Thursday, May 7, 2015

Spark EV efficiency

In previous post, a table with MPG equivalent was posted. The values were based on 4 miles per kWh of electricity used. That value is derived from actual measurement from driving to dog beach and back with my two foster dogs; everyone should foster dogs and cats if they can (see http://fosteringdog.blogspot.com)! Of course, as the battery ages, this number will change, probably lower. But let's not dwell on the future; onwards to present!

I live about 45 miles away, 2000 ft above sea level. Drive is a mix of country road, few miles of city, miles of freeway, some stop and go, generally a good variety of conditions. The trip started at home at full charge. On the way to the dog beach, few stops were made at the post office, Home Depot, McD, Jack in the box. Because the round trip is almost 90 miles, I charged at Blink L2 when I got to the dog beach for 3 hours to get back to full battery. That was 8.97 kWh, $4.40 ($0.49/kWh) after 46 miles.

46 miles / 8.97 kWh = 5.13 mi/kWh

Since going to the beach is mostly downhill, one expects very good mi/kWh, and it didn't disappoint. But Blink public charging is very expensive. Using the value from MPGe table based on 4 mi/kWh and scaling to 5.13 mi/kWh,

$3.50/gallon at $0.49/kWh = 28.6 MPGe
28.6 MPGe * 5.13 / 4 = 36.7 MPGe

Even if the trip is going downhill, charging at public station is worse than driving a Hyundai Elantra. Going back home is mostly uphill, mix of conditions as with downhill, stop off at a friend's house for a total of 45 miles. Monitoring the charge with Kill-A-Watt (KAW for short) for around 15 hours charge at 8A, it showed 13.64kWh.

45 miles / 13.64 kWh = 3.3 miles/kWh
$3.50/gallon at $0.17/kWh (home rate) = 82.4 MPGe
82.4 MPGe * 3.3 / 4 = 68 MPGe

Even if going uphill, it pays to charge at home. This shows independent numbers for good case (downhill) is bit over 5 mi/kWh and bad case (uphill) is bit over 3 mi/kWh. But I'm always coming home, so what I need is aggregate.

(46+45 miles) / (8.97 + 13.64 kWh) = 4.025 mi/kWh, round to 4 mi/kWh

This is actual measured use of energy with respect to the wallet. The car does not measure this, but it does show the energy used from the battery to the motor while running. That number is meaningless in terms of cash out of my wallet, but it is useful to estimate the efficiency. This number was 4.8 mi/kWh. Assuming the number is accurate, the charging efficiency is roughly

(4.8-4) mi/kWh / 4 mi/kWh  = 20% (loss)
100% - 20% = 80% (efficiency)

Note that these are all rough numbers. I have no idea amount of charge the battery is allowed to have; to extend the battery life, I suspect it's not 100% (19 kWh). But I also don't know if electronics would know exactly when to know it's full consistently time after time. I also don't know if the SOC (state of charge) is allowed to vary over time to account for my driving. The truth is somewhere close to this, and it gives a reasonable estimate.

120V, 8A charging

There are other fun things you can find out. Using KAW, I measured 121V when not charging, but 114.5V when charging at 8A. That means the house wiring plus the extension cord (100ft of 12 gauge) is

6.5V / 8A = 0.81 ohms (let's round to 0.8 ohms)

Measuring at socket where extension is plugged in was 117.8 volts

(121 - 117.8 V) / 8A = 0.4 ohms

But there's no way to connect the car to power without an extension cord. So let's use 0.8 ohms for energy loss through wiring. At 80% efficiency, charging 19 kWh battery at 120V and 8A will take

120V * 8A * 80% = 0.768kW (interestingly, about 1 horsepower)
19 kWh / 0.768kW = 24.7 hours, round to 25 hours

Note that 120V is used instead of more accurate 114.5V. This is close enough, especially considering I don't know exactly what the car is doing when charging / discharging and battery deterioration over time. It'll be longer in all cases, but it's close enough for government work blog. 25 hours is to fully charge the battery from empty, assuming the car allows that. Then the energy wasted in wiring is

8A * 8A * 0.8 ohm = 51.2 Watts
51.2 Watts * 25 hours = 1.28 kWh
1.28 kWh * $0.17/kWh = $0.22

120V, 12A charging

I'm wasting about a quarter each time I go through full charge at 8A rate. But how will it be at 12A rate? Car allows for it, and it will be faster, but will it be cheaper? If we assume the charger efficiency at 8A and 12A is the same, electrical engineers already know it's more expensive, but let's do the math.

120V * 12A * 80% = 1.152kW
19kWh / 1.152kW = 16.5 hours, round to 17 hours

Assuming same wiring (which it is),

12A * 12A * 0.8 ohm = 115.2 Watts
115.2 Watts * 17 hours = 1.9584 kWh
1.9584kWh * $0.17/kWh = $0.33

I'll stick to 8A when using 120V and save $0.11 per charge cycle. That's about a dollar a month or a Big Mac every 4 months.

Commercial L2 charger

I'm not in any kind of hurry. But if I'm often in a hurry, I could buy L2 charger that charges at 240V and 12A (about 3kW) for about $600. Of course, installation will cost a lot more, because I have to dig trenches to pull the wires about 150 ft. But let's assume only $600 is needed and the wiring resistance is the same. Let's further assume charger efficiency is the same, although higher voltage should have better charger efficiency.

