Hacker News new | past | comments | ask | show | jobs | submit login
Duck Curve (wikipedia.org)
154 points by B1FF_PSUVM on Oct 27, 2019 | hide | past | favorite | 128 comments

Once smart-grid tech starts to emerge, many possibilities to muffle the quacking of the duck curve present themselves.

The most obvious, and applicable in any climate and weather, is domestic hot water heating. Temporarily suspending power to electric hot water heaters at times of high demand can save tons of short-term load. And, it's potentially responsive enough to support frequency control ancillary services (the very expensive task of keeping the power grid running at exactly 60Hz or 50Hz when power demand is changing fast).

Another is electric vehicle charging time-shifting, either mandatory or voluntary.

A third, more complex, opportunity is HVAC. If heating / airconditioning systems can get ahead of their needs during late afternoons, then temporary power suspension in early evenings is feasible. This kind of application may be easiest to deploy in commercial locations like big-box stores, office buildings, and data centers.

A mindset change required is this: demand time-shifting IS energy storage.

A bigger mindset change: grid operators are in the information business, whether they know it or not. All this proposed timeshifting is based on information.

The saying that's now prevalent at utility conferences is, "In the past, we'd forecast load and deploy generation. In the future, we'll forecast generation and deploy load."


Fuel-driven systems offer dispatchable power.

Resource-driven systems offer dispatchable load.

We're used to thinking of idled loads (factories, etc.) as wasted capital, but idled dispatchable power (peaking plant) is the same thing.

The old approach to this is described by the maxim "make hay while the sun shines". Find a task suitable to your current situation, stockpile for downtime.

(The fact that our present fuels are stockpiles of a previous era, that we're burning through at rates of 5 million times that which they accumulated, also bears heeding.)

IIRC, my Dell laptop has a BIOS option to automatically stop charging at certain times of the day, even if it's plugged into the charger. So that's another opportunity: many people seem to use their laptop plugged into the AC most of the time (and into an extra monitor, and into an external keyboard/mouse, etc). With current laptops lasting several hours while unplugged, running them from the internal battery for an hour in times of high demand is viable.

Let's take optimistic assumptions: 50 watts used by a laptop, ~4 hours a day, plugged in and used every day of the year, one laptop per an American, 300M people. All of that multiplies up to 4 * 365 * 50 * 3e+8 = 2.19e+13 watt hours = 2.19e+10 kilowatt hours affected per year. United States uses about 4.1e+12 kilowatt hours per year, that would have an impact of ~1%. Realistically it would probably be 10-100x less.

Although it does seem easy to deploy at scale, at least in the future models.

There is a tradeoff here since involving the battery will increase the overall power consumption. Cycling the battery also shortens its lifespan, and obtaining a replacement costs more energy. I think some degree of centralized control may be needed for this to pay off. That being said, i wonder if there are any existing programs that look at local electricity demand and adjust things accordingly.

then again you could use solar collectors for hot water and heating, eliminating the need for boiling with electric energy

Each of those has major problems.

Water heating is usually natural gas.

Electric vehicle charging during the day prevents charging at home. So employers would need to come on board paying for charging stations with employee parking. And shopping malls, schools, etc.

Heating is also usually natural gas, and while cooling is almost universally electric, it doesn't take very long minutes to a very few hours) in hot weather to consume any reasonable pre-cooling.

I can see where you are coming from, but you are likely not aware that a lot of this is already happening. Even my small coop gives out a few hundred dollars in rebates for electric water heaters with remote control. They use this for demand control now.

They also offer cheap ecobee thermostats, with an agreement that they can use them for load leveling. Again, rebates make it practical.

I haven't seen anyone do this with electric cars yet, but its a pretty small leap to imagine utilities offering rebates to businesses for the install, if the result is to shift demand in ways that are useful for the utility. Right now that isn't really necessary here, as the currently small demand for EV charging is mostly at night, and that is already off-peak for our utility.

I imagine the electric vehicle charging could be time-shifted to overnight. Between 10pm and 6am

Also, energy for water heaters and home heating vary a bit across the country. Electricity is more common in some parts than in others. Even in places that are overwhelmingly gas or propane fueled, the 10% of systems that are electrically powered represent a small but significant opportunity to time-shift demand.

A/C is definitely the largest culprit here, though. And precooling does work if you lower the temperature of the interior thermal mass of the house.

I used to do it when I was subject to random 4-hour A/C shutoffs as part of my discount plan. If I got the house down to 66˚F before the shutoff occurred, the house would remain comfortable for a couple hours. I may have used an overall larger amount of power to do this, but it was the "right" power: drawn from neighbors' solar rather than generating plants.

You could create a heatsink by freezing a mass of water then thawing it. The problem is space for large amounts of water and the high efficiency cooling and distribution system needed.

The replies miss the point:

Shifting from natural gas heating and water heating cannot help flatten the duck curve, except by overall raising the demand for electricity.

Most car charging is currently overnight, which doesn't match use of solar (the low part of the duck curve). To flatten the duck curve, we need to raise that low part and lower the later part.

And if you drop the temperature of the house way low in advance, yes, you use much more power, you reduce the comfort (at least of some people!) And you still only push your peak demand out a little bit. Again, "use more power" isn't a good option to flatten the curve.

Elimination of natural gas water heaters is necessary.

Workplace charging stations isn't so much an employer thing, as a commercial real estate thing. So it's not so much an inconvenience since borrowing money to pay for upgrades of real estate is what those guys are about.

And of course natural gas heating needs to die as well.

I’m surprised the mitigation discussion didn’t include “more wind generation”. Especially at the coast [1], you have plenty of wind at and around sunset from the large temperature differential. The chart is load minus solar/wind, but the example graph was a low wind day apparently (and IIRC, we are predominantly solar in California).

