek]

The regen numbers would be lower by some constant, the delta between uphill regen and downhill regen would be close to the same. Physics is physics ... it would still require X hWh to lift that mass in earth's gravity, and you'd get most of that back on the way down. But yes, theoretically if you could only coast you should get all of it back, not just 85ish percent; but then you'd be changing a whole bunch of other driving-related parameters, like speed and wind resistance so it's a tough experiment to do properly.

 

[Qtown charger]

Testing regen efficiency in real world situations is difficult because there are so many parameters that matter for consumption and regeneration that it's tough to set up the right experiment. In addition, to make sure it's not a fluke, you have to do the same thing over and over to get reliable data (i.e., statistical averages and estimates of variance, etc.)

So by accident I came upon an idea that would test regen efficiency in a controlled way; at first I was just interested in how much energy it takes to lift the I-Pace uphill, with me and my bike and not much else. Which inspired this thread:

How much battery do you use uphill?

This thread is inspired by Steve's post ONE discharge to 0% causes >5% battery capacity loss Each 1000ft of climbing (roughly 300 meters) uses 1.89 kWh if it's just you. Each additional passenger will be about 1/25 more. You should get back 90% of that on the downhill if you drive at the same...

www.i-paceforum.com

Then it occurred to me that one can measure the regen efficiency by comparing the energy spent going uphill with the energy regenerated while going the same downhill.

I drive to a trailhead several times a week. It's exactly 690 ft above my house, 10.6 miles one way. I go back and forth the same way except for a quarter mile at the end where the way into our neighborhood is different from the way out, but the overall distance is the same and it's flat there anyway.

Distance-wise and regarding all other parameters, the way there and back are identical. On average it takes me 23 minutes to get there and 24 minutes back. Consequently, the average speed going there is slightly higher than coming back (28 mph vs 27 mph, respectively). Not much different, probably reflecting time spent at stoplights which are slightly more forgiving on the way there (and I tend to drive slightly faster on the way downhill.) Anyway, point being that there and back are identical except: the elevation.

Soooo.

I know it takes 1.89 kWh to lift my I-Pace with me and my bike (2270 kg) 304.8 meters (1000 ft). Therefore, it takes 0.69*1.89 = 1.3 kWh to lift my I-Pace 690 feet.

Going there I use 4.7 kWh on average. 1.3 kWh of that is the potential energy required for lifting. Therefore, 4.7 - 1.3 = 3.4 kWh are for the driving.
Going back I use 2.3 kWh on average. Assuming the driving consumption (see above) is the same, I therefore regenerate 3.4 - 2.3 = 1.1 kWh of potential energy. That's 85% (1.1 / 1.3 = 0.85) of what I spent lifting the car. Therefore, by this calculation, the regen efficiency is 85%.

Independently of that calculation, the car also provides the amount of regenerated energy (last column in the spreadsheet). Going there (uphill) I regenerate 1.2 kWh on average. Going back I regenerate 2.4 kWh. The difference should be what I regenerate due to recuperating potential energy going the net downhill of 690 feet: 2.4 - 1.2 = 1.2 kWh. Using this figure, I get back 1.2 / 1.3 = 0.92, implying a regen efficiency of 92%.

Bottom line, there's of course some measurement variance and some minor confounders but it's pretty clear that regen efficiency is very high, somewhere around 80 to 90 percent. 🍺

View attachment 4262

This is a great dataset. Lots of repetitive trips. Confirms your earlier comments regarding regen efficiency. I had thought of something similar, but had no desire to do the repetition, just to prove a point.

I would assume it is human nature to accelerate a little faster downhill than uphill. Do you have any repetitions using cruise control. I was thinking of a quick run up to the Poconos (Interstate highway all the way) with the cruise set at 65mph

 

[sciencegeek]

Ah using cruise control is a great idea. Unfortunately for this specific trip it would be a hassle, too many stop lights much of the way and then changing speed limits. Yeah repeating things a lot is crucial because of changing environment (wind direction, temperature etc) and driving specifics. Cruise control on a highway would be great, but you'd still have to repeat it a bunch of times ...

