While we've seen prices for solar power installations globally fall well below one U.S. dollar per watt, it was only relatively recently (later in 2016, early 2017) that those prices were seen in the United States. And with the effects of tariffs on imported solar panels under Section 201, there are some projects that might have been under that number that won't come to be – yet.
GTM Research did the math as part of a presentation by Research Manager Scott Moskowitz at GTM's Solar Summit, Trends in Solar Technology and System Prices, which projects that utility scale fixed-tilt systems could reach 70 U.S. cents per watt by 2022.
It should be noted that the Section 201 tariff is scheduled to end in early 2022, and that to qualify for the 22% Solar Investment Tax Credit (the last year before it goes to 10% indefinitely) you must break ground in 2021 and complete the project before the end of 2023.
GTM Research projects savings from a range of areas. The largest savings come from the solar panels – expected to fall from from 37¢ per watt (W) today to 24¢/W by 2022, a 35% drop. Much of this price decrease could come from efficiency increases alone. pv magazine staff projects that by 2022, the average utility scale panel efficiency could be 20% – a 17-25% increase over the average 16-17% efficiencies used today.
GTM Research does caution that it will be harder to wring costs out of other parts of the system. For instance they showed inverter pricing slow from the historical 10-20% a year price declines to 5-10% price declines starting in 2019.
However they do sneak in a very optimistic slide – a bifacial solar module, single-axis tracker system whose performance gains versus regular modules mounted on single-axis trackers was 13%. This lines up with pv magazine seeing 12.5% gains earlier this year.
The effect of increased system economics from higher efficiency – not bifacial – panels has already been well documented, and single-axis solar systems are already reaching above 30% capacity factors in the US southwest.
The financial edge from the increased efficiency of bifacial solar panels on single-axis trackers creates compound benefits throughout the installation. If we put that bifacial panel on a single-axis tracker that is optimized to the land beneath it and the geographic location it is in – we get that aforementioned 12.5% bump to an almost 34% capacity factor. That's 20% more electricity from the same piece of soil over a standard solar panel mounted on fixed-tilt racking.
All of these factors paint one clear picture. The price of utility scale electricity is coming down. Research suggests broad impacts across the power system as a whole, with solar and wind expected to bring wholesale prices down 50% by the end of the 2020s.
Getting to cents per kilowatt-hour
Making use of the National Renewable Energy Labs (NREL) Levelized Cost of Energy Calculator (LCOE) we start to get some insight into what the world's largest solar developers are going to be producing electricity at.
We set the period at 25 years, as that aligns well with the current solar panel warranty. And by 2022, it might be set further as developers monetize future repowering of solar farms and longer lived solar modules.
For this analysis, pv magazine chose to increase the system cost above to 75¢/W to account for single-axis tracking. Our opinion is that this price is actually giving an extra penny or two, considering efficiency gains.
For capacity factor – we started with the 30.2% that we're getting in California single-axis trackers last year and the year before, and we added 12.5% for the bifacial panel gain. That brought us to a capacity factor of 34%.
Next, we brought the capacity factor to 38%, an increase of about 11.8%. We did this because 20% bifacial solar panels mean an increase in panel efficiency of 17-25% from today's product, and 38% seemed conservative.**
Next we adjusted O&M costs to $7.50/kW to align with increases expected here as well. Currently, there are contracts sneaking out at $8-10/W – some influenced by the tax credit, some by super dense installation areas.
Solar power plants have no variable O&M costs (PDF).
That leaves us with a simple, levelized cost of renewable energy at 1.5¢/kWh. This price does include profits for the utility scale developers.
And, if the solar power developer were to partner with a strategic tax equity investor who discounted the tax credit and depreciation by 25% – lowering the effective capital cost to 52¢/W to install, we get a price of 1.1¢/kWh. The cheapest electricity on the planet, right here in the United States.
These numbers, while fantastic looking – are not that far from bids we're already seeing float around in some global bids. Saudi Arabia saw 1.78¢/kWh that was declined possibly because they were using bifacial solar panels. Mexico, when accounting for clean energy certificates, saw a 1.97¢/kWh bid. Chile got a 2.15¢/kWh bid.
A reckoning is coming.
**It has been noted that increasing a solar panel efficiency, while it will lower broader system related costs – won't in fact increase the power plant capacity factor, as the power plant's rated capacity will scale with the efficiency gains. There'd need be kWh production gains from the same system size, to gain those benefits – say for instance from bifacial panels (if gains not noted in DC wattage) or single axis trackers.
Altering the LCOE calculation would mean a 1.7¢/kWh value at $750/kW and 1.3¢/kWh at $520/kW.
This content is protected by copyright and may not be reused. If you want to cooperate with us and would like to reuse some of our content, please contact: editors@pv-magazine.com.
Prices this low should change our thinking about how we design our grids in the future. If solar and wind both drop to 2 cents we’ll end up with a mix of new and paid off wind and solar with an average cost well under 2 cents.
At that point we can massively overbuild wind and solar and avoid more expensive storage.
I used the 2017 hourly load, solar production, and wind production for CAISO and ran a ‘how high could penetration go’ model. If California had had 20x as much solar capacity as they did in 2017 and 20x as much wind capacity combined wind and solar penetration could have reached 89% with no storage at a reasonable cost.
I used 2 cents. At 20x solar, 20x wind the cost of generated electricity would have been 8 cents per kWh. If all cars and light trucks were battery powered and charged from ‘surplus’ generation the cost would drop to 5 cents. There would be a lot of surplus to sell to other dispatchable loads, further lowering the cost.
On only four days in the year did wind and solar not produce enough in 24 hours to satisfy demand for that time block.
Do you have data on that 4 day research?
John – how was I to know that you asked me a question? I happened to come back to this article with a different question.
