The Potential for Solar Energy


So today, it’s our great fortune and my
pleasure to be hosting Dr. Jones-Albertus. Becca is the acting Deputy
Director of the Solar Energy Technology Office, which is known to many of us as
the SunShot program in Washington. That office has been instrumental in trying
to do a lot with small resources in terms of bringing solar to a cost
competitiveness with coal. And their… their research impact, and particularly
their impact on creating and informing the community of US researchers and
solar, has been absolutely tremendous. So Becca has been with SunShot, that is
said SETO, solar energy technology office, for four years. Prior to that, she was in
the private sector working on dilute nitride materials for PV. And prior to
that, she received her PhD in material science from Berkeley and undergraduate from Princeton. I’ll mention that—I want to get on with the talk, but I will mention
that papers like this—this is 2016 paper right—in Prague photovoltaics,
Becca is the first author, are absolutely essential. Anyone who wants to get
involved in PV or remain relevant in PV should read these papers to inform
ourselves of how to really move the needle on global carbon. So I find these
very influential and they really help me guide my own research, so I hope this is
informative to everyone else as well. Alright, thank you, Becca. Thanks very much, Raf. It’s my pleasure
to be here talking with you all today. So I’ll be talking…I titled my talk The
Potential for Solar, and I’m excited to give a little bit of historical context
about how much solar has grown and the advances that have occurred really
recently, as well as talking a little bit about where things can go in the future.
So before I dive in, I just wanted to give an overview of solar technologies. I
imagine that most of you in this room are pretty familiar with photovoltaics.
And these are semiconductor materials that, you know, most commonly silicon, that directly convert light into electricity. The cells shown here are integrated and
strung together in modules which then make up systems like you see on this
rooftop. One of maybe the less highlighted but exciting things about PV
is that it maintains its efficiency at any size, effectively. So you can have
very very small systems that have the same power efficient generation
efficiency as very large systems, you know, roughly speaking. And that’s very
different than most power generation systems that are based on generators and
turbines that need sufficient size in order to be efficient. The other
technology which may not be as well known to all of you is concentrating
solar thermal power and that’s the pictures shown here.
So with concentrating solar thermal power—CSP as we call it—you use mirrors
called heliostats to concentrate all the light onto a single receiver. In this
case, the receiver is shown here. It’s a power tower, so you concentrate all the
light onto this receiver where it’s turned into a heat and that heat can
then be stored until it’s needed. Then it’s used to run a conventional turbine
generation system, so again, what…what’s …what’s exciting about CSP is that
it can inherently incorporate storage, so you can inherently have solar
electricity on-demand, not just when the sun is shining.
Today the large dominance of technology that’s been installed that’s solar is PV.
In the U.S., CSP is about 2% of installed solar technology, but again, CSP allows
for some higher value solar energy. As I said, I see this as a really exciting
time for solar. It’s just been amazing how much change has occurred over the
last decade and even…even more recently than the last decade. In the last 10
years, we’ve had a hundred-fold times more solar generation installed in the
U.S., going from less than one hundredth of one percent of U.S. electricity to,
for the first time last year, generating more than one percent of U.S.
electricity. Taking solar from a very expensive niche technology to a viable
electricity source that last year, for the first time, was the largest share of
new electricity generation capacity added to the grid, for the first time. So
39% of all new power generation capacity added last year was solar, was
photovoltaics, with natural gas and wind being the next largest shares. The solar
industry has also been a tremendous source of job creation and that has, you
know, gone commensurately with the increase in installations. So you see here over the same time period the growth in jobs in the solar industry. There were two
hundred sixty thousand jobs in the solar industry last year, and solar has been
growing at a rate that’s 17 times faster than the overall economy. More than half
of these jobs—the ones in dark blue here— are installation jobs, so actually
installing the new PV systems, with the next three largest shares being project
development, sales and distribution, and manufacturing across the supply chain. Now as I said, one percent of U.S.
