As solar capacity grows, duck curves are getting deeper in California | Hacker News Reader

As solar capacity grows, duck curves are getting deeper in California

eia.gov

108 pointsaresant192 comments

> presenting challenges for grid operators.

Don't worry, they've planned their revenge: https://www.10news.com/news/local-news/state-regulators-hear... At least SDGE (and maybe more utils?) are proposing income based fixed rates so that solar homes will have to pay more. I have no issue with fixed grid access fees so that solar users can pay their part, but grid access fees should be equal for every household and the utilities need to figure out how to stop being so profitable and serve their communities

I don't think grid access fees should be equal personally. Why does someone in a condo in San Jose pay the same as someone in a remote Sierra foothill community (ex. Omo Ranch) with extremely high fire danger which is leading to enormous utility costs. Many of those people might be able to be off-grid if there was incentive (ie. high costs of grid access). You find more off grid homes in Alaska where connection is simply not feasible due to remoteness.

The current proposal is the exact opposite. The condo owner with higher income would be paying far more.

The fundamental issue is further complicated by the government involvement in infrastructure expansions and funding.

In your case, if the government passes a funding bill to bring power into remote communities, why should citizens pay again?

Isn't it already more expensive for remote customers?

Electricity bills are typically separated into generation, transmission and distribution.

Remote customers have to pay more to get connected to the grid in the first place, but the higher maintenance costs for low density grid service aren't passed along on an ongoing basis.

The problem is that electricity prices in California historically lumped a bunch of infrastructure costs into the per-unit costs for a kilowatt hour of electricity. This was an implicit subsidy for anyone whose costs to the electricity system came disproportionately more from infrastructure maintenance than from energy consumption. It has become a growing problem in recent years as rooftop solar has proliferated, since solar households are (mostly) still connected to and reliant on the grid but are consuming much less energy from the grid.

The most straightforward solution would be to split out energy consumption and infrastructure charges and bill each customer for their actual infrastructure needs and energy consumption. This is politically unpalatable because it would raise costs more than average for low-energy-consumption customers (typically lower income households).

The better solution that wouldn't hurt lower income households would be to change billing like proposed above but offer state financial assistance to low-income households in proportion to the old implied subsidies. This way the economic distortion would end but low-income groups wouldn't suffer. I don't know why California didn't do this.

The worse solution that California has selected is to charge households a fixed connection charge according to income rather than according to infrastructure costs. This maintains the economic distortions of the old billing system since it means that there is no incentive to avoid wasteful infrastructure. If you're affluent and you're going to be soaked on electricity service either way, you might as well live in a big sprawling rural property since you'll not get any economic benefit from living in more infrastructure-thrifty urban or suburban housing.

California utilities break rates down into territories, so someone in San Jose does have different rates than Sierra foothills.

My understanding is this based more off weather-driven demands (Marin has much less AC needs than Davis), but still, there are regional differences in pricing.

Here's the real crux of what you're suggesting though: utilities are extremely regulated in what they can do, so it's quite possible the utilities commission said they can do price discrimination off of regional load differences but not off of a more nebulous "grid impact / risk score". The latter opens up all sorts of thorny equity questions (eg is what you're describing just redlining poorer rural communities through another mechanism?)

It doesn’t seem that hard. Just calculate the number of end users served by transmission lines and divide the costs accordingly. Don’t even need to factor in risk. This would make denser areas less costly on a per home basis, which is exactly what you want to encourage, and in CA would also generally translate to risk as well.

Redlining was bad because it was explicit racial discrimination.

Setting prices based on costs is not redlining. If we want to subsidize poorer people in rural communities, we can do that through a means-tested program just like we do for the urban poor.

I view safe/stable power, clean water, and waste/sewage systems as bare minimum utilities that everyone should have. I don't live in a third world country and so equal access to those kinds of utilities should be a given for everybody. I really like the idea that everyone in my state is entitled to equal access to the grid and I'd much prefer to have people on the grid rather than having people in poor/remote communities trying to create their own shitty replacements with the dangers and fire risks that'd create.

I see no reason why people who don't live in cities should be punished or burdened with higher access fees, especially if those people can't afford to move/live elsewhere. It may cost a lot more to run utilities out to users in remote areas but it costs very little to run them to users in ultra-dense urban areas and that's where the vast majority of the users are so it should all balance out.

I mean, it costs much more to expand and maintain access for them, why shouldn't they pay more?

Because electricity is a common good. Some subsidize others so we may all have electricity.

Kind of like the postal service. Or insurance.

Just saying… somehow states with huge rural areas manage to have prices 3x or more lower than cali

So do it with marginal prices for electricity. The price for electricity should have nothing to do with the recipient’s income.

If people want wealth redistribution, then do it with wealth taxes and cash.

No, insurance is against things that didn't happen yet that are out of your control.

You can't just buy fire insurance when your house is burning down, so you shouldn't be able to have "electricity insurance" when you buy a far away house

The postal service is necessary to do things like send you IRS communications, drivers licenses, jury duty, etc.

When the government no longer uses paper to communicate with people we can talk about privatizing the postal service

I remember living downtown San Jose last year and being charged $0.75/kWh and it was mostly due to a mandatory 50% upcharge for green power. Totally bullshit.

You're buying PG&E's lies.

Hydro Quebec serves a far, far tinier population than California, and has equal fire hazard, if not more, and far more untamed nature than California. Yet Hydro Quebec doesn't have PG&E's issues, and delivers the least expensive power in North America.

There are reasons for the lower costs, but my point is, they maintain their lines. They cut back brush. They deal with encroachment. The problem is PG&E does not perform the maintenance they need to.

All complaints by PG&E "oh, that's so hard to do, poor us!" is just wavy hand stuff, lies, and bull.

From what I've read, they just see trimming back brush and trees as "cost", and barely do it as a cost saving measure. Don't buy their bull! The problem is 100% them. Other power suppliers do not have this problem!

Roseville power and SMUD, that operate right next to PGE and deliver power at a fraction the price, kinda prove you are right.

I suspect there is a lot correct about what you are saying, PGE sucks. Hydro-Quebec is a crown corporation returning profits to the government unlike privatized PGE. However, the cost of generation is separate, Quebec has tons of hydro power and PGE doesn't generate power.

