Tesla ‘virtual power plant’ second best to real people power

Print Friendly, PDF & Email

The Conversation

The South Australian government and Tesla recently announced a large-scale solar and storage scheme that will distribute solar panels and batteries free of charge to 50,000 households.

This would form what has been dubbed a “virtual power plant”, essentially delivering wholesale energy and service systems. This is just the latest in South Australia’s energetic push to embrace renewables, make energy cheaper and reduce blackout-causing instability.

The catch is that more than a third of the costs of a power system are in the distribution networks, as are most of the faults. A virtual power plant on its own can’t necessarily solve the problems of costly network management.

The bundling of batteries together to power a network doesn’t consider the needs of either households or the network.

To address these problems, we’re trialling technology in Tasmania that intelligently controls fleets of batteries and other home devices with the aim of making networks more flexible, reliable, and cheaper to operate.

The Bruny Island Battery trial

Part of what we need to transition to a more reliable and cleaner grid is better control of power networks. This will improve operation during normal times, reduce stress during peak times, and remove the need for costly investment over the long term.

For instance, sometimes the network simply needs more energy in one particular location. Perhaps a household doesn’t want the grid to draw power from their battery on a particular day, because it’s cheaper for them to use it themselves. Most models of virtual power plants don’t take these different needs into account.

Bruny Island in Tasmania is the site of a three-year trial, bringing together researchers from the the Australian National University, the University of Sydney, the University of Tasmania, TasNetworks and tech start-up Reposit Power.

Thirty-three households have been supplied with “smart battery” systems, charged from solar cells on their roofs, and a “controller” box that sits between the house and the power lines.

Participants are paid when their batteries supply energy to the Bruny Island network, which is sometimes overloaded during peak demand. Their bills will also go down because they’ll be drawing household power from their battery when it is most cost-effective for them.

In a world first, Network-Aware Coordination (NAC) software coordinates individual battery systems. The NAC automatically negotiates battery operations with the household (via the controller box), to decide whether the battery should discharge onto the grid or not.

In these negotiations, computer algorithms request battery assistance at a price that reflects the value to the network. If the price is too low for the household, for example because they are better off storing the energy for their own use later in the day, the controller will make a counter-offer to the network with a higher price.

The negotiation continues until they find a solution that works for the network, at the lowest overall cost.

The NAC-based negotiation is half of the economic equation. Battery owners will also be compensated for their work in supporting the grid. The trial team are working out a payment system that passes on some of the networks’ savings created by avoiding diesel generator use on Bruny Island.

Solving big problems

The problem of co-ordinating Australia’s 1.8 million rooftop solar installations in one of the longest electricity networks in the world is not trivial.

Distributed battery systems, such as in Tesla’s South Australian proposal, represent one possible future. The question that we’re exploring is how to coordinate large numbers of customer-owned batteries to work in the best interests of both the consumer and the network.

The primary feature of virtual power plants, to lump together resources, runs counter to what is required for targeted distribution network support. Nor do virtual power plants necessarily have to act in the best interest of householders.

In contrast, we’re trialling technology that acts in the financial interests of householders, to earn value from their batteries by providing location-specific services to networks, at a time and price that suits the customer.

As currently conceived, the South Australian scheme may not be the most cost-effective solution to dealing with our evolving electricity system’s needs.

The Bruny trial shows a different possible future grid – one which allows people to produce and store energy for themselves, and also share it, reducing pressure on the network and allowing higher penetrations of renewables.

The Bruny trial is funded by ARENA, and is a collaborative venture lead by The Australian National University, with project partners The University of Sydney, University of Tasmania, battery control software business Reposit Power, and TasNetworks.

Source: The ConversationReproduced with permission.  

  • Damien Ashdown

    Yes but having solar and batteries is the first step. What you are talking about is really just control software that buys or sells depending on what’s best for the household, yes your right that lumping everything together is not really an elegant solution to what is a complex problem but given the political reality SA is in I would rather they install an 80% solution than role with business as usual.

    • Jon

      It would be nice if there was a standard protocol for inverter/charger battery controllers, I’ve talked to a couple of companies working on the “magic box” that sits between the house and the power lines and their answer seems to be “we work with 1 or 2 inverter/charges, and aren’t interested in expanding that”

  • Ian

    Conflict of interest in this article? As Damien intimates some batteries are better than no batteries. Even though the virtual battery scheme may be flawed, it is getting on with the job of promoting this technology.

  • solarguy

    Government housing tenants get the system free and they will benefit by using less power from the grid, they won’t use it all but then they won’t be paying for it. Your argument on that score is mute.

    For the remainder who pay for the system sure you have a point. I profit from my Hybrid system because I only sell excess power and none from the battery.

