I have a strong suspicion that there’s a point of solar capacity as a percentage of grid production beyond which you really can’t go because you start having to curtail it to maintain stability.
1 Like
If you don’t curtail output, yes, there’s a certain (fairly low, think 20-30% tops) penetration of solar you simply can’t go beyond and still have a stable grid. Battery systems push this slightly, but you pretty much still have to be able to maintain stability if they can’t charge because they’re already full, so I’m not sure they move the needle much.
However, if you allow intermittent curtailment, you can push this up a lot further, because you can shut down some generation on demand for peak generation vs demand ratios.
On a good late spring day, my system (which is anti-optimized for mid-day peak production but still does peak at solar noon) produces around 10-11kW, when “base house demand” can be down around 0.5-1kW - or a 10-20:1 ratio of generation vs demand. On a spring/fall day with moderate temperatures, that sort of export is actually maintained as well. I’ll see 8-9kWh exported on a peak hour often enough during the spring on 10kW generation, which means that house demands are, in fact, that low. Not everyone will be quite like that, but enough will be that you have to handle it at a grid level. Another (bigger) array is going in up the hill on the same feeder, and between his array and mine, there will likely be times when we’re literally running the entire local neighborhood on our solar exports (if you ignore phases - but the substation should mix those nicely).
So claims of 10% solar penetration being the max you can do without curtailment are at least within the bounds of reasonable. It could be worse, though, if you have partly cloudy days, because cloud edge lensing is a thing (I’ve seen my array peg out both 6kVA inverters and damned near peg out the 3kVA at a time, for almost 15kW instant output), and tends to be paired with rapid drops from the actual clouds. “Puffulus” spring clouds from thermals make for some rather rapid swings in production that the grid has to make up.
Even if you’ve got enough production to meet your demand, there’s no guarantee it will last, so you need other generation plants online (“spinning reserves”) to handle a sudden loss of generation or spike in demand. It’s the reason you can’t run an off-grid system anywhere near peak output on panels without batteries in the system - the spinning reserves are the effective battery of the power grid, able to dump a lot of power in on short notice (hence “spinning” - for traditional generators, they’re at operating RPM, synchronized, and just idling along waiting for a load spike and rapid throttle response). You can meet some of those requirements with batteries, and distributed battery banks help some, but they don’t solve the “10:1 production” issue for very long if you have a lot of solar on the grid.
Remember, home solar is basically “The grid deals with my shit for free.” Not a problem for the lone loon at the end of a feeder, a major problem when you’ve got entire subdivisions (20 houses an acre or more) with significant solar going from export to import at the pass of a cloud.
Curtailment gives you the ability, as a grid operator, to deal with some of this when your solar production is threatening to drive grid demand negative. You can simply start shutting down, or curtailing output, on some arrays, until you have enough net demand to keep your spinning reserves happy. Crank up the battery charge rates, trim spinning reserves, but in some conditions, being able to cut solar output is pretty much required to keep the grid stable - and I’ve no problem with that. The amount of solar you can tolerate on a grid with curtailment is far, far higher than without curtailment.
There are a few ways to do it, but what I’d like to see is a system with current transformers on the man output from a house (hard to do with certain panel layouts, like mine… grumble damned center fed busbars…), with curtailment being a “Ok, thou shalt not export” command for some period of time. Any loads behind the meter can be served by the local solar, but it will reduce output so that it won’t export to the grid (typically you see a slight net import on non-exporting systems, but for curtailment, it really doesn’t matter, just get close to zero and you’re fine). Any loads the house has will be met by solar, so you wouldn’t be importing from the grid, but your net export (which is what matters to the grid) drops to zero and stays there for some period of time. However, this should probably be tied into the 1741-SA/CA Rule 21 style behaviors, so that if the frequency or voltage drop below spec, the inverter does go to full output, in an attempt to help stabilize whatever just happened. Solar has the ability to ramp at absolutely insane rates, so curtailed solar, properly handled, is effectively a form of spinning reserve (you can ramp an inverter from 0% output to 100% output in a handful of cycles, and solar will source the power if it’s well lit).
I expect we’ll see the same sort of things with larger community and utility scale solar arrays in the future - although we don’t currently, in the US, really split out the ancillary services like the UK grid does, this would be of some significant value going forward. A 50MW solar field, on a bright sunny day, could be sourcing 50MW (insert handwaves about DC/AC ratio here) - or, perhaps, it could be bidding into the markets as 25MW solid output (which it can handle in pretty much any condition for that day), with another 20MW of spinning reserves that can come online nearly instantly if needed. This sort of thing also pairs very well with behind-the-inverter batteries that can make up any gaps in production.
“Well, I’ll probably be putting 50MW in, but if we get some clouds going over, that might drop to 30, then back up to 60, and wobble around a bit until it settles in around 50… good luck!” is of far less value to a stable grid than “I can maintain 50MW until 6PM, with a scheduled ramp down in the evening of 1MW/minute.” If you have some batteries behind the inverters, and operate a bit off your peak power point, you can make those sort of guarantees for performance, and that’s much more useful for stable grid operation.
But almost every study I’ve seen about high renewables on the grid includes a good bit of wind/solar curtailment, as a simple requirement for stability. It’s the only way you can make such a grid actually work in a real world environment.
Is rooftop solar + battery the desired endpoint, or a phase we go through that enables consumers get cleaner energy than the power companies (and government?) will currently give them?
Some smallish percentage of consumers having their own clean energy setup might be reasonable, but the photos from the news article showing 100% of suburban houses decked out with solar panels seems pretty inefficient (they don’t all have their own cow, or flour mill etc.), as well as making it more difficult to solve the problems Australia is experiencing.
I would have preferred to see a large buildout of industrial solar that could be reasonably managed and curtailed as needed for stability (a 100MW plant producing 30MW all day is more useful than a 100MW plant whipping up and down as clouds go over), but we are where we are… and unfortunately, there’s no good path through that I see for high solar penetration areas. Substation batteries will help a lot, but there’s still some need to curtail homeowner plants for stability. Home batteries are nice for backup, but I don’t think they’re a sane future for energy systems - they’re radically too expensive to be much more than toys, in most areas.
Another option, which Hawaii has been doing for a while, are non-exporting systems (Idaho Power allows these as well now). This will run behind-the-meter loads, but simply won’t export onto the grid. Pair with batteries if you want, or not, but it’ll turn your house into a non-consumer of energy during sunny periods.
I really don’t have a problem with the power grid being maintained. If everyone has solar on net metering, nobody is paying anything for grid maintenance, which is the bulk of power rates…
1 Like
Or stuff it into storage.
A clean breakout of ‘grid cost’ (transmission) which is bidirectional vs generation cost would be nice. That’s up to your local utilities regulator though.
I mean you push power into the grid? Great, help pay for the grid and you can get the spot price after. You’re just grabbing rid power? Great, pay for the grid along with the energy price.
‘Net metering’ is just rewarding early adopters.
Yeah, breakout the costs for wires/etc as one fee that always needs to be paid, if you want the wire running to your house and be able to sell/buy power from the grid. Then after that’s met, it’s the normal buy/sell. We’ll see how that turns out, we’ll have massive consumer complaint, probably, if that’s the case. Especially with the way a certain SoCal power company has squandered their regular maintenance for decades and caused some wild fires.
The problem with that is that the grid really is most of the costs in many areas, and the goal is to not give everyone $100 power bills with usage making up +/- $20 of that. Energy is typically around $0.04/kWh, the rest is grid costs/profits.