Via Motors, Geely to co-develop medium-duty extended-range electric truck for China and Americas
Tesla awards Musk 10-year 100% at-risk compensation package tied to market capitalization and operational milestones

NREL study finds that uncoordinated plug-in vehicle charging could prove challenging to the grid

An influx of plug-in electric vehicles (PEVs) charging without coordination could prove challenging to the nation’s electric grid, according to research conducted by the US Department of Energy’s National Renewable Energy Laboratory (NREL).

Dr. Matteo Muratori, a transportation and energy systems engineer at NREL, developed highly resolved models of residential power demand and PEV use to assess the impact of uncoordinated in-home PEV charging on residential power demand. He found that although the increase in aggregate demand might be minimal even for high levels of PEV adoption, uncoordinated PEV charging could significantly change the shape of the aggregate residential demand, with impacts for electricity infrastructure, even at low adoption levels.

He also found that clustering effects in vehicle adoption at the local level might lead to high PEV concentrations even if overall adoption remains low, significantly increasing peak demand and requiring upgrades to the electricity distribution infrastructure. This effect is exacerbated when higher in-home power charging is adopted. His paper is published in Nature Energy.

Previous research into the amount of energy required by homes hasn’t taken into account plug-in electric vehicles. Given that more people are choosing to drive these types of vehicles and charging them at home, this additional demand should not be overlooked.

—Matteo Muratori

The simulation concluded that a PEV market share of up to 3%—about 7.5 million vehicles—does not significantly impact the aggregate residential power demand. More than 600,000 plug-in electric vehicles were already on the road at the end of 2016, a figure that includes about 150,000 sold during the year.

Muratori also looked at the impact PEV charging might have on a residential distribution transformer. In this case, a problem arises when motorists gathered in a geographic area began buying these vehicles and plugging them in to recharge upon returning home—a practice known as uncoordinated charging. Even without large numbers of PEVs on the road, this clustering effect will significantly increase the peak demand seen by distribution transformers, thereby potentially requiring the upgrades, he found.

The research also looked at whether the household used the less-powerful Level 1 charging option or the more-powerful—and therefore faster—Level 2 charging option. Muratori found that as more PEVs are added to a neighborhood, and a higher charging power is adopted, the distribution infrastructure might no longer reliably support the local electricity demand. He also noted the higher demand could shorten the expected life of a transformer.

The electric load profiles used in this paper, including household demand and PEV charging with a 10-minute resolution, are available for download in the NREL Data Catalog.

Earlier studies on how PEVs might affect the grid assumed utilities would have some control over when charging occurs, referred to as coordinated charging, which will greatly facilitate PEV integration. Muratori noted that might be true in the future, but not necessarily.

His research didn’t focus on using PEVs to return a battery charge to the grid to increase the reliability of the electric system. Future research, Muratori said, should focus on understanding consumer behavior to determine charging requirements, and the choice between using Level 1 and Level 2 residential charging equipment.

Matteo’s work raises important issues for a world with increasing electrification of the vehicle fleet, and leaves us with clear avenues for additional research. We need to continue looking at the synergies between electric vehicles and buildings, especially to make sure the grid remains safe and resilient.

—John Farrell, NREL’s laboratory program manager for vehicle technologies

Resources

  • Matteo Muratori (2018) “Impact of uncoordinated plug-in electric vehicle charging on residential power demand” Nature Energy doi: 10.1038/s41560-017-0074-z

Comments

Lad

This guy has affiliations with Exxon which raises the question of how biased his announcements are; the trouble is the uneducated populus will make decisions based on this FUD:
https://www.linkedin.com/in/matteo-muratori-6b031222/

mahonj

He has found a potential problem:
That if a load of rich guys buy fancy EVs and all try to rapid charge them at the same time, you will have a problem.
IMO, it is a bit of a corner case, but is worth considering.

The solution is to stagger the charging times across the whole night. You could even imagine a dual charge scenario: a few kwH when you get home in case you need the car soon, and top it off later on during the night.
If the average person drives 34 miles / day, this is about 11 kwH / day, which should be easy as long as you can stagger it.
You need to implement a pricing structure that encourages this.
You might need some kind of connected charging system which negotiates with other local cars as to when they charge (or just charges at a lower rate).
If you have a "big battery" car with more than one day's driving in it, this should not be a problem.
A 60 KwH Tesla 3 should be able to store 5 average day's driving, and so should have considerable flexibility as to when it charges.

SJC

Rapid charging can be done with battery banks, smart grids will charge cars at home in the proper sequence.

