HVB Degradation

[larryh]

The following chart shows the degradation rates for an NCA Lithium-ion battery stored at a constant SOC and temperature (http://www.nrel.gov/transportation/energystorage/pdfs/53817.pdf). A description of the different types of Lithium-ion batteries can be found at the following link:  http://www.batteryuniversity.com/learn/article/types_of_lithium_ion.    This is the type of battery used in the Tesla Model S (http://batteryuniversity.com/learn/article/bu_808b_what_causes_li_ion_to_die).   The type used in the Energi, Volt, and LEAF is NMC/Lithium Manganese Oxide (http://articles.sae.org/11705/).  I suspect that the Energi HVB has higher degradation rates than NCA, but the degradation rates should be comparable.  

 

The top chart shows resistance growth which correlates with power loss.  A 20% increase in resistance results in a 17% power output loss from the HVB.  The bottom chart shows capacity fade rate which correlates with capacity loss.  As an example how to read the bottom chart, consider a battery stored at 25 C and 100% SOC.  The capacity fade rate from the chart is -4.7e-3/sqrt(day).  To determine the time to 20% capacity loss, we compute (0.2/4.7e-3)^2 = 1810 days = 5 years.   If you wish to compute the capacity degradation for a battery stored for one year (365 days) at 100% SOC and 30 C, look at the 100% SOC line in the bottom graph where it intersects a vertical line drawn through 30 C on the x-axis.  This occurs at -5.2e-3/sqrt(day) fade rate on the y-axis.  The degradation is then sqrt(365)*5.2e-3 = 10%.  

 

The dashed horizontal lines indicate the lifetimes for 20% capacity loss or resistance growth.  For example, consider capacity loss.  The 10 year line crosses the 80% SOC line at about 27 C in the bottom chart.  That means the battery capacity loss will be 20% after 10 years if the battery is stored at 80% SOC and 27 C.  The point on the 80% SOC line corresponding to 20 C is above the 10 year capacity loss line.  That means that degradation of the battery will less than 20% after 10 years if stored at 80% SOC and 20 C.  

 

Based on the charts below, if you store the battery at 50% SOC for 10 years, the power loss and capacity loss should be less than 20% for HVB temperatures up to 35 C, or 95 F.  The normal recommended SOC for storing Lithium-ion batteries is between 30% and 40%.  If you store the battery at 30% SOC for ten years, degradation should be less than 20% for HVB temperatures up to 45 C, or 113 F.  

 

To minimize degradation, ideally you would want to store the battery at 0% SOC.  But that is not recommended, especially for batteries that are connected to devices.  Should the cell voltage fall below 3.0 V, the battery will be irreparably damaged.  That can happen if the car draws too much power from the HVB when the SOC falls below 0%.  So I would not want to drain the HVB to 0% SOC.

 

Note that when MFT/MFM reports 0% SOC of the HVB, the actual SOC is about 20%.  So you need to take that into account when using the SOC reported by MFT/MFM.  

 

Edited June 25, 2016 at 09:38 PM by larryh

[larryh]

As an example of how we can use the chart to determine HVB degradation, consider my typical commute to work.  I charge the HVB right before I leave for work.  I park at work for about 9 hours.  Then I leave the battery uncharged until the next day.  While at work, the SOC of the battery is around 70%.  At night, it is around 40%.  The HVB temperature is generally around 10 F warmer than the outside temperature.  So let us assume a typical summer day in Minnesota at 85 F.  The HVB temperature will be around 95 F.   So as an approximation, I have the HVB at 95 F and 70% SOC for 9 hours and at 95 F and 40% SOC for the remaining 15 hours.   Using the chart, the average capacity fade rate will be:

 

9/24*(-3.6e-3)+15/24*(-2.6e-3)= -2.98e-3/sqrt(day).  

 

That corresponds to 12 years until 20% capacity degradation.  Since, it is not summer all year, and temperatures are much colder in the winter, degradation should be less than 20% after 13 years.   Note that this only accounts for degradation when the battery is not cycling, i.e. not charging/discharging.  I would have to compute the degradation due to cycling and add that to the 20% degradation. The actual degradation of my HVB has been 5% in 3 years.  Using a degradation rate of -2.98e-3/sqrt(day), the model predicts degradation is 10% after three years.  The prediction is off because I need to lower the capacity fade rate to account for winter.

 

Now suppose that I charged the battery immediately when I arrived at work and when I returned home, so basically the SOC of the battery is 100% the entire day.    The fade rate from the chart for 95 F and 100% SOC is -5.9e-3/sqrt(day).  At that rate, I would reach 20% degradation in 3 years!!!  Quite a difference from 12 years.  (Of course, the battery temperature wouldn't be 95 F in the middle of January so the actual degradation would be less in Minnesota--but it may be 95 F in some portions of the country in January.)  Don't leave your battery fully charged for any longer than necessary.  Delay charging until the last possible moment.  