electricity comes from solar, but there are some states…parts of the country
where those numbers are much higher, not shown here. These numbers are from last year. Every time, essentially, I look at these numbers they grow and the ones you see
next year will be, in many cases, significantly higher than here. So you have
California, which last year got 13% of its electricity from solar. A little over
1% of that from CSP, commercial and residential systems, distributed PV like
the picture of PV on the rooftop that I had there next, and then this largest
share from utility scale systems like this topaz plant which is a 550 megawatt
plant in California. You see we have states—Hawaii, Vermont, Nevada—
all over 7% electricity from PV and growing. Worldwide the trends are fairly
similar. This data is a year older from 2015, but solar generated about 1% of worldwide electricity in 2015. And again, with some countries in this
case, having much higher penetrations. In this case, Italy, Germany at about 8%
being the…being the leaders. And solar is projected to continue to grow, showing
here baseline projections that are from the National Renewable Energy Labs
regional energy deployment system model. And for 2030 these are a pretty good
agreement with the Department of Energy’s Energy Information
Administration, EIA’s, projections as well. And what you see for the orange
bars here which your solar, is that in sort of the baseline case, baseline
expectation, solar is expected to grow to about 5% of U.S. electricity in 2030 and
then roughly 15%. EIA’s projections are a little lower,
about 12%, by 2050. But again, significantly greater amounts of U.S.
electricity growing over the next few decades. And at the same time, projections
have historically underestimated PV’s growth. These are U.S.
projections from EIA’s annual Energy Outlook and the black
dots are the actual installations in the U.S. and then the lines here are the
projections for solar installations for subsequent…subsequent years. And
basically the trend you see is that the actual black dots in general, you know, are
consistently above the lines where they’re projected. Same has been true
worldwide looking at the IEA’s World Energy Outlook, where the historical..the
actual projections in black continue to exceed historical projections. So why has
solar growth been been growing so rapidly? One of the big factors is it’s
been rapidly declining in cost. So again U.S. installations growing—we’re
seeing deployment or…sorry costs, here’s for system cost,s falling by about a
factor of four at the same time that these installations are growing so
rapidly. Now that was in system price, I’m going to say a few more things about
cost, but switched to a different metric— the levelized cost of electricity. When
talking about costs of solar electricity, I really prefer the LCOE metric to
system costs or other cost metrics and that’s because it’s a lifecycle cost. It
includes the cost of installing the system, the cost of operating…the
operations maintenance costs of the system, tax implications through the
actual taxes paid and depreciation, expense, as well as the cost of the
financing, the cost of the capital, which is a significant factor in overall cost,
in addition to any residual value or decommissioning cost. So it’s the full
costs over the life cycle divided by the power production over the lifetime. So
not just how much power is produced when the system is installed, but it accounts
for how much degradation the system experiences over time, so more reliable
systems produce more power over the lifetime—that’s a lower life cycle cost.
This is also influenced significantly by where these systems are installed. The
actual kilowatt hours per kilowatt, so that’s…that’s actually like the climate,
how much sun is there, what’s the temperature of the region. So now moving
to LCOE and just talking more about cost reduction, very recently we announced
through the Department of Energy the accomplishment of a big metric for us. So
in 2011, the Department of Energy announced the SunShot initiative. At that
time, the big challenge for solar was that it was too expensive. It was about a
factor of four more expensive than conventional electricity sources. So for
solar to become really a viable electricity source, its cost had to fall.
And the SunShot initiative was launched to drive those cost reductions forward.
The goal was by 2020 to achieve, for utility scale systems, the large systems,
six cents a kilowatt hour. We announced in September that that goal was achieved
this year. So again, these…these factor of more than four cost reductions happening
even faster than was..what was seen to be a very aggressive goal when the SunShot
initiative was was announced. Now the SunShot initiative also had goals for the other
sectors—residential and commercial systems. And in red here, you see again
the costs…levelized cost of energy in 2010 in those sectors. And green is the
bars in 2017, so costs have fallen dramatically in these sectors as well,
but we haven’t achieved the SunShot 2020 goals yet in those sectors. And the blue
bars, they still need further cost reduction, particularly in the soft cost—
the customer acquisition, the installation, the interconnection, the
permitting costs—those still need to fall further to reach the 2020 goals. So
cost reduction was a major factor in this increase in deployment we’ve seen.
Policy and incentives have been the other major factor. As you see on this
map, just showing the large number of states in the U.S. that have renewable
portfolio standards that are driving additional investment in
renewables. We also have the federal investment tax credit, which is a 30
percent tax credit, so effectively a 30 percent reduction in the upfront system
costs, not in the full LCOE, but in the upfront system cost of a PV system. Both
of those have also been, you know, very important factors in driving the rapid
growth we’ve seen in solar installations. So as I said before when the SunShot
initiative was launched in 2011, the big challenge was cost. How do we get solar
cost down? Today, solar’s become a viable electricity source. It’s 1% of our electricity
generation and the challenges are changing. Now the challenge is there’s…
our grid integration. How do we get solar power, which is a variable power source,
which is being installed on the distribution side of the grid, how do we
integrate that well into the electricity grid and maintain a reliable resilient
grid? As well as, how do we deal with…I’ll talk more about the declining value of
PV. As more PV’s installed its value decreases. And it turns out that
continuing to focus on cost reduction is an important strategy for that, but this
is no longer the only focus and it’s no longer the only challenge and
opportunity with solar. So saying more about great integration
challenges, I’ll put up the schematic here. So historically our electricity
grid has looked something like this. We have generation systems on the left.
