20
Capacity is another value that solar advocates typically contend is provided by solar 21
PV DG and should be compensated. This claim deserves close examination. The 22
capacity value of a generating asset is derived from its availability to produce 1
energy when called upon to do so. If a generator is not available when needed, it 2
has little or no capacity value. By its very nature, solar PV DG on its own, without 3
its own backup capacity (e.g. storage), can only produce energy intermittently. It is 4
completely dependent on sunshine in good atmospheric conditions. Unless sunshine 5
is guaranteed at all times at which solar PV DG is called upon to produce, it cannot 6
be relied upon to be available when needed. Moreover, even if all days were 7
reliably sunny, the energy derived from the sun is only accessible at certain times of 8
the day. In many jurisdictions, as discussed above, the presence and potency of 9
sunshine is not coincident with peak demand. Frequently, solar PV DG capacity is 10
greatest in the early afternoon, while peak demand occurs later in the afternoon or 11
in early evening.
12
The chart below, from EPRI, illustrates the divergence between solar PV DG and 1
utility peak on a national level:
2
3
The following two charts show the relationship in New England. They are derived 4
from a 2013 report, “Update on Solar PV and Other DSG in New England,”
5
prepared by ISO New England, and illustrate that facile assumptions made that 6
solar benefits include near-term reductions in peak generation are precisely that.20 7
20Black, John. “Update on Solar PV and Other DG in New England.” ISO New England (June
2013).
1
These two charts dramatically demonstrate that on the days chosen as representative 2
of summer and winter in New England, solar PV peak and peak demand are not the 3
same. Solar PV is completely absent during the winter peak, reaches its peak 4
production as peak demand is rising in the summertime, and drops off dramatically 5
during almost the entire plateau period when demand is at peak. It should also be 6
noted that on the days chosen, the sun was shining. The graph, of course, would 7
look very different on cloudy days when solar production is virtually nil.
8
In the western United States, a similar problem is nicely illustrated by the California 9
duck graph (discussed above), with its steepening ramp to peak as it projects the 10
demand trajectory of a solar intensive system.
11
In SRP itself, where the utility’s summer peak rate applies from 1pm to 8pm, and 12
the winter peak rate is applied from 5am-9am and from 5pm-9pm, a similar 13
disjunction between solar peak and demand peak can be seen in the chart below, 14
which illustrates the four-hour summer gap between the solar peak and the system 15
1
% of peak demand compared with percentage of peak rooftop solar production by hour on July 8, 2013
2
in SRP. Data provided by SRP.
3
This difference between the two peaks was commented on for the APS region in 4
report on the value of solar in Arizona prepared for the Arizona Corporation 5
Commission by the technology and engineering consulting firm, Science 6
Applications International Corporation (SAIC):
7
Output from solar PV resources is only partially coincident with the peak demand 8
of the APS load. The APS system peak is somewhat unique, in that it extends past 9
sunset due to the impact from the desert heat. This contributes to a lower 10
- 0.20 0.40 0.60 0.80 1.00 1.20
1 3 5 7 9 11 13 15 17 19 21 23
% of system peak
% of rooftop solar peak production
coincidence with solar PV production than otherwise would be expected with non-1
desert utility service territories.21 2
It seems that SRP, far from being an exception to the overall pattern of the solar 3
peak production coming hours before system peak, is an excellent illustration of this 4
problem—which causes utilities to have to ramp up for peak even faster than would 5
otherwise be necessary.
6
As noted earlier, providers of capacity in the wholesale market may also have 7
availability issues. In their case, however, if they are not available when called upon 8
to produce, they are typically obligated to either provide replacement energy or to 9
pay the marginal cost of energy that they failed to deliver. In short, in order to 10
obtain a capacity payment, a generator has to accept the risks associated with being 11
unavailable when called upon to deliver. Unless a similar obligation is imposed on 12
solar PV DG providers, the capacity value of solar PV DG is virtually nil. In fact, 13
given the fact that solar PV DG production is generally in steep decline as power 14
system demand is at peak, it is extraordinarily unlikely that a solar host would be 15
willing to assume any such risk, Given this reality, it is very difficult to attach any 16
credibility to assertion that solar PV DG necessarily causes reductions at peak, or to 17
attach much capacity value to solar PV.