  • Harry

    These VPP and NAC solutions only address issues arising from transition to renewables. In a world of 100% renewables there is excellent alignment between the storage requirement for making intermittent renewables dispatchable and the idle capacity created if all vehicles were electric in most parts of the world. For example in Australia 150 GW of installed intermittent renewables with a 20% capacity factor could power the grid when paired with 600 GWh of storage, which can be provided by 15 million vehicles with 50 kWh batteries. Rather than creating infrastructure that will become redundant in a few years, government should incentivise the best way to transition, which appears to be electric vehicle uptake. This is outlined in the V2G (vehicle to grid) concept that is gaining traction and for which the upcoming Nissan Leaf is design to facilitate.

    • Mike Westerman

      EVs will also result in battery packs at the end of their EV life with plenty of life left if repurposed as stationary batteries, so the symbiosis is strong, and the duplication somewhat illusory

    • solarguy

      How many of those 15 million vehicles do you think will be on the road at any given time?

      What if you come home and your battery is very low and you need to charge?

    • itdoesntaddup

      You would need of the order of 10 TWh of storage, not 600GWh. You might also need rather more generation if those EVs are ever going to go anywhere, rather than sitting on driveways as expensive storage.

      • David Osmond

        AEMO’s 100% renewable study indicated 100 GWh of storage is required. UNSW indicated 200 GWh. BZE: 700 GWh, ANU: 450 GWh. Care to point at which study indicated 10 TWh was required?

        • Harry

          Your data does put the 600GWh in context and the different values I suspect depend on how you take into account dispatchable renewable such as Hydro. With only 30 GW needed to support grid at worst at any instant of time this would only need 10% of all vehicles being connected to grid which should be easy as at any given time at least 75% of vehicles are idle. What this shows is not only can we have 100% renewable power generation but minimal fuel consumption for mobility. As battery is shared between grid and vehicle, economics are very attractive for $15K for battery and $10K for PV system costs as one can eliminate around $5K in annual power and fuel costs. Much better than economics than separate batteries for home and vehicle, hence why governments should incentivise such an approach to do what is best for their people.

          • David Osmond

            Cheers Harry. The AEMO and UNSW had substantial amounts of dispatchable biofuels (22 GW), generating 24% and 6% respectively of generation. The other studies didn’t rely on an increase in biofuels, which is why they had larger storage requirements.

          • Harry

            It appears various units are being confused. Each GWh of CAPACITY if charged and discharged daily will deliver an annual storage quantity of 365 GWh per annum or 0.365 TWh. I suspect itdoesntaddup is quoting annual quantities rather than capacities.

        • itdoesntaddup

          Try Jacobson 2018. Here’s (Table S4) Case A storage in TWh for Australia:

          PHS 0.12
          CSP 1.08
          Batteries 0.504
          Hydropower 32.6
          CW-STES 0.058
          ICE 0.087
          HW-STES 1.005
          UTES 135.7

          Total 171.154 TWh

          Another order of magnitude greater.

          • David Osmond

            You’ve quoted the figures for Case A of the Jacobson study, which assumes and uses cheap underground thermal storage. Case C, which avoids that, has about 34 TWh of storage, but nearly all of that is from existing hydro. Only 1.064 TWh of storage is additional to what we already have. Now this study is for an Australia in 2050, where everything has been electrified, and average electricity demand is about 160 GW, or about 7 times the current average NEM demand. 1.067 TWh works out as 7 hours of storage at average demand levels. Applied to our current average NEM electricity demand of 22 GW, 7 hours equates to about 150 GWh, which is at the low end of the amount of storage used in the 4 studies I’ve listed above.

          • itdoesntaddup

            If you trust that heat pumps can do your A/C at a quarter of the power consumption of your A/C unit (the Jacobson assumption), then go with case C – which still calls for 34TWh of storage, not 1TWh. Everything I read suggests that A/C and heat pumps are about equal when it comes to cooling, and that heat pump performance degrades rapidly below freezing, which even happens in Australia from time to time. The load for case C is 90.99GW, not 160GW incidentally. The hydropower figures show 32.6TW at a discharge rate of 8.05GW, or 4075 hours, or just short of half a year (plus pumped storage at another 859GW and 100GW of battery capacity at how many bn a GW?). Problem is, much of Tasmanian storage is empty, and there is no adequate lower reservoir to pump back up – just run or river to the sea.

            Anyway, you made a claim, which I disproved, and in your attempt to defend your position you misquoted the figures badly. Lame stuff.

          • David Osmond

            Apologies, load was 90.99 GW, or 4.1 times current average NEM demand.

            Case C calls for 34 TWh of storage, but 32 TWh is existing hydropower storage. So it only requires 1.067 TWh of additional storage. Scaling that down by a factor of 4.1 to be relevant to the current NEM works out as 257 GWh of storage, which is consistent with the other studies I’ve quoted.

            Tassie dam levels are just under 50%. They were run down very low during the carbon price period, and even lower when Basslink was down and Tassie suffered a drought. They’ve recovered pretty well since then, and usually hit peak levels towards the end of winter. Once they’ve built some more wind and solar in Tassie it should be able to keep fairly consistently high dam levels if they desire.

