Hacker News Reader

Comment by paulsutter | Hacker News Reader

They should pump water uphill during these surges. California hydropower is constrained by rainfall, so the dams and generators already exist. Pumps cost less than batteries and this process has about an 80% roundtrip efficiency

View on HN

They do. Many of the reservoirs of the California Water Project also do power storage. They usually pump up at night and generate in the daytime. Sometimes they can't store more energy because they're full of water, as happened at Oroville Dam this year. Sometimes they don't have enough water. A year ago, some hydro plants in California were down due to lack of water.

There's one big pure pumped storage plant in California - Helms, near Fresno.[1] The geography has to be right for pumped storage. You need a high reservoir and a low reservoir near to each other, and since the water levels will change drastically each day, they can't be used for boating or recreation much. Helms has about 1500 meters of elevation difference between the reservoirs. Sites like that are hard to find.

Excess solar power isn't a big problem in areas that use air conditioning. Peak A/C load and peak solar output tend to coincide. The big excess power headache is wind, where output varies 4:1 over a day over wide areas. Excess wind generation late at night is common.

The best way to understand the power business is to read PJM 101, the introductory training material for the US's largest power grid. Unfortunately, PJM recently put that behind a login.

[1] https://en.wikipedia.org/wiki/Helms_Pumped_Storage_Plant

Peak power usage is around 7-8PM, this discrepancy is the source of the duck curve. More storage could allow tons more solar to be installed.

https://cleantechnica.com/2014/07/21/utilities-cry-fowl-over...

(deepsun - the "no sun" critical event happens every night)

7-8pm according to that link is the net peak, after solar production has canceled out the original afternoon peak. The problem is that even though solar supply does roughly coincide with the peak of demand, solar production has increased enough that, since about 2013, it's gone past the point of offsetting peak afternoon A/C usage, to now producing an actual net trough in the afternoon hours (you can see the afternoon extremum flips from a local maximum to a local minimum from 2012 to 2014, and just keeps going further down every year since, as peak afternoon solar grows faster than peak afternoon A/C).

The new net peak around 7-8pm is because demand has a long tail into the early evening, while solar drops off sharply around 5-7pm. But yeah the end result is that it does cause a problem for grid management if you add more incremental solar, since the original afternoon peak it helped smooth out has long since been smoothed and then some.

In California, the peak is actually at 6pm. At that time, Solar is generally producing 6.6 GW from around 9.6 GW at noon.

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

Wind is really variable, and often way out of sync with load. I just looked at PJM's dashboard, and peak wind in the last 24 hours was 4 gigawatts at 4 AM. Wind bottomed out at 0.65 GW at 8 PM. That's total output from wind farms from Illinois to New England. Averaging over a large area isn't that helpful.

Piping energy in from the wind corridor that runs up from the Texas panhandle to Canada has promise. The wind is steadier there. It's quite possible to run megavolt DC power lines now; China has 16,000Km of them. California has had a half-megavolt DC link from Oregon to Southern California since the 1970s.

To be clear, 7-8PM is the daily peak. The annual peak is around 2-5PM. The annual peak is far more relevant for grid capacity planning. So there's still plenty of room for solar in California.

https://www.caiso.com/Documents/CaliforniaISOPeakLoadHistory...

(I used to work for a CA utility.)

Calculations show that it's less expensive to maintain more solar panels than maintaining storage. (And, instead of solar panels -- wind turbines, as they are mostly better than solar panels).

Storage is better to be left for critical events (no sun + no wind + peak consumption). And even then it's better to keep several gas power plants to cover those events.

I'd love to see those studies, that's unfamiliar to me! With the quickly dropping costs of storage, I wonder where/if there's and inflection point.

Its hard to imagine a more efficient solar energy storage medium than oil and gas. I'd be curious how close it is myself.

I'm not sure how efficient hydrocarbons are as storage media; the point is that eras of geological processes have already done the work of storing energy in that form.

I was more interested in the claim of having more solar negating the need for storage.

Oil and gas are horrifyingly inefficient storage mechanisms. Round trip will be less than 50%, best case. That said, power to gas may be worthwhile if there's a ton of excess electrical power certain times of the year and the capital costs of power to gas is low enough.

I'm thinking more in terms of availability, cost, volume and portability. I might be wrong but wouldn't a oil tanker full of diesel fuel contain a great deal more potential energy than the same volume of the best battery cell we can produce for this purpose?

See this comment: https://news.ycombinator.com/item?id=14668163

Long term wise, there is more energy that we can collect store and extract from CSP technologies compared to oil (and lng/coal/etc on the grid/plant level), without all the exposure to market risks being exposed to oil (ie. volatility [0]).

http://www.macrotrends.net/1369/crude-oil-price-history-char...