But even then, this curve doesn’t seem to apply in California in the summer when load demand is highest (look at CAISO demand for say July 24). The amount of built up generation should be looked at in terms of annual peaks and overall efficiency. July has demand up to 45GW range where renewables are only currently able to shave off 10 GW of that (but basically all day long). The forecasted demand for today has a peak of 25 GW.

Is it “just” about how quickly net demand spikes? The grid seems to handle 10+ MW or scheduled maintenance (look at October 1).

[1] https://www.windfinder.com/forecast/pacifica_dumps_san_franc... (ignore the current high winds, go a few days forward).

> I’m surprised the mitigation discussion didn’t include “more wind generation”. Especially at the coast

This is just my conjecture, but I wouldn't be surprised if the very high cost of coastal land, and possible objection by nearby homeowners in CA is part of the reason for lack of wind power on the immediate coast.

If you go just a bit inland, however you'll find the massive new Turtle Island wind farm near the Sacramento River delta.

Hey, it's like this in Washington too! Not a windmill in sight near Seattle, but head East and you'll see them grouped along mountain ridges by the dozen. It does seem very windy out there compared to the coast, though.

It's funny, people seem to dislike the windmills even as they house military firing ranges, hydroelectric dams, nuclear reactors, and similar on the same land. It's less funny that the government coerced people who were resistant to signing treaties or selling land to cede the Columbia River Valley for those things, but the wind turbines - those seem to really bother people.

In fairness, they do sort of remind me of Simon Stålenhag's art[0]. It's a little bit eerie how they loom in the distance and light up in unison at night.

[0]: http://conceptartworld.com/artists/simon-stalenhag/

Coastal land, yes. It's usually unproblematic to build wind farms offshore, and they work even better there.

Offshore wind power started off in areas with shallow seas, it's only now getting around to floating wind that would suit California.

California will have plenty of shallow seas, by and by.

Mostly in places like the Bay counties, Stockton and Sacramento.

CAISO has a great live dashboard that you can play around with on the actual real time demand and supply curves. If you switch to previous days, you'll see the "duck curve" is already here and not just speculation anymore.


The issue nowadays isn't so much the depth of the dip in dispatchable (e.g. traditional) generation, but the slope of the ramp back up after the sun goes down (e.g. the "neck" of the duck). Base load natural gas an other fossil generation have a very difficult time spinning up that quickly (CAISO said it was ~11GW over 3 hours yesterday).

That's why batteries are such a hot topic in renewables right now. Not so much for storing power for "a rainy day", but for helping load shift during the day to prevent such a steep slope in the evenings. Lithium is pretty good at doing that kind of short, daily-burst storage job.

Replying to my own comment with a more meta comment:

I'd also like to point out that the Duck Curve is a good example of the class of problem in the energy transition that seem like a major blocker at first and is often pointed to by detractors to using solar energy, but isn't really.

Since these types of problems take years to manifest, it can be being broken down into smaller prioritized problems (such as addressing the slope problem of the "neck") that we, the cleantech sector, can individually figure out and deploy solutions to without stopping progress decarboninzing the grid. If you just keep solving thousands of these smaller problems over time, eventually you'll end up solving the overall problem.

In a time where it's politically infeasible to do big things (e.g. building hundreds of nuclear generators), the cleantech sector has for the most part been making progress through incremental solutions that over time will get us where we need to go. I have no doubt that we'll solve the Duck Curve problem, but it will likely be one step at a time, not large strategic changes, which to outsiders make it seems like the problem is still a "blocker".

Anyway, this is what fighting climate change looks like, and as an engineer, it's tons of fun to know that if you just keep solving problems, you'll wake up one day in a 100% clean energy world that you helped create.

> Base load natural gas

Sorry, what? Natural gas spins up quickly and is frequently used in “peaking” power plants, while coal and nuclear provide the actual base load.

Not if you look at how it is becoming more common to use in Europe, as renewables are increasingly adopted.

In the UK, with around 20% of capacity possible from nuclear, 20% from wind, gas is the largest source that makes up any shortfall from renewables and nuclear. Coal is generator of last resort kicking in after using the international interconnects. Both of those non-traditional uses are down to simple economics - it's not economic to use coal unless there is literally nothing else available. It's reached a point where the couple of remaining UK coal plants are paid a small subsidy to stay idling, ready for spin up rather than cold start, for days or weeks on end without actually providing power to the grid. Coal will be entirely gone before 2025. Gas is getting to that point, but for now, it's the bulk of UK generation.

Peaking is mostly Dinorwig and the other pumped storage scattered around the UK, along with the secondary ability to disconnect the highest industrial users - around 10% of supply I think - they get a discount on rate for accepting an interruptible supply to help manage the grid.

Yeah, I think the bigger issue with natural gas peaking plants is that it's not an efficient use of capital to build and maintain a plant that only gets used for a few hours a day.

Japan has forty years of experience with energy storage actually:


>one of four battery types selected [...] as part of the "Moonlight Project" in 1980.

They have 2600 MWh of installed grid energy storage capacity as NaS batteries:


>According to NGK, the company has delivered NAS battery systems with a total output of 530 MW and storage capacity

The power generation shortfall on the Duck curve is about 10000 MW. So Na-S capacity has reached 5% of that. But it's a start, and it's cheap, and it's been shown to work. Plus, growth in these kinds of industries can be supralinear.

Couldn't you do other types of storage rather than batteries? Pumped storage can do very fast spin-up because gravity doesn't take any time to warm up, nothing to 1.8GW in a few seconds at Dinorwig (in Wales), if you own half a dozen of those then 11GW in 3 hours feels like child's play.

Not everywhere has the geography for building power stations inside mountains (which is the only way to do pumped storage that isn't horribly ugly), but the US is huge and seems like there are mountains everywhere you look in some parts, surely some of them are suitable for hiding a pumped storage facility?

