> However, because it is physically not possible to charge and discharge the battery at the same time, such trades have to be prevented.
The authors are observing that, if electricity prices are negative and your battery is not perfectly efficient, then you would like to charge and discharge simultaneously to get paid for wasting energy, but you can’t.
This is a silly limitation. Surely the power electronics or even just the control algorithms in a BESS could be slightly modified to consume power, get warm, and not transfer any current to or from the battery cells, effectively taking advantage of the BESS’s heat sink to sink excess power and sell that service.
More seriously, in a world with occasional negative prices, you would want your battery to be able discharge itself, without exporting power, in a controlled and power-limited manner so as to avoid overheating. And the optimization algorithms should factor this in. I wonder if real grid-scale BESS systems have this capability.
Not sure if it's discussed in the paper but apparently in Australia there have already been recorded instances of batteries charging with negative price electricity and then selling back that electricity at a still negative but closer to zero price and so profiting.
When I first heard it, it seemed wild that they couldn't hold on for the price to go back above zero, but I guess if we're talking high frequency trading it makes more sense. They might have bought and sold many times while the price is different levels of negative before switching to charging up in preparation for the later price rises.
As you know, negative electricity prices mean that someone is willing to pay you to dispose of electricity they need to generate for some reason. For example, a conventional steam-turbine-based electricity plant might prefer to just keep running for a brief period of time when demand is low, rather than subject their equipment to a power cycle, which increases their maintenance costs. There's other, dumber, examples based on stupid contracts and badly designed solar... but this example is a reasonable one that exists for good engineering reasons.
The battery provider in this circumstance is profiting from their ability to accept power when demand to dispose of electricity is particularly high. When that need goes down, they can reasonably profit by dumping that energy on someone else who is also able to dispose of the electricity. But at a lower cost. E.g. imagine an big industrial refrigerated storage facility that can consume some excess energy by supercooling their refrigerators. But they can't consume unlimited excess energy, because at some point their warehouse just gets too cold, and they don't have unlimited refrigeration capacity anyway.
So in this simplified example, the battery storage service is getting paid a lot of money to quickly absorb a lot of energy, which they then dump more slowly to the refrigerated warehouse (and similar providers) as the surplus diminishes, in anticipation of another surplus in the near future.
> and then selling back that electricity at a still negative but closer to zero price and so profiting.
How is it not better to discharge the batteries instead? I guess if you don't have that hardware option integrated into the platform maybe, but otherwise...
I don't know why it rankles me to think that generated power should be fed into a dump load just to make the storage owners extra money. Even though it's inefficient at the system level, it shouldn't be harmful releasing energy that would have been eventually dissipated as heat anyways. And yet it still just feel wasteful to me.
I had to go search my bookshelf for this one:
"There has been an increasing awareness among engineers of the last two decades that machines can perform a useful purpose in many applications, even though their characteristics do not conform to the orthodox standards of goodness. The main objective of the engineer is to make money -- to exploit economically the physical properties of materials. Economic considerations, however, do not stop at the first cost of an article, nor at the running cost, but extend to everything connected with that article in the situation in which it is to be used."
Eric R. Laithwaite, Induction Machines for special purposes
My shallow understanding is that utilities and grid operators need to manage the supply/load ratio carefully to keep the grid's operating frequency in a very narrow band, centered around 50 or 60 Hertz. If supply outstrips demand, and assuming supply can't react [quickly enough], the operating frequency starts to rise as all the rotating masses connected to the grid gain momentum from the additional power. If the operating frequency increases too much outside of design parameters that could end badly.
So one solution is to incite demand (with negative rates) for folks to ramp up their use of electricity (into e.g., a dump load resistor bank), bringing demand back in line with supply, and bringing the operating frequency back under control.
I hate the waste, agreed. But it would be irresponsible of the operator to bank that extra supply energy into the momentum of spinning things owned by the consumers just so they could pull it out later by intentionally under-supplying. E.g., an aquarium's big water pumps designed to spin only so fast or produce so much pressure might not like being operated at 110% the rated speed at random times of the day.
>I don't know why it rankles me to think that generated power should be fed into a dump load just to make the storage owners extra money. Even though it's inefficient at the system level, it shouldn't be harmful releasing energy that would have been eventually dissipated as heat anyways. And yet it still just feel wasteful to me.
This is one of those efficient market things where you need to manage the market in order that wasteful things happen sometimes... but that waste is an opportunity.
If you and your competitor are both in the business of dumping energy into heat, you're going to compete with each other for access to that money.
Then one of you is going to try to find a way to make _more_ money with that energy and find something quickly scalable with not-too-high idle overhead costs to do with that energy besides just flowing through a resistor.
Negative prices are a sign of an inefficient market or just the lag time between a changing landscape of resources and someone to utilize them.
If there's a free resource someone's going to figure out how to use it, just let it hang out for a while and the problem fixes itself.
Especially with solar energy, this is just going to be a thing. There's a certain balance where overprovisioning is cheaper than storage and so you just do that. Then you wait for industry (or consumers) to figure out how to take advantage of the intermittent cheap energy.
> Surely the power electronics or even just the control algorithms in a BESS could be slightly modified to consume power, get warm, and not transfer any current to or from the battery cells, effectively taking advantage of the BESS’s heat sink to sink excess power and sell that service.
Unless you specifically design for it (specifically, with a dummy load), the efficiency of the system is inversely proportional to its ability to do this. You need a secondary system.
The power system can connect the battery terminal to in or to out, so if you switch both on at once you effectively bypass the battery. It's called shoot-through current and is generally considered a destructive process. If you can switch on and off fast enough you could limit it to a non-destructive level, but in practice most systems will not switch fast enough. They are designed to operate with the battery load, which is at minimum ~10x higher resistance than the transistor itself. In practice it is often 100s of times higher.
That's where the efficiency comes into it. If a power system is 98% efficient (pretty normal- this does not include power lost to heat in the battery itself), that means the electronics can only burn 1/50th as much power as normally passes through the system. Worse, when you put the switch into shorted position it will try to pass 50x its rated current. You need to switch much faster -certainly more than 50x faster- and that will probably put it outside its operating region.
It is relatively easy to just have a large resistor, but it is not very well suited to use battery power systems. Batteries are very low impedance, and the power system exists to transform to a lower voltage and higher current. Resistors are cheaper when they are higher voltage, so the power system is a hacky kludge.
The overall solution is just more batteries. Oversupply is a transient problem and always will be.
> Unless you specifically design for it (specifically, with a dummy load), the efficiency of the system is inversely proportional to its ability to do this. You need a secondary system.
Many multicell BMS already have this kind of "power shedding" capability. They use it for cell balancing - to equalize voltage between cells with slightly different characteristics. This is desirable despite the power waste, because it reduces wear, increases charging efficiency and allows battery packs to last longer.
Some battery packs are also designed to be able to dump enough power into heat to be able to keep the batteries warm during extreme cold.
I work in the industry making hardware and software for large scale commercial and grid scale storage.
There are several challenges with this, safety, thermal runaway, and life cycle of the asset which has a limited amount of cycles.
Also the architecture of the system for the AC inverters and the DC side can come from very different places in the supply chain and aren’t as vertically integrated leaving you in a position where you can’t actually make this work without compromising something in the supply chain. That being said we are talking about a LOT of energy in these systems and to dissipate that much heat you’d need a load bank.
Our house have geothermal heating (heatpump conncted to 160m drilled hole, pretty common in Scandinavia). The heatpump supports having a coolant loop for cooling the house in the summer. Thus the heat pump pretty much exchanges heat from the house to the well (heating it up ever so slightly). It would certainly be possible to insert a resistive dummy load on that loop and just store that heat in the bedrock as well.
This! Or, if you don't have geothermal heating but have an electric water heater, maybe temporarily increase the temperature it goes to: maybe it's normally set to go to 65C, then when you detect that you have negative prices and your batteries are full and your water already hot, maybe heat the water to 70C and store that little bit of extra energy as heat! If you have thermostatic valves in your bathrooms, you won't even notice the difference except by the fact that your water heater now can apparently hold a little bit more water than usual :)
Yes that is the SBB in Switzerland but they do not turn the lights, instead they turn on the heaters for the track switches.
Not sure if that is all rumours.
