Utility Rates for EVs

[Robert.Boston]

Moving consumption away from the peak is good; but if so much usage shifts that the deemed peak hour-period no longer contains the peak, then you haven't solved the problem, merely shifted it (and inconvenienced a bunch of people in so doing).

 

[Norbert]

Err, no. Current retail rates average the cost-to-serve across all members of a rate class. If you could suss out the actual cost to serve each individual customer, you would find that some customers are paying less than their full cost, while others are paying more. That is, the cheap-to-serve customers are subsidizing the expensive-to-serve customers. When TOU rates are introduced, however, this cross-subsidies begins to unravel: the actual load profiles reveal who is expensive to serve, and who is cheap. Therefore, once TOU rates are introduced, approximately half of all customers will see higher bills, while another ~half will see a rate drop. TANNSTAAFL.

If TOU is optional, however, rational people will only choose TOU if their usage pattern results in a savings. This self-sorting ought to lead to an increase in the flat (non-TOU) rate, because the profile of the average use of people in that rate class is now more costly to serve. Until the utility applies for and receives that higher rate, however, it will be losing money. Which may be one reason why utilities aren't racing to introduce TOU rates.

Of course, obviously!

Everything else begin equal, the result should be studies which show that people are saving money when switching to TOU, but that less than 50% are interested (I say "less than" because it requires some effort to switch). Why don't they show that?

If you look at EV users which are thinking about using timer-based off-peak charging, then the win should be even larger, so perhaps more than 50% of this specific group.

Unless installing a TOU meter requires a fee, and this fee is too costly and falsifies the result of the "experiment".

Until the utility applies for and receives that higher rate, however, it will be losing money.

The question is how to get from A to B. I was talking about B, you about A, so you thought you disagree.

You could start by switching over specify groups only, for example EV owners first.

In the end, money will be saved for all, since the grid and the power plants are used more effectively, with a better return-on-investment.

 

[Norbert]

Moving consumption away from the peak is good; but if so much usage shifts that the deemed peak hour-period no longer contains the peak, then you haven't solved the problem, merely shifted it (and inconvenienced a bunch of people in so doing).

They will all cook dinner at 2 am then?

 

[Robert.Boston]

Depends on how wide your peak hour period runs. Central Maine Power's "peak period" in the afternoon is 2pm to 6pm; it's easy to imagine how a lot of demand could shift into the 6pm-7pm slot.

 

[Norbert]

Depends on how wide your peak hour period runs. Central Maine Power's "peak period" in the afternoon is 2pm to 6pm; it's easy to imagine how a lot of demand could shift into the 6pm-7pm slot.

IIRC PG&E has three "slots": peak, partial-peak, and off-peak. Off-peak is 12 am to 7 am or so. Your example suggests being careful with adjusting the savings for partial-peak.

 

[Dave EV]

Depends on how wide your peak hour period runs. Central Maine Power's "peak period" in the afternoon is 2pm to 6pm; it's easy to imagine how a lot of demand could shift into the 6pm-7pm slot.

Just look at the CAISO Today's Outlook page. Pretty clear how usage peaks in California.

Broad rise centered around 12pm and a sharp rise between 6-7pm with minimum demand between 1-5am.

This is typical for winter months. In the summer months when it's warm, the primary peak is in mid-late afternoon due to air conditioning use. The 6-7pm peak is not as significant. You can use the Renewables Watch page to see historical data.

 

[Norbert]

Just look at the CAISO Today's Outlook page. Pretty clear how usage peaks in California.

Broad rise centered around 12pm and a sharp rise between 6-7pm with minimum demand between 1-5am.

This is typical for winter months. In the summer months when it's warm, the primary peak is in mid-late afternoon due to air conditioning use. The 6-7pm peak is not as significant. You can use the Renewables Watch page to see historical data.

Thanks for posting it. It also shows that consumption between 1:30am and 4:30am is especially low, and that it would make sense to have some consumption balanced specifically into those 3 hours. With a 10 kW charger, that would be 30 kWh, or about 80-100 EV miles, and I know a few people who might drive that amount daily... those living in SF and working in Silicon Valley.

 

[rolosrevenge]

That's not a problem but the point of the whole exercise! Peak shifting (better call it "balancing") is what leads to a more effective use of power plants and therefore to fewer necessary power plants. That means it showed success.

No, peak shifting is when enough appliances shift their usage to create just as high of a power usage, hence a peak, but at a new time that is off peak from the TOU stand point. Then you have the same problem as before but people now pay less to cause it. That leaves you actually worse off than before. Another problem with TOU pricing that most people don't consider is load synchronization. Most of the thermostats and water heater usage is pretty uniformly distributed through time during peak hours. Having a TOU gets everyone to delay using until right after the peak time ends and then they all turn on. This synchronizes the heating cycles and can result in a higher peak than before, and an oscillatory load shape afterwards. It's a shame I can't attach a .pdf to this post or I would upload the module from the smart grid class on demand response and TOU pricing so you could get a better idea of all of the issues surrounding this.