The Bath County Pumped Storage Station in Virginia stores 24GWh. It was built between 1977 and 1985 at a nominal cost of $1.6 billion.[0] If we conservatively take those as 1985 dollars, that's equivalent to $3.9 billion in 2019 dollars, or $162.5/kWh. That's comparable to today's price for lithium ion battery storage (~$176/kWh[1]), even before taking into account the massive real increase in cost of large infrastructure projects over the last 35 years.

That said, Dominion Energy recently selected a site for a planned second pumped storage facility, conceptually entering operation in 2030.[2] It's a long way off from construction, but they claim that the current budget is $2 billion.[3][4]

[0] https://en.wikipedia.org/wiki/Bath_County_Pumped_Storage_Sta...

[1] https://about.bnef.com/blog/behind-scenes-take-lithium-ion-b...

[2] https://www.roanoke.com/news/local/dominion-narrows-to-tazew...

[3] https://www.dominionenergy.com/company/making-energy/renewab...

[4] https://www.nrdc.org/experts/walton-shepherd/dominions-green...

Hmmmm, yes, probably. However the reasons for deploying millions of lithium batteries instead of a few dozen pumped hydro stations are likely more political and business reasons than technical reasons.

First, it's way easier and faster to just park containers full of batteries wherever you need to shift the energy than to find and build new pumped hydro location then build/upgrade the transmission to the locations where the load shifting is needed. And when you need the load shifting ASAP, you're probably going to opt for the solution that can deployed now, especially when the regulatory reaction to this problem is to adjust prices ("time of use" rates) to discourage people from using power during the slope.

Second, battery installers are able to sell directly to consumers, which means they can bypass utility bureaucracy. Many utilities are very slow to adopt "non-wires alternatives" because their business model doesn't incentivize them to building out a distributed, flexible grid, and they are increasingly becoming worried about their generation assets being financially stranded. This means that it's often easier to sell thousands of energy storage units to thousands of homes and businesses, than it is to sell that same thousand units directly to the utility. With pumped hydro, there's no distributed option, so you end up being at the mercy of the utility wanting your product.

So yes, you're probably right that pumped hydro is on paper cheaper for addressing the Duck Curve "neck" problem, but this is a case where the difference between theory and reality is substantial for non-technical reasons.

Pumped hydro needs a pretty specific geographic profile. It's pretty great where it works, but good luck finding good candidate spots.

Some people have been working on an approach pulling railcars full of rock up a slope, which should have more location options, but I don't know how the math and economics works out.

It doesn't. The only gravity based storage mechanisms that work out take their materials from the immediate environment like hydraulic hydro storage which scales up to 2 TWh for a large storage plant.

> Some people have been working on an approach pulling railcars full of rock up a slope, which should have more location options, but I don't know how the math and economics works out.

Have a link?

There's a similar approach that seems promising - dropping a huge weight down disused mine shafts, and winching it back up at times of low demand. It looks to be a lot simpler than the railway wagon approach.

The UK has loads of potential sites, I imagine the US does too.

A fairly detailed look: http://euanmearns.com/short-term-energy-storage-with-gravitr...

There were two or three backup sites within 15 miles of Dinorwig that were being evaluated prior to it being built. There in case the main site proved impractical or too expensive for some reason.

I'd posit there are ample disused quarry sites that could serve a similar purpose scattered around the UK, US and much of the developed world.

> Couldn't you do other types of storage rather than batteries?

Sure. But they all have their own downsides. Pump air into vast underground tunnels. Pump water into towers. Spin up massive drums as flywheels (this one is my favorite. Spinning weights are incredibly power dense). Lift massive weights.

It's all been done. There are companies working on making these things more practical.

Meanwhile the power companies have instituted new standards in grid-tied inverters for solar panels that have effectively solved the problem for them.

It's based on the frequency of the grid. As the grid becomes over powered it's frequency tends to shift a bit higher. As the grid becomes over loaded the frequency tends to shift a bit lower.

By responding to these frequency shifts the grid-tied inverters will either allow electricity to be generated when the frequency falls back slightly, throttle electrical generation when the frequency rises slightly, and then shut off completely when the grid goes down.

So the effect is that those expensive solar panel arrays in places were they are already popular... are effectively shut off during the peak solar power generation hours. And are turned off completely when the grid itself is shutdown.

Lithium batteries represent a unique opportunity for truly distributed power generation. They are just reaching the point, price-wise, were they can be used as a practical way to store energy on a individual small business or household manner.

This way individuals themselves can take charge of their own power needs and no longer has to rely on centralized bureaucracies to save the world for them.

Unfortunately due to the way grid-tied systems are now required to work it has effectively nullified most of the benefits of having private solar arrays. IF you want to be your own power generator you have to be off-grid to do it.

And this the reason this is unfortunate is because almost nobody is building private solar arrays this way.

There is a semi-religious belief that grid-tied systems allow a ROI in the form of selling power back to the grid. Also they are built with the assumption that battery backups for the house is too expensive and impractical. All these things are true if you are looking at low-voltage SLA battery arrays and grid-tied standards from 5-10 years ago, but with high voltage lithium power (it's a lot cheaper to convert 300v DC to 240v AC then it is to convert 48v DC to 240v AC), how accounting for energy usage has changed, and how grid-tied systems are being remote controlled by the power company this is not necessarily the truth anymore.

Can you link a cite for those standards for grid-tied inverters? I've never heard of grid-commanded curtailment for home solar — generally those frequency signals are created by the inverter to tell the (AC-coupled) PV to turn off when a house is off-grid and the battery is full.

Thanks for this pointer! I asked this in my comment about wind, but wanted to ask you: why is a renewal dip (and generation response) so different from several GW of planned maintenance? The renewal dips are actually quite predictable as well.

Hmmmm, I think the difference is that the dip will be ever-increasing to the point where it completely eliminates the need for fossil-based generation. California (and many other places) have 100% carbon-free energy mandates over the next few decades, so this isn't a once-in-a-while situation where you can just ramp up a bit more at other facilities to get by for the maintenance window.

We think that leveraging electric vehicles as a storage device will help solve the duck curve problem.