Yes, I’ve got a pile of data. Send me a message with an email address and I’ll fill you in. Link you to my spreadsheets, whatever you want.
In the meantime I’m wondering if you did your analysis correctly. I’ll put that in another comment.
Agreed; a new study by Berkeley finds that using EVs as mobile power storage could eliminate the need to build costly stationary grid storage for energy from renewable sources.
“Our results show that with its EV deployment target and with only one-way charging control of EVs, California can achieve much of the same benefit of its Storage Mandate for mitigating renewable intermittency, but at a small fraction of the cost. Moreover, EVs provide many times these benefits if two-way charging control becomes widely available. Thus, EVs support the state’s renewable integration targets while avoiding much of the tremendous capital investment of stationary storage that can instead be applied towards further deployment of clean vehicles.”
http://iopscience.iop.org/article/10.1088/1748-9326/aabe97/pdf
Yes, My modeling shows that First Solar’s new series 6 utility scale systems will cost less than 80 cents per watt by 2020. We could see PPA’s at 1 cent per kwh in the sunny locations of the southwest. Here is my article:
https://seekingalpha.com/article/4138338-90-percent-renewable-grid-2-utilities-buy
ESP
Let me address the bifacial panel cost reduction a different way.
First, a LCOE for regular panels based on GTM’s 75 cents estimate and best case single axis tracking in sunny regions.
I used a 20 year period and 5% financing before I noticed what you had used. Is 25 and 3% typical?
Financing 20 years at 5%
Installed cost per kW = $750
Capacity factor w/single axis tracking sunny area = 30%
Opex = $7.50
Cost of electricity $0.026/kWh
Found this –
When bifacial modules are installed on a highly reflective surface (like a white TPO roof or on the ground with light-colored stones), some bifacial module manufacturers claim up to a 30% increase in production just from the extra power generated from the rear.
https://www.solarpowerworldonlineDOTcom/2018/04/what-are-bifacial-solar-modules/
So I’m using a higher efficiency gain (30%) than what you used (20%).
Keep the costs the same but use 30% more efficient bifacial panels and the cost per kWh drops from $0.026 to $0.02.
Am I looking at this wrong?
Using numbers matching yours –
25 years and 3% drops the LCOE to 1.9 cents.
A 20% increase in efficiency with no cost increase would drop the cost to 1.6 cents per kWh.
Also, just a suggestion.
Give a cost range based on the sunny SW and the not-so-sunny NE. Ignoring the most extreme (foggy Seattle coast and Death Valley) there’s a range of 4.2 to 5.5 average solar hours per day.
1.6 cents in the sunny SW outside of Death Valley and about 25% more for the NE. 2 cents. Something like that.
“Next, we brought the capacity factor to 38%, an increase of about 11.8%. We did this because 20% bifacial solar panels mean an increase in panel efficiency of 17-25% from today’s product, and 38% seemed conservative.”
Seems to be saying that, beyond the bi-facility gain, increased conversion efficiency results in Increased capacity factor?
This would be wrong. The capacity factor is the ratio between the average output and the maximum output. An increase in a panel’s conversion efficiency will increase both by the same proportion, so the ratio between them will not change.
Very provoking article. I am a solar developer also, and the prices we are asked to quote is not an “at the fence” price but a price that includes interconnection, transmission, balancing services and the risk of curtailment. Also, three percent interest is not market (for blended loan and tax equity use 7-8 percent) and the loan term will be only 12-14 years. The resulting number is still going to be exciting and more relevant to electricity buyers.
John, when you say “That’s 20% more electricity from the same piece of soil over a standard solar panel mounted on fixed-tilt racking.” does that statement take the increased land required for a tracker versus a fixed tilt system? If not space constrained at a typical latitude, a single axis tracker has a ground coverage ratio (GCR) of about 0.33, while a fixed tilt system may have a GCR of 0.50 or more. Therefore, a single axis tracker will take 150% of the land per DC watt. I think if you factor that in you may find that the best production per unit of land may be with a fixed tilt system.
Am I correct in understanding LCOE does no include the cost of storage, to manage the “Duck Curve”? If we add Tesla Powerwall 2.0 or similar Li technology, so we have solar power for the evening peak, how much does this add to the cost? I’ve done some numbers but they are scary.
The assumption that price declines will slow, which is usual in the renewables industry, doesn’t bear out under the weight, of the evidence, of precedent, the renewables industry is larger, why would an industry, with greater economies of scale, be able to achieve less than it did, when it was a small industry. Wouldn’t funding for research, be dramatically higher, as the market realises, that solar plus storage, is cheaper than current mass energy production techniques, because they want to make money, profit, isn’t that how capitalism works. Isn’t that what happened in the roaring twenties and with the development of the efficient steam engine, which not surprisingly rapidly caused steam shipping, transcontinental railways, the Suez and Panama Canals.
China made 45 GW of solar power last year, to the United States 10 GW, a million Chinese workers in solar, unconventional hydrocarbons, have reduced the United States unemployment, but may be causing you to miss the Clean Disruption boat. With tremendous strategic implications, witness Belt Road, high speed railways pushing out of China, 5G mobile, if China is ahead in transportation, communications and energy cost and amounts, you may be in big trouble.
China’s first industrial revolution, during the early 1950-80 Cold War, second industrial revolution, 1980-2020, have been much faster than the United States, if they continue this pace, they’ll have a third industrial revolution, roaring twenties and outpace the United States economically and therefore militarily. I say this as a concerned ally, Australian, with as much rooftop solar power as the United States, but only a 12th of the population. Your not grasping the simple geometric, exponential, compound growth involved in cheaper energy, transportation, food, (from high rise agriculture.)
And with the pace of research continuing, efficiencies jumping further – game changing aggressively – I think you’re right.