I think they just don't understand how such a market would work. You'd trade futures on the power and if you hold the contracts to maturity you'd be on the hook for delivering/accepting the power. Everything before maturity happens on paper.
The big problem there is you have these intensely capital expensive capture plants sitting idle around 75% of the time. Also the processes may not gracefully start and stop though maybe you could smooth that out by building a huge battery bank along with the CC plant to effectively run a full duty cycle with 'free' energy. That bumps the capital costs up again though so the economics get tricky.
Couldn't the battery just do, as an example, 1 minute long charge then discharge cycles?
For example, if the electricity price is -28€/MWh (like today in Germany), and your battery efficacy is 80%, you could get paid 28€/MWh charging, then only pay back 22€ discharging, generating a 6€/MWh profit.
The wholesale energy markets don't have sub 5-minute granularity anywhere that I'm aware of. In the US, 1-hour is standard in the day-ahead markets and 5-minutes is standard for the spot markets.
There is also the problem that your battery would likely degrade fast depending on the technology.
Balancing.. thats probably 1-0.5% of the BESS capacity. The impedance of LFE cells are so high when charged pretty small amounts of energy can slosh around before a protection disconnect, over voltage for example
I am in the power generation industry and I have honestly wondered why nobody does this. I figure getting the interconnect big enough to make meaningful money is both prohibitively expensive and a lengthy delay.
I know this is grossly pedantic, but not matter what that electricity is used for, it will end up "heating up the planet a bit more". Energy is a waterfall whose base is heat.
Presumably the negatively priced energy came from solar panels, so those sun rays were going to heat the planet anyway. The same still happens with a dummy load, just with extra steps in between to convert to and from electricity.
With enough solar panels deployed, you could still argue that they change the albedo of the Earth and therefore it's temperature.
there are no circumstances under which it makes sense for energy prices to actually be negative. it is a sign of a market behaving very badly. propping up prices by intentionally wasting energy only treats the symptom not the cause.
Some years ago, I helped with battery load tests in a nuclear power station. The constant test load was just a big (~500kW) heater. We burned the battery energy for 5 hours. So it's easy possible to do such things.
I have programmed batteries / algos to do this in some European markets. This is being done right now.
The yield you could make from batteries in the UK dropped from double digits to 2% in 8months once some hedge funds figured out how to build and bid (or commission companies like my employer) batteries in the UK short term reserve market.
There are a few firms in northern Jutland and London specialised in this sort of trading.
Are/where you already trading at a second to sub-second level on the continuous intraday markets? How did you backtest your strategies then, if so? Or is backtesting, e.g. for parametric extensions of the optimization, not yet quite relevant?
Huge part of the reason why negative prices exist in the first place is separation of generation and transmission. With pay-as-clear model negligible-variable-costs generators (i.e. renewables) can bid at zero and pump more into the grid than the local segment can drain, requiring artificial balancing sinks. However, the cost of artificial sinks fall on the grid as transmission costs and are not reflected in the wholesale market.
Battery production rates are creeping up to multiple twh/year now. This is accumulating to a lot of battery sitting around in vehicles, domestic storage, grid storage, etc. Mostly the goal of these batteries of course isn't supporting the grid but some other use case. But if it's plugged in, it could potentially be available for selling power. Right now, most EVs can't contribute power. But that's something that is starting to change. Small experiments with thousands or tens of thousands of vehicles have already been done and seem to work fine. Now imagine tens of millions of vehicles being part of the grid. That's a serious amount of stand by power for absorbing excess power or dispatching power when needed.
Another interesting aspect is that as grid demand fluctuates, a lot of cables are under utilized at least some of the time. Which means there is plenty of capacity for charging batteries provided there is excess generation and cable capacity. A lot of that power currently gets discarded instead. Batteries allow better use of this excess power. And having a lot of local battery means that cable capacity can be freed up as well when needed and then recharged when demand reduces.
And then finally battery prices are coming down. With sodium ion cell production ramping up in several places, things could get quite a bit cheaper. These don't depend on scarce metals or materials. And they last quite long as well (relative to NMC).
Even without talking about batteries, I think a lot can be done using smart grids. For example, your A/C and heaters can run a little later or a little earlier, it won't have a significant effect on your room temperature, especially if your house is well insulated. It means a lot can be done just by remote controlling thermostats with minimal effect on confort. This can apply to other appliances that are not too time sensitive or can have a delayed start, like washing machines, refrigerators, dishwashers, etc... but thermostats are likely to be the biggest ones.
As with batteries, it would require some work with the utility companies, so that they can signal the best times to power on/off, and offer the appropriate financial incentives to do so. We are already doing that in some way with peak/off-peak pricing, but it doesn't help with high frequency, unplanned variations.
Nor should they. People don't want to be cycling their batteries and reducing their life. This use case would be better served by batteries that are designed for that purpose instead of being designed to be light for a vehicle.
EV can grid cooperate without adding charge/discharge cycles. You just intelligently schedule the charging.
This is how most Time-of-Use metering already works. The driver sets a minimum battery percentage to charge immediately (eg 40% range, enough to reach the local hospital), and then schedule a time when the car should be fully charged (eg 80% by 7AM). The software just Does The Right Thing, using the same prediction and bidding algorithms as stationary batteries.
The search term is "V1G" (a cheeky reference to unidirectional V2G).
> People don't want to be cycling their batteries and reducing their life.
More battery cycles just costs money. For the right price, I'd do it.
But more than that: I don't want to be stranded without power in my vehicle because someone in the electric grid made poor power management decisions and decided to offload that decision to consumers.
In general I agree with that, but I think it really would depend on the price. For at least some people it would surely be worthwhile for the right price.
Future: Cycle the power when profitable, replace the batteries when depleted enough? Batteries in cars are many times the domestic demand of a given home, at least in temperate climates.
> A lot of that power currently gets discarded instead.
How is power discarded? I would expect peaking generation to be cut back or perhaps even base load plants can reduce output. (AFAIK "base load" means they are expected to be kept operating continuously whereas "peaking" is designed to start up when needed and shut down when not.)
wind mills are weather vaned (i.e. not broken, but deliberately turned off), solar panels excess energy is curtailed (prevented from going into the grid) and usually transformed into heat on the panels or in inverters.
As for baseload. It's one of those waffly terms that's rarely specified in GW that is needed. Which as it turns out is far less than we used to have given that much of it was replaced with wind and solar over the last decade or so. The real question is how low can we go with this stuff before we need some more solutions. Some would say all the way but the consensus is that the last 5-10% might get very hard and costly.
Either way, having some peaker plants on stand by ready to spin up over the course of hours/days while batteries slowly deplete is probably a good short term compromise. Replacing spinning mass (fly wheels) with batteries seems a particularly popular and very cost effective use for batteries.
> That's a serious amount of stand by power for absorbing excess power or dispatching power when needed.
... and a serious issue should one of the few large manufacturers or remote-control dispatcher/trader companies get hacked. The outage in Spain a few weeks ago was just a small warning, probably caused by a technical malfunction. But now imagine this being used as a side track in an act of war? The first day of the Russian invasion of Ukraine was accompanied by the hack of Viasat satellites, which led to 5.800 wind turbines shutting down due to a lack of remote control capabilities [1]. Now imagine the large Chinese inverter and power bank controller vendors that often enough just white-label for other brands? That's a whole lot of a different game now.
> and a serious issue should one of the few large manufacturers or remote-control dispatcher/trader companies get hacked. The outage in Spain a few weeks ago was just a small warning, probably caused by a technical malfunction.
If everyone agrees, you can use grid frequency/phase to coordinate, and not a separate realtime communication system. Grid interactive demand/response is a proven way to manage supply and load.
When your section of the grid is stressed, supply power or abstain from charging; when your section of the grid is abundant, charge.
Coordination is useful too, of course; predictive charging is helpful, and you wouldn't get that only by monitoring the grid; you also want to know somehow that a supply or load is scheduled to be added or removed at time X, or was unexpectedly removed and will not be reconnected for some time. And the system operator would want to know about capacity in many dimensions.
Of course the Ukrainians are now much less dependent on central power generations out of necessity. It turns out that all those big power plants and electricity distribution centers make for nice drone targets. The Russians did far more damage with that than with their satellite hackery. Those 5800 wind turbines came back online and are mostly still operational.