Utilities are reluctant to invest billions of dollars in energy storage to smooth out their demand curve.

But consumers are already making this investment by switching en masse to EVs.

We're building software that will one day help utilities transact with individual or fleet vehicle owners to leverage their vehicles as energy storage.

Or we could just stop kidding ourselves.

As Bill Gates and others who've done the math have repeatedly said[1], there is no storage solution anywhere on the horizon. Batteries are orders of magnitude too expensive.

You have to consider the problem holistically. Steel and concrete production. Places like Tokyo that go days without power. Countries like India that are not going to pay a premium.

If you're concerned about climate change, please take a serious look at nuclear energy[2][3][4].

1. https://youtu.be/d1EB1zsxW0k?t=520

2. https://www.gatesnotes.com/About-Bill-Gates/Year-in-Review-2...

3. https://ourworldindata.org/what-is-the-safest-form-of-energy

4. https://www.forbes.com/sites/michaelshellenberger/2019/02/05...

According to the U.S. Energy Information administration, nuclear is more expensive than battery storage: https://www.eia.gov/outlooks/aeo/assumptions/pdf/table_8.2.p...

Nuclear builds in the west right now have been boondoggles; the nuclear industry must regain its ability to build effectively to play a role here. This has been done before through standardization of design and serialization of production. This could conceivably be scaled back up if we were serious about decarbonization, with high likelihood of success given previous successes (e.g. France decarbonized 80% of its electricity grid in 15 years by building 58 standardized nuclear plants).

Current prices for variable renewables and batteries are mostly on the current margin. Everyone agrees that costs of variable sources skyrocket as penetration increases due to curtailment and overbuilding. When you get 100% of your electricity on a sunny day (including from storage through the night) from solar, the next solar plant you build will have to be curtailed. Seasonal and crazy-weather variations are much harder and more expensive to fill with variable sources than the daily fluctuations.

These lowlow prices we see headlines about today are about building renewables in a world alongside hilariously cheap fracked natural gas plants that can pick up the slack.

There's good info along these lines in here: https://www.nrel.gov/docs/fy16osti/66970.pdf

Nuclear is unique in that it's the only low-carbon energy source that can run 24/7 for years at a time (followed by a month outage, then 2 more years, etc.). Hydro can kind of do this in certain geographies, but is very hard to scale to the point we need. Nuclear also uses far less land and raw materials than the variable renewables, especially when chemical battery storage is included.

But yeah unless nuclear folks can get costs back down soon, no one is going to be interested. If in the end variable stuff is indeed hard at massive scale (I strongly believe this will be the case), then if we don't have nuclear, fracked and high-carbon natural gas and oil (for transportation) will be around at 50% of our total energy for a long time.

Interesting coincidence that France stopped at 80%. That's the same amount that widely touted as being straightforward with current renewable tech.

I'd guess it's for the same reason, after that you're dealing with seasonal variations that would leave nuclear or renewables unused for most of the year.

Yet in practice let's take a look at France (nuclear) vs Germany (leading solar/wind):

"French electricity costs are just 59% of German electricity prices. As such, according to the prevailing economic wisdom, French electricity should be far more carbon intensive than German's. And yet the opposite is the case. France produces one-tenth the carbon pollution from electricity.

Why? Because France generates 72% of its electricity from nuclear, and just 6% from solar and wind."

France is cheaper and produces 2x more electricity from clean sources compared to Germany, where costs keep going up.

And yet both France and Germany plan to move to mostly renewable. Which the French government thinks will save them money. Something that probably wouldn't have been possible without Germany's far sighted leadership on this issue.

I wonder how you could attribute the carbon saved by all the people choosing solar and wind as the current cheapest options to the people who put their money where their mouth was when that was just a projection.

Right. Far sighted leadership of shutting down perfectly good, clean, nuclear power plants, and trying to replace them with intermittent renewables. That need to be backed up by more reliable alternatives, like strip mining old-growth forests for dirty brown coal to burn in steam generators, or importing electricity from France's nuke plants.

Just brilliant.

You're thinking too small.

Climate change and deforestation are global problems, and Germany has helped fund a global solution.

Solar and wind are growing rapidly, they're currently passing the total yearly generation of nuclear but with 30% yearly growth will soon be adding the equivalent of the total nuclear fleet every year.

It is actually brilliant.

If Germany has sacrificed some of their own woodlands to make that happen then that's just more impressive.

Avoiding nuclear power is just stupid, and shutting down functioning plants is cutting off your nose to spite your face. If you actually care about net effects on the environment, anyway.

I don't know, but I suspect the French government isn't shutting down perfectly good, economical nuclear plants just to spite themselves.

I imagine that the nuclear plants are expensive to maintain since they're getting old. So rather than maintain them, they're probably going to just shut them down. And since new ones are so insanely expensive to build (again, don't know why), it's probably more feasible short term to invest in renewables.

But that's just my suspicion. I know nothing of French politics and very little about energy generation, but a little about finance, politics, and human psychology.

I'd love to know what it costs to maintain the nuclear plants they're shutting down, how much they're spending on renewables, and I'd love to look at it finacially. It's possible it doesn't make sense, and they just want the "green jobs". It's possible none of it makes sense! But I suspect there's some sense in this somewhere.

As far as I know, the French are not shutting down nuclear plants.

The Germans did shut down operating plants, for political reasons, with the excuse that the capacity would be made up in solar and wind. When that fantasy failed to materialize, they had to fall back on burning more dirty coal.


You're just pulling that out of your ass now. Germany already had enough renewables to compensate for the loss of the nuclear energy. The problem with shutting down nuclear instead of coal is that it canceled out the emission reductions of deploying renewables. The amount of energy being produced by coal remained unchanged for 5 years and so did the emissions but they never added more coal capacity.