They flow through transmission lines to substation out onto the distribution
system. Power flows in one direction. With solar, we have solar on the generation
transmission side, but we also see solar on the distribution system. Solar
is not the only distributed energy resource, but it is the most significant
one to date. And with lots of generation on the distribution system, when it
reaches high enough…high enough penetrations, you can actually have cases
where you have power flowing in two directions and not just one. This creates
new sets of challenges for the grid. For example, ensuring that the systems that
are designed to regulate voltage and frequency on the grid aren’t
experiencing reliability issues due to operating more readily. There are
challenges in how does solar begin to support grid reliability itself through
voltage frequency, power quality regulation? As solar becomes a more
significant power source on the grid, there’s also the need to manage and
integrate it and deal with its variability. So we’ve always had load
that has high amounts of variability, but having variability on the power
generation side is something that is being brought from wind and solar.
There’s also a need for new standards for how these devices integrate and
operate and what they’re allowed to do with the grid.
So just some of the challenges…go to a similar schematic, but that illustrates
in addition as we’re moving to what we call a modernized grid, the Department of
Energy has a large initiative called the Grid Modernization Initiative, which is
looking at how we take our our grid and move to a modernized structure. We need
to pay attention with solar in making this integrating into a secure grid with
communication and data, you know, at the forefront, and in addition one that
operates and works within an evolving energy marketplace. So as solar becomes a larger share of our power generation one challenge is can we make it available on
demand? As I said before, you know, as a variable power generation source, there
are certainly challenges. And what’s shown here in blue is sort of a
typical or exemp…an example load profile, what electricity generation
typically might need to meet. In orange is the example of the well known duck curve. So you see solar generation that happens during daytime hours and what it does to
the net load, this is basically the load of minus the solar generation, is it
means that you can have a very low dip, especially in sort of morning midday
times, when the load isn’t quite as high and solar generation can be quite high,
and then as you get toward evening and solar generation is falling, you go back
to that normal load curve, but with a much steeper ramp than you have in the
usual case. So through storage or shifting of load, there are opportunities
to better match solar generation and demand and take what is represented by
this red arm, instead of having this fall down so far, take this area of generation
in red and shift it over to the evening hours where it’s needed. Or conversely,
you could imagine shifting some of the load from these hours into these…these
hours here. But effectively, how do we better match the supply of solar
electricity with the demand and need for… need for solar? And I guess I’ll say
the challenges with the duck curve for those of you who aren’t familiar with it,
are that as this net load falls lower and lower, it can reach the levels of
sort of minimum generation and contracts that exist for sort of base load
generation. And what that means, it’s really just an economic issue, it means
you’ll end up throwing away some of that solar power.
It’s called curtailment. At low levels, when curtailment happens, just in, you
know, certain days in March—like has been the main case of the duck curve to
date, there’s not a huge economic impact on solar. But as you get higher and
higher penetrations of solar, electricity curtailment can have an important
economic impact. And then from the grid operators’ perspective, managing an even
steeper growth of load ramping is an additional challenge. But there are also
important opportunities from the grid integration side with solar electricity,
and in particular, putting solar on the distribution system offers opportunities for resilience that we’re just starting to
look at. When you have solar in the distribution system, if you have a case
where there’s a loss of electricity, where there’s a blackout, it is possible
for solar and other distributed energy resources to restart areas of the grid,
for those areas to begin to work together and supply power while overall
power generation is being restored. So there’s some really exciting
opportunities on the resilience side that are possible by having solar and
other distributed energy resources on the distribution grid. We have a project
that’s just kicking off at Lawrence Livermore National Lab through the Grid
Modernization Initiative, that’s looking specifically at how to use solar and
storage and other distributed resources to be able to provide power during
blackouts and help restart the grid. Another opportunity is with
concentrating solar power, as I talked about before, concentrating solar power
inherently incorporates storage or can inherently incorporate storage to allow
for solar on-demand. It’s also possible with the same components for a
concentrating solar system to create plants that operate more like peakers.