Bernard

How is this different from having everybody in a neighborhood running maximum A/C and cooking dinner at the same time? Electric cars are less of an issue because they can be programmed to charge overnight.

Engineer-Poet

Well, some can.  There's a menu in my Fusion Energi which says when deferred charging is supposed to start...

but...

it is not settable from the vehicle's own touch-screen!  Apparently you have to create an account at Ford and use a smart phone or other interface to actually use this functionality.

Calling this brain-dead is an insult to the brain-dead.

mahonj

@Bernard, it is different because if you want to cook a meal, you need to do it NOW- same with A/C.
However, you usually do not need to charge a battery right now, you just have to have ti charged before you need it the next morning. Thus, you could schedule it to charge from 4-6 AM (or whatever).
What you want is some system where all the local cars agree on when to charge so as to spread it across the night. Some people might want it right away - they would pay more.
Some might want it by 6am, some by 8am. You might pay less if you give more flexibility in terms of time.
You might pay even less if you allow them to take power from you as well as charging.
It sounds simple to me - the sooner you want it charged, the more you pay per KwH.

Roger Pham

The solution is surprisingly simple:
Use a PHEV instead of a BEV. A PHEV does not have the immediate need for charging vs a BEV, the latter must be kept at high charge level in case of emergency need for driving long distance. As such, a PHEV should be MANDATED to charge during grid's OFF-PEAK hours. An obvious advantage of a PHEV over a BEV is that a PHEV is NOT needed to be kept at high charge level all the time.

Additionally, a PHEV should be MANDATED to have an ability to sense the grid's voltage level, and NOT be allowed charge when it senses high power demand from the grid.

Engineer-Poet

Voltage literally doesn't matter, Roger.  You can run voltage almost as low or high as you want by drawing or injecting reactive power.  The thing which gives you the grid's state of power surplus/deficit isn't voltage, it's frequency.

And yes, vehicles could sense this automagically and ramp their charging up or down depending on how the phase is changing relative to a known clock.  The problem there is that if you have a lot of load which is ramping up and down to try to keep frequency stable, you might have unwanted interactions with the grid's own AFC (automatic frequency control) which does the same thing from the generation end.  You really want a second data channel which can tell the loads what to do, preferably by particular zones in the grid.  But that one can have a lot more latency than the direct sensing of frequency.

JMartin

This all seems to point to a need for energy storage. With intermittent generation and intermittent demand, we can no longer look to the grid and expect to produce exactly what is demanded at any time.

Engineer-Poet

As Conley and Mahoney write, we have energy storage.  It's called "fuel".

Roger Pham

Hi E-P and thanks for your feedback. What I have in mind is the following, from Wikipedia page of "Digital Protective Relay" that can be used to detect grid overload/ high-power demand situation to decide to start charging or to stop charging. Thus, a Plug-in EV can have detector of grid power situation to permit charging or stop charging.

"The digital protective relay is a protective relay that uses a microprocessor to analyze power system voltages, currents or other process quantities for the purpose of detection of faults in an electric power system or industrial process system. A digital protective relay may also be called a "numeric protective relay".

Input processing
Low voltage and low current signals (i.e., at the secondary of a voltage transformers and current transformers) are brought into a low pass filter that removes frequency content above about 1/3 of the sampling frequency (a relay A/D converter needs to sample faster than twice per cycle of the highest frequency that it is to monitor). The AC signal is then sampled by the relay's analog to digital converter from 4 to 64 (varies by relay) samples per power system cycle. As a minimum, magnitude of the incoming quantity, commonly using Fourier transform concepts (RMS and some form of averaging) would be used in a simple relay function. More advanced analysis can be used to determine phase angles, power, reactive power, impedance, waveform distortion, and other complex quantities.

Only the fundamental component is needed for most protection algorithms, unless a high speed algorithm is used that uses subcycle data to monitor for fast changing issues. The sampled data is then passed through a low pass filter that numerically removes the frequency content that is above the fundamental frequency of interest (i.e., nominal system frequency), and uses Fourier transform algorithms to extract the fundamental frequency magnitude and angle."

Engineer-Poet

Pretty much that, Roger.  A computationally cheaper way than A/D with Fourier transform is to use a zero-crossing detector and just wait for its output state to flip.  Most microcontroller families have "timer capture" as an option on one or more pins, which grabs the value of the system's clock counter when the prescribed transition occurs (the program can read the value of the capture register at its leisure).  Sampling of system voltage can be done halfway between crossings, at the peak.  More sophisticated functions will require more data, of course.

