[NullPrefix]

> the extreme afternoon demand curves when solar capacity drops and demand jumps

Does the hourly schedule variate day by day or can it be predicted?

If you can predict when and how much power you will need to produce then it can be ramped up slowly, I assume.

[hannasanarion]

"ramping up slowly" isn't an option when the demand curve is changing quickly, because there's nowhere for that extra energy to go. Energy storage isn't really feasible, so all the electrical grids in the world need to match supply to demand 1:1 in real time.

There are electrical dispatchers who are monitoring grid supply 24/7 and instruct plants on how much they are responsible for generating on a minute-by-minute basis.

[dredmorbius]

An alternative to storage is dispatchable demand. This would require high-load activities or uses which can be rapidly cycled.

Storage batteries, pumped storage, and CAES are examples of this, though simple raw thermal banking (hot water heating, typically) is an excellent way to suck up excess Joules or GWh.

1 GWh is roughly the energy required to heat a pool of water 1 hectare * 1 m by 86 degrees Celsius. This scales to multiple GWh by increasing area, depth, or both. Conversion to steam is also possible, though that requires more engineering (pressure is A Thing). Substrates such as molten salt have a lower heat capacity per unit mass, but can be heated to far greater temperatures.

Storage at the scale of entire US generating capacity for multiple weeks using molten salt thermal storage, even accounting for Carnot cycle efficiency losses (about 20-50% depending on specifics, 30% is a good ballpark) is actually a tractable-scale concept. Existing petroleum storage facilities are roughly comperable in size, though molten salt would require somewhat more robust facilities and insulation.

Whilst it doesn't have the net efficiencies of pumped hydro (exceeding 90% round-trip storage efficiency), pumped hydro lacks sufficient developable sites, and has significant environmental impacts.

[nerpderp82]

> Energy storage isn't really feasible

You couldn't be more Dam wrong.

[dredmorbius]

Doesn't scale. Not enough sites.

Mind, where it does work, it's phenomenally effective, efficient, and responsive.

There are a few sites at which seawater-based systems might be possible, in which the ocean forms the "lower reservoir". These are dependent on suitable terrain. Matching terrain to consumption patterns is difficult: the Netherlands and much of Britain are sorely lacking. Some of the best potential sites are along the Balkan coast in Serbia and Croatia. Chile's Atacama Desert, along the Pacific coastline, is nearly ideal geographically, but is far from most use (North America, Europe, Asia). Portions of the US West Coast might be suitable, though would all but certainly face major political resistance for environmental impacts.

And: working with seawater is complex from an engineering standpoint: it's corrosive and sea life has a pronounced tendency to foul large-scale water-handling systems, though this may be tractable. There've been several pilot projects, though those have since been decomissioned, excepting Rance in France, designed as a tidal power plant, though capable of working as a pumped-hydro facility.

[keanzu]

> Doesn't scale. Not enough sites.

Terrible energy density, massive land use.

To store the energy contained in 1 gallon of gasoline requires over 55,000 gallons to be pumped up 726 feet (CCST 2012). [1]

Itaipu Dam has 14,000 MW installed power, 1,350 square kilometres (520 sq mi) were flooded. [2]

[1] http://energyskeptic.com/2015/hydropower-has-a-very-low-ener...

[2] https://en.wikipedia.org/wiki/Itaipu_Dam

[dredmorbius]

Itaipu (or comparable impoundages such as Three Gorges) provide both capacity, indicated as installed GWe of generating capability, a measure of power, and storage, indicated as GWH of energy generation. These are equivalent to the capabilities of an automobile's engine (power) and fuel tank (storage).

Hydroelectric projects also virtually always provide additional services such as flood control, irrigation water, recreation, and waterways management (ensuring water flows, etc.), which also account for design elements including scale, etc.

Pumped hydro should provide a sufficient level of both capacity and storage, though in general the storage requirements are far smaller than traditional hydroelectric dams aimed at generation. There are quite small pumped hydro facilities, the compensation being that water levels can rise or fall considerably, often many metres in a single day's storage/generation cycle.

Gravity is in general a weak force, but water is relatively massive, largely non-toxic, and can be utilised through large-scale pump-generator units (~800 MW per generator IIRC), which is a scale few other options can match. Again, on balance, pumped hydro is a good solution, there's just not enough of it to go around.

[nerpderp82]

Pumped hydro is 70-80% efficient at storing energy. [1]

Converting to sensible units gives

mass = 55e3 * 3.79kg/gallon ~ 208450kg

height = 221 meters

mass * height * 9.8m/s^2 ~ 452043210 joules (watt-seconds)

or 125KwH, 70% of that is 87KwH. What does gasoline have to do with energy storage on the grid? Furthermore, a gallon of gas has 33KwH [2].

What does the Itaipu Dam's installed power have to do with stored energy?

[1] https://en.wikipedia.org/wiki/Pumped-storage_hydroelectricit...

[2] https://en.wikipedia.org/wiki/Gasoline_gallon_equivalent

[keanzu]

> What does the Itaipu Dam's installed power have to do with stored energy?

Nothing. Was there something to suggest otherwise? I listed two weak points for hydro: Terrible energy density, massive land use.

My first reference supported the first point and my second reference addressed the second point. Land use: 1,350 square kilometres were flooded. The installed power was to give context of how big the plant is.

[nerpderp82]

Does it have to scale? What does scaling mean? Could the existing 2400 dams in the United States be used for PSH? [1]

How many sites do we need? Why would the Netherlands need pumped hydro, isn't the European grid interconnected?

See my comment below.

https://www.energy.gov/sites/prod/files/2018/04/f51/Hydropow...

[dredmorbius]

I'm going to suggest you think through what I've written, the general problem of power grids, generation, and storage, energy uses, and consequences of power outages. The answer may reveal itself to you.

Likewise, you seem to have failed to grasp what I'd hoped would be a rather evident reason for mentioning the Netherlands in the context of pumped hydro and terrain.

[dredmorbius]

The demand can be generally predicted for at least several days in advance based on weather. This affects both generating capacity (incident solar, wind) and loads (hours of daylight, heating or cooling load). The process isn't perefect, but converges on experience the closer to present you are. Factors such as predictable human activities (workday, workweek, and seasonal factors) also enter in.

The events you may have noted in news of "negative energy prices" are often failures of prediction -- unexpectedly high availability (more sun or wind), and unexpectedly low demand. Though "pay to take my power" sounds good, it's actually a sign of mismanaged resources.

There are occasional incidental factors -- sudden demand, or more often, equipment or transmission failures which require bringing additional capacity online, or shedding load to prevent under-voltage (and hence: over-amperage), or underfrequency. Grid power frequency is generally 60Hz in the US, 50Hz in the UK, and just for grins, both in Japan, on separate and noninterdependent grids, which made generation capacity loss following the Tohoku earthquake/tsunami and Fukushima incident all the more critical. Loss of synchronisation or deviation by more than a very small fraction from the nominal frequency is considered a Very Bad Thing. Viz the recent UK blackouts.

[tofflos]

The Nordic system operators are in the early stages of transitioning from a reactive to a predictive process. The schedule does vary from day to day but there isn't anyone there that doesn't believe it can't be predicted with an accuracy that is more safe, and economical, than the current reactive process.

You can read more about it at http://nordicbalancingmodel.net.