The lesson here is that distributed power is a good thing in war time scenarios but you might want to pay attention to digital security. Central power generation becomes a tempting target.
Now the good news with Chinese stuff is that a war is not imminent and we have the benefit of hindsight and can do something about that.
I'm not at all familiar with this whole field, but why would you publish a trading strategy if it has potential? Why not sell it to a hedgefund, at least? Or is this research formally publishing what industry is already doing?
Seemingly profitable strategies are actually published all the time in finance literature.
Some also work (usually only for a short amount of time if profitable), most don’t really work at all for various technical reasons (lookahead bias, model doesn’t account for slip/trading costs or assume infinite liquidity or a portfolio too large to realistically rebalance etc.) and some again work, but have unfavorable risk-adjusted return profiles compared to simpler strategies.
Most real-world optimizations for flexible storage assets currently work across multiple markets, sometimes also with more sophisticated boundary conditions. What we show is that high-frequency trading on the continuous intraday market is relevant, especially when training for more optimal parametrized strategies.
It also seems like a sensible idea to publish details and theories about an idea, not necessarily a finished trading product though ;)
On top of what others have mentioned, this paper also sidesteps forecasting electricity prices, which is already a very complicated problem (particularly in U.S markets where we have zonal pricing) needed to build profitable battery systems that actually operate on the grid.
I've had a few chats with some folks working on battery startups, and I think the more conventional approach is to forecast prices + run an optimization to find optimal storage decisions. You could measure the system's performance by looking at how well the algorithm does when it has perfect information about prices (obviously, when you have perfect information about prices it is trivial to optimize the battery).
Our two follow-up papers are addressing exactly this (for Europe)! We are extending our high-frequency continuous intraday approach (CID) with a forecast-based day-ahead bidding stage, and subsequent CID forecast updates.
I'd also be quite interested in strategies for grid-scale BESS trading in the US' real-time markets. Do you know more about it, or could forward me to someone who would be willing to talk about it? ;)
Is the trading strategy making substantial returns (at least a few percent) on the full investment (batteries, electronics, subscriptions)? Otherwise this is only relevant for battery owners that benefit already by other means (using the battery at night for home owners, for example).
There‘s still a lot of work left to do to go from an academic prototype to live trading: real-time data, market access, SCADA, compliance & legal, security, … Also you must be a physical player that owns the battery and/or right to use it and not just do paper trading.
There is "virtual" (paper) trading in the day ahead markets in the US, but it's just for amounts of energy. You can't make a fake battery for the grid operator to optimize.
Just use EVs. EVs are primarily energy storage devices, some people get to drive them about 20 - 30 mins/day. The remaining 23 hours, it is a energy storage device. It can absorb excess power when price is negative, and can even supply power back to the grid when prices are high!
Its more than pennies[1]. By several orders of magnitude. Car batteries now last longer the rest of the car lifespan, it will be millions of miles soon.
That is a great thought and there is a lot of research in using batteries of vehicles for grid stabilization as well as arbitrage trading.
The problem, however, is that a lot of plugged vehicles are required to guarantee a meaningful amount of firm power/energy capacity and owners typically do not want to sacrifice a lot of flexibility when they rent out their batteries to provide such services.
Furthermore, as you can see from the replies, many are scared about battery degradation. Because of this and due to the fact that reverse charging is not possible for most vehicles, this schemes are often restricted to delayed charging which further limits revenue potential.
However, using arrays of old car batteries as stationary batteries is a very viable idea that, in my estimation, will lead to a significant growth in installed BESS in a couple of years.
Yes, exactly this has been proposed with "smart car chargers," along with other things you can do if the grid operator has some control over a bunch of grid-connected equipment. It hasn't taken off as far as I know, probably because that means the car battery wears out more.
The "virtual power plants" are the closest thing to this idea that is actually done in practice. That's individuals who own batteries joining some collective that then sells to the grid the ability to reduce demand a bit. Tesla did a pilot program with its Powerwalls iirc.
In Australia 5 minute spot pricing is now accessible to many residential customers via retailers like Amber electric. With volatile pricing and a large home battery subsidy from the re-elected government, batteries can quickly pay for themselves through arbitrage alone.
EMHASS is an interesting tool to perform the optimisation.
Yep, been using EMHASS for the last couple of years in the UK.
I have a large array (12.8kWp east/west split) but a low export limit of 5kW. In the winter it's charging overnight at 7p per kWh (Intelligent Octopus Go) and then using that stored energy during the day to avoid importing at peak rates, and in the summer it makes sure to discharge most of the battery before the peak generation hours so that battery charges from power which would otherwise be curtailed (discharge to minimise import on my SolarEdge system, but charge from clipped power would also work).
Similarly in Europe; spot market with a big single pay-as-cleared spot auction for every quarter-hour, and then a continuous auction for the same periods closer to delivery, similar to the normal stock market. Millions of residential devices are traded there right now
> “Our method is able to solve over 24 million optimization problems in less than 90 minutes.”
that line is doing heavy lifting. sounds insane until you look closer
they batch out embarrassingly parallel, lowdimensional problems
no live latencies, no network I/O, no grid API jitter. just hammering a static dataset in memory. real markets stall, disconnect, price slippage, queue delays. none of that here. so yeah, 24 million looks cool in the abstract, but under the hood it's just cleanroom compute; feels like they optimised the benchmark more than the actual system
Assuming you add all the annoying details that algo trade execution brings, the algorithm still provides the answer on which position to take within a few microseconds, which is what you want if you trade in a limit order book.
true, you want microsecond decisions at the core, no doubt ; but that’s only half the game. an ideal action in clean memory isn’t the same once it hits fragmented liquidity, stale quotes, partial fills. if the algo doesn’t account for execution drift or book pressure post-placement, the microsecond edge fades fast. so yeah, fast compute’s necessary but not sufficient without modelling the messy tail end too
Imagine software that could run on EVs, Powerwall-type batteries, computers/tablets/smartphones, and so on, which would automatically charge and discharge for passive income. Essentially algorithmic trading, but with power instead of stock. You'd just have to configure any necessary time ranges and charge percentages, e.g. maybe your EV needs to be at 25% by 8am and again by 5pm on weekdays in order to make your daily commute.
Maybe some EVs will start to come with built-in crypto miners to burn negatively priced power when the battery is at capacity. Maybe Lyft/Uber and Waymo/Cruise will take advantage of it by increasing and lowering rates based on the price of power (if they don't already).
Sorry to be the bearer of bad news, but this isn't novel and something that's been talked about for a long time. The industry term I've most heard is "prices to devices". You of course need retail to participate more in the wholesale markets, but there are a lot of barriers - some technological, some regulatory. Some companies did this in ERCOT, but there was a big backlash when customers got $20k bills after Winter Storm Uri as they didn't understand what they were signing up for.
The FERC passed Order 2222 which is a bigger step in that direction by forcing the regional wholesale markets to allow aggregators to aggregate up the smaller stuff that is normally considered noise.
> Some companies did this in ERCOT, but there was a big backlash when customers got $20k bills after Winter Storm Uri as they didn't understand what they were signing up for.
It would be a bit weird but you could have your home supply at a fixed-ish rate and your EV on a separate meter riding the raw market.
And while not making money, there has been a lot of talk around Virtual Power Plants, that is unifying the larger demand devices to help stabilize the grid in times of peak demand.
This type of service is becoming increasingly prevalent among European energy suppliers for their residential customers. Beyond providing a revenue stream for consumers this model aggregates distributed energy resources (home batteries, EV's, PV systems) into a one virtual power plant. This enables the storage of surplus energy generated during solar peaks and dispatch back to the grid during periods of high demand. I find it a fascinating domain to work in!
I wrote a book on this in 2020 and was already somewhat late to the party, as people were running actual pilot programs a decade earlier!
Also large industrial consumers have been participating in similar approaches for decades. See the crazy clever trading schemes that Enron used to do fraud and drive up prices.
I run predbat (https://springfall2008.github.io/batpred/) to achieve something like this with my Home Assistant install to manage my home battery. It can also manage EV charging but I haven't needed to do that yet due to how my tariff works. (Very cheap fixed period overnight).
Thats what bi-directional charging is for and its already becoming political to force the industry to support this.
