Model X Range Degradation Over Time - Estimates and Predictions

MrBoylan · Nov 20, 2015

Since there isn't much else to talk about right now, I thought it would be an interesting exercise to predict and discuss the Model X range and battery capacity degradation over time. Given the fairly extensive research done on the Model S, and the average mileage driven for personal cars, we can make some predictions and assumptions. Many of the EV nay-sayers and FUD-spreading short sellers claim that the Tesla battery will need to be replaced sooner rather than later, and that this significant expense should be considered when purchasing the car (and when considering the overall environmental impact of the car).

Research suggests that the problem with battery degradation over time with Tesla's current generation battery packs is not as bleak as some would have you believe. If anyone has comments or finds a flaw in my math, please let me know.

Assumption 1: Model X battery pack will have similar performance to existing Model S battery pack.

We can assume that the Model X and S use the same cells and battery packs, and that performance of the 2015/2016 Model X will be similar to those in the 2012-2014 Model S. A fairly comprehensive owners survey in the Netherlands plotted out the actual range of several Model S cars, as reported by their owners over time. Based on these survey results, it appears that the Model S cars in the wild as of the beginning of 2015 (cars manufactured between 2012 and 2014) lose on average 6% in battery capacity/range after 50,000 miles. After that initial hit, it appears that range decreases at a much slower rate (1% per 30,000 miles). Again, these are averages with some cars performing better and some worse.

Assumption 2: Model X is driven close to the "average" of other cars.

It may be surprising to some that the average number of miles driven for a passenger car is only 11,244 miles per year (in the United States, data as of June, 2015). The calculations below use that average for estimates. But I'll include actual mileage numbers in the table as well in case people want to do the math for their own driving habits. For kilometer equivalents, you'll also have to do some math.

Could it really be true that a Model X would still have 84% of its range (215 miles) after 30 years of use and 337,000 miles driven? Data seem to suggest that. And this is good news for EVs in general, and Tesla in specific. Again, if anyone finds a flaw in my math, please let me know. Or share any other thoughts or observations.

Model X anticipated range loss over time. 90D with 257 miles (EPA) rated range
Year	Cumulative Miles Driven	Estimated Range (miles)	Range Lost (miles)	% Range Lost	% Range Retained
0	0	257	0	0.00%	100%
1	11244	253.53235	3.46765	1.35%	98.65%
2	22488	250.0647	6.9353	2.70%	97.30%
3	33732	246.59705	10.40295	4.05%	95.95%
4	44976	243.1294	13.8706	5.40%	94.60%
5	56220	239.66175	17.33825	6.75%	93.25%
6	67464	238.70815	18.291854	7.12%	92.88%
7	78708	237.75454	19.245458	7.49%	92.51%
8	89952	236.80094	20.199062	7.86%	92.14%
9	101196	235.84733	21.152666	8.23%	91.77%
10	112440	234.89373	22.10627	8.60%	91.40%
11	123684	233.94013	23.059874	8.97%	91.03%
12	134928	232.98652	24.013478	9.34%	90.66%
13	146172	232.03292	24.967082	9.71%	90.29%
14	157416	231.07931	25.920686	10.09%	89.91%
15	168660	230.12571	26.87429	10.46%	89.54%
16	179904	229.17211	27.827894	10.83%	89.17%
17	191148	228.2185	28.781498	11.20%	88.80%
18	202392	227.2649	29.735102	11.57%	88.43%
19	213636	226.31129	30.688706	11.94%	88.06%
20	224880	225.35769	31.64231	12.31%	87.69%
21	236124	224.40409	32.595914	12.68%	87.32%
22	247368	223.45048	33.549518	13.05%	86.95%
23	258612	222.49688	34.503122	13.43%	86.57%
24	269856	221.54327	35.456726	13.80%	86.20%
25	281100	220.58967	36.41033	14.17%	85.83%
26	292344	219.63607	37.363934	14.54%	85.46%
27	303588	218.68246	38.317538	14.91%	85.09%
28	314832	217.72886	39.271142	15.28%	84.72%
29	326076	216.77525	40.224746	15.65%	84.35%
30	337320	215.82165	41.17835	16.02%	83.98%

References:
Tesla Model S Owner's Survey on range degradation over time: Tesla Model S battery degradation data | Steinbuch
Average miles driven per year by vehicle type (United States): Alternative Fuels Data Center: Maps and Data

David99 · Nov 21, 2015

MrBoylan said:
Could it really be true that a Model X would still have 84% of its range (215 miles) after 30 years of use and 337,000 miles driven? Data seem to suggest that.