Plants shown here when you have, for the same sized power block, for a 50 megawatt generator, if you have a larger…or sorry a smaller solar field and smaller
storage units you can have more of a peaker plant configuration. If you keep
the power block the same and you scale up the size of the solar field and the
size that the storage units, you can move to intermediate or baseload plants. So
basically with the same technology components, you can have flexibility in
the kind of power plant that’s built depending on what the grid needs. And then talking a little bit about
declining value that I mentioned earlier, so as we see the penetration of PV on
the grid increase, there’s a decline in the value of PV. So this is another
challenge to work on. This plot here is showing this—it’s actually showing it
from the sort of an effective cost perspective. So what this is showing is…
is you, for a case study of California, and this is Paul Denholm’s work at NREL,
for a case study of California looking at as you put more solar on the grid,
this is the total amount of California’s load that is met by solar
energy, so as you put more solar on the grid, how does the effective cost of that
solar change due to curtailment, due to solar energy that can’t be used because
there’s too much of it at a time when it’s not needed? Another way to look at
that is it’s sort of the energy value of solar is going down. So the blue
line is the overall cost for all the solar on the grid, but what’s more
important in terms of where’s…where you find an economic limit is the marginal
cost. So for every bit of solar that’s added, what is that marginal cost? You see that in this case, at about 20 percent solar
on the grid, you see the marginal cost going…going steadily up, suggesting that
it would not be economically viable to add more solar onto the grid. This is for
what I’ll call an inflexible grid, where flexibility is the ability of the grid
to rapidly adapt to changing supply and load demands. So this is looking at sort
of maybe the structure of the grid we have today. There’s lots of ways, and this
is probably too small to read, but there’s lots of ways that the…to
increase the flexibility of the grid and this kind of outlines them. Froman
operation perspective, there’s things like better forecasting of when you’ll
have renewable energy generation, having more flexibility reserves.
There’s also allowing for the use of variable renewables of wind and solar
to provide grid services such that there may not be as high a need from sort of a
minimum generation from other generators, so that more renewables can
be used. There is flexibility in when load is provided. Here this focuses
mostly on demand response. I’m personally also very interested in
exploring the overall ability of load to be shifted across not just demand
response, but across general usage of electricity load. And if there’s any of
you in the audience who have done any work on that, I’d love to connect with
you afterwards. More flexible generation, so if other electricity generators are
able to ramp their production up and down more rapidly, then that makes it
easier to adjust to rapid changes in supply and demand.
There’s also transmission expansion. So if the region’s over which we balance
energy supply and demand become larger, then again that can accommodate a more
flexible grid. You know, for example, if you’re able to generate…take your excess
solar generation in Arizona and ship it, you know, far across to an area that
doesn’t have sun at that moment, then you have a natural way to utilize that
excess solar power. And last here is storage. I’ll talk more about storage
later. Storage is probably the biggest lever on overall flexibility, but today
it’s one of the most expensive. And just to show…this is again, the same marginal
cost for adding additional solar to the grid. This orange curve is the one I
showed before for the somewhat inflexible case. As you make the grid
more flexible, you see this just push further out. So now instead of finding an
economic limit at 18 to 20 percent, you’re out closer to 30 percent and with
even additional flexibility which is here represented by a significant build-out of CSP plants with thermal energy storage, you
see the curve push even further. So that… that’s looking at the value of the
energy power from solar has both an energy value and in most markets also a
capacity value ,which is an additional value due to typically solar’s good
match at low penetrations to times of peak load. This study here from Andrew
Mills and Ryan Wiser at Lawrence Berkeley Lab is
from 2012 so it’s a little bit old, but the trend is basically the capacity
credit, the value for PV installed to the grid, due to its ability to reduce the
overall peak demand, goes down as… steadily down as PV penetration
increases. And you know, it’s different for different scenarios, but anywhere
from 5 percent to 15 to 20 percent there’s no longer any value, a capacity value, for
additional PV generation. To explain…to explain what that means or how that
comes from, I’ll show these example load profiles here. And in gray here, this is…
this is load without solar, peak load demands. So you see that in terms of
determining how much capacity you need to meet peak load demands, grid operators
will look at this peak here, you see, in green. Now the lines here are
increasing amounts of solar energy on the grid going from 2% in blue to 14% PV.
And for this example, you see that as you add more solar, where that peak steadily
decreases until the point, in this case, where you get to 10 percent where there’s no decrease in where that peak point is between 10 and 14 percent. So in this
case, the capacity value goes down and it is at 0 above 10 percent PV on the grid.