It seems that numerous remote communities are diesel powered, I know this is the case in the far north (https://www.canada.ca/en/natural-resources-canada/news/2023/...). Are there any communities in California not connected to the grid even in the very remote areas? I suspect the cost of electricity and everything in these communities is very high and also subsidized.

A few other questions I'm wondering is how much of the rural infrastructure in Quebec relies on the connections to the various (remote) hydropower projects (https://upload.wikimedia.org/wikipedia/commons/4/4f/Quebec_M...). Are communities very far from those connections also served. I can't find a map of actual communities connected to the electric grid.

Interestingly California has over 95% of the population living in urban areas compared to 70-80% in Quebec, not really sure which that would favor.

Yes, PGE sucks. However, I still think my argument stands that the infrastructure could be much cheaper if the 95% of people living in urban California weren't paying for the rural infrastructure regardless of the nature of tree trimming costs.

Commercial rates have a demand charge, essentially a charge based on peak power usage.

Meter cost plus demand charge is what would match usage to utility infrastructure cost.

The weirdo progressive rates will push parties to grid detach, which will not be good for the economics of the grid.

The income based fees aren't the utility's idea (although maybe the specific amounts are?).

CPUC was empowered to create the income based fees after AB205 was passed. The individual utilities are just trying to satisfy the government's requirements in a way that won't hurt their profits.

This doesn't seem to be coming from utilities

"The proposal he's referring to comes after California passed a law last year requiring that utility companies establish a fixed monthly fee based on a customer's household income."

The utility companies have absolutely no business knowing the subscribers household income. What a completely ridiculous proposal, they're in a 'cost+' business.

Its not a proposal, It was passed in 2022.

It is simply another example of California introducing back door welfare taxes that aren't included in the state budget.

Taxes aren't increased, but companies are mandated to charge different prices to customers.

This is exactly what it is. It’s actually a tax, basically a back door to raise prices across the board and generate huge revenues, disingenuously passed under the auspices of equity and fair share politics.

Also, now the state government has detailed access to income based billing of every citizen, all set up and ready to go. But I’m sure it will never be used for anything else.

The idea here was to charge an income-based fixed charge to cover the costs of infrastructure, nebulous future “climate investments” and “general operation” (read: offload costs to generate more profits), and then cut the rates for the electricity that’s actually used. Lower rates, the theory goes, will encourage all Californians to buy new electric appliances and electric cars.

But all that’s going to happen is that the cost will go up, and the three companies that sponsored this bill will make massive profits as they shift more of the cost of infrastructure onto rate payers and play the “investment game” they’ve been at for 30 years: the game that’s killed hundreds of people and burned entire cities off the map.

But this time will be different I’m sure.

> but grid access fees should be equal for every household

No, they should vary depending on the grid connection type. Someone with a three-phase max 100A per phase connection should pay more than someone with a single-phase max 50A per phase connection.

Why is that the case if they are next door and tying into the same line? They require identical infrastructure.

Because infrastructure can handle a finite number of users based on their connection types.

If you’re running a 50kW kiln (or more realistically 4 x 12kW) in your back yard then your annual use may not be that high but the grid needs to handle larger demand spikes. Paying per kWh alone doesn’t adjust for the relevant infrastructure.

So your argument is that it is a proxy for use variability, and that determines how much the grid needs to be overbuilt.

I'm not sure that either of those are solid assumptions,but worth some thought.

Demand charges (peak kW rather than kWh) are very common in commercial and industrial.

It's very rare a residential house has a 50kW kiln, so those kinds of spikey loads basically balance out over a subdivision.

Although interestingly, an unintended side-effect of aggressive time-of-use pricing is the entire subdivision has programmed their air conditioners or EV chargers to turn back on right at 9pm when peak pricing ends. That kind of unintended mass-coordination DOES create sometimes-problematic high demand spikes.

Ultimately, there's a need for customer-specific real-time pricing that's responded to by behind-the-meter "power control systems"... managing flexible loads, in a way that doesn't negatively affect occupants.

> If you’re running a 50kW kiln (or more realistically 4 x 12kW) in your back yard then your annual use may not be that high but the grid needs to handle larger demand spikes. Paying per kWh alone doesn’t adjust for the relevant infrastructure.

There are two alternatives here. One is that your spiky demand doesn't correlate with others, in which case it's irrelevant because although 50kW is a lot for one house it's really not a big deal for the power grid. The other is that it does correlate with other spiky demand, at which point you charge everyone a higher price per kWh at those times.

Charging based on regional demand works great on the production side, but residential, industrial, and commercial experiences different peak times. Thus distribution demand at the level of substations don’t necessarily correlate well.

You could model things with individual per customer rates per minute, but charging based on connection sizes is a lot simpler.

It isn't clear that connection size is correlated with any production costs whatsoever, and if it is, what the degree of that cost is.

If the goal is simply to gouge customers, independent of use or cost, you might as well charge by height.

Distribution costs and production costs are different things.

The grid itself needs to be paid for as does electricity production. A random neighborhood that wants a huge number of Christmas lights isn’t an issue from the production side, but it can be a real issue in terms of distribution without using enough kWh to pay for that infrastructure.

Because they have the potential to load the grid up in very different ways.

Is that true in practice and is it more than simply proportional to the power consumed. In the US, most new residential electric services are 200 Amps at 240 Volts. The maximum power that could be drawn would be 48 kW. Seems like the potential variability from a home is already enormous. Conversely, someone with a high capacity connection could have very regular usage patterns

The US is very different from Europe in this sense. My house uses a small fraction of an equivalent family home in the United States. On the coldest days in winter I might go through 10 KWh of electricity and 15 cubic meters of natural gas. During a summer day gas usage will be < 0.5 cubic meters (mostly for shower water) and electricity will be 60 KWh or more returned to the grid.

Someone with smaller service won't be able to create as much variability in load as someone with a much more beefy hookup.

>Someone with smaller service won't be able to create as much variability in load as someone with a much more beefy hookup.

I think we are in dangerous waters when we are are basing public policy on "ability" to have impact opposed to "actual impact. I think it is a genuinely interesting question if and how much this variability contributes to the grid capacity requirements.

You can basically think of individual variability as noise on an analog signal. Does single user variability average out, and if so, on what scale?

How does this variability compare to other amplitude changes, like aggregate or seasonal daily use patterns?

I think it is entirely possible that this noise could be negligible at most scales, but obviously dont have the data.