Storage doesn't much care about volume and portability; batteries are good enough. Availability is important for pumped storage, but after that everything scores pretty similarly. So you're left with cost dominating the equation, along with the lesser factor of round-trip efficiency. And it costs a lot of money to store excess energy as hydrocarbons.

Easy.

If you have a ton of panels you have a ton of excess electricity during peak production hours and whatever you skim off the top is basically free. You're better off inefficiently converting what you skim off the top into something you can convert back into energy efficiently or cheaply/easily (pumping water up a hill, running grow lamps on algae that turn into bio-fuel or some other Rube Goldberg operation) than using an expensive, efficient storage mechanism.

Basically, putting excess capacity to use is basically free so you don't care about the round trip efficiency (in dollars, joules in/out, or whatever metric you're using) of your storage medium, just the return portion of the trip.

Heavy metals are a pretty damn good storage medium. The bang for your buck you get when nuclei fall apart is unbeatable.

why pump at night? wouldn't it be better to pump while the solar panels are active?

do they cover the water to stop it from evaporating?

Electricity demand is lower at night, so they can use the excess electricity at that time to pump the water, and release it during the day to provide extra power during peak periods.

so from the fossil fuel produced electricity. now it makes sense.

So they'll eventually flip it over to day.

If they priced residential power in line with real market rates (in this case negative), I suspect this would not be a problem for long. But we need smart meters first.

I would love to be able to have my heat pump run during steep discounts, but in Colorado we currently have a flat 12 hour off-peak (at night -- the opposite of this) with a mild 20% discount.

It's possible to get a smart meter in Illinois, and I've thought about rigging up appliances to run when rates are negative. However the price feed is lagging by 15 minutes or so. So while you can have a pretty good idea rates would be negative, you can't actually be sure.

It's not a huge deal if you are doing something productive with the energy, but if your goal is to make money it's probably not very effective.

That being said, it might be productive to rig up your gaming pc to mine bitcoins or fold proteins when the price is close to zero.

I'm not aware of any utilities that pass on spot pricing on power to their customers on a minute-by-minute level. Even with smart meters, most just tier the pricing into 2-3 levels per day that stay constant for 6 months or so (here in CA, it resets every summer/winter)

Com Ed in Illinois only has two types of metering: standard flat rate, and hour-by-hour.

https://hourlypricing.comed.com/

https://www.comed.com/SmartEnergy/SmartMeterSmartGrid/Pages/...

Most customers don't want this. You have to realize that slightly negative costs in some slots (and your power cost would not actually be negative, as it would be more than balanced by the transmission and service costs) would come with eye-poppingly high rates on other days, like 50x the normal rate.

Imagine the horrible news stories and "NBC on Your Side" pieces that will come about when someone pops their A/C on at the wrong time and gets a $1000 power bill.

Even with smart meters (which we all have already in CA) I don't think residential customers will ever see more than three tiers of time-of-use cost per day, although those hours may shift around. However, utilities already contract with major users like universities and large businesses to shape demand, in return for deep discounts. And, smart home devices like EV chargers and battery storage units will be connected to the grid to provide near-instant capability to suck up excess power, or curtail usage.

Well, solar farms in CA are mostly in deserts, where hydro reservoirs are not that feasible.

Power can be transmitted to areas inside of CA that already have hydro reservoirs or ones at higher elevations can be built.

Since it's already wired up to the same grid, does the storage have to be near the generating plant?

I think the point of paying AZ to take excess power is to lower the load on CA grids.

So long distance transmission cannot solve the immediate problem.

> So long distance transmission cannot solve the immediate problem.

Then how is it getting to AZ?

In most cases the answer is "no, but it helps". Close physical proximity reduces transmission loss (both inward and outward). And hydro storage requires a very specific sort of topography. If you have fortunately built solar energy plants near these sites, great! Otherwise you will probably have to build new transmission infrastructure (likely HVDC), which costs money.

> California hydropower is constrained by rainfall, so the dams and generators already exist.

But there's nothing to pump, at least a good part of the year. The dammed river beds are practically dry during the summer/fall months.

Plus there are environmental issues where you are pumping water out of places where endangered species are living or spawning. Removing water has an environmental impact, which California already feels due to water already being diverted in large amounts for agriculture or drinking water.

>this process has about an 80% roundtrip efficiency

How is that possible? Even if you assume frictionless pipes and lossless power transmission, that's assuming 90% efficiency for both the pump and the generator.

From https://en.wikipedia.org/wiki/Pumped-storage_hydroelectricit...

> The round-trip energy efficiency of PSH varies between 70%–80%,[4][5][6][7] with some sources claiming up to 87%.[8]

Well we need all the reasons we can get for adding dam capacity.

Reply on news.ycombinator.com