The currently announced policy is to reduce nuclear to 50% (from 75%) by 2035 as nuclear plants age out because it's cheaper and easier to replace with renewables. But within the last week they've been saying they haven't made a final decision on building any new plants and still could go 100% non-nuclear renewable if they don't get an explanation of the current cost overruns and are sure it won't happen again.


Most of the difference in prices is due to Germany having very high taxes: https://ec.europa.eu/eurostat/statistics-explained/index.php...

And those are still (non-household) consumer prices. That means they don't just reflect the cost of generation, but also what consumers are able and willing to pay. Unfortunately I didn't find any information on generation costs.

The problem is that the cost of elecricity in germany is going down but industry is excempt from the EEG fee which is used to pay for the renewable energy plants. The end result is that households end up paying the electricity bills of industrial companies and so the retail price rises as the renewable share goes up.

Germany (leading solar/wind):

Huh? Germany leading? Their fabled Energiewende was about moving away from nuclear. It had the unfortunate effect of moving to more coal plants to ensure base load is met.

One problem with wind power in Germany is that they don't have sufficient capacity for north-south power transmission. They've got wind in the north, they need it in the south... problem solved, you'd think. Alas! They need far, far more capacity over very long distances (1000km).

Anyway, I would not have called Germany leading in solar xor wind, but i didn't look into this. Perhaps the rest of the world is even worse? Seems unlikely though.

Germany never added coal, the amount of power being generated by coal remained unchanged from 2010 to 2014. Of course one should asking oneself what the point of renewable energy is if it isn't being used to reduce dependence on fossil fuel but that is a different topic. The problem with the Energiewende is that the lazy government isn't deploying renewables fast enough. The renewable technologies that are available are more than sufficient until Germany hits 80% renewables.

I don’t know how they compare these costs, the document doesn’t say. Battery storage doesn’t generate power after all. Nuclear power’s great feature is that all the externalities tend to be included in the price. This is also makes it look more expensive than almost anything else. Do these wind/solar prices include decommissioning and full life cycle maintenance costs?

A short, back-of-the-envelope calculation gives a price of 600billion dollars for Germany's consumption over a whole year when the cost drops by just two orders of magnitude. So yes, it is still expensive, but the technology is still in its early years and it could plausibly work out at a truly grand scale.

Here's another link for you: https://en.wikipedia.org/wiki/Dungeness_Nuclear_Power_Statio...

Summary: Dungeness B reactors were scheduled to be complete in 1970, but finally cam online in 1983 and 1985 at 4 times the original budget, but now "have been non-operational since 2018 due to ongoing safety concerns."

> Or we could just stop kidding ourselves.

Speaking of kidding ourselves,

> If you're concerned about climate change, please take a serious look at nuclear energy.

Please take a serious look at human political structures over the last decade. I don't believe that any nation in the world has sufficient stability to safely manage nuclear waste for the next few decades, let alone hundreds of years.

Electric cars add storage at almost no cost to the utility.

Consumers in North America bought 260,000 battery electric vehicles in 2018. This is over 15 GWh of storage. And utilities paid $0 for it. As many others pointed out in this thread, EV penetration is minimal--so imagine how much storage potential will be added as adoption grows exponentially in the next 5 years.

There is no way for utilities to access this storage yet--but this is what we are building (at least the software part of the equation).

I agree that nuclear should be taken way more seriously.

But V2G is also game-changing since it lets utilities "rent" capacity instead of "owning" it.

I'm not putting additional power cycles on my battery and necessitating early replacement unless you pay me for the privilege.

Most importantly, where do you place the discharging posts? Definitely you do not want the cars discharged during night so you cannot use them in the morning, so you'd have to somehow predict how much the given car is driven. Plus it would have to be connected to the grid during the middle of the day to charge with solar. Which means reworking all parking spaces...

Lithium Ion battery cells right now cost ~$100USD/KWh and are good for approximately 1000 cycles. That works out to ~.10USD/KWh of provided capacity, which is too expensive, but not "orders of magnitude" too expensive. Obviously there are a lot of additional costs, but I don't think they add up to making them orders of magnitude too expensive either. For perspective, Natural gas comes in at ~.025USD/KWh here.

There is also some evidence that cycle count may improve on cells in the near future as prices fall. a 5x improvement in cell lifespan would make a huge difference.

> But consumers are already making this investment by switching en masse to EVs

"En masse"? Maybe if you live in Norway or something like that.

According to the data I was able to find online, only about 2% of the new cars sold worldwide in 2018. are EVs. If we take the number of all cars currently on the roads worldwide, it's around 0.25%.

As for the "Vehicle-to-grid" thing in the United States, the Wikipedia page lists 3 experiments which are in progress and 2 claims with "citation needed" flag:


I might be saying "en masse" 18 months prematurely. But at the speed utilities move, that's a blink of an eye.

I'm often reminded of Steve Balmer in 2007 laughing at the iPhone: "it's $500 and it doesn't even have a keyboard!?"

Less than 10 years later smartphones are ubiquitous. Not popular--ubiquitous.

That's where EVs are right now: "they cost more and they can't even go 250 miles!?"

EVs will be cheaper than gas cars by 2022, and a few years after that nobody will buy another gas car.

$500 smartphones are only ubiquitous in the most wealthy parts of the most wealthy countries in the world.

Smartphones are ubiquitous, not high-end iPhones.

EVs will be ubiquitous, not Tesla Model S.


I know we're thinking focused on cars, but fwiw: "China has 99% of the world’s 250 million electric two-wheelers."

I'm not sure that I'd say "en masse" either, but the current numbers to understate the phenomenon by quite a bit. Just a year or two ago it was <1%, and I'd expect it to hit >4% by the end of next year with a continuing high growth rate for at least another couple of years after that.

I do not think it is at all obvious what will happen afterwards, but its going to start looking really rapid to people in another year or two.

Most homes have multiple cars.

A single high-end car could support multiple homes without the grid for over 24 hours.

So I think even at adoption rates under 5% you would see enough capacity to shave those peaks and make it easier to respond to changes in demand.