 

[Pollo de muerte]

No offense, but I don't think your calculation is measuring regen efficiency. You are combining regeneration and energy avoided by coasting downhill. Every non-lab test of regen efficiency that I've seen has been on level roads with no wind or a dyno. The EPA sets EV efficiency at 77% from grid to wheels, so assuming that regen is similarly efficient (a generous assumption since EV motors are designed for propulsion efficiency and not regen efficiency), the theoretical max is about 60% once you try to put the recaptured power back to wheels. I have not seen a proper test that reaches this level (highest is about 50%).

I also check (but don't record) my daily commute. It's 5.4 miles each way in flat Florida. Without running A/C, it's consistently 1.1 kWh with light regen and 1.3 with heavy regen. Obviously this is with an old hypermiler as the driver (me) so technique is a factor.

I also get much better combined numbers on the Volt by avoiding the cruise control, but that's a special Volt thing and not really applicable.

 

[sciencegeek]

Modern electric motors are >90% efficient and there is virtually no difference between running them 'forward' or 'reverse'. My results are consistent with that. EPA gives 77% as guidance because there is also loss from the grid to the battery (>10% on Level 2 and about 20% on Level 1).

Flat road tests ask a somewhat different question: given that I'm spending energy on forward propulsion, how much am I getting back with braking? The denominator in that is much larger because it's all the energy used for propulsion. It's similar to the question of how much better mileage a hybrid gets vs its ICE-only counterpart.

 

[Pollo de muerte]

Your methodology is simply not what I've ever seen used to measure regen efficiency and is much higher than what I've seen produced in other real world tests.

In the end, these regen debates give off more heat than light. I'm just going to say that your data is interesting, but is not "regen efficiency" as I've seen that term used elsewhere.

I'll bow out of this discussion with the following: folks should drive in what ever mode is most comfortable and safe ... regen is not going to make enough difference to override those other concerns IMO.

 

[sciencegeek]

Your methodology is simply not what I've ever seen used to measure regen efficiency and is much higher than what I've seen produced in other real world tests.

In the end, these regen debates give off more heat than light. I'm just going to say that your data is interesting, but is not "regen efficiency" as I've seen that term used elsewhere.

I'll bow out of this discussion with the following: folks should drive in what ever mode is most comfortable and safe ... regen is not going to make enough difference to override those other concerns IMO.

I totally agree, I'm measuring something different from other 'real world' tests. One needs to get beyond the semantics ("what does 'regen' mean?") and look at the nature of the tests. I'm testing how much potential energy you get back into the battery, and I'm making the inference that it should be the same for kinetic energy, which is much much harder to measure because it's so confounded with losses during driving. The title of the thread is too generic, I should have been more specific. I also agree that folks should drive the way they want, I certainly do 🍺

 

[Qtown charger]

Ah using cruise control is a great idea. Unfortunately for this specific trip it would be a hassle, too many stop lights much of the way and then changing speed limits. Yeah repeating things a lot is crucial because of changing environment (wind direction, temperature etc) and driving specifics. Cruise control on a highway would be great, but you'd still have to repeat it a bunch of times ...

Just had time to look back thru my trip log. One pair of trips stand out: Queensbury NY to Montreal and back again - approx 175miles each way. Up and over the High Peaks region of the Adirondacks in NY. Both were using cruise control for ~160mile each way. Regen numbers for the two trips were 1.6 and 1.3kWh. Thus, not much regen when using the cruise control. For comparison, I see similar magnitude regen numbers on my 17mile commute w/o cruise set
In hindsight, not surprising, coming down hills the car just reduces the energy supplied to the motors. In other words, the energy gained coming downhill was not greater than the energy used to propel the car. Perhaps an ICE freewheels better than an EV?