We don't have data from a 30 year old Tesla battery, so I wouldn't say the 'data seems to suggest that'. We actually have very few people with more than 100k miles and even those cars are just 3 years old. Time is a significant factor that we just don't know much about so far. Some research showed that some Lithium battery can do great for a while and then suddenly go bad when the build up of material (over time) closes microscopic gaps inside the cells. There is a lot of testing you can do with cells, but you cannot accelerate time and test 30 years in a matter of less than 30 years. Well according to relativity you can but it's not practical for a test lab.

Aljohn · Nov 21, 2015

It seems your chart is assuming the any degradation of the battery is linear. I am not sure with the 2+ years on the Model S, this assumption is valid. At this point we don't know. Assuming your chart is correct... there is a 15% lost in ability in 30 years. I have never had a ICE for 30 years. Two reasons: They don't last that long, and they degrade overall far greater than 15% due to valve, cam, piston, drive shaft, muffler and other degrading parts. So, if the chart is a worst case scenario, Tesla is better than any ICE owned.

vandacca · Nov 22, 2015

David99 said:
We don't have data from a 30 year old Tesla battery, so I wouldn't say the 'data seems to suggest that'. We actually have very few people with more than 100k miles and even those cars are just 3 years old. Time is a significant factor that we just don't know much about so far. Some research showed that some Lithium battery can do great for a while and then suddenly go bad when the build up of material (over time) closes microscopic gaps inside the cells. There is a lot of testing you can do with cells, but you cannot accelerate time and test 30 years in a matter of less than 30 years. Well according to relativity you can but it's not practical for a test lab.

There are many techniques used to accelerate longevity testing like increasing temperature, increasing charge/discharge rate, constant cycling, etc. They can perform decades worth of testing in a matter of months. BTW, earlier this year, Tesla announced a 5-year research plan with Jeff Dahn (in the previous link) to research Lithium Ion battery longevity.

With current Lithium Ion batteries, the degradation typically occurs due to parasitic reactions at the cathode that causes a solid to form, "gunking-up" the cathode. A very fast charge actually significantly reduces the time that those parasitic reactions can happen and is actually good for the battery, hence using SuperChargers is good for the battery's longevity.

Many battery manufacturers add special inert chemicals that help reduce the build-up of solids on the electrodes, and hence increase the life-span of the batteries, but none have been found so far to completely eliminate the build-up. If they could prevent the build-up, then some other process would be the weak link, and batteries could last almost indefinitely. What a wonderful world that would be.

Jeff Dahn gave a good talk at the University of Waterloo describing all this stuff:

MrBoylan · Nov 22, 2015

Aljohn said:
It seems your chart is assuming the any degradation of the battery is linear. I am not sure with the 2+ years on the Model S, this assumption is valid. At this point we don't know. Assuming your chart is correct... there is a 15% lost in ability in 30 years. I have never had a ICE for 30 years. Two reasons: They don't last that long, and they degrade overall far greater than 15% due to valve, cam, piston, drive shaft, muffler and other degrading parts. So, if the chart is a worst case scenario, Tesla is better than any ICE owned.

The chart only looks at battery performance over time, based on surveys done with existing owners. And yes, we don't have data past 2-3 years. But the battery degradation is definitely not linear. According to the studies, there is more loss toward the beginning and it slows down after the first 50K miles. But we'd need a lot more time to see if it gets better or worse over longer time periods. My estimates don't take time into consideration as it is not a known factor (yet).

And yes, there are many many other parts that could wear out (and will wear out, and may or may not be cost feasible to replace). But so far, the batteries seem to be holding up pretty well.

- - - Updated - - -

David99 said:
We don't have data from a 30 year old Tesla battery, so I wouldn't say the 'data seems to suggest that'. We actually have very few people with more than 100k miles and even those cars are just 3 years old. Time is a significant factor that we just don't know much about so far. Some research showed that some Lithium battery can do great for a while and then suddenly go bad when the build up of material (over time) closes microscopic gaps inside the cells. There is a lot of testing you can do with cells, but you cannot accelerate time and test 30 years in a matter of less than 30 years. Well according to relativity you can but it's not practical for a test lab.

I wouldn't say "data seems" anything because "data" is plural.

But to get back on track, do you know of any long term studies of similar Li Ion battery cells/battery packs over time? The only studies I've seen are based on charge/discharge cycles, but even there, there is a huge range in performance.