However, someone with the actual data could easily do an ANOVA evaluation, and see what the actual numbers are.

It's simple physics, actual impact follows ability. In other words: you don't ask for a hookup larger than the one that you intend to use because you already pay more per month for that larger hookup.

Except that it doesn't, necessarily.

You could have some unusually heavy usage during off-peak hours and that doesn't require any additional grid infrastructure because there is already plenty of capacity during off-peak hours. Whereas if you want to use the same amount of power during peak hours, that would require more grid capacity, but in general the way to handle that is by charging a higher price per kWh during peak hours, giving everyone the incentive to use less then (and charging them appropriately if they don't/can't).

Impact does not follow ability, it is the exact opposite when we are talking about variability.

Total demand for power increases linearly with the number of users. Percent variability of demand decreases with the number of users and approaches a limit of zero.

If we are talking with sizing power infrastructure, capacity required increases with the number of users, but safety factor required decreases with the number of users.

At the margin, infrastructure cost scales with per capita power usage, not with individual variability. The variability cancels out.

Not California, but in general terms this is not accurate. My local grid connection provider (the largest in the country), does not differentiate between 1x25A and 3x63A in cost. It's the same price.

That's a pretty big difference in available power for the same price. (5.8~43.5 kW)

But the only differing cost in that scenarios is the 2 extra conductors to the local transformer, and more likely just to the edge of the property. The 3-phase power is still present in the area, as ideally alternating houses are on alternating phases to balance the load over the phases. The difference in cost would be a one-time hit during installation, and the ongoing maintenance would be the same as 3 separate houses (1 house per phase). That maintenance could be rolled into the cost per watt, so the more you have available the more you can use and the more you could pay.

The installation cost should vary based on what the house wants access to, and the ongoing cost should be the same as every household. A standing charge for the cost of the infrastructure existing is ridiculous when that same infrastructure is what the power company relies on to deliver their chargeable commodity. It's effectively double dipping - how is it any different from ISPs charging for access and then charging for data on top?

I think an important consideration is that the overall grid is not designed for all houses to have the higher capacity connections. So enough connections and they're forced to make massive changes to the infrastructure.

Not to say utility companies don't make obscene profits instead of reinvesting much of that into the infrastructure.

Regarding the ISP, its really the same argument. If I want 2GB/s on a neighbourhood line that supports a max avg service size of 1GB/s then they would be forced to upgrade their lines to service me. Granted, unlike power grids they're liable to just not do that and let service quality degrade and quote the "Up to X Gbps" clause

Solar panels and batteries are so cheap now that I expect a lot of SFHes in California to go off-grid when the income-graded access charge hits. My SFH is in the south bay and I disconnected in November.

Grid access fees should be paid by the hour rather than the month. If I need the grid for 72 hours a month, then bill me at a 90% discount.

Battery backup seems like a great way to divorce yourself from the grid and solve this duck problem, except for the cost.

Whole house backup batteries are, for some reason much, much more expensive than electric vehicles despite not needing any of the car parts (motor, wheels, seats, chassis, etc.)

Comparing the Tesla Powerwall $9,200/13.5kWh ($681/kWh) vs Model 3 Battery: $14,000/75kWh ($186/kWh)

Why is there such a premium for house batteries, when there are even less engineering constraints on them?

Obviously there are inverters to pay for, but that can't explain a 4x difference in price?

Not sure why the prices are so much higher when compared to EV batteries, but now that Na-ion batteries are in production we should see a significant reduction in prices for house backup battery systems over the next few years - 1/2 to 1/3 the cost of Li-ion. And while Na-ion has less energy density it's probably better suited for home power backup (less likely to catch fire, works better in low temps, for example) where you don't care about the extra weight.

Sodium ion batteries don't have the same longevity as LFP in their current form (2-3k cycles vs 4-8k cycles), so they are not really an economical replacement at this point.

Wouldn't batteries used in a home power backup situation be rarely cycling? (unless you completely go off grid as the earlier poster was suggesting)

If the battery is being used to address the duck curve, it cycles approximately every day.

If you can buy twice the capacity for the same price a cell cycles every second day, except on the rare days where you would have had a problem with the old solution.

I expect the answer is some combination of economy of scale (they sell lots more cars than walls) and subsidies....

I don't know the particular reasons for the battery discrepancy, but going off-grid is pretty complicated (assuming you want to connect to the grid sometimes). Generally, solar and/or battery installations stop working when the grid goes out, so as not to be feeding power into a potentially dangerous downed circuit. I'm not sure of every scenario, but the most common grid+off-grid usage lets you wire a seperate set of outlets for off-grid and on-grid usage. This adds a fair bit to the cost, as well as decreasing convenience.

Hybrid inverters have auto shutoff (islanding I think?) so you can merge the grid power with your solar power and there is no difference when the grid is down. Off the shelf, installed, and magically works. It acts as a glorified whole house UPS.

Solar is one of those areas where there is a lot of bad info floating around promoted by installers and certain manufacturers (ahem Victron).

Interesting; it might be different in different jurisdictions, or my information might be out of date.

Going off grid is easier than keeping the grid around when you have batteries. You need an automatic cutoff that prevents backfeeding to the grid during outages for safety. Off grid you do not have such a constraint.

Doesn't the power wall include an inverter that allows for hot switching between battery and main line supply? That would explain a large part of the price disrepancy.

An inverter is a single speed motor controller. Every EV has at least one.

I suspect the better availability of V2L setups recently and in the next years will make those prices closer. Unless you drive lots, unexpectedly or every day, you essentially have a huge battery sitting idle in your garage. Why bother getting a separate one then? (Or maybe just a tiny one integrated with hot water)

And with the vehicle battery, if you need extra energy you can drive to a fast charger somewhere and bring it back to your home.

It seems so inconvenient. My wife has to shut off the air conditioning if I want to drive the car? I can't store my solar when I'm at work during the day? I can't go for a drive at night because the house used all the power that was stored in the car?

Yes, the grid connection disappears the moment you install a battery at your home, and exactly as I wrote in the comment there's no possibility to have a smaller battery as well. Everyone drives their own and only car to work for a whole day in 2024. And no car has a "charge/discharge to threshold X" setting. It's a terrible solution that won't satisfy anyone. /s

My guess is higher manufacturing cost.