Local problems can cause major undervoltages and connections going bad from e.g. overheating can cause flicker.  The latter is cause to shut down and raise a fault flag; you can just pause and wait for the former to go away.  The broader grid surplus/deficit condition is telegraphed by frequency, in other words the time between those zero crossings.  If that time suddenly starts getting longer, there's an energy-deficit condition on the grid and backing off on the charging rate might be in order.

You can potentially respond to such conditions within 1/2 cycle, or 8.33 msec at 60 Hz.

Arnold

Are there any applications known aside from grid support services?
@E.P.
" And yes, vehicles could sense this automagically?? and ramp their charging up or down depending on how the phase is changing relative to a known clock"

I wonder if this technique has been used more widely.
I've mentioned the possibility as rational or logical for years but E.P's comment is the first time I've seen any reference. From my perspective is simply the mirror end of e providers or grid operators supply quality assurance.
On the numerous occasions I've tried to bring this question, the response has been silence. 0..?..No! can't be done. Not necessarily in that order.

If the grid operator uses frequency shifting as a response to under capacity and that the supply guarantee is within +-1Hz that is corrected over 24hours,
That a device capable of comparing grid frequency and trending to asses capacity sounds very simple.
Any device informed in this way could adjust it's demand.


Engineer-Poet
I wonder if this technique has been used more widely.

I don't know if it's being used anywhere yet.

I've mentioned the possibility as rational or logical for years but E.P's comment is the first time I've seen any reference.

I've seen references to it before, so it's not a new idea.  I may incorporate it in some patent claims that I'm revising almost as I write this.

From my perspective is simply the mirror end of e providers or grid operators supply quality assurance.

Yes, precisely.

On the numerous occasions I've tried to bring this question, the response has been silence. 0..?..No! can't be done. Not necessarily in that order.

The mentality of grid operators is that frequency control is done by generators, as it has been... to date.  They have not yet internalized that grid frequency is a balance between generation and load, and dynamic load control has just as much influence on frequency.  This will require an awakening on their part.

If the grid operator uses frequency shifting as a response to under capacity and that the supply guarantee is within +-1Hz that is corrected over 24hours

IIUC the frequency limits are far tighter than that, OTOO 0.15 Hz or maybe narrower.  Parts of the grid can have wider immediate tolerances than the average, because as power flows over lines reverse the phase relationship between the two ends necessarily reverses with them.  This means a shift in frequency over time at at least one end and probably both.

mahonj

It looks like you need a way to stagger the charging if there is high demand.
This can either come from frequency signals, or some direct approach, such as a controller run by the power supply company.
Frequency signals are IMO not a good idea as the same signal is given to all participants at the same time, which isn't what you want + the operator may try to keep the frequency as stable as possible irrespective of what is happening on the grid.
Ideally, you want a schedule which is automatically worked out such that everyone who wants to charge can do so. You could charge in small time slices (say 5-20 minutes at a time) so that no-one gets left out at the end, in case of shortage.
However, this requires that someone is responsible for running the charging timing algorithm and that everyone agrees to use it.
So: frequency signals, or direct algorithmic signalling or PHEVs.

Philip Davis

"Muratori also looked at the impact PEV charging might have on a residential distribution transformer.Even without large numbers of PEVs on the road, this clustering effect will significantly increase the peak demand seen by distribution transformers, thereby potentially requiring the upgrades, he found."

I'm sure the utility companies will have zero problem upgrading these transformers with the extra revenue from all that extra electricity they will be selling. That's how business works.

Of course they will try to put the onus on EV owners and require them to pay an extra fee and try to force them to charge only at off peak times so they don't have to upgrade equipment. That way they get to sell extra electricity AND don't have to spend any extra on upgraded equipment.

Arnold

Proposals around delayed power or synchronised schedule have suggested prioritising devices. Either by programmable device including V2G for large battery devices or I would suggest a simple device able to recognise the grids condition in real time will also have an adjustable set point for switching.
While the variation in manufacturing tolerance on its own will be of some help, I.E. there will be some spread, when aggregated across many different devices from a variety of manufactures and with the possibility for setting the switching point, I would't think that the grid oscillation would be as hard as first glance may suggest.

The other consideration is that in some countries and areas many premises have wireless 'smart meters installed and of course changes to meters on new installations or if vehicle charging is added will likely be required.
The average house will need >2x supply.

I guess the whole point of the term 'smart grid' as a work in progress is to foresee and adapt by various means for changing demands (as it must today) more efficiently using the existing infrastructure as well as cope with higher expectations and demands.