 

[Norbert]

No, peak shifting is when enough appliances shift their usage to create just as high of a power usage, hence a peak, but at a new time that is off peak from the TOU stand point. Then you have the same problem as before but people now pay less to cause it. That leaves you actually worse off than before. Another problem with TOU pricing that most people don't consider is load synchronization. Most of the thermostats and water heater usage is pretty uniformly distributed through time during peak hours. Having a TOU gets everyone to delay using until right after the peak time ends and then they all turn on. This synchronizes the heating cycles and can result in a higher peak than before, and an oscillatory load shape afterwards. It's a shame I can't attach a .pdf to this post or I would upload the module from the smart grid class on demand response and TOU pricing so you could get a better idea of all of the issues surrounding this.

None of that would be a reason not to charge EVs between 12am and 7am and to get the better rate which it would deserve. This discussion is simply missing the key points.

 

[Lloyd]

None of that would be a reason not to charge EVs between 12am and 7am and to get the better rate which it would deserve. This discussion is simply missing the key points.

Just a note that PG&E E6 TOU rates are partial peak 5 to 8 pm November through March. After 8 pm and weekends and holidays are off peak.

 

[Norbert]

No, peak shifting is when enough appliances shift their usage to create just as high of a power usage, hence a peak, but at a new time that is off peak from the TOU stand point. Then you have the same problem as before but people now pay less to cause it.

Do this differently in two (or more) areas served by the same power plant, and you have cut the relative height of the peaks in half. Voila, you have "balancing" instead of "shifting". Also, in your example, the start of the partial-peak hours might be too early.

Earlier you claimed that TOU would not have enough of an effect on behavior. Now you claim it has too much of an effect. I do not believe that would be a real problem, except maybe in teaching material of a class or at the beginning of an early pilot project.

 

[Norbert]

Just a note that PG&E E6 TOU rates are partial peak 5 to 8 pm November through March. After 8 pm and weekends and holidays are off peak.

It seems in 2011 the start of off-peak time varies between 8pm and 10pm. I think it was 12am at some point in time.

However AFAIK, each utility is free to choose hours as they wish. Or at least different utilities seem to have different times. It will be easy to "prove me wrong" by dragging the discussion to that level of detail, since I'm not working in this field.

EDIT: If the graph posted by drees (above) also applies to PG&E, then it would seem that 8pm is far too early for off-peak.

 

[Dave EV]

EDIT: If the graph posted by drees (above) also applies to PG&E, then it would seem that 8pm is far too early for off-peak.

The chart applies to all of California - not sure how the demand varies by utility...

 

[GSP]

Unidirectional V2G doesn't have any discharging and thus does nothing to reduce cycle life. It allows EVs to be dispatched while charging around a given set point to perform frequency regulation, spinning reserves, non-spinning reserves, and peak shaving. Using proper control and revenue optimization can allow you to charge using real-time prices and actually be paid to charge. If you don't want control, the best sort pricing scheme is a double auction real time nodal price based on optimal power flow. The nodes in this case are determined by any distribution segment limited by congestion. But, it only works if enough appliances have a bid curve and your larger appliances like EVs are not price inelastic. If there are too many price inelastic customers and no control, you need to figure out what rate to charge people to pay for a bigger transformer. For Duke Energy's system, EVs can be charged the same rate as any residential customer as long as they charge at 6.6 kW or less.

RR - I think you are on the right track here with uni-drectional V2G. Hopefully SAE's new J1772 PLC protocol can support this and will be supported by all of the automakers, including Tesla, for AC charging as well as fast DC charging.

Are you saying that the 6.6 kW chargers require uni-directional V2G to work with Duke Energy's existing transformers, or is V2G only needed for more powerful chargers, like the Model S's 10kW and 20kW units?

GSP

 

[TEG]

Moving consumption away from the peak is good; but if so much usage shifts that the deemed peak hour-period no longer contains the peak, then you haven't solved the problem, merely shifted it (and inconvenienced a bunch of people in so doing).

Perhaps this is obvious, but I think the goal would be to equalize usage so that it is consistent all through the day & night, and no more "peak" period.
It would take a lot of overnight EV charging to balance things so that it matches the daytime peak usage.

And by the time EV saturation starts to become a big factor, I think they will have the 'smart grid' worked out so that they can stagger charging and even out the load so that the peaks are controlled.

 

[slcasner]

We're on the E7 plan, not the E9 plan. Which would be better? Not quite sure because it all depends on when you use the juice, but my bet is that E9 would be a bit more.

[snip]

So, I'd have to say, the relationship is somewhat adversarial - I don't trust them, they seem to do the minimum necessary to provide the service, often only when the PUC forces them to. Some honesty from them and some genuine moves to help customers and improve efficiency would go a long way. They are just not good corporate citizens.

Scott, definitely keep E-7. PG&E is proposing to change the E-9 tariff in a way that will significantly increase the cost for those who have solar to cover the usage.

I definitely agree with you for the second part I excerpted.