          • itdoesntaddup

            There is no capacity to store the required amounts for Case C. That requires not only a store, but also the ability to put energy into store and extract it at the required rate. Ten Snowy 2s would get you there.

          • David Osmond

            The study does not assume that you can pump water back into the hydropower storage (you are confusing hydropower with PHS). As the study says:
            “Although batteries are currently a relatively expensive form of storage, they can be charged and discharged at will, whereas hydropower can only be discharged at will since it is charged naturally.”

          • itdoesntaddup

            The problem is that if you need to use that storage you will have to wait years to refill it again. That does not make it storage suitable for Case C operation.

          • David Osmond

            It certainly is storage suitable for Case C operation. When PV is reduced during winter, and if we have an extended lull in wind generation, then it is possible to run extra from hydro, run the dam levels down, and then let them recover by using less hydro in summer or windy periods.
            You seem to be confusing hydro storage with pumped hydro storage. Storage levels in the latter can recover much more quickly by pumping, and the study lists these separately.

          • itdoesntaddup

            How much rainfall do you get in oneseason to refill Lake Pedder etc. if you empty them? Especially if it is a drought year? The point is that the intention is that the storage should be used – i.e. run down to minimum – annually. You might have to wait 3-4 years for refill through rain. That is why you need a lower reservoir as well, and pumping.

          • David Osmond

            What makes you think that the intention of the hydro storage is that it should be used annually? It is clear that they haven’t done that, as they have not assumed any increase in annual hydro generation. A vast increase would be required to empty the hydro storage in a year.

          • itdoesntaddup

            Yes they have. They are assuming over 8GW of hydro generation. Do keep up.

          • David Osmond

            I’m well aware that they have assumed the (existing) 8GW of hydro. My question was where does it say that they run down the entire 32 TWh of hydro storage each year?

          • itdoesntaddup

            p 39. Total Case C hydro generation is 209.79 TWh over 5 years or about 42TWh per year, or running the hydro at an average of 4.8GW.

          • Mike Westerman

            Not if wind generation is ramped up as is happening at the moment and needs to continue. If wind firm capacity is high enough to replace all the energy stored the all the storage is available .

          • itdoesntaddup

            But it is no longer available once it is used. Jaconson plans on using 42TWh p.a. of hydro in his case C. I assume he has a hot line to some rain god.

          • Mike Westerman

            Maybe but for us mundane engineers we rely on increasing sources of cheap energy, which in Tas means wind – some of the best wind resources in the world – calm <1%, greater than light almost 75% of the time at Ulverstone for example.

          • Mike Westerman

            You don’t seem that worried about your own misquotes and errors, so maybe you own that “lame” moniker yourself? As David points out, your claim of Tas storages being empty is demonstrably wrong, and the causes of shortfalls in the past are in the process of being overcome thru its ample access to cheap windpower. But your claim of no adequate lower reservoir is also ignorant of current options studies, several of which could deliver several TWh of storage.

            If I were to critique many of the estimates of storage it would be to express surprise that there is still support for the notion that demand will not radically change in the light of very cheap and ubiquitous solar power. In combination with intelligent controls, the scope for changes that don’t require storage is enormous. It is axiomatic that investors make choices on the balance between capital plant and operating costs, of which energy was a significant opex in the past. When it isn’t and when you have the capability to intelligently control processes, you have scope to operate in ways not possible previously. A very simple example is one of our major water companies, which has cut its power costs to a 1/3 and expects them to go much lower, by using an advanced monitoring system to allow it to buy power when it is at low or negative prices.

            But then you believe any form of demand management is a retreat to the dark ages, hence this sort of thing not adding up.

          • itdoesntaddup

            Tas storage is not yet full – and once you use it one year, how do you refill it in time for the next? You have no lower reservoir to fill it from – even if you installed another 6GW of pumping/generation and 7.5GW of Basslink expansion and onward grid transmission.

          • itdoesntaddup

            I presume the water company did its sums on beefing up its pumps so that it can take advantage of cheap prices, and that there is no requirement to pump away sewage when the electricity price is high, or to refill water tanks. Not sure it really passes the smell test.

          • Mike Westerman

            Doesn’t change the pumps, the main change is storage which is cheap, as is thermal storage. Your “status quo” thinking doesn’t let you see how elastic demand is.

          • itdoesntaddup

            I am sure demand is very elastic at $14,200/MWh, and more so when there’s no supply.

          • Mike Westerman

            It’s called a market for a reason tho a failed one when these signals are contradicted by gov policy or failure to reform.

  • itdoesntaddup

    So we have an admission that a VPP “might not be the cheapest”. Neither will this. Bet it’s hackable too. Your neighbour tells your unit to bid low, so they get power and you don’t.