18
21 SAIC, 2013 Updated Solar PV Value Report, Prepared for Arizona Public Service, May 10, 2013: 2-18. https://www.azenergyfuture.com/getmedia/77708c68-7ca6-45c1-a46f-84382531bae3/2013_updated_solar_pv_value_report.pdf/?ext=.pdf
Good pricing policy would suggest that DG prices should be fully reflective of the 1
value of the type of capacity that is actually provided. Determining the actual 2
capacity value of solar PV DG is a fact-specific question that should consider 3
capacity availability resulting from timing of generation and the frequently less than 4
optimal placement of photovoltaics. What is important to note in regard to the SRP 5
proposal, however, is that abstract debate over solar’s value in terms of availability 6
at peak is largely irrelevant. That is because, under the SRP pricing proposal, this 7
kind of analysis is really not necessary--to the extent solar PV DG customers do 8
consistently provide power at peak, or at least reduce their demand for power at 9
peak hours, this will be reflected in lower electricity demand charges and higher 10
prices being paid for energy provided to the grid at peak hours. In short, SRP’s 11
proposal effectively matches the price with the value provided. Indeed, the SRP 12
proposal might be called “green pricing,” because it pays the highest price to solar 13
energy produced at peak demand, at which time it is safe to assume the “highest 14
emitting” plants are likely to be running and would be displaced.
15
Reliability
16
Reliability is another supposed source of value for solar PV DG that deserves 17
careful examination. Many solar advocates assert that solar PV DG enhances 18
overall reliability because the units are small and widely distributed and are close to 19
load and not reliant on the high voltage transmission system. It is argued that they 20
are less impacted by disasters and weather disturbances. These claims are highly 21
speculative and, as explained below, not necessarily accurate. It would be a serious 1
mistake to simply assume that solar PV DG improves reliability.
2
Solar PV DG is subject to disaster as much as any other installation. Strong winds, 3
for example, can harm rooftop solar as much as any other facility connected or not 4
connected to the grid. Cloudy conditions can disrupt solar output while not affecting 5
anything else on the grid.
6
Solar PV DG has more reliability benefit in some places than others. In Brazil, for 7
instance, a system that relies on large hydro plants with large storage reservoirs, 8
solar has considerable long-term reliability value because whenever it generates 9
energy it conserves water in the reservoirs associated with baseload units, thereby 10
adding to the reliability of the system. However, in a thermal dominated system like 11
SRP, even though some pumped storage is available, reliability has to be measured 12
on more of a real-time basis. Therefore, solar’s intermittency makes it unable to 13
assure its availability when called upon to deliver energy. Indeed, it is far more 14
likely that a thermal unit will have to provide reliability to back up a solar unit than 15
the other way around.
16
It is also important to examine rooftop solar reliability issues in two contexts: that 17
of the individual customer and that of the system as a whole. Solar vendors, as part 18
of their sales pitch, claim that reliability is increased for a customer with a rooftop 19
solar unit because on-site generation provides the possibility of maintaining electric 20
power when the surrounding grid is down. When the sun is shining, this claim is 21
likely to be true.
22
Conversely, without the sun, the claim has no validity. That argument, however, 1
only applies to the solar host. During a system outage the power inverter, an 2
electronic device or circuitry that converts direct current to alternating current, is 3
automatically switched off to prevent the backflow of live energy onto the system.
4
That is a universal protocol to prevent line workers from encountering live voltage 5
they do not anticipate. Thus, if a solar PV DG unit is functioning properly, when the 6
grid goes down, the solar PV DG customer’s inverter will also go down, making it 7
impossible to export energy. If the solar PV DG unit is not functioning properly, 8
then the unit may be exporting, but will do so at a considerable risk to public safety 9
and to workers trying to restore service. The result, of course, is that the solar panel 10
provides virtually no reliability to anyone other than perhaps to the solar host. There 11
are virtually no reliability benefits for the system as a whole in SRP, and therefore 12
no basis for calculating a payment for such service.
13
In general, attributing reliability benefits to an intermittent resource is a stretch. By 14
definition, intermittent resources are supplemental to baseload units. The only 15
possible exceptions to that are, as noted above, where there are individual reliability 16
benefits, where the availability of the unit is coincident with peak demand, or where 17
the units are part of a system dominated by hydro power (which provides storage).
18
Absent those circumstances, and absent other storage, it is almost certainly the case 19
that the system provides reliability for solar PV DG, rather than the other way 20
around.
21
From an investment perspective, any kind of pricing that requires utilities to pay for 1
the reliability “value” solar PV DG brings to the grid are detrimental to reliability 2
because they deprive utilities of the revenue needed to maintain high levels of 3
service. For utilities, the diversion of funds leaves them with the Hobson’s choice of 4
either delaying maintenance and/or needed investment, or seeking additional funds -5
- in effect, another cross-subsidy from non-solar PV DG users. It is also relevant to 6
reliability to note that the prevalence of intermittent resources on the grid, including 7
solar PV DG, may well cause new, cleaner, and more efficient generation to appear 8
less attractive to investors (as noted above in discussing the environmental impacts 9
of solar PV DG). Over the long term, that effect could lead to reliability problems 10
associated with inadequate generating capacity, especially at times of peak demand.
11