See Australia's tests with grid-smoothing batteries. Immensely successful and it has paid for itself.


Countries with significant lake hydroelectricity can "store" the solar power (no batteries needed) by reducing flow during the daytime (and increasing flow at nighttime if required).

This is because hydroelectric dams are essentially stores of electricity. No need for pumping or other expensive storage schemes, if you already have one!

This isn't mentioned in the Wikipedia article on the duck curve: I'm not sure why as it is a legitimate storage technology (just an existing one).

Largely true, in many places. Hydro has a few problems though when scaling: It takes lots of land, and hydro reservoirs become massive sources of biogenic methane, in some cases to the tune that building them is worse than building coal plants (!). Also, hydro is statistically far more dangerous than things like wind, solar, and nuclear. When a big dam fails, lots of people die.



I dont think you can just say Hydro is far more dangerous than nuclear. When a nuclear accident happens not only a lot of people lose their lives but the effects are felt for centuries. Lands become inhabitable, generations of people carry the burden.

You might be surprised! Contrary to popular belief, nuclear energy is among the safest forms of energy production we know. Yes, this includes short and long-term deaths from Chernobyl (50/4000), Fukushima (0/<=1), and Three Mile Island (0/0).

The "Chernobyl" of hydro, of course, was the Banqiao Dam failure, which killed up to 230,000 people. For some reason, very few people know about it, and there are no HBO specials on it. Go figure.



Considering that fossil fuel kills about 4 million people per year via air pollution, year after year, both hydro and nuclear are really safe. Nuclear net saved 1.8 million lives by 2013 (and counting). Hydro probably has a good number like that too.


Graphical comparison of nuclear safety vs fossil (not hydro): https://twitter.com/EnergyJvd/status/1183663682552827904/pho...

Hydro plants are already going to be running at full output during the peak demand time when solar is not working. Just because you have solar during that day, which means you get to use less of the stored hydro water during the day, does nothing to increase the generating capacity of the hydro plant at peak time. The turbines and generators are already certain size and turn a certain speed. They don’t get bigger because you used less water at midday.

So to make hydro work as a storage solution you need to build more hydro than what you currently have already to replace fossil generation and make sure you can meet all your evening demand. That is absolutely not going to happen anymore, in developed countries hydro is well tapped and nobody wants more dams or lakes. In developing countries there may be some potential but it won’t meet all demand.

There needs to be another storage solution for solar to become viable in this way.

> Hydro plants are going to be running at full output during the peak demand time when solar is not working.

How do you know that?

During drought periods surely it's not true (if there's not enough water to run them).

The flow of rivers is controlled to some extent (which may prevent saving during the day and generating in evening).

And presumably it might be true in one country, but not another, depending on a raft of factors.

I'm guessing there is data somewhere showing utilisation of hydro during evening.

Part of it is based on previous experience working for the national grid in a hydro rich 80-90% clean generating country. I can’t speak specifically for the US but think about it like this. You need to have enough generating capacity on your grid to meet the peak demand. Which according to this duck graph is when solar is not generating. But we have not overbuilt capacity by some large margin. And hydro is only some portion of that total capacity today. The rest is coal, nuclear, increasingly gas, and some small amount of renewables. Therefore how can hydro alone meet the all of the demand when you remove the goal, gas, etc.

Can your reference some live data?

I had a look at https://www.transpower.co.nz/power-system-live-data

And that showed hydro was running at less than 50% of capacity for one relevant data point.

I am guessing the data series of hydro usage versus capacity is available for NZ. Finding out the constraints and understanding them is more difficult.

Your point of maximum capacity may be true, but we are talking usual daily usage (duck graph) and whether solar can use hydro for that.

In think you are arguing about occasional peak network capacity (e.g. heat wave) which is an outlier and you retain power stations with extremely low utilisations for those abnormal peaks (and ignore green issues).

Yeah, I'm curious if the number of "batteries" already connected to the grid could mitigate this quite a bit. For example, large industrial freezers (or perhaps even "smart" home freezers) could take their temp extra low during the day so they wouldn't need to use power during the evening.

Also, with the current discrepancy in electricity costs between the day and evening, seems like this could make something like Tesla's Powerwall much more price competitive if it meant you only had to use the grid during the cheapest times.

Yes they can. The company I work for does that using a few halls with industrial freezers.

This seems like the sort of claim that can only be checked with math, not intuition. Without further ado: https://colab.research.google.com/drive/1fcwuNd2XANIXmKGPRN_...

Basically, I got some raw data by tracing the graph in Wikipedia[0], as well as the current amount of hydroelectricity generated per day. I made the (highly pessimistic) assumption that aside from hydro, solar would be the sole energy supply for the grid by scaling the production numbers in the graph until the area under the solar curve matched the area under the total, minus the production of hydro. I don't actually know that anybody is seriously advocating for pure solar + hydro, but I wanted to run the numbers.

The result, assuming cars with 85kWh batteries that can be 70% available for grid balancing (leaving ~80-100 miles of range for driving and such) is that California would need 4.3 million cars acting as grid balancing.

For reference, that is just under 30% of the number of cars currently in California, where electrics made up 7% of new car sales last year.

Ultimately, I think the bigger problem is battery degradation, not battery capacity. Being used this heavily for grid balancing would at least double total battery usage on a car, and apart from Tesla with their million-mile drivetrain, no other manufacturer seems to be over-designing their batteries enough to support that. Perhaps that will change if the economics of feeding power back into the grid become compelling enough.

[0]: https://apps.automeris.io/wpd/

Wouldn't EV increase electricity demand in the evenings when people get home and plug in their cars?

Not if you program the cars to charge overnight. It requires more smarts, and grid interaction, but the cars could be made to obey instructions from the grid operator.

There would have to be compensation, of course.

Individual EV drjvers can set the start time for charging. It's polite to set it for sometime after midnight. This isn't fancy smart-grid stuff, just a new kind of basic common sense that will emerge as EVs get more popular.