And we already have energy provider which provide a tarif for exactly this.
The only idea i hate is the mentioning of crypto. Not only is it waste, it converts the energy in heat which needs to get disipated and potentially wastes even more energy to get this heat away from the current location (ac).
A world where individuals are incentivized to use some wasted space to place low maintenance automated trading batteries to make a little money on the side seems like it'd be an interesting solution to the renewable energy storage problem. Put a few units in otherwise useless locations like on roofs or in between highways and make some cash, sounds like a decent investment.
Would there be any expected problems in doing such distributed power storage on a very large scale around the grid that you'd have to account for? Perhaps issues with synchronization, power flow or the possibility of large scale drops in avaialble stored power at times?
Outside of the grid level, maintenance, security, and safety of batteries is important. Many Americans have a garage to securely store their $30k+ car, bike/etc - similar security may be necessary (literally or for the feels) for expensive batteries.
pretty sure nodal energy market trading, duck curve arbitrage is the whole profit play for the tesla power wall/autobidder and the Base Power startup coming out of ERCOT. There is definitely a land rush from energy firms and GIS guys to front run land purchasing near solar buildouts and to build as much grid scale battery storage as possible. The percent swings can be huge. Also a huge rush for this in the Northeast now that the offshore wind contracts have been cancelled.
> seems dumb to have electricity needing to be wasted when there is seawater to desalinate
That's a much more complicated problem. On an energy market, you have only one price to look at, and the battery operator can always buy, sell, or hold energy. The article here talks about optimizing this problem at 5-minute to several-hour intervals.
If you drop excess power into desalination, however, now you have two prices to worry about: energy and water. I also doubt we have 5-minute spot markets for water, so the operator must probably commit to some medium-term water delivery regardless of price.
This means that a desalinating firm takes on much more risk. This might still be profitable, but it's a long-term play based on a deep model of expected energy prices (i.e. knowing that energy is "always" almost free at noon in summer) rather than short-term time-shifting.
Desal plants are also extraordinarily expensive and need to operate at very high 'capacity factors' in order to payoff the capital investment that was required to build them. Operating for a a few hours every day because your operating costs are low/negative only works if you don't have a hugely expensive piece of infrastructure depreciating as you wait for those prices to come down.
Industrial processes like desalination tend to call for some optimal amount of near 24-7 utilization (barring maintenance and such) for capex reasons and efficiency. You want to use it as much as possible to get the most bang for your buck. The entire reason why there are these excess power periods is because we cannot predict accurately how much power we would really need.
Desalination plants really don’t like being ‘throttled’, and are quite capital intensive. Stopping production for any length of time can even destroy the plant, if not done very carefully. Similar for geothermal, though the specific details are different.
Even free power would likely not be worth using if it was sporadic, and it’s extremely energy intensive. So that really is saying something.
That kind of association is why you can buy futures in just about every crop except onions. The fact that you can structure fraud around a specific kind of future doesn't make them special or a bad idea. One major fraud should not bias us forever.
That's what the infamous company Enron was running: simultaneous energy futures brokerage and market making. Their market making part blew up and they hid it from everyone for a while until the rest of the company did. It was a massive scandal. It was pretty similar to what FTX was doing just with energy instead of crypto.
If anyone is interested in Batteries, I highly recommend following NAATBATT.org on LinkedIn and joining their newsletter.
Obama set this org up as a senator to help bring lithium ion batteries supply chain to the US and it since evolved into the trade association for all things batteries.
Jesus! Why finance people are so hell-bent in extracting rent from every single thing, pervert it, make sure the incentives are all pointed to the shortest run while socializing all the costs to the rest of us?
If this was implemented in full scale, then it would .. stabilize energy prices towards the mean, and make the energy transmission more stable and resilent in general. I'd happily pay some overhead for that, as it also mean the network can accept more cheap energy sourced from wind and solar.
This is the kind of thing that is funding the massive expansion in renewable energy build-out at the moment. The whole reason there's an energy transition happening is that solar and wind and batteries are cheap enough you can make a lot of money building them, and that'll remain true until basically the whole grid is renewable (finally kicking out the expensive gas turbines), and the average price drops to reach cost of supply.
(And the kind of optimisation that happens with this kind of paper is really in the margins stuff. It generally helps the predictive power of the grid, and usually doesn't make much money once more than one group starts doing it, since it's pretty cheap to run and the margins shrink quickly)
No need of speculators. A utility purchases cheap off peak power and extremely costly peak power and pass the costs to the consumer. If the utility pays homeowner with BESS something comparable to peak power rate, they can recover their investment quickly.
Except this is exactly what you want to happen. The reason electricity prices range is because of supply and demand. Batteries help smooth out the supply curve.
In the short term adding more batteries may allow someone to generate income using this strategy but long term what it will do is push electricity prices down and prevent power generation from being overwhelmed. As the battery "market" gets crowded profit margins will fall and everything will reach an equilibrium.
This is a great demonstration of how capitalism works and why it's beneficial.
Smart People with money are willing to give it to people if they can make them more money.
Why would anyone give them money if they were just going to throw up their hands and go ‘well, nothing we can do I guess!’.
There is of course the risk that the money gets burned instead of more money getting made, which is the risk in risk/reward.
Rent seeking type behavior tends to happen when there are no obvious ‘green field’ type endeavors to invest in. Or when risk appetites are trending negative.
Note - many of those people with money that want to use it to make more money are retirees, pension funds, etc.
Capitalism is the only form of communism that works on longer time frames. Periodic resets are still needed, otherwise the monopolist becomes the ruler.
Negative prices are mostly caused by unreasonable terms towards generation plants. E.g.: requiring the grid to take every kWh generated and paying a fixed price over a 20-year term. This of course encourages capacity to be built with no flexibility. Why not dump your solar power into the grid? You're getting paid, the state guarantees for that... the negative prices are someone's elses problem.
Energy overproduction is going to become a serious viability problem for baseload generators, which in time will significantly affect grid reliability. Rolling blackouts will become the norm unless we figure out a serious scalable solution to this.
> However, because it is physically not possible to charge and discharge the battery at the same time, such trades have to be prevented.
The authors are observing that, if electricity prices are negative and your battery is not perfectly efficient, then you would like to charge and discharge simultaneously to get paid for wasting energy, but you can’t.
This is a silly limitation. Surely the power electronics or even just the control algorithms in a BESS could be slightly modified to consume power, get warm, and not transfer any current to or from the battery cells, effectively taking advantage of the BESS’s heat sink to sink excess power and sell that service.
More seriously, in a world with occasional negative prices, you would want your battery to be able discharge itself, without exporting power, in a controlled and power-limited manner so as to avoid overheating. And the optimization algorithms should factor this in. I wonder if real grid-scale BESS systems have this capability.
Not sure if it's discussed in the paper but apparently in Australia there have already been recorded instances of batteries charging with negative price electricity and then selling back that electricity at a still negative but closer to zero price and so profiting.
When I first heard it, it seemed wild that they couldn't hold on for the price to go back above zero, but I guess if we're talking high frequency trading it makes more sense. They might have bought and sold many times while the price is different levels of negative before switching to charging up in preparation for the later price rises.
And the round trip inefficiency helps too.
That's not as ridiculous as it sounds!
As you know, negative electricity prices mean that someone is willing to pay you to dispose of electricity they need to generate for some reason. For example, a conventional steam-turbine-based electricity plant might prefer to just keep running for a brief period of time when demand is low, rather than subject their equipment to a power cycle, which increases their maintenance costs. There's other, dumber, examples based on stupid contracts and badly designed solar... but this example is a reasonable one that exists for good engineering reasons.
The battery provider in this circumstance is profiting from their ability to accept power when demand to dispose of electricity is particularly high. When that need goes down, they can reasonably profit by dumping that energy on someone else who is also able to dispose of the electricity. But at a lower cost. E.g. imagine an big industrial refrigerated storage facility that can consume some excess energy by supercooling their refrigerators. But they can't consume unlimited excess energy, because at some point their warehouse just gets too cold, and they don't have unlimited refrigeration capacity anyway.
So in this simplified example, the battery storage service is getting paid a lot of money to quickly absorb a lot of energy, which they then dump more slowly to the refrigerated warehouse (and similar providers) as the surplus diminishes, in anticipation of another surplus in the near future.