Assumption: manufacturing batteries in a different form factor is a significant cost driver - a big slice of the manufacturing cost pie (the other being common raw materials). Powerwall is lower volume, therefore doesn’t leverage manufacturing economies of scale as effectively as the Model 3, making it significantly more expensive to manufacture (hypothesis; I haven’t done the math).

Idk why is this being downvoted, looks like a very probably explanation to me

This company sells rack-mounted LI batteries and inverters that are sufficient to build a Powerwall-like system (in nice enclosures) at a significantly lower cost [1]. I think Tesla is banking on wealthy customers and installers wanting the premium product. The problem is that (unlike cars) it’s not likely that brand name is going to win here, which means long term this product category is going to be a race to the bottom.

[1] https://www.currentconnected.com/product/sok-sk48v100-ncbt-u...

Warranty is probably a big factor as I believe Tesla powerwalls came/come with a 10 year warranty and considering these batteries are basically fully drained daily in normal usage that’s a bit of wear.

Still doesn’t justify 4x however!

>Whole house backup batteries are, for some reason much, much more expensive than electric vehicles

It could be the invisible hand of the market. Cynically or un-cynically.

Not being cynical I mean that is the market could be smaller, so they're pricing in the additional overhead per-unit due to having to spin up additional production lines for everything but the storage medium itself.

More cynically, even if the price difference driven by differing, smaller product delivery workflows isn't enough to explain the markup, their market analysis may have shown that the list price that sector can bear is higher. And there could be rebates, tax incentives, etc that they're pricing in so they can "offer" while maintaining their desired profit margin even after these.

It also bears mention that the construction sector is less competition driven. The "payoff" price isn't necessarily the price that a consumer will balk at and find a competitor but the price that the consumer will bear before "value engineering" it out. When you're dealing with home construction prices a relatively large line item like this can slip through unnoticed.

LiFe02 batteries for home grid-tie systems run about $150/kwh (10kw grade A 200ah cells) at the consumer level. Inverter capacity runs about 3k$ for an 18kw grid tie inverter. So, in order-able parts it costs about 5000 total to build a 10kw “power wall”.

It seems like there should be (and probably are) comparable integrated solutions around the 5k price point, which would be something like 500 per KWh.

At any rate, at sizes under 20kwh the battery is likely to be the smaller part of the cost of the unit….which also makes it possible to build an 80kwh system with 18kw capacity for around 18000 dollars, so there’s that.

So in short, yes, it’s probably the inverters and the branding that make up the vast majority of the powerwall price.

I'm interested in a 50-100kWh system, and (I think) I'd only need one inverter? Maybe I have that wrong.

It just depends on how much energy you will need to use at any given time and how quickly you want to be able to charge your batteries. If you are somewhat careful, 5-10kw is enough for many homes, depending upon whether appliances are electric or gas, as well as heating/cooling loads.

Worth bearing in mind that as of 2022, California also has the second highest retail electricity prices in the US [0] (behind Hawaii, which seems fair enough). Their policies are a bit suspicious, they seem to be leading to bad outcomes.

[0] https://www.eia.gov/electricity/state/

We are also the richest state by a lot of measures so higher prices from that higher standard of living should almost be expected

New York is the richest state by GDP per capita

Income is a better proxy for “richer state” than GDP per capita:

https://dqydj.com/income-by-state/

At every percentile I checked, Californians have higher incomes than New Yorkers.

Except for 70th percentile and below down to the 40th percentile

Thanks for the correction! The graph disappears for me around the 80th, so I didn’t know you could click in the invisible part and still see numbers. I assumed the trend continued.

The differences, however, are pretty minor (a few hundred dollars to $1,500 at most from what I spot checked), whereas in the differences are thousands and tens of thousands of dollars after the 80th percentile. So it seems like the top 20% in CA earn sufficiently more than the top 20% in NY to far outweigh whatever extra 30% to 70% in NY earn.

Yes, that's true for the average, but doesn't affect the median, which is basically exactly equal.

https://en.wikipedia.org/wiki/State_tax_levels_in_the_United...

if you look at taxes, it depends on what kind of taxes you count, for example North Dakota's severance taxes makes it rank #1 in the nation, but it's from drilling for oil/gas

But if you look at the highest state spending per capita

https://www.kff.org/other/state-indicator/per-capita-state-s...

#1 is DC, so it's the "richest state" even though it's not a real state, with #2 Alaska spending the most per each resident of which there are not much

The conclusion is that California in reality doesn't spend the most on its residents

Title: "As solar capacity grows, duck curves are getting deeper in California"

HN title is exactly backwards, but gets at the insight that the operator demand in California is almost at zero -- or neutral -- during the day due to behind the meter solar.

Right, I read this as the duck curve had inverted back to a neutral load curve. To be fair though I feel like I may be fairly familiar to the duck curve compared to most people.

This graph (and "net load" generally) includes the subtraction of both behind and in front of the meter solar, as well as wind. Doesn't really affect the point though.

Worth noting that this graph represents the single worst/best day of the year, when demand was lowest and supply highest.

The trend over years is obvious but sadly we're not as close to carbon free power in California as this would suggest if you assumed it was an average day.

Note that demand being at zero doesn't mean we are shutting down power plants. Nuclear, for example, is baseload and basically never shuts down. But California does export a lot of energy when the net demand is low. And of course charging batteries is an even bigger part of it.

>> is baseload and basically never shuts down

https://www.rstreet.org/commentary/understanding-negative-pr...

Renewables like solar and wind have already changed our grid so dramatically that "baseline" is sometimes finding ways to dump power off the grid. Texas going negative is a harbinger of things to come. The duck curve, the uneven generation of solar and wind, means that "baseline" solutions like nuclear (and coal) aren't responsive or flexible enough.

The CAISO (California independent system operator) website has really good visualization that's updated frequently. For example, yesterday (2024-01-25) the net demand in the afternoon dropped to as low as 3.3GW. Then within about three hours the net demand increased to 23GW.

It's also obvious that the net demand trend only subtracts utility-scale solar that is directly visible to the utilities. Home solar has definitely dampened demand and this is apparent in the demand curve.

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

https://www.caiso.com/TodaysOutlook/Pages/default.aspx#secti...

is the specific chart that corresponds, showing "net demand".

And note the graphs truncate above zero depending on content. Activating the "show solar and wind" option adjust this in this case.