So above 10 percent, the peak…where the peak load position is, no longer changes
as you add more solar and the capacity value
declines. So the opportunity here to sort of combat declining value of PV as
generation increases is to continue to reduce costs. And last year, about a year
ago, we announced new cost targets for 2030 for PV systems. Cost targets are
three cents a kilowatt hour for utility-scale PV, four cents for
commercial systems, and five cents for residential PV systems by 2030. So this
is a 50% reduction compared to the 2020 cost targets,
so for utility-scale PV it’s a 50% reduction from where we are today. For
the commercial and residential sectors, it’s about a factor of three reduction
from green where we…where we are today. And as I talk about these costs, I want
to point out that as I…as I mentioned earlier, LCOE, the levelized cost of
energy, depends on the climate, depends on where the system’s installed. The same
system installed in a really sunny area will generate more power. That generation
is in the denominator of LCOE, so the overall cost per kilowatt hour goes down.
Same system installed in a less sunny area will by the same factor, be
more expensive per kilowatt hour that’s generated. And at DOE, we are focused on
enabling solar energy for all Americans across, you know, the U.S. and so we use
average climate for our cost targets. So the three cents a kilowatt hour we
calculate is for average climate which we represent by Kansas City, Missouri.
That same system installed in really sunny Arizona or California would be
almost two cents a kilowatt hour. Similarly in Seattle, one of our least
sunny regions, it would be four cents a kilowatt hour. And one of the reasons
this is important is if you’re thinking about costs that you hear reported for
solar, typically when folks are talking about really low costs
of solar and PPAs that are being signed at four cents a kilowatt-hour,
that’s almost always for very sunny regions and it’s also typically
including the…if it’s in the U.S., the investment tax credit. So that will cause
you…you know, you could hear numbers today of four or five cents a
kilowatt-hour, but those numbers reported in that way aren’t comparable to the
targets I’m talking about. And you know, how much does this matter? How much would cutting costs 50% matter? So to get a sense of that, use again projections from
NREL’s regional energy deployment system. And this shows this business as
usual case, where again, you know, solar is expected to be about 5 percent of
generation in 2030. With cost being half that in 2030 instead, so halving the
cost, you get more than double the solar deployment. In this case, you know, lower
solar costs again come they..they counteract the decline in value. At three
cents a kilowatt-hour solar also would be cheaper than many existing…operating
many existing power plants. So it could actually be cheaper to install and use
solar energy than to run some existing generators, which could lead to an
overall reduction in electricity prices, making electricity more affordable. So
I’m going to talk about a pathway, how would we, you know, moving to the
technology side, how would we get from where we are today to three cents a kilowatt- hour? Is that practical, like what would that look like? So to walk you through this
waterfall, and to get from the six cents cost to the three cent target, here’s one
example pathway. Of course there are others, but this illustrates what it
would take. In this case, you’d be looking for module prices to fall to about 25
cents a watt. This bucket here about 0.4 cents, might be too small
because it’s based on sort of reported prices on the
market today, which may not be what we call sustainable, so it may actually be a
little bit larger. What we’re targeting when we target reductions in the module
costs or we’re targeting what we call sustainable reductions, which allows for enough
profit for module manufacturers and in the supply chain for them to continue to
grow capacity. So it’s not what they’re selling in cases of oversupply on the
market, but it’s what they would be selling at with enough sustainable
profit to continue to grow their businesses. So ways in which we can get
there, you know, one key way is improving the efficiency of systems without
increasing cost or while decreasing costs. Next bucket here, larger bucket, is
lowering the balance of system hardware and soft cost, so this is everything from the
inverter, the wiring, the racking of the system, to the installation and our
connection and permitting costs. Improving the overall system efficiency
does help with this bucket as well. And things like speeding up, standardizing
installation and interconnection processes can help this as well.
Third buckets, also a large bucket in this scenario, and this is an area of
particular interest to us at the Department of Energy’s solar office
right now, is improving the reliability of the systems. It’s a big lever. If you
take systems from 30 year lifetimes to 50 year lifetimes, you reduce degradation
rates to 0.2 percent a year from a half to 0.75 percent a year, that can be a big
big lever for cost reduction. And I’ll show that in a different way on the…on
the next slide. So better understand… ability to understand what causes
degradation and also better ability to predict that, so that as structures and
devices change, we can immediately understand how that impacts their
reliability. Last is lowering of operations and maintenance costs. It’s an
important bucket in this case too. And here, employing automation and better data
analytics to better predict what maintenance is needed and more quickly
diagnose issues improves characterization tools help as well here. But as I said before, there are a number
of pathways to get to these targets and I’ll illustrate the number of
pathways here just from a module technology perspective. So in this plot,
it’s looking at module price on the y-axis and module efficiency on the
x-axis and different reliability cases on the curves I’ll show. Everything else
in this scenario is held constant. So in this case, if you just look at the
trade-offs between module price and efficiency for a system that is a 50 or
a lifetime 0.2 percent per year degradation rate, you see that you can
hit the SunShot 2030 cost targets with a module that’s 25% efficient that costs 30
cents a watt. If you have a more efficient module, it can cost more. In
this case 35% efficient, it can cost an extra five cents a watt.