The charging time is half of 120V at 12A, 17/2=8.5 hours. Power wasted in wiring is the same, because the current is assumed to be the same. Because the charge time is half, the energy waste is also half, 0.33/2=$0.17 per full charge. Compared to 120V 8A charging, that's $0.05 savings per full charge. But how many charge cycles in lease period of 30000 miles? Assuming 80 miles per full charge,

30000 miles / 80 miles = 375 charge cycles
$0.05 * 375 = $18.75

Even if the charger with L2 is 100% efficient, I'm only getting a penny or two per charge cycle, still around $20 overall. Rather than spending $600 to save $18.75, I think I'll stick to 120V 8A charging and get a Big Mac every 4 months. For me, I wish they offer even lower current charging mode (2A for 100 hours), so I can buy that Big Mac sooner.

This is only true in my case where I'm on normal electric plan, because I use so little electricity to begin with. If one gets time based tier plan (aka, EV plan), this is definitely not true. In that plan, only about 8 hours are low rate, and the rest are very expensive. In such case, L2 charger may make sense to save money. Charging plans will be discussed in future blog post.

Edit June 2015:

Chevy offers $500 rebate on Bosche L2 charger. Bosche was nice enough to call me to take advantage of the deal. Their 3.3kW L2 charger is $500, their 6.6kW L2 charger is $625 ($125 more). I took the 3.3kW L2 charger with rebate, making it free!

I also found NEMA 6-50 socket in one of my barns. I made a short extension cord using 12 gauge wire and NEMA 6-50 plug for about $15. I looked into wiring cost, and it can get expensive! I was able to use 12 gauge SJOW wire, but 6.6kW would need much thicker wire, 10 gauge or 8 gauge. Since 12 gauge is far more common, price difference was substantial. Unless there's extreme reason, I think 3.3kW would be cheaper all around.

I'm saving $3.75 if L2 electricity savings add up to $18.75. But what I read is that L2 is more efficient than L1 (120V) since the charging circuit doesn't have to run as long. While I haven't measured it, it sounds plausible. Maybe I'm saving far more than $3.75 ($7.50?) Still, I'll continue to use 80% as conservative estimate for L1/L2 charging effciency.

6 comments:

  1. As I understand it, there is a certain amount of overhead that the charger incurs regardless of your charge rate - I have figured it out to be around 350W or so for my Leaf - power from wall is approx. 120Vx12A = 1,440W (pretty close to that using my KAW), based on a number of different charging sessions and seeing how much my state of charge % increases for an hour of charging, I am pretty comfortable saying that my battery receives about 1,100W. That is about 1.1/1.44 = 76% efficiency at 120V/12A trickle charging with my Leaf (I can't change the A it draws).

    For 240V charging, I have a charger that will, at maximum, deliver 240Vx30A = 7.2kW (confirmed by two different Teslas that have charged at my home). My car will charge at the rated rate for the charger (6.6kW) on that circuit, so at worst, the efficiency of L2 charging (at max rate) is about 92% for a >2013 Leaf. I have no way to determine what the actual amp draw is on my charger, so I can't tell directly (I don't have a fancy enough energy meter, nor do I have any diagnostics that work at the voltage [no clamp meter etc]).

    If in fact we assumed the same power overhead for the charging circuit for L2, then it would be 6.6 + 0.35 = 6.95 kW total power draw. So efficiency would be 6.6/6.95 = 95%. Lower bound is therefore 92% and upper bound 95%. I threw a dart and chose 93% as the value to use for my own calculations and blog.

    The wrinkle in all of this is what happens when the battery is close to 100% and is balancing the cells? I rather suspect the efficiency goes way down in this process (below even the L1 charging), so I avoid charging to 100% for the most part. I set my charge timer to finish somewhere in the 90's, and when on road trips, I unplug usually in the mid 90's, unless I really think I need the extra few %. I do still leave it to charge completely (and therefore balance) at least once per month, just to be on the safe side.

    ReplyDelete
    Replies
    1. 92%/95% L2 seems awful high. I know from step down switchers I've built that they're typically low-mid 90% at peak efficiency, but less for step up (mid-upper 80% range). L2 is step up process. Since I assume same electronic parts are used for L1/L2 (probably true to save cost), slew rate would be comparable, so the difference in efficiency would be mainly parasitics. Since L2 takes shorter time to charge than L1, the savings in efficiency would be just for shorter parasitic time.

      That's only for electronics. If your L2+battery is assumed 93%, what would be your battery efficiency in taking the charge? If you use DCFC, you can find the battery efficiency, which is what I do for freeway efficiency post. In case of liquid cooled SparkEV, battery seem to be bit over 90% efficient, and I would think air cooled Leaf battery to be less efficient. Then the overall L2 being >90% seems unlikely.

      Delete
    2. I plan to take my next holiday at home, during which I will get some readings from my meter to try and collaborate energy used in a drive followed by a recharge. My daily routine involves charging at both locations, so I can't do that now. Thank you for your insights into the electronics side - I have no theory to draw on, only what I think I have observed. Cheers.

      Delete
    3. I found a webpage today where the person has a meter on their charger etc, and their L2 efficiency calculated out to 87% http://www.casteyanqui.com/ev/year03/index.html

      Delete
    4. Web site doesn't show charging efficiency. 87% is in ballpark, although bit on high side. If battery is 95%, and charger electronics 91% efficient, that'll work out to 87% overall. I think safer estimate might be 85%, and very conservative estimate to be 80%. That's what I use, and there's no excuses with that; they can't say I'm using "optimal efficiency for charging" in any of my numbers.

      Delete
    5. This might be of interest to 2013 Leaf owners.

      http://avt.inl.gov/pdf/fsev/fact2013nissanleaf.pdf

      It shows 22.7kWh was sourced from EVSE, but 19.6kWh went to battery, charging efficiency of about 86%. Battery efficiency to be 95% to 98%. Overall trip efficiency (charge + discharge) to be 82% to 85%.

      Delete