When solar is abundant, homes should be pre-cooling as low as any occupants can tolerate. Example: 85 degree day, homeowner is at the office, cool the home to 64 degrees when the sun is shining, so that there's less cooling necessary when the owner walks back into the home at 5:30.

How quickly does heat naturally increase in a house that's been cooled like that?

I've only lived in apartments, but I guess it gets back to equilibrium with outside temperature in a few hours once AC is turned off. I guess a couple hours of cheap cooling in the middle of the day might save an hour of expensive cooling later, if it's timed right. I'm interested if there have been any studies or maths to back it up and quantify the effectiveness.

depends on insulation

In part, but that's not all of it. The bigger driver is the amount of thermal mass.

Using the house as a thermal battery (heat or cold, dealer's choice) works great in my house but my house is so well insulated that we have CO2 problems.

Planning to setup a fresh air intake with an ERV to mitigate this.

I mean that's going to depend heavily on insulation, light insulation, how many device you have generating heat (computers, servers, refrigerator, tv), and how many times you enter/leave your place.

In a well insulated place with airlock-style entrance (even just a mudroom with solid insulation on both doors) where you're blocking out sunlight and not running a ton of electronics, it could easily take hours and hours.

Some research here: https://www.pv-magazine.com/2023/08/14/mitigating-residentia...

> I guess it gets back to equilibrium with outside temperature in a few hours once AC is turned off.

I've been in houses where we shut the windows as soon as possible after sunrise, as temperatures started increasing, and the house never caught up with the outside temperature - it was cooler all afternoon, cooler until sundown. No AC needed.

I think that tactic might work well generally. Just let in the cool air at night, shut the windows and trap it. At least you can save a good part of the day's AC. It's obvious, free, and simple, but I never see it discussed.

My dad has holes under the window for the water to roll off when you clean it. Those holes let in cold and hot air readily. He doesn't even wash the windows. It's a really bad design for what you're saying

Read on about decrement delay and thermal buffering if you're curious.

https://www.greenspec.co.uk/building-design/decrement-delay/

Nearly have my own solar, I'm wondering if there's a way to do a home thermal battery like this (although for me might not be necessary, as I've also got a battery...)

Like - if I put a water tank under the house, can I cool it (summers) / heat it (winters) as both a backup water supply, and thermal sink? And have that also work to help maintain house temperatures?

You are talking about hydronic heating, which is old 1970s tech.

Imo, it would make more sense to store the power in battery form, and cool or heat it on demand.

That's true from an energy perspective but from an economics perspective it is very hard to beat a well insulated tank filled with water.

I get the physics, the question that economics when were talking about installing a well-insulated cistern under a house, and retrofitting it with hydronic Heating.

Disposing of the cistern capacitor chemical (H2O) is also better for the environment than the battery chemicals. Manufacturing is probably better too.

Sure. I have had hydronic heating in several houses. You don't even need storage. Soler thermal is common throughout the world.

But on a price basis, you can get a 5kw battery for 2k. Good luck running water plumbing and radiators throughout your house for 20k

> you can get a 5kw battery for 2k. Good luck running water plumbing and radiators throughout your house for 20k

That's not quite apples-to-apples; the energy from the battery also needs distribution through the house. Of course, every house has wiring, not all already have radiators.

Especially if it's a large tank. You've got the square-cube law on your side here.

Better still would be to use a truly huge cavern as a thermal battery for your whole city.

Using it for both heating and backup water supply is probably a really bad idea: you'd basically be storing water in the ideal temperature range for legionalla growth.

The workaround is to use a heat exchanger so that you're not drinking the stored water, just transferring heat from it on-demand.

They're doing similar things at larger scales in Sweden: https://www.bbvaopenmind.com/en/technology/innovation/underg...

Yes, but OP wanted to use it as both heat storage and emergency drinking water supply. If you're using a heat exchanger you'd still need separate tanks for hot water storage and for emergency drinking water - and that's the very thing they were trying to avoid.

Water under your house is hard to use as a backup water supply. You would need a pump to use it in the house when your utility water had no (or low) pressure, and you'd want to bypass the pump in normal operations, etc: mechanically complex. It might be easier to have the water tank above the house, but only if your structure can manage the extra weight, and you need to be well prepped for leaks. And yeah, microorganisms.

It can be done but it's not that practical since you need a huge insulated tank and if you're also using the water you it's even less practical to store heat in it. One can use a pool as a heat sink during summer though. Water stores 1.16 kWh/m³K. The heat capacity could be increased with phase changing materials.

> 85 degree day, homeowner is at the office, cool the home to 64 degrees

That sounds miserable, and realistically no one is going to do that. If you want to leverage that sort of idea, you need new technology (like a big insulated cold water tank or other buffer device).

Our 50 year old CA house is typical---wood framed with a vented crawl space, so no slab and minimal link to the ground mass. It has only 1970s fiberglass insulation in the 2x4 framed walls and ceiling, with none beneath the floors. It also has 1970s single pane glass and many southwest-facing windows. We make effective use of the house's thermal inertia in both summer and winter. We lack central air-conditioning and have an inefficient 1970s gas furnace.

I can definitely feel the radiant temperature of the house and its content around me. This is linked to the thermal mass and contributes to comfort along with actual air temperature. Our biggest improvements have been modern roofing and DIY heat-rejecting window films. Both seem to reduce the radiant connection between the outdoors and interior, helping in both summer and winter.

In summer, nighttime ventilation and daytime closure means our indoor temperature can start low and lag behind the outdoor heat by many hours. At some point in the evening, the falling outdoor temperature meets that indoor and we start ventilating again. During multi-day heat waves, we will see that the house retains heat, increasing both morning lows and evening highs. But, it still lags behind the outdoors and smooths out the peak.

In winter, we heat to a more comfortable target for a few hours in early morning and after dusk. We let it coast with a lower baseline target otherwise. During clear days, we get enough solar gain to either hold it steady or even warm the house above our desired set point. Then, the evening heat cycle either holds or brings it back up as the light fades and the walls start to feel a lack of radiant warmth. By the time the house is feeling chilled, we're snug in our bedding and it coasts lower until the next dawn cycle.

> modern roofing and DIY heat-rejecting window films

What do you use?

