Battery Degradation Behavior
- Thread starter Ray B
- Start date
- Replies 37
- Views 7K
Ray B
Well-Known Member
Hi @MrFixit
My loss estimate was based on my recharging data which showed an approximate kWh/deltaSoC at ~0.165 kWh when the battery was new. I have seen a variety of numbers for the kWh for a full recharge when new, which could put that number anywhere from ~0.160 - 0.168 kWh I think. But in the Excel table you can make the adjustment in the calculation of the estimated capacity.
Also I haven't had too many issues with HondaLink's SoC readout. Sometimes it doesn't update quickly and I may need to refresh it to ping for a new value. If it doesn't update (looking at the date and time of the last update) I may need to reboot the car for a minute to get another HondaLink update or alternatively just use my OBD-II/TorquePro. In general, I seldom have that issue and 19 times out of 20 I can get the current SoC from HondaLink without a fuss.
Ray B
Well-Known Member
Ray B - your work on this is incredible- thanks again! Based on the data I previously shared, how would my Clarity Battery longevity compare using your revised assumptions. Also let me know if it would be helpful to continue to report my chargepoint KW charge activity when the battery is fully depleted.
MrFixit
Well-Known Member
@Ray B,
I am seeing something very weird with HondaLink. Yesterday (4/10), I pulled into the garage and plugged in. The charge began at 3:25 PM. The charge ended at 05:23 PM.
When I looked at HondaLink this morning (4/11), it showed 100% with a last update at 5:23 AM on 4/10, but the vehicle location was last updated at 5:23 PM on 4/10 (plausible as this was the end of charging).
After playing around for maybe 15 minutes attempting to update on both the main page and the location (to no avail), then it suddenly did update both to the current time (10:49 AM today). Never touched the vehicle throughout this process. It just seems to always be unresponsive or VERY sluggish for me. I believe we have a good signal, but not sure how to know.
I am seeing something very weird with HondaLink. Yesterday (4/10), I pulled into the garage and plugged in. The charge began at 3:25 PM. The charge ended at 05:23 PM.
When I looked at HondaLink this morning (4/11), it showed 100% with a last update at 5:23 AM on 4/10, but the vehicle location was last updated at 5:23 PM on 4/10 (plausible as this was the end of charging).
After playing around for maybe 15 minutes attempting to update on both the main page and the location (to no avail), then it suddenly did update both to the current time (10:49 AM today). Never touched the vehicle throughout this process. It just seems to always be unresponsive or VERY sluggish for me. I believe we have a good signal, but not sure how to know.
Ray B
Well-Known Member
Thanks for the endorsement, @jpkik96 !
I filled in 31000 miles, 11% SoC at Start, and 100% SoC finish, with 13.01 kWh, and it estimated 11.1% loss currently, and estimated to hit 33.3% loss at 280,368 miles.
Just a 'back-of-the-envelope' model but I think it is probably closer to what will happen compared to the previous model. The more data that is collected over the years or miles, the more confident we can make the projections, and the better we can tweak the model.
Ray B
Well-Known Member
Hmmm - not sure what the trouble could be. I think there was an update to HondaLink in the fall and any update problems I was having seemed to be resolved. Perhaps you need to check for an update.
Also, when I was occasionally having server update problems I would clear the data (Android settings>Apps>HondlaLink>Storage> Clear cache or Clear Data). If you clear data you may need to login again.
If you're using iOS I can't help.
Ray B
Well-Known Member
In this update to the Capacity Loss Estimation model I include a line that involves the scaling factor that determines how quickly the loss gets depleted. I call this a 'Stress Factor' and was the "a" in my modeling function. This Stress Factor 'a' (SFa for short) is dependent upon the driving patterns, the charging behavior, not to mention the quality of the battery itself.
So there are many things that may lessen the stress on the EV battery - moderate temperatures during use and charging, limiting depletion (depth of discharge, 'DoD') and maximum charge, avoiding extended time with full SoC or very low SoC, moderate use of hard acceleration and aggressive regen, etc.
