quantified: how much power is used to charge an EV

[majbthrd]

Say a driver wants to figure out the electrical power cost of making a trip using an EV. (Alternately, say they want to figure out the CO2 generated by the power grid to make the EV trip possible. Or, maybe they are trying to figure out how many solar panels are needed to offset their anticipated journeys in an EV.)

It would be a mistake in to blindly believe the efficiency number displayed on an EV or advertised by the manufacturer.

Say a driver gets 4 miles / kWh after driving 12 miles. Surely, 12 / 4 = 3kWh was consumed for that trip, right? Not really. It may be that only 3kWh was taken from the battery during the trip, but more than 3kWh was consumed to put that 3kWh in the battery.

I've been collecting data with my 2011 Nissan LEAF for the past nine months, and so I now have some answers to quantify what the electrical losses are. I would be interested in how other vehicles compare, but this is the only EV that I have to test with. It seems a pity to me that this information is not more readily available.

The overall conclusion in my nine months of testing is that, at least for my 2011 Nissan LEAF, an average of 30% more power was consumed than what actually made it into the EV battery.

I used data from two types of trips: one was 9 miles and one was 16 miles. On the 16-mile trip, I sometimes charged the vehicle to 100% instead of the normal 80%.

One of the conclusions apparent from the data is that there is an efficiency disadvantage of charging the vehicle past 80%. If you compare "16 miles 80%" to "16 miles 100%" in the chart below, you can see yourself that more power is lost when charging to 100%.

Reasoning through it, this makes sense. If you have a power meter connected to your EVSE, you will have already observed that the LEAF tapers down the power as it charges the battery from an 80% to 100% state of charge. (Even without a power meter, most EV users will be well aware that charging that last 20% takes much longer to happen.) As the on-board charger reduces its output, its efficiency is reduced. Furthermore, parasitic losses like vehicle electronics and the coolant pump take up an ever-greater pie-slice portion of the overall consumption.

An unexpected second conclusion and what I find curious is that the charging efficiency seems worse on shorter trips than longer ones. This can be seen in "9 miles 80%" versus "16 miles 80%". I am speculating that parasitic losses like vehicle electronics remain on for a period of time after the vehicle has finished charging. These losses would be the same regardless of the amount of power put into the battery during the charge, so this extra consumption would tend to swamp out smaller battery charges.

Below is the graph. For comparison: "9 miles 80%" averaged 38.7% loss, "16 miles 80%" averaged 13.3% loss, and "16 miles 100%" averaged 31.6% loss.

For those interested in the nitty-gritty details on how this data was obtained, here is my procedure:

My home EVSE has a power meter, so I can measure the exact power going to charge the EV. Prior to the trip, I start with the vehicle charged to a known state of charge (80% in my case). If I choose to charge the vehicle above this state of charge prior to the trip, this additional power consumption is included in the overall trip measurement.

As I start the journey, I reset the vehicle efficiency gauge. At the end of the journey, I jot down the vehicle efficiency gauge value.

I then charge the vehicle back up to the same known state of charge as it was (before any additional charging just prior to the trip). I then jot down that value too.

I also know the trip distance (by noting the odometer both at the beginning and end of the journey, by using a reset-able trip odometer, or by existing measurements if the journey is very consistent).

All this results in three figures (per journey):

trip distance (miles)
efficiency reported by vehicle (miles/kWh)
overall kWh measured by EVSE power meter

From this, we can compute:

vehicle reported consumption = miles / (miles/kWh)

and this resultant kWh can be compared against the actual power consumption measured by the EVSE power meter.

The loss is calculated as:

loss = (overall kWh) / (vehicle reported kWh) - 100%

 

[cornbinder89]

I think the only real thing that is of import, is power used for miles driven. Power in can be measured, from in the EVSE, yours is like mine. Miles driven is easy. What the gauges on the car usage show is like the "Guess-o-meter". a rough indication.
Like trying to be sure of MPG on a liquid fueled vehicle, the longer you take data and adv it out the more accurate the results will be.