> nighttime ventilation and daytime closure

I do that, it works very well (though depending on the house, of course), but I'm surprised I don't see it as standard advice for homeowners. Even better, an input at a cool spot (near shady grass, etc.), and output with an exhaust fan at the top; auto open/close based on thermostats outside and inside.

> In winter, we heat to a more comfortable target for a few hours in early morning and after dusk.

Is the early morning heat for efficiency somehow (or just it's too chilly when you wake up after an unheated night).

Also: Our perception is of it being colder when it's colder outside, despite the thermostat being set to the same temperature - that is, we need to set the thermostat a bit higher to not feel chilled. I wonder if our perception of temperature depends on something like the coldest (micro-)draft more than the ambient average.

It saves energy to let the house get colder overnight when we can rely on bedding to keep us warm. With a lower differential between indoors and outdoors, there is less total energy lost to the outside than if we held a constant target. I'm only dropping the set point about 5 degrees, not disabling heat entirely.

Then the morning heat is to make it easier to get out of bed and start the day. We stop it again to take advantage of the common solar gain. But also, it doesn't take as much heat to feel comfortable once we've eaten and gotten more active for the day.

For the roof, it is just one of the "cool roof" compliant asphalt shingle systems, with full plywood sheathing. We did not retrofit radiant-barrier sheathing, as there is fine print that this could overheat the shingles. As far as I can tell, the same performance to reject solar load in summer also reduces radiant loss in winter. Vaulted ceilings feel less exposed on a cold winter night.

For windows, I applied Gila "heat control" film found at the local Home Depot. It is a modestly cool gray tint that looks only a little bit mirror-like from the outside, not too different from modern low-e glass. It takes a little practice to get satisfying results with the DIY application. I just used a spray bottle filled with tap water with a few drops of dishwashing soap, rather than any proprietary application fluid.

Regarding heat perception, it is air temperature, air movement, and radiant effects. In winter, you may feel colder facing an exterior wall or ceiling than with an extra room or story between you and the outside. Or facing a window versus having heavy drapes pulled shut. This is even if the air in contact with your body is the same temperature.

Thanks! One thing that works for me: I just turn the heat all the way off at night (or down to ~12 C to protect pipes, but I don't think it's ever gotten that cold inside). I just use more bedding if I need it - I'm generating 37 C, much warmer than desireable, so I just need to trap enough of that heat. Mornings are delightful, and my body retains the nighttime heat from the bedding (in its thermal mass, I suppose). Usually I don't turn on the heat until sundown.

You can get a better effect with the phase change materials. CaCl2 hydrate has a transition temperature of 29 C = 84 F. That's good for keeping warm in the winter, but not so great for the summer. On the other side, acetic acid has a transition temperature of 17 C = 62 F. But that will be below the dew point sometimes, so you need a clever way to distribute the cold and stay dry.

64F is 17.7C - what is miserable about that? In Scotland in winter my heating is set to back off to 17C during the night and when I’m not in. Is that an extreme temperature for some?

My apartment rarely goes above 19C in winter, but for many people that is much too cold.

In most parts of California you don't need active cooling at night. Just open all the windows at dusk and let the cool air blow in.

And then well insulated buildings will stay comfortable a good portion of the day too. Even during Sacramento 100+ heat wave I was amazed how cool a modern home can stay without A/C

Works well unless you are a renter (as many in CA are) and do not have screens on your windows thanks to the “landlord’s special” effect.

Screens should come with the rental, but adjustable window screens are easy to come by if this is a problem you have.

Where are these well insulated California homes? Coming from somewhere with real winter, I am aghast at the insulation standards here.

The air in a house has relatively little mass so for heating it works well but if you switch off the heater the house will get cold relatively quickly compared to water based systems and the reverse is true for cooling. What you could do is to cool down the floors if they are made of concrete, but then you have to be very careful to stay above the condensation point.

The contents of the house (furniture, counters, floors, walls, etc.) also act as thermal mass. I would guess many times the thermal mass of the air.

Large ceramic indoor potted plants are wonderful thermal mass vessels.

Do you mean real plants in a ceramic planter? Why is that better than furniture or rugs, for example? Water content?

Three dimensional volume for one. Ceramic and soil also will hold the heat you get from sunlight and radiate the heat back out into the environment.

Thermal mass is the key and potted plants provide that with the large volume of soil combined with the ceramicpot that if placed into sunlight will absorb the energy and radiate it back out and keep the room warmer.

Isn’t this the case for energy storage technologies? Store excess generation duri bc periods of low demand and sell them back during peak load?

Grid attached storage is growing and is also a good use of used cells repurposed from EVs (vs. direct recycling and recovery of metals) when their capacity is degraded beyond usefulness for mobile applications they remain valuable for stationary use.

Report for CAISO is here: https://www.caiso.com/Documents/2022-Special-Report-on-Batte...

Looks like capacity has grown to >16 GWhr last year.

Where do you see 16gwh?

I am always confused by how CA reports energy storage in GW, not GWH.

The number I see most often is 6GW for California storage capacity.

Right side of Figure 2.2.1

I still don't understand why GW would be capacity, and GWh is duration.

I would think capacity would be gigawatt-hours, and max delivery rate would be gigawatts

From a grid operator perspective its capacity https://www.energy.gov/ne/articles/what-generation-capacity

Capacity = Maximum Power Output

Capacity is the amount of electricity a generator can produce when it’s running at full blast. This maximum amount of power is typically measured in megawatts (MW) or kilowatts and helps utilities project just how big of an electricity load a generator can handle.

Thanks, that makes a lot more sense then storage capacity. What's the rationale for calling actual energy storage capacity duration?

duration is how long the storage will last which is related to the energy storage which you could call capacity but they already used that word so I guess duration makes sense... I'm not a grid operator though

Most energy is stored with pumped hydro and building hydro isn't exactly cheap nor fast and at this point nuclear doesn't sound that bad)))

Look at the Time of Use plans that CA IOUs have on offer and you can see why this is happening. Most of them only have 2 modes: high and low, with only 4-9 as peak. If they actually wanted to manage the duck curve, they could easily put a steep discount at the hours when the problem is most pronounced, and promote EV charging, laundry, etc be done during those hours. This is just another example of the IOUs failing to innovate, their captured regulator doing nothing about it, and ratepayers continuing to pay more for all of it.

Market solutions based on price signals?!? It’ll never work in California!