My driving is fairly sedate (limited EV on the highway, ECO mode, easy on the throttle usually), and I normally don't go all the way down to 0 EV before recharging, and I limit most of my recharges to 93% SoC. This plot from the paper I cited earlier explains why.
By doing a full discharge of the battery and then a full recharge, the use pattern would probably equate to the green line in a best case scenario, or perhaps worse than that. I believe the BMS buffer at the top and bottom prevent the more extreme curves. By recharging before 0 EV and topping off at 93% instead of 100%, I think I am easing the stress on the battery and following more along the light blue or magenta lines. The end result should extend my capacity longer than if I went from 0 EV to 100% routinely. Note that we don't recharge at 1C, but the key takeaway from the graph is the effect of the breadth of the recharge cycle affects the capacity loss.
Right now my SFa = 0.048
Attached below is my updated Excel model with the SFa calculated at the bottom.
So there are many things that may lessen the stress on the EV battery - moderate temperatures during use and charging, limiting depletion (depth of discharge, 'DoD') and maximum charge, avoiding extended time with full SoC or very low SoC, moderate use of hard acceleration and aggressive regen, etc.
My driving is fairly sedate (limited EV on the highway, ECO mode, easy on the throttle usually), and I normally don't go all the way down to 0 EV before recharging, and I limit most of my recharges to 93% SoC. This plot from the paper I cited earlier explains why.
By doing a full discharge of the battery and then a full recharge, the use pattern would probably equate to the green line in a best case scenario, or perhaps worse than that. I believe the BMS buffer at the top and bottom prevent the more extreme curves. By recharging before 0 EV and topping off at 93% instead of 100%, I think I am easing the stress on the battery and following more along the light blue or magenta lines. The end result should extend my capacity longer than if I went from 0 EV to 100% routinely. Note that we don't recharge at 1C, but the key takeaway from the graph is the effect of the breadth of the recharge cycle affects the capacity loss.
Right now my SFa = 0.048
Attached below is my updated Excel model with the SFa calculated at the bottom.
Last edited:
Cash Traylor
Well-Known Member
@Ray B
Dude, that rocks!
I just realized that during my late night early morning diatribe post I read the research paper, while drinking my Scotch - THEN posted on the forum. Have to remember to read paper, post, then drink... In doing so I failed to mention an important part of my "plan." I never intended to "train" this model using a lot of people. We are not looking for that kind of accuracy. I meant for 5-10 to take it, get the battery caps and input data, adjust the curve to approximate the results experienced by a limited but representative set (2018's, 2019's 2020's) and then just let it run and see what happens. Mr. Fixit is right, this is just to give someone an idea when it is time to get a real check - or make sure you don't miss your warranty. Putting purchase data and (WARRANTY EXPIRES "HERE") at the top of the sheet is likely a good idea. Either way I know it would never replace real live data. I'll keep working on the decoding of the PID, someone else may beat me to it at this rate.
Ray, that version two includes some nice refinements! So, just curious - what do you do for a living or are you just a maths person!?! Either way I am impressed. Sorry folks, never intended for this to turn into a deep machine learning project - just a forum's attempt at a very slick educated guess!
Cheers - and yes my head hurts,
Cash
Dude, that rocks!
I just realized that during my late night early morning diatribe post I read the research paper, while drinking my Scotch - THEN posted on the forum. Have to remember to read paper, post, then drink... In doing so I failed to mention an important part of my "plan." I never intended to "train" this model using a lot of people. We are not looking for that kind of accuracy. I meant for 5-10 to take it, get the battery caps and input data, adjust the curve to approximate the results experienced by a limited but representative set (2018's, 2019's 2020's) and then just let it run and see what happens. Mr. Fixit is right, this is just to give someone an idea when it is time to get a real check - or make sure you don't miss your warranty. Putting purchase data and (WARRANTY EXPIRES "HERE") at the top of the sheet is likely a good idea. Either way I know it would never replace real live data. I'll keep working on the decoding of the PID, someone else may beat me to it at this rate.
Ray, that version two includes some nice refinements! So, just curious - what do you do for a living or are you just a maths person!?! Either way I am impressed. Sorry folks, never intended for this to turn into a deep machine learning project - just a forum's attempt at a very slick educated guess!
Cheers - and yes my head hurts,
Cash
@Ray B -
Thanks for providing your revised estimate. Assuming I can maintain 70-80% of the Clarity's all-electric range for up to 8 years/120M miles (when the HondaCare warranty expires certainly makes a compelling argument to consider buying the Clarity after the lease expires at the $16,500 residual price.
I will share additional battery data the next time I fully charge it and will also have my local dealer complete a load test the next time it is deemed safe enough in NJ to have it serviced and share that data as well.
Thanks again!
Jay
Thanks for providing your revised estimate. Assuming I can maintain 70-80% of the Clarity's all-electric range for up to 8 years/120M miles (when the HondaCare warranty expires certainly makes a compelling argument to consider buying the Clarity after the lease expires at the $16,500 residual price.
I will share additional battery data the next time I fully charge it and will also have my local dealer complete a load test the next time it is deemed safe enough in NJ to have it serviced and share that data as well.
Thanks again!
Jay
Cash Traylor
Well-Known Member
@jpkik96
Ok, this is all fun and games until you are talking about money!
First, the work Ray did is awesome, especially since all I have done is talk and vomit links and data, he built a tool! However, if you are going to buy the vehicle off lease, you need a real battery capacity check/report from the dealer's i-HDS system (see my thread here about the report).
https://www.insideevsforum.com/comm...-test-reports-ongoing-data-for-warranty.8317/
I would consider this critical "condition of my PHEV vehicle" type stuff before I take ownership. Just because the PDI was part of a new vehicle purchase, the concept for a lease holder planning to buy is the same - you need to know the current state of the HV battery "for realzz man." Get that data before you sign for the car!
Second, if (and I hope you do) get that report. Please forward the data to us either here or on my other thread. We need year, age, mileage, and capacity data. It would help compare Ray's model you are already looking at to the hard data point that is the "only thing that matters" regarding your HV pack replacement warranty from Honda.
Third, I am working up a new concept for determining the pack wear level and your data would be very useful. I have to reach out to Ray about this and plan to tomorrow. However his work got me to thinking about the calculation of what he termed "SFa" in a pretty cheap and easy way, without the need for a data reporting charger.
So, your information is appreciated and good luck with the change from lease to own!
Cheers,
Cash
Ok, this is all fun and games until you are talking about money!
First, the work Ray did is awesome, especially since all I have done is talk and vomit links and data, he built a tool! However, if you are going to buy the vehicle off lease, you need a real battery capacity check/report from the dealer's i-HDS system (see my thread here about the report).
https://www.insideevsforum.com/comm...-test-reports-ongoing-data-for-warranty.8317/
I would consider this critical "condition of my PHEV vehicle" type stuff before I take ownership. Just because the PDI was part of a new vehicle purchase, the concept for a lease holder planning to buy is the same - you need to know the current state of the HV battery "for realzz man." Get that data before you sign for the car!
Second, if (and I hope you do) get that report. Please forward the data to us either here or on my other thread. We need year, age, mileage, and capacity data. It would help compare Ray's model you are already looking at to the hard data point that is the "only thing that matters" regarding your HV pack replacement warranty from Honda.
Third, I am working up a new concept for determining the pack wear level and your data would be very useful. I have to reach out to Ray about this and plan to tomorrow. However his work got me to thinking about the calculation of what he termed "SFa" in a pretty cheap and easy way, without the need for a data reporting charger.
So, your information is appreciated and good luck with the change from lease to own!
Cheers,
Cash
Cash - I am sorry I neglected to acknowledge your many insightful posts yesterday along with many others I have followed since I joined this forum. The fact of the matter is my brain usually hurts after reading some of the research you add to your posts! However, the fact that you, Ray and so many others provide useful insight in how the Clarity works and what steps one can take to improve its future operation has given the confidence to consider purchasing it!
As a long-time car enthusiast (my former car was an Acura TL 6speed and my other car is 2015 Mustang GT Conv. 6 speed), I am secretly hoping one of you figure out in a few years how to retrofit a 50 KW solid-state battery that is lighter so I can mod into the ultimate PHEV...thanks again for all you and so many others contribute to this forum - Happy Easter all!
Jay
As a long-time car enthusiast (my former car was an Acura TL 6speed and my other car is 2015 Mustang GT Conv. 6 speed), I am secretly hoping one of you figure out in a few years how to retrofit a 50 KW solid-state battery that is lighter so I can mod into the ultimate PHEV...thanks again for all you and so many others contribute to this forum - Happy Easter all!
Jay
Clarity_Newbie
Well-Known Member
@Ray B
Below is c/p of data off v2a spreadsheet you devised.
Charging conditions:
Ambient temp start of charge: 58 F...end of charge 61 F (rare night where temps didn't vary which is why I choose it)
SoC start of charge: 8.7%...end of charge 100%
kWh: 14.5
Length of charge 11 hours 29 minutes
EV miles at end of charge 56
Level 1 OEM charger used
Car Age 20 months
Odometer 15900 miles
EV SoC at Start of Charge 8.7 %
EV SoC at End of Charge 100 %
Level 2 Energy Used 14.5 kWh
Estimated kWh/deltaSoC 0.159 kWh
Estimated Capacity 53.1 Ah
Estimated Loss of Capacity 3.4 %
Estimated odometer at 33.3% EV Capacity Loss: 1531615 miles
Interestingly, the BCT performed after one year ownership in September 2019 showed capacity at 54.3Ah and ~7 months later v2a reports a 2.2% drop to 53.1Ah which seems inline with expectations of loss over time.
As I've stated before...I personally question the BCT as an absolute...it may be more of the old adage "close enough for government work"...(ie) if variance is ~0.5% either way...that changes the number significantly. Dunno...just a thought.
Link to earlier post reference Clarity ownership with driving stats n habits etc. Post #3 discusses BCT last September.
https://www.insideevsforum.com/community/threads/one-year-ownership-stats.6853/
Also found in the previous thread is miles per kWh which came in at an average of 3.69...I did a quick calculation and figure over the past 5900 miles (3850 miles on EV)...the average miles per kWh = 3.71. Does this represent another sign of decent battery health??? Dunno.
The numbers for the Clarity I drive should not be a surprise given the relative low mileage.
What I would love to see is data from 1-2 year old Clarity's where the owners drive exclusivity on EV/rarely use ICE. That would be great and very helpful!!!
My prognosis for the Clarity I drive: I fully expect to see a return to 12 hour charge and 15+ kWh when ambient temps increase in the coming months. I will update the thread in a few months with actual numbers.
Great job on the spreadsheet and hope this info helps future discussions.
Below is c/p of data off v2a spreadsheet you devised.
Charging conditions:
Ambient temp start of charge: 58 F...end of charge 61 F (rare night where temps didn't vary which is why I choose it)
SoC start of charge: 8.7%...end of charge 100%
kWh: 14.5
Length of charge 11 hours 29 minutes
EV miles at end of charge 56
Level 1 OEM charger used
Car Age 20 months
Odometer 15900 miles
EV SoC at Start of Charge 8.7 %
EV SoC at End of Charge 100 %
Level 2 Energy Used 14.5 kWh
Estimated kWh/deltaSoC 0.159 kWh
Estimated Capacity 53.1 Ah
Estimated Loss of Capacity 3.4 %
Estimated odometer at 33.3% EV Capacity Loss: 1531615 miles
Interestingly, the BCT performed after one year ownership in September 2019 showed capacity at 54.3Ah and ~7 months later v2a reports a 2.2% drop to 53.1Ah which seems inline with expectations of loss over time.
As I've stated before...I personally question the BCT as an absolute...it may be more of the old adage "close enough for government work"...(ie) if variance is ~0.5% either way...that changes the number significantly. Dunno...just a thought.
Link to earlier post reference Clarity ownership with driving stats n habits etc. Post #3 discusses BCT last September.
https://www.insideevsforum.com/community/threads/one-year-ownership-stats.6853/
Also found in the previous thread is miles per kWh which came in at an average of 3.69...I did a quick calculation and figure over the past 5900 miles (3850 miles on EV)...the average miles per kWh = 3.71. Does this represent another sign of decent battery health??? Dunno.
The numbers for the Clarity I drive should not be a surprise given the relative low mileage.
What I would love to see is data from 1-2 year old Clarity's where the owners drive exclusivity on EV/rarely use ICE. That would be great and very helpful!!!
My prognosis for the Clarity I drive: I fully expect to see a return to 12 hour charge and 15+ kWh when ambient temps increase in the coming months. I will update the thread in a few months with actual numbers.
Great job on the spreadsheet and hope this info helps future discussions.
90% in EV mode, charging all at ~7.15kwh L2. The daily commute (pre-virus) used to be basically an entire battery (+/-) depending on temperature. Seems like the 1st year degradation is pretty steep.
Car Age 14 months
Odometer 13701 miles
EV SoC at Start of Charge 9 %
EV SoC at End of Charge 100 %
Level 2 Energy Used 13.5 kWh
Estimated kWh/deltaSoC 0.148 kWh
Estimated Capacity 49.6 Ah
Estimated Loss of Capacity 9.8 %
Estimated odometer at 33.3% EV Capacity Loss: 159722 miles
Car Age 14 months
Odometer 13701 miles
EV SoC at Start of Charge 9 %
EV SoC at End of Charge 100 %
Level 2 Energy Used 13.5 kWh
Estimated kWh/deltaSoC 0.148 kWh
Estimated Capacity 49.6 Ah
Estimated Loss of Capacity 9.8 %
Estimated odometer at 33.3% EV Capacity Loss: 159722 miles
Cash Traylor
Well-Known Member
So, maybe a predictive performance calculator to balance out a curve estimation calculator? Still working on a way to get what matters, the BCT on OBDII...
I'm not going to create a new thread yet, although @Ray B 's tool here deserves it's own thread in the "Clarity Useful Tools" read first sticky! A few folks here helped with the initial shakedown of a formula I was working on (thanks Ray). I want to include it as a "page two" on Ray's excel sheet, as a way to pre-load performance and allow folks to play with some variables. I think my formula still needs work, but it is close to modeling some industrial Panasonic NMC cells and although we are missing specific data I hope to get it soon. I was really bummed this week when I checked to find out that AZ battery had sold his last Clarity pack to a local walk up. I was supposed to get one on order, as soon as shipping resumed, but apparently I am going to have to wait for another module to become available. The problem is I need a pretty new module for my tests to mean anything. I have a full LIB test rig on my bench to categorize the SOC/DOD/EOL/ESR of the cell (once I cut it out of the pack). Don't worry, I have uses for the batteries after the tests, but figured I could kill two birds with one stone if I bought a Clarity HV module versus something else.... ugh - missed opportunities!
Anyway, here is a performance calculator that I have been modeling for later use in another project (battery performance tool). I put the math in this online calculator that I found by accident. I am not subscribed to this calculoid thing, as I have no use for it long term and it is not cheap. I have no idea how long this "calculator" will remain online and available after my free 30 days are up. So, feel free to go play with it. @MrFixit recommended I do a "basic" and advanced version so you can try it out. However note the assumptions, it bases everything on starting at zero. I didn't bother to put starting mileage on it. You can reduce the starting HV Capacity, as technically you can get a new car with a PDI that is well, anything above 36.6 Ah and that is ok - of course I wouldn't buy it... However if your PDI was 53.5Ah, then that is where you put that. If it is more than 54.6 then your battery is slightly above nominal OEM capacity as far as I know.
I am more interested in if you select the variables that fit an imaginary starting point, how accurate are the results to where you happen to be today? You can put your last BCT as the "EOL point" and see if the age and mileage match yours. If you want to use it for a "I am here" to see where you will be, then you have to adjust the data and add your current mileage to the reported. It only reports EV mileage so if you drive HV, then you will have more miles than reported at failure. This is just trying to estimate the minimum mileage you would have if you drove ONLY in EV mode. Since there are a fair number of people here that do that (not me, range is to low for my driving style) this may be a tool those can test and report back.
https://app.calculoid.com/#/calculator/78363
I plan to put this in a spreadsheet at some point so it can be used off line, this was just easier to test and share.
Folks this is in beta, and likely not within 10% of actual. We/I just don't have enough data on the Clarity yet. No one has run into a battery warranty event yet....
Anyway, before anyone writes back to say how wrong they think it is... This software tool comes with no warranty, stated or implied, as to its accuracy or fitness for purpose of intents of use. User assumes all risk for its predictions and at no time does the author make any claim as to its value. You get what you pay for, and blah blah blah...
Cheers,
Cash
I'm not going to create a new thread yet, although @Ray B 's tool here deserves it's own thread in the "Clarity Useful Tools" read first sticky! A few folks here helped with the initial shakedown of a formula I was working on (thanks Ray). I want to include it as a "page two" on Ray's excel sheet, as a way to pre-load performance and allow folks to play with some variables. I think my formula still needs work, but it is close to modeling some industrial Panasonic NMC cells and although we are missing specific data I hope to get it soon. I was really bummed this week when I checked to find out that AZ battery had sold his last Clarity pack to a local walk up. I was supposed to get one on order, as soon as shipping resumed, but apparently I am going to have to wait for another module to become available. The problem is I need a pretty new module for my tests to mean anything. I have a full LIB test rig on my bench to categorize the SOC/DOD/EOL/ESR of the cell (once I cut it out of the pack). Don't worry, I have uses for the batteries after the tests, but figured I could kill two birds with one stone if I bought a Clarity HV module versus something else.... ugh - missed opportunities!
Anyway, here is a performance calculator that I have been modeling for later use in another project (battery performance tool). I put the math in this online calculator that I found by accident. I am not subscribed to this calculoid thing, as I have no use for it long term and it is not cheap. I have no idea how long this "calculator" will remain online and available after my free 30 days are up. So, feel free to go play with it. @MrFixit recommended I do a "basic" and advanced version so you can try it out. However note the assumptions, it bases everything on starting at zero. I didn't bother to put starting mileage on it. You can reduce the starting HV Capacity, as technically you can get a new car with a PDI that is well, anything above 36.6 Ah and that is ok - of course I wouldn't buy it... However if your PDI was 53.5Ah, then that is where you put that. If it is more than 54.6 then your battery is slightly above nominal OEM capacity as far as I know.
I am more interested in if you select the variables that fit an imaginary starting point, how accurate are the results to where you happen to be today? You can put your last BCT as the "EOL point" and see if the age and mileage match yours. If you want to use it for a "I am here" to see where you will be, then you have to adjust the data and add your current mileage to the reported. It only reports EV mileage so if you drive HV, then you will have more miles than reported at failure. This is just trying to estimate the minimum mileage you would have if you drove ONLY in EV mode. Since there are a fair number of people here that do that (not me, range is to low for my driving style) this may be a tool those can test and report back.
https://app.calculoid.com/#/calculator/78363
I plan to put this in a spreadsheet at some point so it can be used off line, this was just easier to test and share.
Folks this is in beta, and likely not within 10% of actual. We/I just don't have enough data on the Clarity yet. No one has run into a battery warranty event yet....
Anyway, before anyone writes back to say how wrong they think it is... This software tool comes with no warranty, stated or implied, as to its accuracy or fitness for purpose of intents of use. User assumes all risk for its predictions and at no time does the author make any claim as to its value. You get what you pay for, and blah blah blah...
Cheers,
Cash
Cash Traylor
Well-Known Member
Added starting mileage to the calculator. However, it should only be inputted if you have a BCT report (with corresponding current mileage at that time) to permit the results to remain plausible, as those are interrelated factors. So, advanced mode, input adjusted BCT Ah and mileage for a current report on "your" vehicle. I would LOVE for @StickWare to get a BCT and put his miles in, that data would be really cool as he is the first person out of battery warranty! But wish in one hand and ..... 
Cheers,
Cash
Cheers,
Cash