 

[knightmb]

This is certainly a lot of work, so my critique is just some observations for improvement. It's true that the Leaf isn't using lab level scientific equipment to measure power, but from what we can tell via LeafSpy, it is pretty good down to the thousands place in most measurements. That works very well for trying to figure out the power in and power out like you have done. The issue you will encounter is the dash readings are doing a lot of rounding.

The first thing that would help with these measurements is LeafSpy. The reason being, you can get more precise measurements of where you are taking data from 80% SOC for example. Or you can even use LeafSpy to record a history of measurements for a trip to know pretty accurately how much power that trip used. That you would save you from having to reverse calculate the power used via the dash numbers and charging numbers from your EVSE.

So, when you take a measurement from your EVSE, then do a trip drive, then charge it back up to 80% with your EVSE, you'll have a lot more accurate numbers because 80% SOC is a big round. With LeafSpy you can see that if you left your home with 51.0 kWh of battery remaining, did the drive, then came back to charge, you could easily charge it right back up to 51.0 kWh exactly to get even better measurements and calculations.

This would help to smooth out the anomalies in your data where the distance driven seems to produce wide variations in power loss. For all intents and purposes, the power loss should be nearly the same relative to how much power is loss to heat in the battery and the charging system itself (cooling pumps, fans, internal heat, etc.)

Say a driver wants to figure out the electrical power cost of making a trip using an EV. (Alternately, say they want to figure out the CO2 generated by the power grid to make the EV trip possible. Or, maybe they are trying to figure out how many solar panels are needed to offset their anticipated journeys in an EV.)

 

[majbthrd]

This is certainly a lot of work, so my critique is just some observations for improvement.

Trust me, I'm an engineer.

I do already have LeafSpy; I principally use it to periodically gather SOH. I don't dispute its measurements might *marginally* narrow the widths of the peaks in the histogram, but it is also a major hassle to get it running and connected each time just to collect data.

There is a difference between printing out figures with many decimal points and actual precision. At best, all LeafSpy can do is report what the BMS tells it, and the coulomb counting ability of any BMS is limited.

The whole point of my doing "a lot of work" is do deliberately gather many data points. Individual measurements (whether done through LeafSpy or not) can and will be subject to measurement error, but aggregating many similar measurements statistically improves the quality.

Also, my suggested methodology would work on any EV, allowing vehicle-to-vehicle comparisons.

 

[knightmb]

Well, you are in the right place because this forum is full of over a decade worth of data of people that have already done exactly what you are doing now and very long discussions about what the loses are, etc.

I would be curious myself to see how the 2011 differs from the 2011 tested way back when it was still young and new given the battery age. I would imagine there are higher charge losses due to the age difference, change in resistance, etc. I will have to read up on some of those older topics to see what power losses they recorded.

 

[Bombastinator2]

Say a driver wants to figure out the electrical power cost of making a trip using an EV. (Alternately, say they want to figure out the CO2 generated by the power grid to make the EV trip possible. Or, maybe they are trying to figure out how many solar panels are needed to offset their anticipated journeys in an EV.)

It would be a mistake in to blindly believe the efficiency number displayed on an EV or advertised by the manufacturer.

Say a driver gets 4 miles / kWh after driving 12 miles. Surely, 12 / 4 = 3kWh was consumed for that trip, right? Not really. It may be that only 3kWh was taken from the battery during the trip, but more than 3kWh was consumed to put that 3kWh in the battery.

I've been collecting data with my 2011 Nissan LEAF for the past nine months, and so I now have some answers to quantify what the electrical losses are. I would be interested in how other vehicles compare, but this is the only EV that I have to test with. It seems a pity to me that this information is not more readily available.

The overall conclusion in my nine months of testing is that, at least for my 2011 Nissan LEAF, an average of 30% more power was consumed than what actually made it into the EV battery.