Conversely, if it’s less efficient, if it’s more like sort of the average
efficiency on the market today of 17% well, it has to be cheaper, it has to be 20
cents a watt. Illustrating the importance of reliability, here’s the same curve for
a system that is more like the lifetimes and degradation rates assumed today—30
years, half a percent per year performance degradation. Here, looking at
the 25 percent module now instead of 30 cents a watt, if all of the change is
borne by the module price, it now has to be 13 cents a watt, so it has to be
substantially cheaper for this lower lifetime case. And if you look at even
lower yet still not that low lifetimes, in this case, 20-year lifetime,
one percent per year, it becomes very difficult to achieve these cost targets.
Here you’ve, you know, if your module were free, it would have to be 27% efficient.
If it was 40% efficient, it could only cost about six
cents. This is holding everything else constant. Lots of folks who get
interested in new materials and new system possibilities that could be a lot
cheaper, might not last as long, are also looking at possibilities that would
reduce installation costs and some of the other costs. Changing some of the
other costs, if this kind of a system enabled that, would push this curve out
like this, but I think one of the messages I want to drive home on terms of how important reliability is it’s going to be very hard to achieve the substantial
cost reduction targets without systems that are at least as long-lived as today,
without getting to 30 year lifetimes…25 to 30 year lifetimes. We also have cost
targets for residential PV systems as I showed. This is just the waterfall here.
Again, similar buckets, but what you see here is a much much larger bucket than
any seen on the previous slide and this is for the soft costs. As I mentioned
before, the commercial and residential systems have a much larger soft cost
component. Just blowing this up to get this kind of a reduction would require
major reductions in customer acquisition costs, permitting interconnection taxes,
installation labor, the supply chain costs, as well as the profit and overhead
costs of the installers and developers themselves. The other bucket that’s on
here that wasn’t on the other waterfall is lower finance rates. The cost of
capital, the financing of the system, is actually a major lever in the overall
levelized cost of energy. And if residential systems were able to obtain
lower financing rates, if they could be tied into mortgages, things like that,
that could also be a big cost reduction lever. Similarly, we have cost targets for
CSP systems. CSP systems are utility scale systems and you’ll notice that the
2030 cost target here is 5 cents compared to for PV, the utility scale
system is 3 cents. The higher costs target is possible
because CSP systems have a higher value due to their incorporation of energy
storage, so they can be competitive at comparatively higher costs. For getting
costs down with CSP systems, here’s some example buckets. A big fraction of
the cost of a CSP system is the mirrors, the heliostats, a solar field, we have a
number of different terms used, but finding cheaper ways to collect the
sunlight and concentrate it is an area of critical need. rRducing the cost of
the power block and improving the efficiency of the power cycle are other
other large buckets. CSP technology is…a great opportunity for that is to
incorporate supercritical CO2 power cycles that are currently under
development. These power cycles are of great interest not only for CSP, but also
nuclear and fossil energies. And they offer opportunity for higher
efficiencies, as well as higher efficiencies at smaller size, so you
don’t have to have as large of a plant to reach efficiencies and lower costs.
Also, we need to see the remaining parts of the system, the thermal energy storage,
the receiver, and the operations and maintenance come down. So the next
section I’d like to talk about is, I’m gonna call here the solar storage
synergy, and if you remember this plot I showed earlier on grid flexibility,
one of the biggest levers, probably the biggest lever for increasing the
flexibility of the grid, which in turn enables greater solar deployment, is
storage. However today, storage costs are really too high to see large-scale
deployment of storage, but one form of storage, battery costs are falling
dramatically. And so we’ve looked at what would happen if these large cost
reductions in battery storage or other forms of storage, but batteries are the example here, continued. So here’s the projections of solar deployment, the
percent of U.S. electricity from solar in the low-cost solar case of hitting the
SunShot 2030 goals where we get to 3 cents a kilowatt hour for PV in 2030 and
2 cents a kilowatt hour PV in 2050. If we throw a low-cost storage on top, where
low-cost storage in this case is getting to a system install cost of $100 a kilowatt
hour in 2040, that’s about a factor of 4 reduction in storage costs by 2040, what
you see of course is that there’s dramatically greater solar deployment. This is, you know, somewhat of an
arbitrary choice of scenario, but just illustrating what a large lever for
solar deployment at the same cost of solar, you see substantially more
deployment with low cost storage. This is just illustrating what the assumed
battery storage capital costs are for this scenario. So the bold lines are
from Wesley Cole at NREL’s study, looking at overall projections for energy storage
cost, so those are mid-case projections are the solid lines, and the dashed lines
here is the low-cost scenario, where for utility scale storage it’s hitting $100
a kilowatt hour and 2040. That’s for an 8 hour battery,
whereas commercial and residential assume 3 hour storage. So as I just
mentioned, more storage leads to more solar. Storage does this by providing a
sink for curtailed solar, so rather than just throwing away excess solar, you can
charge the batteries and then use it at times when power is needed. But I say
this is the solar storage synergy, because it actually goes in the other direction
as well. As you put more solar on the grid, there’s more market opportunity for
storage and I’ll explain why on the next slides. And in addition to that, actually
coupling solar plus storage systems together, offers some opportunity for
cost reduction in some of the soft costs, and possibly even in some of the
hardware like the inverter. So instead of deploying them independently when
they’re actually deployed together, you may have cost reduction opportunities.
But in terms of understanding why increasing solar increases the market
opportunity for storage, I’ll come back to this plot here where we see…here’s the
net load and how the net load changes when you add more solar. Now just going to the case of no solar and the case of solar with about 10% of PV, and
what you see here is it’s looking effectively at capacity value. So to have
about a two gigawatt reduction in peak capacity, the capacity value for storage,
in the case without solar, requires in this particular illustration, requires
about a five to six hour storage life time. With solar, this load peak decreases,
load peak narrows, so now it’s two and a half to three hours storage that are
needed, which for the same overall peak reduction. So for the same capacity value,
you don’t need a battery that can last as long, which is an easier requirement
on storage costs. And Paul Denholm from NREL has looked at this across
electricity markets and found this trend holds broadly even across markets for
which load shape can vary significantly. So in all these cases, these bars here
represent going from zero solar penetration up to 20 percent, and
after the initial, in some cases there’s an initial decrease, basically in all
cases you see that as solar goes up, the market for storage, in this case, it’s four hour storage, the market for storage… market size for storage goes up. And as part of
getting to a conclusion, I want to talk or just sort of end with a discussion of
really what I see the tremendous potential for solar. I’ve shown some of this
already, but just to talk a little bit more about what could be possible
with aggressive innovation in solar and synergistic technologies like storage,
but that allow for a greater grid flexibility. I’ll be talking a little bit
more about this deployment projection modeling that we’ve been looking at and
just before I do, these are models, modeling tools at the National Renewable
Energy Lab, at NREL, it’s the regional energy deployment system is the primary
tool here. And this is an optimization model that looks at what is the lowest
cost way to basically meet electricity needs, to balance supply and demand,
maintain power quality across the U.S. and it explicitly deals with renewable
energy integration issues and variability by having a large number of
time slices over which it does its balancing, and looks at transmission
build-outs as well across 134 balancing areas. What it does not do well is
project deployment on the distribution system. And so it’s coupled with a model
called D Gen that looks at adoption of distributed solar. And these are the same
projections I showed earlier. Based on these models, baseline…again baseline
sort of expectations here, low cost solar, low cost solar, and this particular case
of low cost storage, showing again by 2050 in these scenarios, low class solar
alone could enable 30% of electricity demand by 2050. With a more flexible grid,
this number only goes up and in the case of low cost storage, could be over 50%.