ToU isn’t even available for everyone, I was rejected even though I could benefit.

Wow! Who is your utility?

This will only increase worldwide. We are moving from a demand driven to a supply driven grid, as predicted years ago by people like Amory Logins etc.

Especially because in most countries the actual cost of electricity is only a small fraction of the actual price. If you can use rooftop solar to charge your car it's several times cheaper.

In Europe 40% of the population live in dense cities, usually in flats. No solar or even home charging for them, it's going to be a mess

I don't think so. Solutions don't need to be perfect for them to not be a mess. All buildings have roofs.

Hopefully people are looking for productive ways to use intermittently free electricity.

I guess it depends on your opinions of what's productive, but bitcoin mining is obvious.

Any large scale interruptible compute, for that matter.

You would think so, but not really, because you still have to pay for the hardware but then you're only using it a small fraction of the time.

Where it works well is in the opposite direction: You buy the hardware and use it 90%+ of the time but shut it down during a supply shortfall. Then you're making good use of the hardware and helping to amortize the cost of additional generation capacity but save the grid from running out of power in an emergency.

Was just thinking that!

I know that, for example, Iceland does a lot of the world's aluminum smelting (ore gets shipped in from Australia, for example) b/c the cost of electricity is so low (all the geothermal).

Edit: For socal, desalination seems like a good option. Also maybe something involving the recent updates to lithium mining in the Salton Sea...? No clue how much power that takes, and, it sounded like it was paired with geothermal plants, so...

iceland has so much electricity then run warmers under the roads to melt the snow

I've been firmly of the opinion that after a transition period the pricing structure is going to inverted and contain seasonal and episodic events. Long term I don't think this is a problem. Because customers will appear to utilize cheap/zero cost power. I also think a lot of 'baseload' customers are just chasing lowest cost. Those guys will happily shift their demand.

And we're no longer looking at what you could believe 20 years ago, a world where energy costs on average 4-5 times more than it does today. Actually looks like energy will cost less. We definitely have issues around transition, utilities don't want to get caught out and go bankrupt.

Utilities have zero risk of bankruptcy from dropping prices. People need the power and pay the cost of transmission.

The real problem they face is an inability to grow revenue with a cost+ model and dropping production costs.

They have to be as inefficient as possible to drive up costs further, but this opens the door for competition that isn't playing this game.

I think this decade will see many more municipalities getting into the game, and a huge regulatory fight as utilities try to stop them.

A sibling comment proposes cooling the houses to reduce demands later during the night.

Carbon Capture?

I don’t believe this is the best use of the periodic excess energy, given that you can prevent CO2 directly and not have to deal with as much inefficiency as recapture. Perhaps someday, though.

https://caseyhandmer.wordpress.com/2022/02/03/terraform-indu...

Obvious solution: incentivize people to charge their EV at work!

I bet plenty of businesses would be more than willing to equip their parking lot with chargers given the right tax incentive. Have the electricity company give out charging cards so that EV charging can be combined with your residential power bill, and add a 5-9 Duck Curve Discount. Make the charging speed electricity company controlled for additional demand shifting possibilities.

PGE has chosen another path: frequent blackouts and higher prices.

You could also:

incentivize e-mopeds and ebikes, both of which have far lower energy use than a car ever will.

incentivize trains and buses, both of which have far lower energy use as long as they're even slightly filled (a 15-ton bus with only 7 passengers will still have better energy-efficiency than a 2-ton car).

The latter might require building out more public transport in the first place, though.

Of course, those would be far superior, but it'd require people to significantly change their habits.

On the other hand, plenty of people already drive to work and park there. Plugging in your EV while parking at work vs. plugging in your EV while parking at home is an absolutely minuscule change for most people, which means it is quite likely to actually be picked up.

> intermittently free electricity

Does California permit variable-cost electricity to consumers? If so, just charge a battery in the day and let it run at night.

It does. I pay $0.51/kwh at peak, and less off-peak.

I tried to convince my power company when moving back into CA to allow me to have time of use billing, but they required >x MWh over two(?) years of usage history “to make sure customers benefit from time of use billing” or some other red tape . I have off-grid solar generation for some types of appliances, and battery storage, so ToU would be economical for me, but I wasn’t permitted.

They want to make sure you can afford the power bill and know what you are doing. The average household isn't structured around flexible pricing at all and your appliances are even less likely to support it.

PV solar has been pulling away from CSP in terms of cost, but I wonder if the deepening duck curve means CSP could claw its way back? AIUI one advantage of CSP is that in theory you can heat the working fluid at one point of the day but actually use the energy later.

Dumb question: Why can't we store the excess capacity midday for consumption during the evening?

E.g. using using solar to pump reservoirs to a higher elevation then hydropower during the day? I know they do it with natural lakes[1] but couldn't we simply do this more?

https://www.nytimes.com/interactive/2023/05/02/climate/hydro...

Feels like this is an opportunity to offer consumers attractive export rates in the evening peak so they force-discharge excess battery storage

For the most part, I believe most folks with solar in California don't have batteries. Those that do, don't have enough battery to last themselves through the night. The correct policy almost certainly seems like the one which incentivizes them to use their battery at night rather than the grid. As much as people hate it, NEM 3.0 provides those incentives. A policy which has folks exporting battery energy to the grid rather than using it themselves would simply be less efficient unless they have more solar+ battery than they need for themselves, which is really just rewarding rich folks who can over build. In particular rich neighbors will end up generating more power than they need, still destabilizing the grid.

How it works in the UK at the moment is there are particular times when the grid knows it is going to be under severe load. Typically it gives about 6 hours warning. Consumers can then charge up their batteries from the grid and reduce load/force discharge during a (usually) 1 or 2 hour slot. It's designed to reduce the need to crank up old coal-fire power stations. I realise this a bit different to the every day duck-curve, but interesting nonetheless

So what's the answer to smoothing out supply and demand peaks here? More and better batteries? Wind, hydro, etc? All the above?

Dynamic pricing! Get a little smartness in front of large consumption appliances and charge much less when there's too much power. The answer will create itself without any need for breakthrough technologies. (I hope)

The current low hanging fruit for demand response would be industrial customers who have some ability to scale up/down power usage.

It seemed like we were inching towards this with smarthome tech, but now we’re losing ground. Nest was pretty revolutionary at the time with their “learning” thermostat, but aiui most smart thermostats don’t bother with real smartness.