I suspect part of what makes the Tesla packs hold up better over time than (say) the Nissan Leaf batteries is probably better thermal management and I'm guessing there is also some internal charge management that prevents the individual cells from being at extremes (100% or 0%) for an extended period of time. Of course, it's impossible to prevent the battery from getting to 0% if it's left unplugged for an extended period of time, but it is certainly easy enough for Tesla's battery management to report 3 miles of range, when there are really 30 miles of range or "100% charge" when the battery pack is actually 95% charged, thereby preventing the battery from reaching its extremes. In other words, couldn't the 85kWh pack really be a 90 kWh pack which is software limited to 85kWh? Wouldn't this prolong the battery life?

Maybe all this has already been published in a white paper somewhere. Or maybe it's a trade secret. But so far it does seem like the Tesla battery packs are holding up better than expected over time/miles, and certainly better than most of the competition. Again, if there are studies or data to the contrary, please post a link.

-CB

MrBoylan · Nov 23, 2015

vandacca said:
There are many techniques used to accelerate longevity testing like increasing temperature, increasing charge/discharge rate, constant cycling, etc. They can perform decades worth of testing in a matter of months. BTW, earlier this year, Tesla announced a 5-year research plan with Jeff Dahn (in the previous link) to research Lithium Ion battery longevity.

Thanks for the comment and the links. I recall watching this video shortly after the partnership announcement was announced. Dahn seems like a really smart, reasonable guy so it's great to see him involved with Tesla on battery research. The battery systems in the Tesla in particular, and in EVs in general will only get better over time.

-CB

MrBoylan · Nov 23, 2015

vandacca said:
Jeff Dahn gave a good talk at the University of Waterloo describing all this stuff:

I watched it again and am even more excited that this guy is working directly with Tesla on battery research. Even here, before his affiliation with Tesla, he gives kudos to Tesla for their research, choice of battery chemistry and thermal pack management. And in the video he documents that the Nissan Leaf problems with accelerated capacity loss are a combination of poor choice of electrolyte additives and a lack of thermal management. It's amazing how different the results can be just by using the right additives in the electrolytes. There are apparently some Medtronic Lithium Ion batteries used internally (in the human body) that have gone through over 20,000 cycles of testing over 8 years at 37 degrees Celsius (body temperature), which are still able to get about 60% of their original capacity. And Dahn said that the cells in Tesla packs are similar in chemistry and "at least as good" as these.

vandacca · Nov 23, 2015

MrBoylan said:
I watched it again and am even more excited that this guy is working directly with Tesla on battery research. Even here, before his affiliation with Tesla, he gives kudos to Tesla for their research, choice of battery chemistry and thermal pack management. And in the video he documents that the Nissan Leaf problems with accelerated capacity loss are a combination of poor choice of electrolyte additives and a lack of thermal management. It's amazing how different the results can be just by using the right additives in the electrolytes. There are apparently some Medtronic Lithium Ion batteries used internally (in the human body) that have gone through over 20,000 cycles of testing over 8 years at 37 degrees Celsius (body temperature), which are still able to get about 60% of their original capacity. And Dahn said that the cells in Tesla packs are similar in chemistry and "at least as good" as these.

I know, it's a really good talk. But just a point of clarification: Tesla hasn't yet started working with Prof Dahn, probably sometime next summer. Here is a snippet of the earlier news article.

Dalhousie said the exclusive partnership with Tesla will begin in a year's time. In the meantime, Dahn will continue to work as the Industrial Research Chair in Materials for Advanced Batteries, which is funded by 3M Canada and the Natural Sciences and Engineering Research Council of Canada.

Canuck · Nov 23, 2015

MrBoylan said:
And in the video he documents that the Nissan Leaf problems with accelerated capacity loss are a combination of poor choice of electrolyte additives and a lack of thermal management.

Plus, add to that a much smaller battery to start with in the Leaf and it's really sad for the Leaf. This is why they had to provide the $5k battery replacement. It's also the reason I went with the 85kW battery for the Tesla. Even with better additives and thermal management degradation will happen. If many years down the line, I end up with the capacity of a 60kW that's fine. I will have upgraded before then anyway but the larger battery allows for more degradation. It's an all out failure after the warranty expires that is the main concern.

MrBoylan · Nov 23, 2015

vandacca said:
I know, it's a really good talk. But just a point of clarification: Tesla hasn't yet started working with Prof Dahn, probably sometime next summer. Here is a snippet of the earlier news article.

Doh. My original post said "will be" but then I edited it to say "is" because I assumed it already started. Thanks for the clarification.

-CB

PoweredByRain · Nov 25, 2015

vandacca said:
There are many techniques used to accelerate longevity testing like increasing temperature, increasing charge/discharge rate, constant cycling, etc. They can perform decades worth of testing in a matter of months.

I don't think that's accurate. They can perform many cycles, or what they think will be the same effect as decades of use. But there really is no guarantee that all variables have been accounted for, and all possible mechanisms understood. Heck, maybe too many years of cosmic rays cause sudden failure... Just being silly, and perhaps pedantic, but the point David99 made I think still stands: we don't have data from a 30 year old Tesla battery. I personally don't think there is any way you can possibly extrapolate that far out at this point. Not with really high confidence.

BornToFly · Nov 26, 2015

When I sold my 2 year old Model S, there was no loss of range that I could detect. In fact, my new P85D that replaced it shows 12 miles less range every day than my P85+ did when I sold it.

vandacca · Nov 26, 2015

PoweredByRain said:
I don't think that's accurate. They can perform many cycles, or what they think will be the same effect as decades of use. But there really is no guarantee that all variables have been accounted for, and all possible mechanisms understood. Heck, maybe too many years of cosmic rays cause sudden failure... Just being silly, and perhaps pedantic, but the point David99 made I think still stands: we don't have data from a 30 year old Tesla battery. I personally don't think there is any way you can possibly extrapolate that far out at this point. Not with really high confidence.

I take it you didn't watch Prof. Jeff Dahn's 73 minute lecture? This is exactly what he is saying, that up until now, nobody was doing the proper research into why/how cells degrade. He has decided to tackle this exact problem. In doing so, they now understand the factors that affect cell longevity. He had to build a huge, very precise, coloumbic machine to do that. Cell characterization is a long and hard problem to solve, but he is doing it for all types of lithium ion cells.

A number of his team members are now Tesla employees and starting middle next year, Prof Dahn is going to be directly working with Tesla.

With all the cell characterization work done, they can accurately predict the life-span of the cells. The trick is to have a smart battery controller and heat management system so that the cells always stay in their ideal zone when charging, so that the above predictions stay true. This is one of Tesla's areas of competence. The Nissan Leaf didn't have such a controller/system so their batteries degraded very quickly (in some cases) and a lawsuit ensued.

Knowing that a battery only degrades during charging, you can focus on that part of the equation. In 30 years, if you charge every other day, you'll looking at about 5,475 charging cycles. You can probably discharge/charge a single battery cell in 15 minutes. You can then reduce 30 years to 60 days of testing. But if you precisely measure and graph the cell's characteristics during that time, you'll see a very predictable curve and you can use that to extrapolate even further. Therefore, in the future, you could get away with even less testing and more extrapolation.

If one takes the time to research and understand the physics behind the processes, the possibilities are endless.

MrBoylan · Nov 26, 2015

vandacca said:
This is exactly what he is saying, that up until now, nobody was doing the proper research into why/how cells degrade.

Yeah, the breakthrough (I think) in his thinking was that the parasitic reaction that causes cells to degrade generates heat, so if you can very precisely measure exactly how much heat is generated during charging, you know how much of the parasitic reaction is happening during the charge and you can extrapolate from that precisely how quickly (or slowly) the batteries will degrade. His projections for specific electrolyte chemistries matched up incredibly closely with the *actual* longevity tests done by the manufacturers over time.

This drastically shortens the development cycle by allowing manufacturers to test different electrolyte additives and have the results in weeks instead of years. What was interesting to me was that his research shows that faster charging times actually *reduce* that parasitic reaction, thereby reducing capacity degradation. This is counter to what some have speculated that using super chargers too often would lead to premature battery failure.

David99 · Nov 26, 2015

MrBoylan said:
What was interesting to me was that his research shows that faster charging times actually *reduce* that parasitic reaction, thereby reducing capacity degradation.

If I remember correctly he explained it as 'just beating the clock'. Faster charging and discharging reduces overall time. The better results were not due to faster charging but less time overall. Maybe I'm wrong. Do you remember the time in the video where he talks about it?

vandacca · Nov 26, 2015

David99 said:
If I remember correctly he explained it as 'just beating the clock'. Faster charging and discharging reduces overall time. The better results were not due to faster charging but less time overall. Maybe I'm wrong. Do you remember the time in the video where he talks about it?

It was near the beginning of the talk. He started explaining the concept at ~6:06 and mentions "beating the clock" at ~6:25.

Search

Model X Range Degradation Over Time - Estimates and Predictions

MrBoylan

Member

David99

Well-Known Member

Aljohn

Member

vandacca

ReActive Member

MrBoylan

Member

MrBoylan

Member

MrBoylan

Member

vandacca

ReActive Member

Canuck

Well-Known Member

MrBoylan

Member

PoweredByRain

Member

BornToFly

Well-Known Member

vandacca

ReActive Member

MrBoylan

Member

David99

Well-Known Member

vandacca

ReActive Member

Similar threads