Engineer-Poet
Frequency signals are IMO not a good idea as the same signal is given to all participants at the same time

Except it isn't.  The phase you measure at your end of a feeder depends on the phase shift over the lines between you and the generator.  That phase shift depends on the load.  If the load is increasing near your end you'll measure a frequency drop, and vice versa.  Changing transformer taps can cause abrupt phase jumps.

the operator may try to keep the frequency as stable as possible irrespective of what is happening on the grid.

That's one of the operator's jobs; there is only so high/low frequency can go before generator specs call for them to trip off the grid.  One of the possibilities with frequency-responsive charging is that chargers could back off as frequency sagged toward the control limits.  This automatic load-shedding could happen much faster than the grid's AFC, letting slow machinery catch up to rapidly-changing conditions.  This is a substitute for mechanical inertia which is being lost as wind and PV increasingly displace conventional generation.

Ideally, you want a schedule which is automatically worked out such that everyone who wants to charge can do so.

Handling of the broad-scale grid condition and the load on the local distribution transformer are related but different problems.

Engineer-Poet

(dafuq? where my comment go?)

Frequency signals are IMO not a good idea as the same signal is given to all participants at the same time

Except it isn't.  The phase you measure at your end of a feeder depends on the phase shift over the lines between you and the generator.  That phase shift depends on the load.  If the load is increasing near your end you'll measure a frequency drop, and vice versa.  Changing transformer taps can cause abrupt phase jumps.

the operator may try to keep the frequency as stable as possible irrespective of what is happening on the grid.

That's one of the operator's jobs; there is only so high/low frequency can go before generator specs call for them to trip off the grid.  One of the possibilities with frequency-responsive charging is that chargers could back off as frequency sagged toward the control limits.  This automatic load-shedding could happen much faster than the grid's AFC, letting slow machinery catch up to rapidly-changing conditions.  This is a substitute for mechanical inertia which is being lost as wind and PV increasingly displace conventional generation.

Ideally, you want a schedule which is automatically worked out such that everyone who wants to charge can do so.

Handling of the broad-scale grid condition and the load on the local distribution transformer are related but different problems.

Engineer-Poet
Frequency signals are IMO not a good idea as the same signal is given to all participants at the same time

Except it isn't.  The phase you measure at your end of a feeder depends on the phase shift over the lines between you and the generator.  That phase shift depends on the load.  If the load is increasing near your end you'll measure a frequency drop, and vice versa.  Changing transformer taps can cause abrupt phase jumps.

Remember that grid frequency is what the generators use to sense if they should ramp power up or down.

the operator may try to keep the frequency as stable as possible irrespective of what is happening on the grid.

That's one of the operator's jobs; there is only so high/low frequency can go before generator specs call for them to trip off the grid.  One of the possibilities with frequency-responsive charging is that chargers could back off as frequency sagged toward the control limits.  This automatic load-shedding could happen much faster than the grid's AFC, letting slow machinery catch up to rapidly-changing conditions.  This is a substitute for mechanical inertia which is being lost as wind and PV increasingly displace conventional generation.

Ideally, you want a schedule which is automatically worked out such that everyone who wants to charge can do so.

Handling of the broad-scale grid condition and the load on the local distribution transformer are related but different problems.

Engineer-Poet
Frequency signals are IMO not a good idea as the same signal is given to all participants at the same time

Except it isn't.  The phase you measure at your end of a feeder depends on the phase shift over the lines between you and the generator.  That phase shift depends on the load.  If the load is increasing near your end you'll measure a frequency drop, and vice versa.  Changing transformer taps can cause abrupt phase jumps.

Remember that grid frequency is what the generators use to sense if they should ramp power up or down.

the operator may try to keep the frequency as stable as possible irrespective of what is happening on the grid.

That's one of the operator's jobs; there is only so high/low frequency can go before generator specs call for them to trip off the grid.  One of the possibilities with frequency-responsive charging is that chargers could back off as frequency sagged toward the control limits.  This automatic load-shedding could happen much faster than the grid's AFC, letting slow machinery catch up to rapidly-changing conditions.  This is a substitute for mechanical inertia which is being lost as wind and PV increasingly displace conventional generation.

Ideally, you want a schedule which is automatically worked out such that everyone who wants to charge can do so.

Handling of the broad-scale grid condition and the load on the local distribution transformer are related but different problems.

Engineer-Poet

Can I even comment on this post?  3 comments I've tried to post have vanished.

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Working...
Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.

Working...

Post a comment

Your Information

(Name is required. Email address will not be displayed with the comment.)