When pitting politeness against convenience (having the most charge in your car as is possible at any moment in time, in case you need to use it for something unanticipated), I don't have high hopes.

To really incentivize this behavior properly, economic incentives need to be strong.

I used the word "politeness" because my coal-age utility (National Grid Massachusetts) doesn't offer time-of-day pricing. They seem to have no interest in doing so beyond occasional small-scale government-funded trials. If they did offer TOD pricing you bet I'd be on it.

But don't underestimate the power of social pressures: "politeness". I don't throw my trash in the next street over, because to do so would be rude. The integrity of the entire Internet depends upon an altruistic TCP endpoint algorithm called "exponential backoff".

The incentives are strong. My electricity is 50% more expensive before 9pm. So I set my timer (a one time setting) for then.

Problem solved.

California has time of use rates (relatively new and uncommon). And we're beginning to see deep discounts in the middle of the day consistent with the solar. So people may strategically charge during the day. If it's a commuter car on a reasonable commute and you've got 300 miles of range, it's feasible to wait till midday.

You can see the power supply / consumption at http://www.caiso.com/TodaysOutlook/Pages/supply.aspx

Let's I have an all electric car and sign up for your program. How would I calculate the cost to me of the shortened lifespan of my batteries? Would the manufacturer honor their warranty if I used the car for not only driving but also as part of a battery bank?

> consumers are already making this investment by switching en masse to EVs

I don't think we are there yet. The US fleet of electrics is in the range of about one million units. Total cars on the road is about 300 million. In other words, electrics represent 0.3% of the total fleet. That is a very long way from "en masse".

Electrics, without a doubt, in some form, are the future. We all know this to be true. However, the adoption curve requires solving a bunch of problems and an expansion of the infrastructure that is also nowhere near being adequate.

This translates to a simple fact: We need far more competition and far more entrants into the segment. Not announcements, but rather real vehicles you can buy.

Car makers have been announcing and showing models that are vaporware for years now. I learned a long time ago that this is bad business. I used to do this kind of thing during my early days as an entrepreneur. It frustrates the customer base to no end and makes them lose trust in the company. It isn't quite lying, but it's close.

I believe the inflection point will come at the intersection of a new energy storage technology (let's call it "batteries" for now) as well as the infrastructure to support it.

The fires and power issues we are having in in California highlighted the weaknesses: You can't adequately charge your electrics during emergencies and range issues make them either inconvenient or just plain dangerous. I'll bet a lot of people have been made to reconsider electrics in CA precisely due to this experience.

Power self-sufficiency by means of a large enough solar array at home could mitigate some of this, however, this does nothing for the person caught in an emergency away from home. We had friends who had to endure three hour trips to go get their kids after forced evacuations due to fires. When you are concerned about the safety and well-being of your loved one's, particularly kids, having range and charging issues isn't a joke at all. In fact, it's a powerful deciding factor against electrics.

For context, as soon as the fires started we drove 5 minutes to the local gas station and filled-up our tanks. We got hundreds of miles of range pretty much instantly. Even better, we didn't have to think about extending that range at all, because the infrastructure is ubiquitous and available without issues outside the affected radius (and within a very small fraction of the vehicle's range).

I'm not down on electrics at all. Just being realistic. It will happen when all the conditions are met and there are a dozen companies offering real products, rather than three.

Lateral thought: My thinking is that the future of electric transportation will require an energy storage system based on LIQUID charging rather than plugging in. We need to be able to go into the equivalent of a gas station, pump out the spent portion of the electrolyte (or whatever) and pump in fresh "charged" electrolyte. It needs to be 5 minutes in and out for 0 to full range.

I realize this might not be as "green" as people would like it to be. Not that electric cars are green at all (just wait until we have 300 million battery packs to deal with as waste). Going from 0 to full range quickly by plugging --assuming a charge storage system that could handle this-- requires very dangerous voltages and currents. I am not sure what this would look like if we had 300 million vehicles on the road with a need for that kind of energy to be delivered that quickly. I think I can say that our electrical grid is likely not designed to deliver at this rate (I haven't researched this but I have a sense this statement is likely true).

Anyhow, not a simple problem. At a personal level, we were about to invest in one or two electrics by the end of the year. The fires and power outages have painted a real negative light on the practicality of these vehicles when things matter most. We'll have to rethink. No conclusion yet.

The Duck picture alone makes this page worth checking out

I went into the article based on your recommendation, and I was not disappointed

According to the Wikimedia page, the duck is from an article [1] by Jordan Wirfs-Brock.

Some quick internet stalking suggests is the daughter of Smalltalk and OOP pioneer Rebecca Wirfs-Brock. Illustrious family!

[1] http://insideenergy.org/2014/10/02/ie-questions-why-is-calif...

Indeed! Veering (far) off topic, but I cannot resist posting this while we're on the topic of ducks:


With PG&E shutting off power in California, I think we'll see a growth in battery packs (like Tesla Powerwall) -- at least among the ~1 million impacted customers.

There is now a much stronger incentive to get off the grid, with the shut-offs expected to continue for 10 years.


> "This now feasible, intercontinental network would integrate America, Asia and Europe, and integrate the night-and-day, spherically shadow-and-light zones of Planet Earth. And this would occasion the 24-hour use of the now only fifty per cent of the time used world-around standby generator capacity, whose fifty per cent unused capacities heretofore were mandatorily required only for peakload servicing of local non-interconnected energy users. Such intercontinental network integration would overnight double the already-installed and in-use, electric power generating capacity of our Planet."

> Telegram to Senator Edmund Muskie, Earth, Inc., 1973, Fuller.

This applies to Photovoltaics only, not to wind or concentrated solar, and looks like something that is pretty specific to California's energy mix and policies.

If we replace fossil fuels during the day and turn on hydro and other during the night, I still think the duck curve is not an argument against adding much more PV capacity.

Could we use space mirrors to illuminate solar farms at night?