Edited June 25, 2016 at 11:24 PM by larryh

[larryh]

Now consider someone in Phoenix, AZ.  During the summer, the HVB temperature will be around 45 C and during the winter around 27 C.  At 100% SOC, that corresponds to capacity fades rates of -7.4e-3 and -4.8e-3 (the value for 45 C is off the chart so I have to guess).  Assuming 50% of the time is winter and 50% is summer, the average degradation rate is the average of those two numbers or -6.1e-3/sqrt(day).  At that rate, the battery capacity degradation will reach 20% in 2.9 years.   The actual degradation will be greater than 20% after 2.9 years since I have not included degradation due to cycling.  A test of the Nissan LEAF driven in Phoenix, AZ can be found here:  https://avt.inl.gov/sites/default/files/pdf/vehiclebatteries/FastChargeEffects.pdf.  In this test, the actual battery degradation was 25% after 1.5 years.  Cycling was probably responsible for 8% of the 25% degradation.  

 

Now suppose that the person in Phoenix does not charge the car and drives it as a hybrid so the HVB SOC is approximately 20% all year round.  The average summer/winter capacity fade rate will then be approximately -2.4e-3/sqrt(day).  Battery degradation will reach 20% after approximately 19 years, i.e. the battery lifetime is 8 times longer at 20% vs 100% SOC!.   Again, I have not included cycling effects.  

 

An actual test of a Fusion Energi driven in Phoenix, AZ can be found here:  https://avt.inl.gov/vehicle-button/2013-ford-fusion-energi.  They charged the cars once per day in the evening so the HVB was at 100% SOC during the night (and probably all day Sunday).  The cars were driven as part of a courier fleet for legal documents.  After the HVB was depleted, they continued driving the car in hybrid mode the rest of the day.  As a result, the SOC of the HVB would have been below 20% most of the day.  After 100,000 miles and 1.5 years, actual battery degradation was 8%.  

 

If I assume the car is at 100% SOC for half the day and 20% for the other half, I come up with an average fade rate of 4.5e-3/sqrt(day).  According to the model, degradation will reach 20% after 5.4 years.  The model predicts 10% degradation after 1.5 years corresponding to a fade rate of 4.5e-3/sqrt(day)

Edited June 25, 2016 at 11:33 PM by larryh

[bdginmo]

Great information. I had to park my car at the airport in St. Louis this week where high temperatures were 95-100F all week. I made sure to park in covered parking (minimizing direct sunlight) with about 60% SoC. I always make it a habit to store my car for an extended periods with less than 70% SoC. Based on this data, however, the lower the SoC even down to hybrid the better.

[larryh]

I would try to keep the HVB in a state that is above the X=10 years line on the bottom chart and below the X=10 years in the top chart.  The HVB temperature should be somewhere between the high and low daily temperature if the car is parked for a week.  Assume a mean temperature of 85 F or about 30 C for the days the car was parked and that the HVB temperature was around that temperature.  At 60% SOC and 30 C, you are well within the desired regions of the chart.  If you were in Phoenix, you would not be. 

 

However, note that if MFT/MFM is reporting 60% SOC, the actual SOC of the battery is 20% + 80%*0.6 = 68%.  If this is the case, you are just barely within the desired regions.  MFT/MFM reports the HVB SOC as 0% when the actual SOC is 20%.  MFT/MFM reports HVB SOC as 100% when actual SOC is 100%.  So you have to adjust the MFT/MFM SOC to get the actual SOC to use in the charts above. 

 

Also note, that if you drove the car during the day, the HVB temperature would be about 10 F above the outside temperature.  So if it is 100 F, the HVB might be 110 F or about 45 C.  In that case, you would want the SOC to be 30% or less when you park.  MFT/MFM would report 12%.  But if not, the HVB will eventually fall and degradation will again be inside the desired limits. 

Edited June 25, 2016 at 11:34 PM by larryh

[larryh]

Consider the Chevrolet Volt.  It has an active liquid cooling system that attempts to maintain the battery temperature around 25 C.  Provided the car is either running or plugged in, the battery temperature should be around 25 C.  In addition, the battery is never charged above 80% SOC.  Looking at the charts above for 80% SOC and 25 C, that operating point is within the desired region, i.e. below the X=10 year line in the top chart, and above the X=10 year line in the bottom chart.  Disregarding cycling of the battery, the battery should last for at least 10 years before 20% degradation. 

 

However, the car is likely to remain off and not plugged in for long periods of time, so degradation may be a worse than predicted.  You can see the actual degradation for a Volt operated in Phoenix (under similar conditions to the Energi) for three years and 120,000 miles here: https://avt.inl.gov/vehicle-button/2013-chevrolet-volt.  Capacity degradation was about 9%.  But again, the car was operating in hybrid mode for most of the time. 

 

For the Volt, the excessive degradation observed by some for the Fusion Energi after only 3 years should not occur.  Unlike the Volt, the Energi is charging the battery to 100% SOC (it actually charges to about 95% SOC, but that is not much of an improvement) and does not actively manage the HVB temperature when the car is off or plugged in.  I suspect this is the main reason people are observing high degradation rates.  The average temperature and SOC of their HVB is too high. 

Edited June 25, 2016 at 11:38 PM by larryh

[larryh]

In addition to the Calendar (storage) degradation discussed above, further degradation occurs due to cycling of the battery.  The total degradation of the battery is the sum of these two effects.  Each time you charge/discharge the battery, you are adding additional degradation to the battery above and beyond the calendar effects discussed above.   You can find the results of cycling the Chevrolet Volt's battery for 4323 cycles in the paper "Investigation of battery end-of-life conditions for plug-in hybrid electric vehicles" by Eric Wood, Marcus Alexander, and Thomas Bradley.  Just search the internet.

 

In that paper, each cycle consisted of driving the car in EV mode until the battery was depleted to a specified SOC (EV range for the Volt is 38 miles).  Then the car was driven in hybrid mode for a total trip distance of 50 miles.  They did not actually drive a car.  Instead, they applied a simulated load to the battery for a driver who drives aggressively.  The battery was allowed to rest for 15 min and then was charged.  After charging completed, the battery was allowed to rest for one hour and then the next cycle began.  Each cycle took 7.3 hours.  

 

They did 4323 cycles or approximately 10 years worth of driving if the car is cycled approximately once per day.  They tested three batteries.   For the first battery, the battery was only discharged from 80% to 20% SOC, i.e. 60% DOD (depth of discharge)--which is close to what the Volt actually does.  The second was cycled with 70% DOD, and the last battery was cycled with 80% DOD.  The battery capacity degradation never exeeded 20% for any of the batteries.  Degradation was linear, i.e. a constant amount of degradation occurred with each cycle.

 

Power degradation was about 15% for 60% and 70% DOD after 4323 cycles.  Degradation was linear  However, the power degradation did exceed 20% for 80% DOD at 3650 cycles.  It was linear to about 2400 cycles with 80% DOD and then degradation began accelerating.  Cycle degradation increases with increasing DOD.  You want to keep DOD to less than 70%.   DOD greater than 70% will shorten the life of the battery.  

 

The DOD for a fully depleted Energi battery is 100% - 15% = 85%.  That is way too aggressive.  If you constantly deplete the HVB every day, the battery will reach end of life before 2400 cycles, or before 6.6 years due to cycling alone.  If you do it twice a day, then it will reach end of life before 3.3 years due to cycling alone.

 

As a consequence of power degradation due to cycling, acceleration times will be reduced.  Normally, the Volt accelerates to 60 mph in a little over 9 seconds.  For the battery with 80% DOD after 4323 cycles, the acceleration time was still just over 9 seconds when the battery was fully charged.  However, as the battery is depleted, acceleration becomes sluggish.  When the battery SOC reaches 30%, acceleration times increase to 15 seconds.  

 

Note that for my round-trip commute, I discharge the battery from approximately 100% to 40%, or 60% DOD.  So cycling degradation isn't going to be a concern for me.  However, anyone who has a long commute and charges the HVB immediately after arriving at work and at home is going to experience rapid degradation due to Calendar effects from high HVB temperature and SOC and from Cycling effects due to high DOD.  

Edited June 25, 2016 at 11:41 PM by larryh

[larryh]

The following chart shows the relative importance of Calendar and Cycling degradation effects for three different locations for an EV:  Minneapolis, Los Angeles, and Phoenix.  This is from a presentation that can be found at http://www.nrel.gov/docs/fy13osti/58550.pdf.  

 

The top two charts are Calendar and Cycling capacity degradation after 10 years.  The bottom two charts are Calendar and Cycling resistance growth after 10 years, i.e power degradation.

 

Each marker represents one of 317 different trip histories (DOD).  The different colors of the markers indicate a low, median, and highly aggressive driver (charge/discharge rate).  

 

To compute total capacity degradation due to Calendar and Cycling losses, you need to add the degradation for a trip on the top left chart to the degradation for the corresponding trip on the top right chart.  In Minneapolis, the minimum overall 10 year capacity degradation is 21% and the maximum is 30%.  In Phoenix, the minimum is 29% and the maximum is 45%.  

 

Climate has the greatest effect on battery degradation.   Drivers in Phoenix will experience greater battery degradation no matter what they do as opposed to drivers in Minneapolis or Los Angeles who treat their batteries poorly.  Trip history has the next greatest affect.  And finally, driver aggression.  

 

With respect to capacity degradation, Calendar losses have a far greater impact than Cycling losses.  In Minneapolis, Calendar losses will be about 22% after 10 years.  Cycling losses will be about 1.5%.  So Calendar losses are 14 times greater.  Similar results apply to the other locations.  If you want to preserve the capacity of the HVB, your main priority should be keeping a low average temperature and/or a low average SOC.

 

With respect to power degradation, Calendar losses still have a greater impact than Cycling losses.  But now Calendar losses are only twice Cycling losses.  To reduce power degradation loss, you now need to be concerned with cycling losses due to DOD, temperature, and initial SOC of the HVB.  

 

Driver aggression, i.e. charge/discharge rates, has only a minor impact on degradation over the 10 year period, i.e. about 2% for capacity and 4% for power loss.  

 

I'm not sure how directly these results apply to the Energi.   So according to this paper, the factors that determine degradation of the battery in priority order are climate, commute, and driver aggression.  

 

Edited June 26, 2016 at 03:34 PM by larryh

[larryh]

Notice from the charts in the previous post, compared to Calendar (storage) degradation, Cycling degradation (charging/discharging) seems to be less dependent on HVB temperature.  

 

It is clear from the left two charts, that the warmer the battery, the greater the Calendar degradation (all the Phoenix trips result in greater degradation than all the Minneapolis trips).  The increase in degradation with temperature is exponential.  Temperature impact on resistance growth is especially evident (bottom left chart), implying high battery temperatures degrade the maximum power output of the battery much faster than lower temperatures.  The same is true for battery capacity (top left chart), but the impact of high temperatures, while significant, is not quite as great.

 

From the right two charts, it appears that the main affect of higher temperatures with respect to Cycling degradation is to increase the variability of both capacity loss and resistance growth.   Rather than shifting the boxes upward as on the left side, the height of the boxes are expanded.  For moderate trips and driving, there is a small increase in the rate of degradation with increasing temperatures.    For an average battery temperature of 285 K, the average degradation due to resistance growth is about 6%.  For an average battery temperature of 300 K, the average degradation due to resistance growth is about 8%.  However, at higher temperatures, more aggressive trips and driving results in disproportionately more degradation with temperature.  The maximum degradation due to resistance growth is 12% when battery temperature is 285 K.  The maximum degradation due to resistance growth is 17% when the battery temperature is 300K.   In effect, higher temperatures are magnifying poor driving/charging behavior.  

 

So driving/charging the car with a hot battery does increase degradation a small amount.  The degradation increase is minimal for short  trips/shallow charging.  It becomes more significant with longer trips/longer charging.  But the main reason for the increase is degradation of the HVB at high temperatures is due to the the fact that the battery is simply hot (and at a high SOC) and not because you are driving/charging it when it is hot.  

Edited June 26, 2016 at 03:45 PM by larryh

[larryh]

The following chart is the same as the bottom chart in the first post, except now the contour lines are HVB temperature.  This chart emphasizes the importance of maintaining a low average SOC for the HVB.  You should never leave the HVB fully charged for an extended period of time.  

 

If the HVB temperature is 100 F, then follow the blue 100 F contour line.  At 100% SOC, the battery will last 2.8 years.  If instead, you keep the HVB temperature at 100 F and SOC at 20%, the battery will last 18 years.  If the HVB temperature is 50 F, then it will last more than 10 years no matter what SOC is maintained.  You want the battery to last the life of the car, which generally considered around 10 years.

 

When outside temperatures are hot, you especially want to be sure that you delay charging the battery until the last moment, that you charge the car quickly, and that you drive off immediately after charging completes and start using up the charge in the battery.  The less time the battery spends at a high SOC, the better.  

 

When it is extremely hot, you might want to consider driving the car in hybrid mode and not charging the battery.  

 

 

 

Edited June 29, 2016 at 10:38 PM by larryh

[larryh]

This chart explains why you want to delay charging until the last possible moment and to charge as fast as possible.  This chart assumes a typical summer day in Minnesota.  I arrive home from work at 5:00 pm and leave for work at 5:00 am.  When I arrive home, the HVB temperature is 91 F.  The garage temperature during the night is around 70 F.  The battery cools at a rate of about 1.3 F degrees per hour when not being charged.  If I charge with a Level 2 charger, the HVB temperature increases at a rate of about 3.6 F degrees per hour.  If I charge at a slower rate, the temperature rises proportionately slower.  

 

The chart shows five choices for charging:

1. Charge at plug-in as fast as possible using a Level 2 charger (takes 2 hours to charge). [Plug-in 2 hour charge time]

2. Charge at plug-in at a rate such that charging completes in 6 hours (Level 1 charger). [Plug-in 6 hour charge time]

3. Charge slowly at plug-in all night long so charging takes 12 hours and completes at 5:00 am. [12 hour charge time]

4. Wait 6 hours to begin charging and then charge at a rate such that charging completes in 6 hours at 5:00 am (Level 1 charger). [Delayed 6 hour charge time]

5. Wait 10 hours to begin charging and charge using a Level 2 charger (takes 2 hours to charge) so that charging completes at 5:00 am. [Delayed 2 hour charge time]

 

The chart plots the the capacity fade rate for each option at each time throughout the night.  Time 0 corresponds to 5:00 pm when the charger is plugged-in.  Time 12 corresponds to 5:00 am when I leave for work.  Consider the curve for option 1 for which charging completes in 2 hours after plug-in.  The SOC of the battery at plug-in is 20% and the temperature is 91 F.  This corresponds to a fade rate of 2.2e-3 as indicated by the light blue line at time 0.  The battery now charges and the capacity fade rate increases.  At time 2 hours, the HVB is at 100% SOC and the temperature is 98.6 F.  The fade rate is 6.2e-3 as indicated by the light blue line at time 2.  The battery now cools at a rate of 1.3 F degrees per hours and the fade rate decreases to 5.2e-3 at 5:00 am.

 

Option 1 degrades the battery the most, followed by option 2, 3, 4, and finally, option 5 degrades the battery the least.  You can see that the capacity fade rate for option 1 equals or exceeds that for all the other options at any time.  Similarly, the capacity fade rate for option 4 equals or exceeds that for options 3, 2, and 1 at any time.  I can compute an upper bound on the battery lifetime for each option by computing the average fade rate for each of the curves (area under curve divided by time).  The result provides an upper bound on the battery life since I am missing what happens to the battery for the remaining 12 hours of a day that it is not in the garage.  Whatever happens then will only degrade the battery further.

 

1. 7.6 years

2. 10.4 years

3. 20.2 years

4. 30.7 years

5. 44.5 years

 

If you plug in and start charging the car immediately after arriving home using a Level 2 charger, the battery will last less than 7.6 years.  If you wait until 3:00 am to charge using a Level 1 charger, the upper bound on battery lifetime is 44.5 years--it will be less because I have not taken into account degradation of the battery occurring during the day from 5:00 am to 5:00 pm.  If you live in a warmer state, the upper bounds will be significantly less.  For Phoenix, the upper bound on battery lifetime for option 1 is about 3.5 years.  

 

 

Edited July 2, 2016 at 01:28 AM by larryh

[larryh]

This chart shows the capacity fade rate for the HVB during a typical summer weekday when I drive to work.  Rather than showing the actual fade rate, instead the graph shows the battery lifetime at that fade rate.  The higher the fade rate, the shorter the battery lifetime.  

 

I charge the HVB starting at 3:00 am.  The HVB SOC is 45% and the temperature is 84 F.  The capacity fade rate is at its minimum for the day (battery lifetime is at its maximum).  If the HVB SOC and temperature were fixed at these values, the battery would last about 20 years--the chart shows the battery lifetime is 20 years at 3:00 am.  Charging completes at 4:30 am.  The SOC is now 98% and temperature is 88 F.  The capacity fade rate is now at its maximum for the day (lifetime is at is minimum).  If the HVB SOC and temperature were fixed at these values, the battery would now last less than 5 years--the chart shows the battery lifetime is less than 5 years at 4:30 am.  I want to use up the SOC quickly to prevent rapid degradation to the HVB.   Do not leave the battery at a high SOC for very long.  

 

I start for work around 5:20 am and arrive around 5:40 am.  You can see the battery lifetime going up during this period as the battery is discharged.  At the end of the commute, the battery SOC is 68% and temperature is 93 F.  I leave work around 4:00 pm.  During the day the battery cools to 88 F and the battery lifetime rises slightly from just below 10 years to just above 10 years.  I arrive home at around 4:20 pm.  The battery SOC is now 45% and temperature is 93 F.  During the night, the battery cools and the lifetime increases from 15 years at 4:20 pm to 20 years at 3:00 am.

 

If I compute the average lifetime over the entire day, I get 11.3 years.  If I drove the car under these conditions every day for the next several years, the battery would last for 11.3 years.  The average high temperature for the week was 80 F.  Had the high been 90 F, lifetime would have been reduced to 8.5 years.  Had the high been 110 F, lifetime would have been reduced to 4.4 years.  At 60 F, lifetime is 20.7 years.

 

If instead of charging the HVB to 100% SOC, I charged to 70% SOC, the lifetime would increase from 11.3 to 20.7 years.  Unfortunately, there is no easy way to stop charging at a given SOC.

 

 

Edited July 3, 2016 at 03:56 PM by larryh

[rbort]

Makes sense, but that's only for this (11.3 years), then you have to cut into that for usage, recharge cycles, etc which all dip into the battery life.   What is life defined as, 80% of initial capacity, 70%?  Less?  

 

Here's a side question, you drive from 100% to 68% to get to work, then you say when you get home its at 45%.  When you return to the car at 4pm after sitting outside the 1/2 the day, is the charge level still at 68% or has it changed?

 

Would you benefit more if you started to charge the car at 4am instead of 3am and never got to 100% before you leave to work?  Since you have 45% remaining you could dig into that further on the low end getting home and be at a constant lower charge level average overall this way.  Your 11.3 years might go up a little.

 

-=>Raja.

Edited July 3, 2016 at 04:02 PM by rbort

[larryh]

Suppose I extrapolate the results plotted above to determine the overall degradation rate for the calendar year.  The following shows the average high temperature in Minnesota for each month and the corresponding degradation rate that I estimate for that month based on the commute to work described in post 12.  

 

Month Avg High (F)Capacity Fade Rate (1/sqrt(day))x1E-3

Jan           24               1.26

Feb           29               1.38

Mar           41               1.69

Apr           58               2.23

May           69               2.64

Jun           79               3.06

Jul           84               3.29

Aug           81               3.15

Sep           72               2.76

Oct           58               2.23

Nov           41               1.69

Dec           27               1.33

 

The average capacity fade rate for the year is computed as the average of the values for each month.  (Actually you have to take the square root of the average of the squares.)  I come up with an overall capacity fade rate of 2.34E-3/sqrt(day).   The following chart shows the predicted degradation of the HVB with time based on an average fade rate of 2.34E-3.  The actual degradation after 3 years is approximately 5%.  The predicted degradation from the chart is about 7%.  The prediction is going to be off due to many simplifying assumptions.  For one thing, I don't drive the car to work 7 days a week.  It usually gets a rest over the weekend.  Note that if the fade rate is a, then the predicted HVB degradation after d days is degradation = a * sqrt(d).    After 15 years, degradation will be approximately 17%.  The HVB should last the life of the car.  

 

Edited July 3, 2016 at 05:00 PM by larryh

[larryh]

Makes sense, but that's only for this (11.3 years), then you have to cut into that for usage, recharge cycles, etc which all dip into the battery life.   What is life defined as, 80% of initial capacity, 70%?  Less?  

 

Here's a side question, you drive from 100% to 68% to get to work, then you say when you get home its at 45%.  When you return to the car at 4pm after sitting outside the 1/2 the day, is the charge level still at 68% or has it changed?

 

Would you benefit more if you started to charge the car at 4am instead of 3am and never got to 100% before you leave to work?  Since you have 45% remaining you could dig into that further on the low end getting home and be at a constant lower charge level average overall this way.  Your 11.3 years might go up a little.

 

-=>Raja.

 

End of life is defined when the HVB capacity drops to 80% of its initial capacity.  The results I have posted above do not include degradation due to cycling.  But from post 8, the degradation due solely to cycling will be about 1.5% after 10 years.  Degradation due to cycling is much smaller than calendar fade degradation.  In that post, degradation due to calendar fade is 14 times larger than degradation due to cycling.  Since cycling degradation has such a small impact on overall degradation, I ignore it.  The errors in my estimates for calendar fade are going to be much larger than degradation due to cycling.

 

The SOC of the HVB as reported by the BECM changes between the time I park at work (5:40 am) and the time I leave (4:10 pm).  The BECM estimate of SOC is not 100% accurate and the battery cools down.  It typically falls from 74% to 68%.  At night the opposite happens, it rises from 43% at 4:20 pm to 45% at 3:00 am (even though the battery cools).  

 

Yes, it would be better to start charging later in the morning.  But unfortunately, I do not always end up with 45% SOC from the previous day.  Sometimes it as low as 15% depending on what I did the previous day.  I will have to see if I can figure out how to adjust charging that is not too difficult to manage.  If I started charging at 4:00 am, the model predicts a lifetime of 12.6 years instead of 11.3 years.

Edited July 3, 2016 at 06:16 PM by larryh

[rbort]

Well just a simple change to 4am from 3am and try it for a week.  See if you have enough battery to come and go in the summer, in the winter you can move it back to 3am.  As you well know you get more MPGe's in the summer than winter, so you need less of a charge then to get around.  1 hour charge will add 50% to the car's battery.  If you need a little bit more since value charge only works in 1 hour increments maybe you can fast forward the car's clock 15 minutes.  That would add about 63% charge which would be more than enough to cover your daily trip.

 

-=>Raja.

Edited July 3, 2016 at 06:25 PM by rbort

[larryh]

MFM and Value Charge are hard to work with.  They seem to have a mind of their own and do what they want--not want you want or what is best for the battery.  I have a timer attached to the charger, I can adjust that to determine when charging starts.

 

The industry standard criteria for end of life of a battery for an EV is established by the United States Advanced Battery Consortium (USABC).  A battery has reached end of life when either:

 

  1. The net delivered capacity of a cell, module, or battery is less than 80% of its rated capacity when measured on the DST (Reference Performance Test); or

  2. The peak power capability (determined using the Peak Power Test) is less than 80% of the rated power at 80% DoD.

 

DoD (depth of discharge) is defined as:

 

  The ratio of the net Ampere-hours discharged from a battery at a given rate to the rated capacity.

 

I am only considering the first criteria regarding capacity of the battery.  To be complete, I would also have to model the resistance growth rate of the battery, i.e. the top chart in the first post.  But since the Energi is a PHEV, peak power is of less concern to me.  Unlike a pure BEV, we have an ICE to provide additional power if needed.  In most of the papers I have read, the second criteria regarding peak power occurs first.  The tests are performed at 30 C.  For details, see:  https://inldigitallibrary.inl.gov/sti/4655291.pdf.

Edited July 3, 2016 at 09:08 PM by larryh

[larryh]

Here's a side question, you drive from 100% to 68% to get to work, then you say when you get home its at 45%.  When you return to the car at 4pm after sitting outside the 1/2 the day, is the charge level still at 68% or has it changed?

 

 

Here's another example of the battery SOC falling after a 60 mile commute.  I arrived home at 4:14 pm and let the battery sit for a couple of hours.  I measured the HVB SOC and temperature at 4:48 pm and then again at 5:56 pm.

 

Time   SOC    Temp

16:14  15.80% 98.6 F

16:48   9.29% 98.6 F

17:56   8.86% 96.8 F

 

The first SOC drop from 15.8% to 9.29% was probably due to BECM estimation errors.  The drop from 9.29% to 8.86% was probably due to a temperature drop in the HVB.  I wonder how far it would drop if I let it continue to cool.  It better not fall below 0% SOC.  I then charged the battery up to 15.99% at 6:08 pm.  The difference in the degradation rate between 8.86% and 15.99% is small, 21.7 years vs. 20.5 years.

Edited July 3, 2016 at 11:31 PM by larryh

[rbort]

It wouldn't fall more than 10%, mine never has, though they do lose power when you let them sit and cool down.  If you arrive to some destination with say 2 miles left after driving 30 miles EV, wait a half hour and those 2 miles will be gone, down to the hybrid battery.  That's why if you need to get somewhere far 100% ev, do not stop 3/4 of the way there and get some lunch expecting to find more charge later.  I also don't like my battery to get too low, that can damage the cells, so If I get home on the hybrid battery then I want to plug it in and bring it up to 5%.  If you charge it like that there seems to be a variation in the battery level calculation later, that 5% can become 10% the next morning.  If you don't charge it, it can crash bad enough that next time you power up the car you see the hybrid battery shown at 10% and the engine starts right away (bad).

 

So you got the timer there is no need for you to start charging at 3am for usual work trip, change it to 3:30 or 3:45am and it will still be enough if you don't mind living on a lower average HVB SOC.

 

-=>Raja.

Edited July 4, 2016 at 03:19 AM by rbort

[larryh]

Below is an updated table from my previous post to add average cell voltage, variation between cell voltages, and variation between cell SOC:

 

             Avg Cell  Variation   Variation

Time   SOC    Volts    Cell Volts  Cell SOC  Temp

16:14  15.80%  3.44    0.072       0.6%      98.6 F

16:48   9.29%  3.44    0.062       0.9%      98.6 F

17:56   8.86%  3.45    0.057       1.1%      96.8 F

18:08  15.99%  3.55    0.006       1.1%      96.8 F

 

It seems strange that SOC went down at 17:56 but cell voltage went up.  In any case, all the cell voltages at time 16:14 are at least 3.44-0.072= 3.37 V (above the 3.0 V low voltage threshold for damaging the cell).  They are also above the threshold at the other times.  So the cells should not be harmed by the low SOC.

 

The one thing that charging at time 18:08 seems to do is to balance the cells.  Before charging, the variation in the cell voltages was slowly decreasing from 0.072 volts.  After charging, it was reduced to 0.006 volts.  Unbalanced cells puts stress on the weaker cells.   So it might be prudent to charge the HVB for a few minutes after depleting it to bring the cells into balance.  During the week, when I return home from work and the HVB SOC is around 40%, the variation in cell voltages is at most 0.009.  So fully depleting the HVB causes the cells to become more unbalanced.  If you don't go deeply into hybrid mode, the cells should be relatively well balanced at the end of the trip.  

Edited July 4, 2016 at 10:48 AM by larryh

[rbort]

The reason you lose capacity in the pack is because as the pack gets older some cells get worse than others.  Those cells crash in voltage towards the end of the charge faster than the other cells, so for those people who are getting 4.5kwh out of a charge many of their cells may still have 1kwh to give out, but some have crashed low and the voltage of the entire pack drops enough to make the system show that its empty and switch to hybrid mode.  In this case the bad cells will have a lower voltage than the good cells and hence what you see the bigger difference in cell voltages - that's why the balance goes out further and further as you discharge the cells lower and lower, or also it can happen if you draw high currents from the cells, in other words using more than 2 bars of power and draining the pack quicker.

 

When you charge balancing happens all the time so the system starts draining the higher voltage cells while waiting for the lower voltage ones to catch up.  That brings everything into balance eventually.  But it doesn't change the fact that the lower cells are worse than the higher voltage cells, at this point that's now a given and a fact that won't change.  Things can get worse from there, but not better.  At the end of the day, your battery pack is as good as the worst cell in there.  

 

I'm not sure what you mean "Unbalanced cells puts stress on the weaker cells", I can't make any sense of that statement.  Weaker cells fall out of balance with the rest of the pack but they don't put any stress on the good cells.  They are just worse and their voltage drops quicker, they also heat up more, and they can puff.  Best to keep current draw on the whole pack (which includes them) low to make them live as long as possible.  Due to the chance of the voltage getting too low for the bad cells, its a good idea to charge the pack up some from the ground if its into the hybrid portion when you get home.

 

I charged mine up yesterday for 42 minutes on 120v.  It shows on MFM 13% charge last night when I unplugged it, and this morning it still shows 13% charge.  I find a difference when the pack is hot and you charge it, from a long drive.  Yesterday I only drove 8.5 miles to the fireworks and back 8.5 miles, so total of 17 miles EV at night.  If I had driven the entire HVB and it was hotter, then if I charge to 13% at night and disconnect, this morning I might have found the charge level to be at 20%.  That seems to be the factor for me Larry.

 

-=>Raja.

[Hybridbear]

But from post 8, the degradation due solely to cycling will be about 1.5% after 10 years.  Degradation due to cycling is much smaller than calendar fade degradation.  In that post, degradation due to calendar fade is 14 times larger than degradation due to cycling.  Since cycling degradation has such a small impact on overall degradation, I ignore it.  The errors in my estimates for calendar fade are going to be much larger than degradation due to cycling.

This is why I'm not a big fan of Tesla telling owners to charge to 80% or 90% every night. If I had a Tesla, I would likely choose to charge only to 50-60% on most days based on this info about capacity loss due to calendar fade to keep the average SOC as low as possible, while not going too low where cell voltage variation spikes. Calendar fade degradation should be a concern to anyone buying a used EV. It is a concern for me as I consider a used Model S as a possible replacement for the Focus Electric when its lease ends in under 6 weeks.

Edited July 5, 2016 at 01:37 PM by Hybridbear

[larryh]

The Chevy Volt has active liquid cooling that attempts to maintain the battery temperature at 25 C (77 F) when the car is on or plugged in. I’m not sure how Tesla manages battery temperature, but I assume it does something similar. At 80% SOC and 25 C, the battery should last 10 years. At 50% SOC and 25 C, the battery should last 20 years. Using active liquid cooling avoids the high battery temperatures that occur in cars with passive or active air cooling (such as the Energi) during the summer. As a result, battery degradation is not as much of a problem with the Tesla.

 

In the winter, cold temperatures aggravate battery degradation. A battery may have 5% capacity loss in the summer, but in the winter, the capacity loss may be twice as much or 10% (versus when the battery was new). This is on top of capacity loss due cold temperatures, i.e. a cold battery stores less energy than a warm battery. In the summer, when my car was new, I was able to get about 5.9 kWh of energy out of the battery. In the winter, it was around 5.5 kWh. So due to the cold, the battery stored 0.4 kWh less energy. After three years, I get 5.6 kWh during the summer and 4.9 kWh during the winter. The battery now stores 0.7 kWh less energy in the winter. So during the summer, the loss due to battery degradation after three years is 5.9 – 5.6 = 0.3 kWh. During the winter, the degradation loss is 5.5 – 4.9 = 0.6 kWh, i.e. twice as much.

 

Tesla warms the battery in Winter (when plugged in), so both capacity losses due to higher degradation and lower storage capacity with colder temperatures are eliminated. In theory, you should be able to get about the same amount of energy from the battery in the winter as in the summer. I wonder if that is what owners observe.

 

Also, I wonder how much energy Tesla consumes to heat and cool the battery. Hopefully Tesla has GO times similar to the Energi that allow you to precondition the car. In the winter time, there is no point wasting energy to heat the battery 24 hours a day when plugged in if you are not going anywhere. Also, it would be advantageous to leave the battery cooler than normal in the Winter to take advantage of the cooler temperatures and slow down battery degradation. Leaving the battery at 25 C and 80% SOC yields a lifetime of 10 years. However, if you left the battery below 50 F, lifetime goes up to 25 years.

 

Tesla should allow you to enter your driving schedule on a calendar and your time of day electric rates, so that it can optimize the cost of energy and minimize battery degradation.

Edited July 5, 2016 at 04:40 PM by larryh

[murphy]

Tesla does not have GO times.  They have smart preconditioning which attempts to learn your schedule and then have the car ready to go for you.  It's of no value to anyone that doesn't have a regular schedule.  Heat or cooling can be activated from the smartphone app.

[larryh]

Tesla does not have GO times.  They have smart preconditioning which attempts to learn your schedule and then have the car ready to go for you.  It's of no value to anyone that doesn't have a regular schedule.  Heat or cooling can be activated from the smartphone app.

 

That doesn't seem like "smart" preconditioning to me.  The car is trying to guess your schedule and you have no control over it.  The car is going to waste energy preconditioning the car when you are not going anywhere and will fail to precondition it when you do want to go somewhere.  They need to allow you to enter your schedule on a calendar.   Can you at least control the times of day when the car will charge, i.e. similar to MFM enter cost windows for time of day electric rates?  Perhaps there are third party applications that give you more control over preconditioning and charging the car?

Edited July 6, 2016 at 10:34 PM by larryh