Cost breakdown for Solar PV with lithium-ion storage | Hacker News Reader

Cost breakdown for Solar PV with lithium-ion storage

pv-magazine-usa.com

38 pointsepistasis22 comments

> PV system with one-axis tracking

In the South West US, this reaches 30% capacity BEFORE any storage.

Add 4hr storage and we reach 50% capacity. This is true for about 50% (pure guess) of humanity that lives in sunny climates.

Even better, poorer/populous countries are often closer to the equator with even better solar conditions.

We are literally half-way to solving our climate problems.

Now, about those Windows updates...

Climate problems are not about price, they're about CO2 émissions. Turns out producing and installing batteries produces a lot of CO2, and they're not even enough on a large scale so they have to be complemented with more pilotable electricity, which unfortunately mostly is carbon-based. As long as we depend on those intermittent sources (that's the real name we should give them not to be confused with other renewable), we rely on carbon.

>Turns out producing and installing batteries produces a lot of CO2,

That is counter to anything I've ever read on the subject. Can you source your claim?

Batteries are mostly produced in countries where a huge portion of electricity comes from CO2 (http://www.worldstopexports.com/lithium-ion-batteries-export...). On top of that, each battery needs to be transported to its final destination, which is yet another source of pollution.

I don't think this is a good argument. What you are describing is the chicken and egg problem. Once one part is solved, the other one is too.

4 hours of storage is enough for a PV system to serve the early-evening period when demand is at its highest.

For example, here's the demand curve for California's CAISO today:

http://www.caiso.com/TodaysOutlook/Pages/default.aspx

The peak demand is between 6 and 7 PM. Since the sun sets at 4:55 PM today in Los Angeles, and solar output dips significantly starting a couple of hours before sunset, solar cannot supply any of the peak-demand load at this time of year. Even with later sunsets in the summer, solar can supply only a fraction of peak-period demand. But solar coupled with 4 hours of storage can cover the most important hours just before and just after sunset. By the later evening, demand is lower and wind output has often increased.

The two most interesting cases:

Standalone 100-MW PV system with one-axis tracking ($111 million) -- this is a pretty typical configuration for a new solar farm being built today.

Co-located, AC-coupled PV (100 MW) plus storage (60 MW/240 MWh, 4-hour duration) system ($188 million) -- this is a battery system plus the solar farm at the same facility. AC coupling is a little bit more expensive than DC coupling but it is more flexible; it can be installed as part of the original solar project or as a retrofit. Most utility scale PV projects that are already built have ample room for a battery system on the same site.

A year ago, XCel Energy in Colorado received median bids of $36/MWh ($0.036/kWh) for PV plus 4 hours of storage:

https://www.greentechmedia.com/articles/read/record-low-sola...

However, those bids are for projects to be completed in 2023 and NREL's cost estimates are for projects built now. That can explain part of why NREL's cost premium for adding storage appears to be higher than the premium assigned in recent auctions.

Shifting power east / west is another option. Supplying Florida with solar from California can significantly reduce the need for storage. This also allows you to supply morning power in California with solar in Florida.

The interactions are complex enough I think production, transmission, and storage all need to be dealt with separately.

Long-distance grid connections are technically viable and can reduce the need for storage. The biggest problem in the US is the very slow and uncertain process of getting state-by-state approval for interstate transmission projects. A Florida-California transmission project would need approval from at least 8 states.

The proposed Grain Belt Express transmission project goes through only 3 states and received its first approval (in Kansas) 7 years ago. It is still trying to get the necessary approvals in Missouri. Due to this drawn-out process in Missouri, opponents in Illinois were able to reverse the project's earlier-obtained Certificate of Public Convenience and Necessity. Even if Missouri gets on board the project will have to re-obtain approval in Illinois, and perhaps by the time that's over Kansans will have new objections.

JOOI, who is the opposition, and why are they opposing this?

Appears to be the usual mix of NIMBYs and anti-eminent-domain activists: http://enquirerdemocrat.com/grain-belt-express-project-halte...

There is a long-distance transmission line in California that is used a bit like this, allowing the use of power from the North in the South in summer when peak demand is there, see https://en.wikipedia.org/wiki/Pacific_DC_Intertie

The Pacific DC Intertie is not nearly as long as a connection between CA and FL would be. In that scenario, transmission losses would be quite significant. The cost of building and maintaining the line would also be quite high.

Although transmission does help alleviate the problem, the nature of building new linear infrastructure is that it will take much more time and risk to complete a transmission megaproject than to simply build more storage where it's needed.

Meanwhile NJ has 50% of all electricity carbon free.

As of 2017 California is at 50.45% carbon-free electricity:

https://www.energy.ca.gov/almanac/electricity_data/total_sys...

That's summing up nuclear, large hydro, and renewables excluding biomass from the California Power Mix column.

> or the standalone systems, a constant per-energy-unit battery price of $209/kilowatt-hour (KWh) is assumed, with the system costs vary from $380/kWh (4-hour duration system) to $895/kWh (0.5-hour duration system).

Still seems very low to me. I'm looking at my own 2x powerwall estimate (24 kWh), and it's $450/kWh after all rebates and credits. Maybe they're averaging some state rebate in this methodology.

This is looking at utility scale storage, not small residential systems. Single home installs will be far more expensive per unit of capacity than industrial scale systems.

> For the standalone systems, a constant per-energy-unit battery price of $209/kilowatt-hour (KWh) is assumed, with the system costs vary from $380/kWh (4-hour duration system) to $895/kWh (0.5-hour duration system)

“The 19.4 cents per kWh New York households paid for electricity in November 2018 was 45 percent more than the national average of 13.4 cents per kWh” [1]. The four-hour duration system thus costs over 280,000% more than the baseline. How do those numbers make sense?

[1] https://www.bls.gov/regions/new-york-new-jersey/news-release...

I think you're confusing operating costs and installation costs. The $380/kWh is referring to the cost for a battery that can store 1kWh of energy. That battery will store and release energy many times.

> * you're confusing operating costs and installation costs*

Yup–thank you!

LCoE for utility-scale solar-plus-storage is very roughly $30/MWh:

" Xcel Energy in Colorado saw bids for solar-plus storage at $36 per megawatt-hour, compared to $25 per megawatt-hour for standalone solar, and NV Energy has seen even lower bids in its solar and solar-plus-storage RFPs.

That equates to about a $6 to $7 per megawatt-hour premium for solar projects that are partially dispatchable — if not completely dispatchable in the manner of a power plant. By 2023, when the ITC window closes, the levelized cost of energy (LCOE) of combined solar-storage projects will put them in direct competition with traditional generators, Manghani said."

https://www.greentechmedia.com/articles/read/five-prediction...

What does the `hours of duration` mean in this context?

It's the energy capacity of the battery divided by the power of the battery; i.e. if you discharge the battery at maximum power, how long does the battery last?

Reply on news.ycombinator.com