Now of course, the one consistent thing about projections is they’re always
wrong. They account for many many many many
many factors in the baseline cases so as I said, only guaranteed to be wrong. You
know, it’s looking at expected costs for the whole suite of energy generation
technologies, expected electricity demand, and how that changes over time—
a large number of factors, all of which are subject to change. So in this case, we
did a sensitivity analysis that looked at a wide number of potential changes
and came up with sort of a wide range of possible outcomes, all for the same low
cost solar, low cost storage case. The general trends, however, remain clear that
decreasing the cost of solar leads to substantially more solar deployments,
making the grid more flexible. Decreasing the cost of storage leads to more solar
deployment for the same solar costs. And this is just a plot looking at the
build-out of PV for these sets of projections and so here we see
projections to date, you know, where we’re in the less than ten…tens of
gigawatts. You see, you know, a large expected growth in solar deployments
over the years around 2020. This is due to cost continuing to fall while the
investment tax credit makes solar effectively even lower cost. And then you
see some drop-off and then another build out here expected when, as solar cost
continue to decline and reach a level at which solar becomes competitive even
with generation from some existing sources. And here’s where you see, at this
point, the value of solar is falling in the case of an inflexible grid and
you’ll see these orange bars, that solar deployment would fall, but in the case of
a more flexible grid, low cost storage, that solar deployment can stay…could
stay at this higher rates. Again, projections with all the usual caveats
here. But one thing I do want to show from these projections is that solar
build out is not just occurring in a few places in the country. When we get to low
cost solar, low cost solar plus storage, you see deployment across the country
and this is colored where darker colors mean
that more of that state’s electricity demand is met by solar, lighter is less.
In this case, goes from about three percent to 60 percent depending on the
state and here these numbers are even… even higher. So solar, you know, you
really do see solar deployment across the country. And I’ll end just talking a
little bit more about the solar office at DOE and some opportunities that might
be of interest to some of you. So just introducing what we do. As Raf said, we
were commonly known as SunShot until recently and, you know, with the
accomplishment of the utility scale solar goal and with cost reduction no
longer being the only important factor for solar, we’re using more…our solar
office name. And so what we do is we support early-stage research and
developments of solar technologies that strengthen grid reliability, resilience,
and security. We do this through primarily through funding opportunities.
We have what we call funding opportunity announcements or FOAs that are
competitive, that are open to applications, certainly from MIT and
generally to the general community. They’re open for applications. They fund
different critical research areas which we identify, but broadly, you know, aiming
at continuing to lower electricity costs, integrate solar energy into the grid, and
enhance the use and storage of solar electricity. We have what’s being called
science technology policy opportunities. We have what’s been for the last several
years a great fellowship program in our office that I wanted to talk about and
would love to talk to any of you who might be interested and please do spread
the word. This is a…it’s a fellowship opportunity to come and work in our
office. It’s both for recent graduates as well
as folks who have been in the field for awhile. Recent graduates at the
undergraduates or PhD level, as well as folks who have more experience. We also
have a senior fellow equivalent program. And you come, you work directly in our
office and have the opportunity to really be exposed to this broad, big
picture, to think about what are the big challenges in solar energy research, and
then to work with our team to define and design new funding strategies as well as
to help manage our funding program. So it’s a great opportunity to get involved
in, you know, looking at and shaping where millions of dollars of research will go
and new research directions. We’ve had really fantastic folks come through this
program. It’s a…it’s a two-year program, and, you know, people have done all
kinds of things afterwards. A lot of them have stayed in our office and have
grown into all kinds of different roles and people have done a variety of other
things as well. But I, you know, I’d be delighted to talk with any of you who
are interested or feel free to send me an email. Our next application deadline
is January 15th and the program is run through Orise. There’s a…there’s a
website here and again you can feel free to follow up with me if you don’t get
that down and you’re interested in more info. We do have one funding opportunity
open the moment and this is actually a new area of funding for us—solar
desalination. This funding opportunity, it’s a 15 million dollar funding
opportunity and it’s looking at utilizing concentrating solar technology
to desalinate water, so leveraging solar thermal technology either to directly
drive desalination processes or coupling thermal desalination processes with
solar thermal generation to utilize some of the the waste heat of the soil
thermal generation and more cost-effectively generate clean water. So
this is an exciting new area for us and concept papers which are short, I believe
five page descriptions of a proposal and an idea, are due in about a month if you’re
interested. The way our funding cycles work is we have a concept paper cycle
and then concept papers are either encouraged or discouraged to submit full
applications. So you’ll get some initial feedback on that idea and then full
applications can be submitted whether you’re encouraged or discouraged,
you can submit a full…a full application based on that feedback, but we provide
that feedback to help folks know whether or not it may be worth their time to
develop…a develop a full proposal. So I will end here. I just want to acknowledge
we have a fantastic team at DOE that’s contributed to all the work I’ve talked
about, as well as some really excellent analysts at NREL whose work has been
what I’ve been highlighting here. Wesley Cole, Paul Denholm, Dave Feldman, Robert
Margolis and Mike Woodhouse. Thank you very much!

Leave a Reply

Your email address will not be published. Required fields are marked *