Wouldn't it be cheaper just to whip a cable over to Australia?

We could bore a series of holes through the center of the planet and then use gigantic orbital crystals to create a array of lazers to focus the sun's light through to the dark side of the planet.

I'm a fan of using magical heat-producing rocks to boil water and spin turbines, but admittedly that has fallen a bit out of fashion.

This could be nuclear or coal

There's an argument to be had that burning coal doesn't count as magic while fission does, but in any case, good catch, let me amend that to: magic heat-producing rocks that don't fill the atmosphere with dinosaur squeezings.

It's temperature would have to be maintained around absolute zero to compensate for the transmission loss, for that we would need a continuous source of cheap, clean energy.

Would it?

I thought about a bunch of small satellites with mirror foil.

The solution to balance it is batteries or any other energy storage solution.

Why not just put the farms in space and beam the power back?

This gives me a flash back to SimCity. I wish someone told me the clean microwave plant occasionally malfunction, sending an orbital ion cannon through my downtown.

I thought about this too, but had the impression that only mirroring a bit of sunlight around would be cheaper and safer.

It is because of effects like the "duck curve" that make me wish I was an electricity trader. Being able match these patterns and events and correlate them to different geographies would probably be quite profitable.

I briefly looked into trading on some of the popular ISOs (NYISO for example) - but it quickly became clear to me that my hobbyist approach to these markets was not going to cut it as they are way more technical and require a bit more capital than your typical equity markets.

EDIT: https://www.nyiso.com/ << if you want to learn more about the US de-regulated and open electrical grid, this site has 100s of technical pages and material. They seem to want and encourage people to be part of the market.

Sounds like capacitors would help here, even regular ones, since they'd only need to hold excess solar energy for a few hours in order to smooth out that evening demand.

In the graph on the page, the solar output isn't even meeting the current demand. No need to store it.

The point is that, without storage, no matter how much solar you install you don't lower the required amount of fossil fuel plant capacity at all because the peak occurs when solar is at 0.

The yearly peak, in summer, occurs when solar is running.

The daily peak, on spring and autumn days, falls outside the solar peak, but also well below the yearly peak.

So solar does help displace the least efficient and most expensive part of your gas fleet.

But it's not the peak that's the point of the duck curve, it's the solar falling faster than gas can switch on and that difference is very amenable to being fixed by even a small amount of battery storage.

> it's the solar falling faster than gas can switch on

Is "switch on" really the problem? As Germany showed during the solar eclipse (https://www.dw.com/en/german-power-net-survives-solar-eclips...), as long as it's a predictable event, you can have the necessary generation already "switched on" and running at a low power, and the limit becomes only how fast you can ramp up their power. Which is still an issue, but not as much as it would be if they had to turn on all the gas generators with zero notice.

Yeah, it's just the ramp up, it's predictable in advance.

Also the worst case scenario is curtailing some solar because you need to bring the gas plants online a little earlier.

Storing it would help reduce the peak in the evenings.

The duck curve is a non-problem, just like all the predictions from the same era that 5, 10, 15, 20, 25% was the absolute limit that renewables could provide without totally overhauling the grid.

That number has continually risen as grids just did it and blew past the imagined limits.

California now has a mandate for 100% clean energy which makes "we can't turn our gas plants on fast enough a few days a year to avoid curtailing some renewables" seem ridiculously parochial.

On those high renewable, low demand days in the future there will be plenty of responsive demand to soak it up. Free car charging for anyone plugged into a socket for example.

I drew a duck graph on a whiteboard the other day. Wish I had known it had a name, could’ve gotten an easy chuckle from the room.

See Also: Bart pattern in crypto trading

So there is no wind in california?

This problem is blown out of proportion.

The energy storage is not a "do or die" issue for the industry by any extent.

Technical solutions for short term energy storage are known, and been used for decades. Were it economically sound, it would've been adopted already. In fact, it is being done when the need is genuine, just without much publicity.

People who spin this topic feel to me to be in the business of selling tech companies.

The problem today is solved, almost universally, with high-carbon fracked natural gas plants. This is cheap in current markets and incredibly effective from a technological standpoint. As we recognize climate change as a problem, this solution loses credibility, but yet, there it is.

I don't think people are recognizing how hard it is to deeply decarbonize with intermittent energy sources, and so I have to disagree. I think this problem is too often brushed aside as a non-issue by renewable energy optimists.

Hmmm, so I work in the distributed energy resources sector, and I disagree that people in this space aren't recognizing how hard it will be to deeply decarbonize using intermittent generation. You're completely correct that as we add more and more intermittent generation to the grid, keeping the grid always online becomes crazy difficult.

However, everyone I know in the space looks at that challenge as opportunity rather than discouragement. It's not like we're going to hit high penetrations of intermittent renewables overnight. It's going to be a gradual ramp over years that allows us to also deploy complimentary systems to deal with the ever increasing intermittency. Piece-by-piece we will deploy flexibility technologies (DR, DERMS, storage, etc.) and move to new regulatory incentives and utility business models (performance-based ratemaking, etc.) that allow us to reach our 100% goals.

There's a ton of super smart people already working on the intermittency problem, and entire businesses are being built to solve it (disclosure: many of them use our platform). So I guess while I agree with your assessment of the level of difficulty, I don't agree with your level of despair. There's way too many smart engineers working in the space now for intermittency to be in impossible problem to solve.

I'm sure people in the space are well aware of intermittency issues. I think the interested public isn't well aware of it. Many people seem to think batteries solve the problem, and it's just Big Oil and Big Coal in the way.

I'm not despairing that it can't be done. I just think people who say it's a solved problem are undercutting the challenge, and hand-waving away serious problems with the 100% intermittent renewable route while proclaiming confidently that other massive low-carbon energy sources aren't desired or needed. Full disclosure: I'm a nuclear proponent.

I know we can decarbonize rapidly; it's been done at scale in small regional areas many times.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact