Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Garbage Article at MIT Technology Review

This site may earn commission on affiliate links.
Does the Model S not have a 'charge based on departure' setting? As GSP notes, this would randomize the load to the grid and local transformers.

Would it though?

If we allow the assumption that the majority of folks are going to start charging within an hour of midnight... we also should then probably allow the assumption that the majority will need to end within an hour of 7am.

So we just staggered the start time instead of the end time... instead of everybody coming online at midnight and saturating the grid for a couple of hours at midnight and then gradually dropping off as the night progresses... everybody gradually comes on and the grid is saturated for a couple of hours leading up to 7am.
 
Would it though?

If we allow the assumption that the majority of folks are going to start charging within an hour of midnight... we also should then probably allow the assumption that the majority will need to end within an hour of 7am.

So we just staggered the start time instead of the end time... instead of everybody coming online at midnight and saturating the grid for a couple of hours at midnight and then gradually dropping off as the night progresses... everybody gradually comes on and the grid is saturated for a couple of hours leading up to 7am.

Yep. The better option would be to set a start and end time and have the car draw whatever amps needed to do the job. If you set start at 1am and finish at 6am and the battery is nearly full the car might charge at 6 amps. If it's quite depleted, maybe 35 amps. This would spread the loads out over the whole off-peak period
 
Yep. The better option would be to set a start and end time and have the car draw whatever amps needed to do the job. If you set start at 1am and finish at 6am and the battery is nearly full the car might charge at 6 amps. If it's quite depleted, maybe 35 amps. This would spread the loads out over the whole off-peak period

I THINK you are agreeing with my response, and not saying "yup" to my question?

The variable current-draw idea is a nice one...
 
Would it though? .... instead of everybody coming online at midnight and saturating the grid for a couple of hours at midnight and then gradually dropping off as the night progresses... everybody gradually comes on and the grid is saturated for a couple of hours leading up to 7am.

At least one study, linked by GSP, has found otherwise:
http://newsroom.edison.com/internal....com/166/files/20136/SCE-EVWhitePaper2013.pdf

Drivers have less overall power grid impact when they program electric vehicle charging to be complete by a specific time, as this seems to randomize the start time of their charging. This prevents a large number of vehicles from coming online at the same time and avoids power-load spike.

I think the reason that there is less of a demand spike when setting charge to be complete at a specific time is because there are multiple variables that play into when your car will start charging, including:

1) How many miles did you drive, and thus how long does your car need to charge?
2) How fast are you charging your car? Standard Charger? High-Powered-Wall Connector? Twin Chargers?

Both of these factors will vary the start time greatly, but I'm sure there would be a bell curve (with a short tail) for overall demand to the grid. The other advantage of programming for your charge to be complete at a specific time is that I'm pretty sure the high point on the bell curve would be in the early morning hours. That's the way it would definitely be for me. With my Volt, I usually have 3-4 hours to charge at 240/15AMP, and when I have used the 'charge by departure time' setting, I have set it for 6:30 AM. That means that I start charging at 2:30 AM to 3:30 AM, when electrical demand is the lowest.

But you're right that there probably would be saturation just before the 'short tail' of the bell curve around 6:00-7:00 AM. It would be good if EV's could randomize the departure time somewhat to prevent this. That is, if you ask for your car to be fully charged by 7:00 AM, the car radomly pulls that back an hour, or two, or even three (or none), so that it is fully charged a few hours early (and of course therefore fully charged by 7:00 AM).

I'm rather surprised that Tesla doesn't have either a "Time of Departure" setting and/or a programmable "Priority Charging". I would think those would be simple software upgrades.

On a related note, I remember reading that GM (and maybe Nissan) did a pilot program with utilities where the utilities could send out a blanket "Stop charging" command via OnStar or Carwings to deal with a spike in demand. If you had 'priority charging' and your car was under your desired range, the 'stop charging' command would be ignored.
 
Last edited:
Though there's a lot of good discussion here about the national grid, the article's point is that the specific local infrastructure for a small neighborhood "could" become seriously strained... and on that score, it's spot on. My house is large, and I'm in Miami which means high A/C, and we go through 5000-6000 KWh of total consumption in a month... and my peak current draw from the utility (sans EV) was 22A on average over two years. Adding a single 80A load does indeed quadruple the load. Adding the Model X next year and charging both at max current would be a big hit on my local transformer.

We don't actually cause anywhere near that load, of course. I've thought out my usage patterns and I've set the Model S to start charging at 11pm -- when the load from the rest of the house is down to minimum -- and charge only at 20A to maintain that "peak" usage period. Even then, because I don't drive much, my car is done charging by 3-4 am usually. Over time, though, EV charging algorithms in the actual car will have to get smarter: add "finish charging by this time" and variable-rate charging and some priority level of charge, and so on. But while all of this does need to get done, IMHO this is the right time to do it... we're not behind the curve yet.

Overall, the article is shallow and somewhat alarmist: precisely what I expect from most media since it gets them more attention, more views, and thus more ad revenue. It ignores or lightly covers some significant counterfactors, like the fact that EV's are charging at night (actually evening out the base/peak ratio and helping the grid rather than harming it), and that they'll get smarter about their charging algorithms over time. But in general, it's correct that some neighborhoods will get upgraded proactively and some will blow their transformers causing inconvenience. Progress is rarely without pain, whether caused by the progressives or the resistors (pun intended).
 
I'm sure the specific points of the article are accurate. I'm just tired of being told that I'm going to blow up grid because I'm A)feeding back in too much energy with my solar panels and B)drawing too much energy with my electric car. Yes, infrastructure needs to be upgraded. Luckily not everybody is going install solar panels and buy EVs overnight. I'm pretty sure the utilities will figure it out.
 
OK maybe I'm missing something. So help me out here. Less than 2% is not matching what I'm estimating.

239,800,000 cars * 12,500 miles/year/car * 0.350 kWh/mile(1) = 1,049 billion kWh/year World Vehicle Population Tops 1 Billion Units | News Analysis content from WardsAuto
Total electricity consumption in the U.S. (2011) = 3,856 billion kWh/year Electricity - U.S. Energy Information Administration (EIA)
Percent electricity usage if all cars in U.S. were BEVs = 1,049 billion kWh/year / 3,856 billion kWh/year = 24%. Not less than 2%.

For one thing, you are not considering that night time charging does not, up to a certain (but large) amount, increase the burden on the grid, since it needs to be designed for the peak at day time.

The MIT article discussed in this thread actually mentions that:
And researchers at the U.S. Department of Energy’s Pacific Northwest National Laboratory have calculated that the grid has enough excess capacity to support over 150 million battery-powered cars, or about 75 percent of the cars, pickups, and SUVs on the road in the United States.

Then, the question is what 100% refers to, which depends on the larger context of a discussion. In the post I was responding to, for example, it referred to "energy consumed in the US".

- - - Updated - - -

And even if I take the 24% you calculated, as per electricity consumed: The MIT article mentions 75% of that is covered by excess night time capacity. That leaves 6%. We can assume by the time electric cars consume that much energy, sophisticated distribution of charging time (over 24 hours) is not an issue. That's not a large number either, considering that we are doing a lot to reduce electricity consumption.
 
For most power companies, there is no "normal grid upgrade plan". The majority of transformers out there stay on the poles until they cause some type of problem, then they're replaced to the new standard based on their new calculations. The deployment of significant numbers of continuous loads will accelerate that normal aging process (read: transformers blowing up) quite significantly. There are transformers on poles near me that have been rusting away since the 1980's.

.

This is ABSOLUTELY true! I went to my local power company and showed them that their transformer is being oversubscribed 2 to 3X at peak periods. TOU Car charging does not allow the transformers to cool at night. They will replace it when it blows, and not before then!! No big deal to me! I thought I was helping by showing them the oversubscription.
 