I used data from two types of trips: one was 9 miles and one was 16 miles. On the 16-mile trip, I sometimes charged the vehicle to 100% instead of the normal 80%.

One of the conclusions apparent from the data is that there is an efficiency disadvantage of charging the vehicle past 80%. If you compare "16 miles 80%" to "16 miles 100%" in the chart below, you can see yourself that more power is lost when charging to 100%.

Reasoning through it, this makes sense. If you have a power meter connected to your EVSE, you will have already observed that the LEAF tapers down the power as it charges the battery from an 80% to 100% state of charge. (Even without a power meter, most EV users will be well aware that charging that last 20% takes much longer to happen.) As the on-board charger reduces its output, its efficiency is reduced. Furthermore, parasitic losses like vehicle electronics and the coolant pump take up an ever-greater pie-slice portion of the overall consumption.

An unexpected second conclusion and what I find curious is that the charging efficiency seems worse on shorter trips than longer ones. This can be seen in "9 miles 80%" versus "16 miles 80%". I am speculating that parasitic losses like vehicle electronics remain on for a period of time after the vehicle has finished charging. These losses would be the same regardless of the amount of power put into the battery during the charge, so this extra consumption would tend to swamp out smaller battery charges.

Below is the graph. For comparison: "9 miles 80%" averaged 38.7% loss, "16 miles 80%" averaged 13.3% loss, and "16 miles 100%" averaged 31.6% loss.

For those interested in the nitty-gritty details on how this data was obtained, here is my procedure:

My home EVSE has a power meter, so I can measure the exact power going to charge the EV. Prior to the trip, I start with the vehicle charged to a known state of charge (80% in my case). If I choose to charge the vehicle above this state of charge prior to the trip, this additional power consumption is included in the overall trip measurement.

As I start the journey, I reset the vehicle efficiency gauge. At the end of the journey, I jot down the vehicle efficiency gauge value.

I then charge the vehicle back up to the same known state of charge as it was (before any additional charging just prior to the trip). I then jot down that value too.

I also know the trip distance (by noting the odometer both at the beginning and end of the journey, by using a reset-able trip odometer, or by existing measurements if the journey is very consistent).

All this results in three figures (per journey):

trip distance (miles)
efficiency reported by vehicle (miles/kWh)
overall kWh measured by EVSE power meter

From this, we can compute:

vehicle reported consumption = miles / (miles/kWh)

and this resultant kWh can be compared against the actual power consumption measured by the EVSE power meter.

The loss is calculated as:

loss = (overall kWh) / (vehicle reported kWh) - 100%

View attachment 3226

Those that don’t have such can charge on phase 1 through their “granny” charger and a kill-a-watt. I can check one out like a book from my local library.

I generally charge on phase 1 because it’s enough for what I do. I describe that to others as “being about the same draw as an 8,000btu air conditioner”

 

[Bombastinator2]

Trust me, I'm an engineer.

I do already have LeafSpy; I principally use it to periodically gather SOH. I don't dispute its measurements might *marginally* narrow the widths of the peaks in the histogram, but it is also a major hassle to get it running and connected each time just to collect data.

There is a difference between printing out figures with many decimal points and actual precision. At best, all LeafSpy can do is report what the BMS tells it, and the coulomb counting ability of any BMS is limited.

The whole point of my doing "a lot of work" is do deliberately gather many data points. Individual measurements (whether done through LeafSpy or not) can and will be subject to measurement error, but aggregating many similar measurements statistically improves the quality.

Also, my suggested methodology would work on any EV, allowing vehicle-to-vehicle comparisons.

The problem with that is YouTube doesn’t help the trust level much. A lot of the people on that who claim to be engineers don’t seem to be real mental giants. Rather a bit the reverse. Of course they are making videos rather than being off somewhere actually working on something. The engineers I’ve known were generally smart as heck.