25 replies →
> and then selling back that electricity at a still negative but closer to zero price and so profiting.
How is it not better to discharge the batteries instead? I guess if you don't have that hardware option integrated into the platform maybe, but otherwise...
I don't know why it rankles me to think that generated power should be fed into a dump load just to make the storage owners extra money. Even though it's inefficient at the system level, it shouldn't be harmful releasing energy that would have been eventually dissipated as heat anyways. And yet it still just feel wasteful to me.
I had to go search my bookshelf for this one:
Eric R. Laithwaite, Induction Machines for special purposes
My shallow understanding is that utilities and grid operators need to manage the supply/load ratio carefully to keep the grid's operating frequency in a very narrow band, centered around 50 or 60 Hertz. If supply outstrips demand, and assuming supply can't react [quickly enough], the operating frequency starts to rise as all the rotating masses connected to the grid gain momentum from the additional power. If the operating frequency increases too much outside of design parameters that could end badly.
So one solution is to incite demand (with negative rates) for folks to ramp up their use of electricity (into e.g., a dump load resistor bank), bringing demand back in line with supply, and bringing the operating frequency back under control.
I hate the waste, agreed. But it would be irresponsible of the operator to bank that extra supply energy into the momentum of spinning things owned by the consumers just so they could pull it out later by intentionally under-supplying. E.g., an aquarium's big water pumps designed to spin only so fast or produce so much pressure might not like being operated at 110% the rated speed at random times of the day.
related links:
https://fnetpublic.utk.edu/frequencygauge.html (you can watch the grid frequency fluctuate in real-time, here!)
https://en.wikipedia.org/wiki/Stuxnet
https://en.wikipedia.org/wiki/Utility_frequency
2 replies →
>I don't know why it rankles me to think that generated power should be fed into a dump load just to make the storage owners extra money. Even though it's inefficient at the system level, it shouldn't be harmful releasing energy that would have been eventually dissipated as heat anyways. And yet it still just feel wasteful to me.
This is one of those efficient market things where you need to manage the market in order that wasteful things happen sometimes... but that waste is an opportunity.
If you and your competitor are both in the business of dumping energy into heat, you're going to compete with each other for access to that money.
Then one of you is going to try to find a way to make _more_ money with that energy and find something quickly scalable with not-too-high idle overhead costs to do with that energy besides just flowing through a resistor.
Negative prices are a sign of an inefficient market or just the lag time between a changing landscape of resources and someone to utilize them.
If there's a free resource someone's going to figure out how to use it, just let it hang out for a while and the problem fixes itself.
Especially with solar energy, this is just going to be a thing. There's a certain balance where overprovisioning is cheaper than storage and so you just do that. Then you wait for industry (or consumers) to figure out how to take advantage of the intermittent cheap energy.
2 replies →
> Surely the power electronics or even just the control algorithms in a BESS could be slightly modified to consume power, get warm, and not transfer any current to or from the battery cells, effectively taking advantage of the BESS’s heat sink to sink excess power and sell that service.
Unless you specifically design for it (specifically, with a dummy load), the efficiency of the system is inversely proportional to its ability to do this. You need a secondary system.
The power system can connect the battery terminal to in or to out, so if you switch both on at once you effectively bypass the battery. It's called shoot-through current and is generally considered a destructive process. If you can switch on and off fast enough you could limit it to a non-destructive level, but in practice most systems will not switch fast enough. They are designed to operate with the battery load, which is at minimum ~10x higher resistance than the transistor itself. In practice it is often 100s of times higher.
That's where the efficiency comes into it. If a power system is 98% efficient (pretty normal- this does not include power lost to heat in the battery itself), that means the electronics can only burn 1/50th as much power as normally passes through the system. Worse, when you put the switch into shorted position it will try to pass 50x its rated current. You need to switch much faster -certainly more than 50x faster- and that will probably put it outside its operating region.
It is relatively easy to just have a large resistor, but it is not very well suited to use battery power systems. Batteries are very low impedance, and the power system exists to transform to a lower voltage and higher current. Resistors are cheaper when they are higher voltage, so the power system is a hacky kludge.
The overall solution is just more batteries. Oversupply is a transient problem and always will be.
> Unless you specifically design for it (specifically, with a dummy load), the efficiency of the system is inversely proportional to its ability to do this. You need a secondary system.
Many multicell BMS already have this kind of "power shedding" capability. They use it for cell balancing - to equalize voltage between cells with slightly different characteristics. This is desirable despite the power waste, because it reduces wear, increases charging efficiency and allows battery packs to last longer.
Some battery packs are also designed to be able to dump enough power into heat to be able to keep the batteries warm during extreme cold.
1 reply →
I work in the industry making hardware and software for large scale commercial and grid scale storage.
There are several challenges with this, safety, thermal runaway, and life cycle of the asset which has a limited amount of cycles.
Also the architecture of the system for the AC inverters and the DC side can come from very different places in the supply chain and aren’t as vertically integrated leaving you in a position where you can’t actually make this work without compromising something in the supply chain. That being said we are talking about a LOT of energy in these systems and to dissipate that much heat you’d need a load bank.
Our house have geothermal heating (heatpump conncted to 160m drilled hole, pretty common in Scandinavia). The heatpump supports having a coolant loop for cooling the house in the summer. Thus the heat pump pretty much exchanges heat from the house to the well (heating it up ever so slightly). It would certainly be possible to insert a resistive dummy load on that loop and just store that heat in the bedrock as well.
This! Or, if you don't have geothermal heating but have an electric water heater, maybe temporarily increase the temperature it goes to: maybe it's normally set to go to 65C, then when you detect that you have negative prices and your batteries are full and your water already hot, maybe heat the water to 70C and store that little bit of extra energy as heat! If you have thermostatic valves in your bathrooms, you won't even notice the difference except by the fact that your water heater now can apparently hold a little bit more water than usual :)
2 replies →
I know a country, the national train company turns lots of outside lights on (daytime), if the price is negative.
Yes that is the SBB in Switzerland but they do not turn the lights, instead they turn on the heaters for the track switches. Not sure if that is all rumours.
Great for the local insect population.
I think they just don't understand how such a market would work. You'd trade futures on the power and if you hold the contracts to maturity you'd be on the hook for delivering/accepting the power. Everything before maturity happens on paper.
Wouldn't that discharge the battery and hence contribute to battery wear, by wasting a charge cycle?
If you have free energy the obvious thing to use it for is carbon capture.
The big problem there is you have these intensely capital expensive capture plants sitting idle around 75% of the time. Also the processes may not gracefully start and stop though maybe you could smooth that out by building a huge battery bank along with the CC plant to effectively run a full duty cycle with 'free' energy. That bumps the capital costs up again though so the economics get tricky.
2 replies →
The real low hanging fruit is energy use you were going to do earlier/later anyway but where timing isn't important.
Heating water, cooling water, pumping water, charging batteries, running power hungry machines.
It's half century old tech and usually the only thing missing is a financial incentive to do so.
I wonder if desalination would be another good use. But, yeah, it is probably just a matter of how fast the processes can absorb extra power.
6 replies →
Couldn't the battery just do, as an example, 1 minute long charge then discharge cycles?
For example, if the electricity price is -28€/MWh (like today in Germany), and your battery efficacy is 80%, you could get paid 28€/MWh charging, then only pay back 22€ discharging, generating a 6€/MWh profit.
The wholesale energy markets don't have sub 5-minute granularity anywhere that I'm aware of. In the US, 1-hour is standard in the day-ahead markets and 5-minutes is standard for the spot markets.
There is also the problem that your battery would likely degrade fast depending on the technology.
Balancing.. thats probably 1-0.5% of the BESS capacity. The impedance of LFE cells are so high when charged pretty small amounts of energy can slosh around before a protection disconnect, over voltage for example
Supercapacitors ready to soak up power to charge batteries whenever rates stabilize.
Supercapacitors are much more expensive then batteries with much shorter lifespans to boot (years, versus decades).
Find area near shore, stick two big electrodes in water a mile apart, energize circuit when price is negative, profit!
I am in the power generation industry and I have honestly wondered why nobody does this. I figure getting the interconnect big enough to make meaningful money is both prohibitively expensive and a lengthy delay.
1 reply →
The solution is actually what's called a "dummy load". Get paid to waste energy and heat up the planet a tiny bit more, gotta love it.
I know this is grossly pedantic, but not matter what that electricity is used for, it will end up "heating up the planet a bit more". Energy is a waterfall whose base is heat.
4 replies →
Can’t we power a big laser and point it at space or something instead? Anyone got a dumber idea?
8 replies →
Presumably the negatively priced energy came from solar panels, so those sun rays were going to heat the planet anyway. The same still happens with a dummy load, just with extra steps in between to convert to and from electricity.
With enough solar panels deployed, you could still argue that they change the albedo of the Earth and therefore it's temperature.
4 replies →
>"dummy load"
You mean crypto miner.
9 replies →
Could we use it for some kind of carbon capture process?
2 replies →
Isn't this the basic description of what a gravity battery should provide?
7 replies →
there are no circumstances under which it makes sense for energy prices to actually be negative. it is a sign of a market behaving very badly. propping up prices by intentionally wasting energy only treats the symptom not the cause.
It’s only ever brief and essentially a penalty to power generators for not turning off the generation in a reasonable time
1 reply →
Good luck trying to get rid of such an amount of heat anyway near your batteries.
Some years ago, I helped with battery load tests in a nuclear power station. The constant test load was just a big (~500kW) heater. We burned the battery energy for 5 hours. So it's easy possible to do such things.
6 replies →
Well in times of negative energy prices wouldn't it even be good if the air conditioning ran at full capacity?
17 replies →
I have programmed batteries / algos to do this in some European markets. This is being done right now.
The yield you could make from batteries in the UK dropped from double digits to 2% in 8months once some hedge funds figured out how to build and bid (or commission companies like my employer) batteries in the UK short term reserve market.
There are a few firms in northern Jutland and London specialised in this sort of trading.
Are/where you already trading at a second to sub-second level on the continuous intraday markets? How did you backtest your strategies then, if so? Or is backtesting, e.g. for parametric extensions of the optimization, not yet quite relevant?
At least the grid is better off and it’s not all approximately zero sum…
That's a very bold statement.
Huge part of the reason why negative prices exist in the first place is separation of generation and transmission. With pay-as-clear model negligible-variable-costs generators (i.e. renewables) can bid at zero and pump more into the grid than the local segment can drain, requiring artificial balancing sinks. However, the cost of artificial sinks fall on the grid as transmission costs and are not reflected in the wholesale market.
I'd make the same argument for the other markets they operate in.
1 reply →
Sometimes the efficient market happens to you (and that's good)
That’s yield over what base?
Battery production rates are creeping up to multiple twh/year now. This is accumulating to a lot of battery sitting around in vehicles, domestic storage, grid storage, etc. Mostly the goal of these batteries of course isn't supporting the grid but some other use case. But if it's plugged in, it could potentially be available for selling power. Right now, most EVs can't contribute power. But that's something that is starting to change. Small experiments with thousands or tens of thousands of vehicles have already been done and seem to work fine. Now imagine tens of millions of vehicles being part of the grid. That's a serious amount of stand by power for absorbing excess power or dispatching power when needed.
Another interesting aspect is that as grid demand fluctuates, a lot of cables are under utilized at least some of the time. Which means there is plenty of capacity for charging batteries provided there is excess generation and cable capacity. A lot of that power currently gets discarded instead. Batteries allow better use of this excess power. And having a lot of local battery means that cable capacity can be freed up as well when needed and then recharged when demand reduces.
And then finally battery prices are coming down. With sodium ion cell production ramping up in several places, things could get quite a bit cheaper. These don't depend on scarce metals or materials. And they last quite long as well (relative to NMC).
Even without talking about batteries, I think a lot can be done using smart grids. For example, your A/C and heaters can run a little later or a little earlier, it won't have a significant effect on your room temperature, especially if your house is well insulated. It means a lot can be done just by remote controlling thermostats with minimal effect on confort. This can apply to other appliances that are not too time sensitive or can have a delayed start, like washing machines, refrigerators, dishwashers, etc... but thermostats are likely to be the biggest ones.
As with batteries, it would require some work with the utility companies, so that they can signal the best times to power on/off, and offer the appropriate financial incentives to do so. We are already doing that in some way with peak/off-peak pricing, but it doesn't help with high frequency, unplanned variations.
"Right now, most EVs can't contribute power"
Nor should they. People don't want to be cycling their batteries and reducing their life. This use case would be better served by batteries that are designed for that purpose instead of being designed to be light for a vehicle.
EV can grid cooperate without adding charge/discharge cycles. You just intelligently schedule the charging.
This is how most Time-of-Use metering already works. The driver sets a minimum battery percentage to charge immediately (eg 40% range, enough to reach the local hospital), and then schedule a time when the car should be fully charged (eg 80% by 7AM). The software just Does The Right Thing, using the same prediction and bidding algorithms as stationary batteries.
The search term is "V1G" (a cheeky reference to unidirectional V2G).
1 reply →
> People don't want to be cycling their batteries and reducing their life.
More battery cycles just costs money. For the right price, I'd do it.
But more than that: I don't want to be stranded without power in my vehicle because someone in the electric grid made poor power management decisions and decided to offload that decision to consumers.
3 replies →
In general I agree with that, but I think it really would depend on the price. For at least some people it would surely be worthwhile for the right price.
2 replies →
Future: Cycle the power when profitable, replace the batteries when depleted enough? Batteries in cars are many times the domestic demand of a given home, at least in temperate climates.
They are almost the same batteries. Different form factor but same thing. They are rolled from the same lines.
1 reply →
> A lot of that power currently gets discarded instead.
How is power discarded? I would expect peaking generation to be cut back or perhaps even base load plants can reduce output. (AFAIK "base load" means they are expected to be kept operating continuously whereas "peaking" is designed to start up when needed and shut down when not.)
wind mills are weather vaned (i.e. not broken, but deliberately turned off), solar panels excess energy is curtailed (prevented from going into the grid) and usually transformed into heat on the panels or in inverters.
As for baseload. It's one of those waffly terms that's rarely specified in GW that is needed. Which as it turns out is far less than we used to have given that much of it was replaced with wind and solar over the last decade or so. The real question is how low can we go with this stuff before we need some more solutions. Some would say all the way but the consensus is that the last 5-10% might get very hard and costly.
Either way, having some peaker plants on stand by ready to spin up over the course of hours/days while batteries slowly deplete is probably a good short term compromise. Replacing spinning mass (fly wheels) with batteries seems a particularly popular and very cost effective use for batteries.
> How is power discarded
It isn’t, not at scale in any traditional sense.
> That's a serious amount of stand by power for absorbing excess power or dispatching power when needed.
... and a serious issue should one of the few large manufacturers or remote-control dispatcher/trader companies get hacked. The outage in Spain a few weeks ago was just a small warning, probably caused by a technical malfunction. But now imagine this being used as a side track in an act of war? The first day of the Russian invasion of Ukraine was accompanied by the hack of Viasat satellites, which led to 5.800 wind turbines shutting down due to a lack of remote control capabilities [1]. Now imagine the large Chinese inverter and power bank controller vendors that often enough just white-label for other brands? That's a whole lot of a different game now.
[1] https://en.wikipedia.org/wiki/Viasat_hack
> and a serious issue should one of the few large manufacturers or remote-control dispatcher/trader companies get hacked. The outage in Spain a few weeks ago was just a small warning, probably caused by a technical malfunction.
If everyone agrees, you can use grid frequency/phase to coordinate, and not a separate realtime communication system. Grid interactive demand/response is a proven way to manage supply and load.
When your section of the grid is stressed, supply power or abstain from charging; when your section of the grid is abundant, charge.
Coordination is useful too, of course; predictive charging is helpful, and you wouldn't get that only by monitoring the grid; you also want to know somehow that a supply or load is scheduled to be added or removed at time X, or was unexpectedly removed and will not be reconnected for some time. And the system operator would want to know about capacity in many dimensions.
3 replies →
Of course the Ukrainians are now much less dependent on central power generations out of necessity. It turns out that all those big power plants and electricity distribution centers make for nice drone targets. The Russians did far more damage with that than with their satellite hackery. Those 5800 wind turbines came back online and are mostly still operational.
The lesson here is that distributed power is a good thing in war time scenarios but you might want to pay attention to digital security. Central power generation becomes a tempting target.
Now the good news with Chinese stuff is that a war is not imminent and we have the benefit of hindsight and can do something about that.
1 reply →
I'm not at all familiar with this whole field, but why would you publish a trading strategy if it has potential? Why not sell it to a hedgefund, at least? Or is this research formally publishing what industry is already doing?
Seemingly profitable strategies are actually published all the time in finance literature.
Some also work (usually only for a short amount of time if profitable), most don’t really work at all for various technical reasons (lookahead bias, model doesn’t account for slip/trading costs or assume infinite liquidity or a portfolio too large to realistically rebalance etc.) and some again work, but have unfavorable risk-adjusted return profiles compared to simpler strategies.
Or in the case of batteries: Requires a whole lot of hardware to be bought to get a reasonable economy of scale.
1 reply →
Most real-world optimizations for flexible storage assets currently work across multiple markets, sometimes also with more sophisticated boundary conditions. What we show is that high-frequency trading on the continuous intraday market is relevant, especially when training for more optimal parametrized strategies.
It also seems like a sensible idea to publish details and theories about an idea, not necessarily a finished trading product though ;)
1) It doesn't actually work
2) It's worth more to your career
3) Selling a strategy to a hedge fund as a stranger is, unless they hire you, probably a good few months of umming and ahhing and paperwork.
On top of what others have mentioned, this paper also sidesteps forecasting electricity prices, which is already a very complicated problem (particularly in U.S markets where we have zonal pricing) needed to build profitable battery systems that actually operate on the grid.
I've had a few chats with some folks working on battery startups, and I think the more conventional approach is to forecast prices + run an optimization to find optimal storage decisions. You could measure the system's performance by looking at how well the algorithm does when it has perfect information about prices (obviously, when you have perfect information about prices it is trivial to optimize the battery).
Our two follow-up papers are addressing exactly this (for Europe)! We are extending our high-frequency continuous intraday approach (CID) with a forecast-based day-ahead bidding stage, and subsequent CID forecast updates.
I'd also be quite interested in strategies for grid-scale BESS trading in the US' real-time markets. Do you know more about it, or could forward me to someone who would be willing to talk about it? ;)
2 replies →
Is the trading strategy making substantial returns (at least a few percent) on the full investment (batteries, electronics, subscriptions)? Otherwise this is only relevant for battery owners that benefit already by other means (using the battery at night for home owners, for example).
There‘s still a lot of work left to do to go from an academic prototype to live trading: real-time data, market access, SCADA, compliance & legal, security, … Also you must be a physical player that owns the battery and/or right to use it and not just do paper trading.
There is "virtual" (paper) trading in the day ahead markets in the US, but it's just for amounts of energy. You can't make a fake battery for the grid operator to optimize.
This isn't a trade "just on paper". You need real hardware integrated into the grid.
The authors are from universities. Publicly funded.
And its not like they can just do that and get rich. Its particular for/with battery storage systems.
Basically making battery storage systems more interesting for investors to invest into.
Thank you all for the interest in our paper, it is cool to see that people are interested in the topic!
Just use EVs. EVs are primarily energy storage devices, some people get to drive them about 20 - 30 mins/day. The remaining 23 hours, it is a energy storage device. It can absorb excess power when price is negative, and can even supply power back to the grid when prices are high!
Sounds like deprecating a >10k battery pack on a >30k vehicle and reducing your max range with power cycles to earn pennies.
Its more than pennies[1]. By several orders of magnitude. Car batteries now last longer the rest of the car lifespan, it will be millions of miles soon.
Tesla Electric customers report making as much as $150 a day https://electrek.co/2023/07/05/tesla-electric-customers-repo...
4 replies →
That is a great thought and there is a lot of research in using batteries of vehicles for grid stabilization as well as arbitrage trading.
The problem, however, is that a lot of plugged vehicles are required to guarantee a meaningful amount of firm power/energy capacity and owners typically do not want to sacrifice a lot of flexibility when they rent out their batteries to provide such services.
Furthermore, as you can see from the replies, many are scared about battery degradation. Because of this and due to the fact that reverse charging is not possible for most vehicles, this schemes are often restricted to delayed charging which further limits revenue potential.
However, using arrays of old car batteries as stationary batteries is a very viable idea that, in my estimation, will lead to a significant growth in installed BESS in a couple of years.
Yes, exactly this has been proposed with "smart car chargers," along with other things you can do if the grid operator has some control over a bunch of grid-connected equipment. It hasn't taken off as far as I know, probably because that means the car battery wears out more.
The "virtual power plants" are the closest thing to this idea that is actually done in practice. That's individuals who own batteries joining some collective that then sells to the grid the ability to reduce demand a bit. Tesla did a pilot program with its Powerwalls iirc.
This seems like building more batteries (just, with extra hardware).
An EV could be good for this sort of thing, but I guess it would have to sit around at less than 100% charge, to have the capacity.
We almost never charge our EV to 100%, to not degrade the battery faster.
1 reply →
In Australia 5 minute spot pricing is now accessible to many residential customers via retailers like Amber electric. With volatile pricing and a large home battery subsidy from the re-elected government, batteries can quickly pay for themselves through arbitrage alone.
EMHASS is an interesting tool to perform the optimisation.
Yep, been using EMHASS for the last couple of years in the UK.
I have a large array (12.8kWp east/west split) but a low export limit of 5kW. In the winter it's charging overnight at 7p per kWh (Intelligent Octopus Go) and then using that stored energy during the day to avoid importing at peak rates, and in the summer it makes sure to discharge most of the battery before the peak generation hours so that battery charges from power which would otherwise be curtailed (discharge to minimise import on my SolarEdge system, but charge from clipped power would also work).
Similarly in Europe; spot market with a big single pay-as-cleared spot auction for every quarter-hour, and then a continuous auction for the same periods closer to delivery, similar to the normal stock market. Millions of residential devices are traded there right now
> “Our method is able to solve over 24 million optimization problems in less than 90 minutes.”
that line is doing heavy lifting. sounds insane until you look closer they batch out embarrassingly parallel, lowdimensional problems no live latencies, no network I/O, no grid API jitter. just hammering a static dataset in memory. real markets stall, disconnect, price slippage, queue delays. none of that here. so yeah, 24 million looks cool in the abstract, but under the hood it's just cleanroom compute; feels like they optimised the benchmark more than the actual system
Assuming you add all the annoying details that algo trade execution brings, the algorithm still provides the answer on which position to take within a few microseconds, which is what you want if you trade in a limit order book.
true, you want microsecond decisions at the core, no doubt ; but that’s only half the game. an ideal action in clean memory isn’t the same once it hits fragmented liquidity, stale quotes, partial fills. if the algo doesn’t account for execution drift or book pressure post-placement, the microsecond edge fades fast. so yeah, fast compute’s necessary but not sufficient without modelling the messy tail end too
For those of you suggesting we use the extra energy for $pet_topic I suggest reading https://www.moderndescartes.com/essays/factobattery/
Tldr: most applications of free energy have capital costs that far outweigh the free energy harvest potential.
This sounds similar to something I suggested at one point: https://news.ycombinator.com/item?id=38669706
Imagine software that could run on EVs, Powerwall-type batteries, computers/tablets/smartphones, and so on, which would automatically charge and discharge for passive income. Essentially algorithmic trading, but with power instead of stock. You'd just have to configure any necessary time ranges and charge percentages, e.g. maybe your EV needs to be at 25% by 8am and again by 5pm on weekdays in order to make your daily commute.
Maybe some EVs will start to come with built-in crypto miners to burn negatively priced power when the battery is at capacity. Maybe Lyft/Uber and Waymo/Cruise will take advantage of it by increasing and lowering rates based on the price of power (if they don't already).
Sorry to be the bearer of bad news, but this isn't novel and something that's been talked about for a long time. The industry term I've most heard is "prices to devices". You of course need retail to participate more in the wholesale markets, but there are a lot of barriers - some technological, some regulatory. Some companies did this in ERCOT, but there was a big backlash when customers got $20k bills after Winter Storm Uri as they didn't understand what they were signing up for.
The FERC passed Order 2222 which is a bigger step in that direction by forcing the regional wholesale markets to allow aggregators to aggregate up the smaller stuff that is normally considered noise.
> Some companies did this in ERCOT, but there was a big backlash when customers got $20k bills after Winter Storm Uri as they didn't understand what they were signing up for.
It would be a bit weird but you could have your home supply at a fixed-ish rate and your EV on a separate meter riding the raw market.
If you can prevent too much cheating.
And while not making money, there has been a lot of talk around Virtual Power Plants, that is unifying the larger demand devices to help stabilize the grid in times of peak demand.
5 replies →
Interesting, thanks. That doesn't sound like bad news at all.
1 reply →
This type of service is becoming increasingly prevalent among European energy suppliers for their residential customers. Beyond providing a revenue stream for consumers this model aggregates distributed energy resources (home batteries, EV's, PV systems) into a one virtual power plant. This enables the storage of surplus energy generated during solar peaks and dispatch back to the grid during periods of high demand. I find it a fascinating domain to work in!
I wrote a book on this in 2020 and was already somewhat late to the party, as people were running actual pilot programs a decade earlier!
Also large industrial consumers have been participating in similar approaches for decades. See the crazy clever trading schemes that Enron used to do fraud and drive up prices.
I run predbat (https://springfall2008.github.io/batpred/) to achieve something like this with my Home Assistant install to manage my home battery. It can also manage EV charging but I haven't needed to do that yet due to how my tariff works. (Very cheap fixed period overnight).
Thats what bi-directional charging is for and its already becoming political to force the industry to support this.
And we already have energy provider which provide a tarif for exactly this.
The only idea i hate is the mentioning of crypto. Not only is it waste, it converts the energy in heat which needs to get disipated and potentially wastes even more energy to get this heat away from the current location (ac).
A world where individuals are incentivized to use some wasted space to place low maintenance automated trading batteries to make a little money on the side seems like it'd be an interesting solution to the renewable energy storage problem. Put a few units in otherwise useless locations like on roofs or in between highways and make some cash, sounds like a decent investment.
Would there be any expected problems in doing such distributed power storage on a very large scale around the grid that you'd have to account for? Perhaps issues with synchronization, power flow or the possibility of large scale drops in avaialble stored power at times?
Outside of the grid level, maintenance, security, and safety of batteries is important. Many Americans have a garage to securely store their $30k+ car, bike/etc - similar security may be necessary (literally or for the feels) for expensive batteries.
Or preferably your EV? Requires no extra space or extra investment...
pretty sure nodal energy market trading, duck curve arbitrage is the whole profit play for the tesla power wall/autobidder and the Base Power startup coming out of ERCOT. There is definitely a land rush from energy firms and GIS guys to front run land purchasing near solar buildouts and to build as much grid scale battery storage as possible. The percent swings can be huge. Also a huge rush for this in the Northeast now that the offshore wind contracts have been cancelled.
I find Autobidder fascinating, they've been at this for awhile. When I saw this HN post I thought the paper may have come from Tesla themselves.
https://www.tesla.com/support/energy/tesla-software/autobidd...
https://www.tesla.com/support/energy/tesla-software
https://octopusenergy.group/kraken-flex This is already being done at an industrial scale in the UK
seems dumb to have electricity needing to be wasted when there is seawater to desalinate
> seems dumb to have electricity needing to be wasted when there is seawater to desalinate
That's a much more complicated problem. On an energy market, you have only one price to look at, and the battery operator can always buy, sell, or hold energy. The article here talks about optimizing this problem at 5-minute to several-hour intervals.
If you drop excess power into desalination, however, now you have two prices to worry about: energy and water. I also doubt we have 5-minute spot markets for water, so the operator must probably commit to some medium-term water delivery regardless of price.
This means that a desalinating firm takes on much more risk. This might still be profitable, but it's a long-term play based on a deep model of expected energy prices (i.e. knowing that energy is "always" almost free at noon in summer) rather than short-term time-shifting.
Desal plants are also extraordinarily expensive and need to operate at very high 'capacity factors' in order to payoff the capital investment that was required to build them. Operating for a a few hours every day because your operating costs are low/negative only works if you don't have a hugely expensive piece of infrastructure depreciating as you wait for those prices to come down.
2 replies →
Industrial processes like desalination tend to call for some optimal amount of near 24-7 utilization (barring maintenance and such) for capex reasons and efficiency. You want to use it as much as possible to get the most bang for your buck. The entire reason why there are these excess power periods is because we cannot predict accurately how much power we would really need.
Desalination plants really don’t like being ‘throttled’, and are quite capital intensive. Stopping production for any length of time can even destroy the plant, if not done very carefully. Similar for geothermal, though the specific details are different.
Even free power would likely not be worth using if it was sporadic, and it’s extremely energy intensive. So that really is saying something.
Energy futures you say? Oh boy I've heard this one before...
That kind of association is why you can buy futures in just about every crop except onions. The fact that you can structure fraud around a specific kind of future doesn't make them special or a bad idea. One major fraud should not bias us forever.
explain?
That's what the infamous company Enron was running: simultaneous energy futures brokerage and market making. Their market making part blew up and they hid it from everyone for a while until the rest of the company did. It was a massive scandal. It was pretty similar to what FTX was doing just with energy instead of crypto.
If anyone is interested in Batteries, I highly recommend following NAATBATT.org on LinkedIn and joining their newsletter.
Obama set this org up as a senator to help bring lithium ion batteries supply chain to the US and it since evolved into the trade association for all things batteries.
https://www.linkedin.com/company/naatbatt-international/
[dead]
[dead]
Jesus! Why finance people are so hell-bent in extracting rent from every single thing, pervert it, make sure the incentives are all pointed to the shortest run while socializing all the costs to the rest of us?
If this was implemented in full scale, then it would .. stabilize energy prices towards the mean, and make the energy transmission more stable and resilent in general. I'd happily pay some overhead for that, as it also mean the network can accept more cheap energy sourced from wind and solar.
This is the kind of thing that is funding the massive expansion in renewable energy build-out at the moment. The whole reason there's an energy transition happening is that solar and wind and batteries are cheap enough you can make a lot of money building them, and that'll remain true until basically the whole grid is renewable (finally kicking out the expensive gas turbines), and the average price drops to reach cost of supply.
(And the kind of optimisation that happens with this kind of paper is really in the margins stuff. It generally helps the predictive power of the grid, and usually doesn't make much money once more than one group starts doing it, since it's pretty cheap to run and the margins shrink quickly)
More money made with batteries means more batteries installed. How is that a bad thing?
Profit motive commonly has obscene consequences, like destroying food instead of using it to feed the hungry.
11 replies →
No need of speculators. A utility purchases cheap off peak power and extremely costly peak power and pass the costs to the consumer. If the utility pays homeowner with BESS something comparable to peak power rate, they can recover their investment quickly.
1 reply →
Except this is exactly what you want to happen. The reason electricity prices range is because of supply and demand. Batteries help smooth out the supply curve.
In the short term adding more batteries may allow someone to generate income using this strategy but long term what it will do is push electricity prices down and prevent power generation from being overwhelmed. As the battery "market" gets crowded profit margins will fall and everything will reach an equilibrium.
This is a great demonstration of how capitalism works and why it's beneficial.
Smart People with money are willing to give it to people if they can make them more money.
Why would anyone give them money if they were just going to throw up their hands and go ‘well, nothing we can do I guess!’.
There is of course the risk that the money gets burned instead of more money getting made, which is the risk in risk/reward.
Rent seeking type behavior tends to happen when there are no obvious ‘green field’ type endeavors to invest in. Or when risk appetites are trending negative.
Note - many of those people with money that want to use it to make more money are retirees, pension funds, etc.
Capitalism is the only form of communism that works on longer time frames. Periodic resets are still needed, otherwise the monopolist becomes the ruler.
Negative prices are mostly caused by unreasonable terms towards generation plants. E.g.: requiring the grid to take every kWh generated and paying a fixed price over a 20-year term. This of course encourages capacity to be built with no flexibility. Why not dump your solar power into the grid? You're getting paid, the state guarantees for that... the negative prices are someone's elses problem.
Energy overproduction is going to become a serious viability problem for baseload generators, which in time will significantly affect grid reliability. Rolling blackouts will become the norm unless we figure out a serious scalable solution to this.