Plus, the question is, [regarding my previous post] whether some additional amount can be covered by off-peak day time charging. Seems likely to me. Maybe even all of it? Leaving 0% ? That is, except for charging done at explicit times, during peak time, for example when traveling.

Perhaps it comes down to evaluating how much charging will be really necessary during peak day time.

Furthermore, if concepts for vehicle-to-grid technologies work out, electric cars might even reduce the peak load on the grid? Those are questions one can't do on the back of an envelope anymore.

- - - Updated - - -

This is ABSOLUTELY true! I went to my local power company and showed them that their transformer is being oversubscribed 2 to 3X at peak periods. TOU Car charging does not allow the transformers to cool at night. They will replace it when it blows, and not before then!! No big deal to me! I thought I was helping by showing them the oversubscription.

I think the problem is really the so-called "last mile", the connections and transformers close to the consumer. (Not the larger "grid" itself).

Neighborhood transformers are often outdated in any case. From the MIT article:

Utilities say that the upgrades they’ve performed so far would have been made anyway as part of routine grid modernization
 
For one thing, you are not considering that night time charging does not, up to a certain (but large) amount, increase the burden on the grid, since it needs to be designed for the peak at day time.

The MIT article discussed in this thread actually mentions that:


Then, the question is what 100% refers to, which depends on the larger context of a discussion. In the post I was responding to, for example, it referred to "energy consumed in the US".

- - - Updated - - -

And even if I take the 24% you calculated, as per electricity consumed: The MIT article mentions 75% of that is covered by excess night time capacity. That leaves 6%. We can assume by the time electric cars consume that much energy, sophisticated distribution of charging time (over 24 hours) is not an issue. That's not a large number either, considering that we are doing a lot to reduce electricity consumption.

right. power is a funny thing. it's not like we'd all be charging simultaneously at the peak of day. that's actually a pretty hard thing to do.

seriously though, that's like our panels rated at 200amps yet I have 350+ in breakers. Most (if not all) electrics use the most power the instant they are turned on. after that, its only a fraction. it's pretty damn hard, if not impossible, for me to turn on every single electric device in my house simultaneously to come anywhere nears close to tripping my main breaker.

now as for 240 million electric cars, maybe if we had a "national EV 4:20 charging day" where the entire nation agreed to plug in simultaneously at 4:20 GMT in the middle of the hottest day of the summer, well that would be a moment to remember. remember the NYC black out? hahaha but as for general usage, I doubt 240million electric cars would be any problem for the national grid to handle.
 
For one thing, you are not considering that night time charging does not, up to a certain (but large) amount, increase the burden on the grid, since it needs to be designed for the peak at day time.

I suggest reading this report, it's a lot more in depth than the MIT article http://funginstitute.berkeley.edu/sites/default/files/Impact%20of%20Widespread%20Electric%20Vehicle%20Adoption%20on%20the%20Electrical%20Utility%20Business%20%E2%80%93%20Threats%20and%20Opportunities.pdf

Here's an example from the report showing the load on California's grid for various adoption rates for EVs, if EV owners are all on the TOU rate that encourages charging between midnight and 7 am.

Figure A-8.jpg


If 25% of California cars were EVs, the load on the grid when EVs charge at night would be about the same as the peak daytime load. At EV adoption rates significantly higher than 25%, the nighttime load on the grid from EV charging would exceed the peak daytime load and at times could exceed the supply capacity of the system. The report indicates that California cannot support anything close to a 75% adoption rate for EVs with its current electrical supply capacity. Obviously more generating capacity can be added to match an increasing demand due to EVs, which won't happen quickly.

The feasible EV adoption rate varies regionally. Most states do not have an electrical supply as constrained as California.
 
The feasible EV adoption rate varies regionally. Most states do not have an electrical supply as constrained as California.

We were discussing national numbers. I've heard the "140 million are covered by nighttime charging" a lot, for many years. That's been a general statement in many expert discussions.

Are you saying the numbers in CA would be different? Or that the MIT report, and those generally used numbers, are wrong? Or both?

If the Berkeley report really reflects an untypical situation in CA, the first thing I'd say is there doesn't seem to be a reason not to extend the nighttime charging to before 12am and/or after 5/6 am, possibly using grid communication in the future. There is a lot of unused space below the dashed max peak line above 70 GW. In fact, it would look to me as if 100% could fit under that dashed line. (Of course in reality that would require optimal distribution and probably still be a stretch, but just pointing out what the graphic shows).

CA has big plans to add sustainable energy generation, so a lot will change there in any case. Glancing over the Berkeley study, I get the impression they are over-emphasizing a lot of little practical problems. As electric cars become a mainstream solution, this development will go along with significant political action, changing many of the assumptions made in the report, and the context and for example the regulations by which utilities are operating today.
 
If 25% of California cars were EVs, the load on the grid when EVs charge at night would be about the same as the peak daytime load. At EV adoption rates significantly higher than 25%, the nighttime load on the grid from EV charging would exceed the peak daytime load and at times could exceed the supply capacity of the system.

That part of the analysis seems to be predicated by examining the effects of PGE's E9 Rates. So only of very specific significance.

The report indicates that California cannot support anything close to a 75% adoption rate for EVs with its current electrical supply capacity. Obviously more generating capacity can be added to match an increasing demand due to EVs, which won't happen quickly.

Here your interpretation of the Berkeley report actually agrees with the MIT report, regarding 75 or 74%. (Both reference a PNNL study.)

More generating capacity may have to be added when EV adoption increases above 75%, but that will take quite a while in any case.
 
We were discussing national numbers. I've heard the "140 million are covered by nighttime charging" a lot, for many years. That's been a general statement in many expert discussions.

Are you saying the numbers in CA would be different?

Yes, the Berkeley report says that California is worse off than the national average in terms of electrical generating supply to support EVs. But there is a region that is worse off than California. Yet some other regions could accommodate 100% EV adoption with generating capacity to spare. The point is, the 74% national average is not necessarily meaningful. Each region needs to be looked at.

Table B-1.jpg

Table 2.jpg


Also, the report suggests that to achieve the 74% national average assumes that EV charging demand is managed to perfectly fill in the low demand periods throughout a 24 hour day, so everything matches daily peak demand. It assumes a flat demand cycle across 24 hours a day. No juggling of the TOU periods and TOU rates can accomplish that.

If EV charging is managed so that the 6 pm to 6 am period is filled in to match peak demand, the study found that the feasible national average EV adoption rate drops to 43% with our current generating capacity.

To support high adoption rates for EVs, energy storage might be needed in a big way. Or a Smart Grid. Or more generating capacity in many regions. The problem can be solved of course. Generating capacity has grown a lot in the last 50 years. It can keep growing if needed.

It would have been interesting to model the case where 39% of EV owners have rooftop solar, which I understand is the norm in California.
 
2. Drivers have less overall power grid impact when they program electric vehicle charging to be complete by a specific time, as this seems to randomize the start time of their charging. This prevents a large number of vehicles from coming online at the same time and avoids power-load spikes
I have trouble with this assertion. For me, for example, my plug-in time (at night, or early morning) is far more random than my unplug time (morning for work). Furthermore, I would argue most employed folk have similar unplug time (between 7 and 10 am) so it doesn't seem like it's very randomizing at all.
 
Yes, the Berkeley report says that California is worse off than the national average in terms of electrical generating supply to support EVs. But there is a region that is worse off than California. Yet some other regions could accommodate 100% EV adoption with generating capacity to spare. The point is, the 74% national average is not necessarily meaningful. Each region needs to be looked at.

The map you posted suggests 23% for CA (or most of it) with 24-hour valley filling. However, the graph you posted earlier shows that much more is easily achieveable.

To me, the CA curve actually looks quite good, once you do away with the report's arbitrary 6pm - 6am restriction, and the misguided assumed absence of intelligent load balancing. (See discussion below.)

Also, the report suggests that to achieve the 74% national average assumes that EV charging demand is managed to perfectly fill in the low demand periods throughout a 24 hour day, so everything matches daily peak demand. It assumes a flat demand cycle across 24 hours a day. No juggling of the TOU periods and TOU rates can accomplish that.

This is wrong, I think. Of course, at the point at which the EV adoption will be *exactly* 74%, at that point it would require achieving a perfectly flat demand curve, to avoid even the smallest additional generation.

However, while the EV adoption is less than that, as well as once it will be more than that (meaning most of the time), it will *not* require a perfectly flat demand curve to make optimal use of the currently existing off-peak capacity.

The report seems to assume that all load-balancing would have to be achieved with fixed TOU rates. It does away with grid communication more or less by calling it "hype" and making a list of difficulties which apparently the authors imply are unsolvable.

They seem to forget, for example, that each car built by Tesla has an internet connection, and knows exactly how much electricity it consumes. And is able to lower and raise its charging rate by remote control.

The report is simply wrong in thinking that these things are that difficult.

If EV charging is managed so that the 6 pm to 6 am period is filled in to match peak demand, the study found that the feasible national average EV adoption rate drops to 43% with our current generating capacity.

The restriction to charging only in the period 6 pm to 6 am is justified with nothing more than the following sentence (as far as I can tell):

Assuming electric vehicles can be recharged at any time of day is probably unrealistic, as most recharging will occur at home in the evenings and overnight. As such, a second scenario was considered whereby valley filling could occur only between the hours of 6 p.m. to 6 a.m.

What's wrong with charging at work, after the morning commute, which in many cases is half the day's consumption? What about delivery trucks, taxis, buses, and many other business-related vehicles which can (or even have to) be (re-)charged during the day? Those things are not unrealistic, they are more of a practical necessity.

To support high adoption rates for EVs, energy storage might be needed in a big way. Or a Smart Grid. Or more generating capacity in many regions. The problem can be solved of course. Generating capacity has grown a lot in the last 50 years. It can keep growing if needed.

We already have a smart grid: The combination electric car and internet connection, courtesy of Elon.

It would have been interesting to model the case where 39% of EV owners have rooftop solar, which I understand is the norm in California.

Now that sounds more like it. ;)

I think that report is, to summarize it, wrong. :)

- - - Updated - - -

I have trouble with this assertion. For me, for example, my plug-in time (at night, or early morning) is far more random than my unplug time (morning for work). Furthermore, I would argue most employed folk have similar unplug time (between 7 and 10 am) so it doesn't seem like it's very randomizing at all.

The idea is to avoid a big spike at midnight (or whenever the local off-peak time/rate starts), which many use to set the start time.

With more sophisticated software (which currently isn't really necessary), setting a need-by time would allow the on-board computer to do something fancy, including randomizing both start and stop time. Given cooperation between Tesla and utilities, the car could even get real-time info about the charging time and rate that is best for the grid. All within the confines of making sure the charge is complete at the specified need-by time. Of course, the driver should still have a charge-ASAP option when needed.
 
Al, did the utility company foot the bill for the new transformer, or was it all on your nickel?

Can't say for absolute certain. Only because I don't really know what the transformer and installation should cost. Our PoCo is a Coop. But, yes I paid. Seemed like a lot also. They thought about it and figured up a price. Seemed as though negotiation WAS possible especially since my electrician thought the transformer was already too small before the 14-50's and the HPWC. I just didn't feel like it at the time. Yes, I paid. Maybe in the 3K range all in. Tax deductible I'm hoping.
 
I have trouble with this assertion. For me, for example, my plug-in time (at night, or early morning) is far more random than my unplug time (morning for work). Furthermore, I would argue most employed folk have similar unplug time (between 7 and 10 am) so it doesn't seem like it's very randomizing at all.

Agreed... and I made the same point HERE earlier in the thread.

After seeing several replies, I'm not convinced it doesn't simply move the saturation point to the end of the charging cycle rather than the beginning.