Google/nest just spun off their energy team, Ecobee was bought out by a home energy company, so hopefully we’re seeing a new generation of smart energy companies coming.

Exactly, or at least have more tiers of pricing based on the time of day. I visited another state and was pleasantly surprised to see four time windows that shift based on season. In Southern California, there are only 2, and I doubt most customers have even been transitioned to a TOU plan. So dumb.

Batteries, hydro, and natural gas (peaker plants) all help directly. Wind isn't really "dispatchable" in the sense that the grid operator can't ask for more wind.

I don't believe nuclear power ramps up very easily, and in any case it has to run as much as possible to pay for high capital costs. Power that's going to be on standby much of the time needs to be cheap to build.

Anything that usually generates power in the evening will help some, though.

Also, the grid operator being able to schedule demand helps too. Things like water heaters and air conditioning can be run in advance and turned off at peak times.

Modern Nuclear ramps up very quickly, in minutes.

The challenge is almost 100 of the cost is up front, so the economics favor using the power all the time, because it is basically free once built.

The point is that ramping up isn't really a thing because it will have already been producing at 100%. But that doesn't mean it isn't useful to address this, because it's still a generation method that generates power during the hours that solar doesn't. You would also then have even more generation during daylight hours, which isn't as valuable, but it still has economic value and can be sold for some smaller but still non-zero price per kWh.

In California the price actually is hitting zero sometimes during the day [1], which means that at that time, it's not economically valuable at all - nobody wants it. Rather, cutting back is valuable. Other kinds of electricity generation are easier to shut down, though, so they go first. This is graphed as "curtailment" [2].

Surplus electricity could hypothetically be put to use, but something would need to built to take advantage. (Perhaps batteries or some other load.)

[1] https://pv-magazine-usa.com/2023/04/20/driven-by-solar-calif...

[2] https://www.gridstatus.io/datasets/caiso_curtailment

> In California the price actually is hitting zero sometimes during the day

Sure, sure, but these are unusual times and not the common case. And if it was the common case that the price was zero during most daylight hours, it would be solar that goes bankrupt, since that's the only time it's generating. The other generation methods would still have the revenue from generating at night, which was the bulk of their revenue to begin with because of the expected price differential.

> Surplus electricity could hypothetically be put to use, but something would need to built to take advantage. (Perhaps batteries or some other load.)

Charging electric vehicles is an obvious one which will increase with their adoption. It's also likely that the grid will need some kind of storage because even if you have baseload plants to carry most of the nighttime load, the peak load of the day is just after sunset, and picking that up is going to require some kind of storage or peaker plants but the existing peaker plants are fossil fuels.

What baseload plants give you is to only need enough storage to make up the difference between what the baseload plants generate and that peak, and only for that couple of hours instead of the entire night.

Changing evs will become more popular, but is in time conflict with demand. The discharge time is at night, and people don't want to start the day on low, even if you do all your charging at worksites, which is the other challenge.

The average EV on a workday won't need to get even close to below 50% battery if regularly charged on worksite

It's not basically free, the cost of nuclear waste disposal is gigantic and unfortunately mostly ignored.

Same for solar panels and windmill manufacturing and recycling, except you're dealing with thousands of time the amount of waste

Solar panel recycling is a hot topic, that already works for some parts and even if we don't progress at all it will not cost hundreds of billions per country like it does now for nuclear already. Per country is vague but mostly large countries have nuclear now anyway and dozens if not hundreds of reactors. Yeah, "same".

Comparison should be made per kw generated. How much money for reprocessing 1kg of nuclear that produced x amount of energy vs how much money for solar panels that should be recycled and produced the same x amount. There are other options too like building breeding reactors to reduce the costs over time but that's another topic

Let's take Germany as an example. The German nuclear plant operators made a deal with the state in 2017 for it to take responsibility of the waste. They paid 23 billion to wash their hands of their own product, and the estimate is that it will cost over 100 billion. But again, it's the tax payers' problem now. [1]

Let's take Britain as an example. It has a nuclear waste storage facility in Sellafield that costs billions every year to run and the price for the final solution is anyone's guess, again we are talking about hundreds of billions - 263 billion pounds is the latest estimate to be exact. [2]

Let's take Finland as an example, which is a smaller country. It plans to dispose of the waste in a specially built nuclear facility that cost billions. They will then depose 300 caskets there made of thousands of tons of iron and copper, again costing billions.[3]

Now whatever the cost of solar panel disposal will be, I am pretty sure the number will not be 100 billion and it will not be 23 billion either. And it will not be a really big issue what to do with it, and the vast majority of it will likely be recycled.

[1] https://www.dw.com/en/german-government-does-nuclear-waste-d... [2] https://www.theguardian.com/business/2023/dec/04/sellafield-... [3] https://www.trtworld.com/europe/finland-to-bury-nuclear-wast...

Can't rely on hydro power in California. In 2021, for example, a serious drought caused water levels to drop so low behind reservoirs they could no longer be used for pumped storage. Batteries are the best solution, quick to charge, high efficiency and can be brought online instantly.

> More and better batteries?

Yes, batteries. From yesterday:

https://news.ycombinator.com/item?id=39144978

Grid operators need some power intensive process on the demand side that they can switch on in the middle of the day and switch off at night. Something they can switch on only when needed and ramp up quickly.

That or energy storage.

This article is from last June, with statistics on the graph only through May of 2023.

My solar panel inverters have this wonderful bug where they refuse to generate power for the rest of the day after a 5-minute power outage. I hope that's the exception rather than the rule...

Weirdly, it's not seen as a problem at all in Europe while we have same or higher proportion of electricity from variable renewables, and lower from fossil fuels here.

The problem is that we don't have enough transport infrastructure to distribute it where it's needed, so what do you do with the excess capacity, smelt aluminium?

I think eventually this problem will be solved mostly by reducing transportation distances as solar will be wider distributed so on average power will travel a shorter distance.

It's not some terribly fantastic amounts of power we are talking about. Full decarbonisation of EU will need only about 5x more renewable electricity than today.

In any case, 2050 plan does not call for any drastic increase in transmission capacity. Don't forget that power output from hydro can be greatly increased if needed - total amount of energy produced can't because water in dams is finite - but peak power can be increased rather cheaply.

Reply on news.ycombinator.com