EV Charging won’t destroy the grid

I’ve been working on EV charging since 2014, and in that time have heard some deeply weird opinions on EV charging and the various ways it will destroy the grid.

There are three main classes of EV Charging weirdness; energy problems, capacity problems and technical problems. I’ll start with the technical because there’s only one major example:

Technical Problems

In the early days, distribution network service providers (DNSPs, like Essential Energy in NSW) would look at our connection applications and mostly wave them through as just another big load. But every now and then the application would land on the desk of someone a bit more curious, who would wonder why 500kVA was being connected to an abandoned parking lot and go looking for more details.

Big DC chargers have big rectifiers, and any electrical engineer worth their salt knows that (old) rectifiers produce harmonic distortion. This changes the shape of the AC voltage curve, often by sort of reflecting smaller waves back through the network, hence the harmonic part of the term. Instead of a beautiful, perfect sine curve, something might reflect a sine curve at 100Hz, rather than 50Hz, so the waves add up some times and subtract at others. The grid gets impacted by the way the device uses the power supply.

The most classic example of something that introduces harmonic distortion is a cheap hair dryer. At full power the hair dryer draws as much current as it can and turns it into heat, the motor and heater in perfect sync with the grid. But that gets a bit hot, so there’s a half power on the hair dryer as well. To achieve half power, this switch drops a diode into the power circuit which means the dryer only uses half the sine wave now. Instead of using the whole curve, above and below zero, it only uses the bit above zero, like a series of speed bumps. If the dryer is only using the positive, where does the negative go?

Back out to the grid, as harmonic distortion.

50 years ago the only big rectifiers were making power for train and tram lines, and engineers knew they needed to take precautions to manage them. Typically they would try and isolate the harmonics by using a transformer, which dampens those weirder edges transforming the curve between coils. The train supplies are manageable because they’re big, dedicated sites with all the engineering bells and whistles. What is going to happen when they start plonking big rectifiers everywhere, will it poison the sine curve for everyone else?

Fortunately, charger manufacturers are across this problem and have implemented fixes and tests to show their harmonic distortion performance. It’s usually a self solving problem; if the big rectifier is separated from the rest of the grid by a transformer, say a dedicated 500kVA distribution transformer, the distortion will largely be limited to that little bit of the grid and no problem for any other users. Some manufacturers include an isolation transformer in the design of their charger, separate to any potential distribution transformer. And all of them now perform “Total Harmonic Distortion” tests as standard and can issue reports saying that “THD<5%” should anyone start asking questions. But, I think the industry has caught up with this in parallel, because I haven’t been asked THD questions for a few years now.

Energy Problems

“Where is all this electricity going to come from! We will need to quadruple generation! And it’s all coal anyway!”

Starting with coal, yes it is a problem that we still make a lot of electricity from coal, but we can solve it in parallel with rolling out EVs by transitioning to clean energy for our electricity. Even with current power technology EVs are a better result for the climate, and will keep getting better over time. For me the gold standard study into this is from Auke Hoekstra, helpfully summarised and linked in this RenewEconomy article.

How much electricity does an EV need? In Australia the average daily drive is about 40km, roughly 15,000km per year, spread out over 365 days. Assuming 20kWh/100km for an EV seems about right, it’s been my rule of thumb for 5 years now, so to drive 40km your EV will need about 8kWh per day.

The average Australian house uses about 20kWh/day. So adding 8kWh to a house load increases it by about 40%. Not multiples of normal house use, a 40% increase.

Where will this energy come from? The grid, but we need to make sure it’s clean power. EV drivers can ensure their power is clean by adding solar and plugging in at the right times. To deliver 8kWh/day in Australia requires only 2kW of solar, plus another 5 to cover the house load. This nice graph from APVI shows the average household system in Australia is now over 8kW, so I’m taking that as evidence that adding enough solar to cover your house and car load is feasible, and fairly practical. It’s what we’ve done.

Solar, and home charging, will be difficult for some people, particularly those renting, living in apartments or in one of those nightmare streets in Sydney where you park wherever you find a spot and that could be 6km from home. These folks will rely on public charging, and how clean their electricity is depends on the procurement practices of those running the network. I work at Chargefox, Australia’s largest public charging network, and we work hard to make sure all the power we buy is clean power. When I worked at Tesla AU they did the same thing, and I’m hearing that other networks building in Australia are doing the same thing.

So on energy, yes we can supply the additional electricity and yes it can be clean power. There’s work to do, but definite pathways to get there.

Capacity Problems

These are probably the trickiest for lay people to get their heads around, because capacity problems are time dependent and can occur anywhere. I banged on at length about capacity problems in the Battery Post, but I’ll summarise here.

Capacity problems occur anywhere in the electricity network where too much power tries to flow. They can occur in your home switchboard when you try and run the dryer, oven and iron at the same time. They can occur in suburb-scale substations when every house in the suburb is trying to pump solar back into the grid, exceeding the capacity in reverse.

The key thing to remember here, and it brings us back to our friend at the start, is that EV chargers are just another electrical load. They are not special. It’s just like installing a new air conditioner or water heater. It’s a thing that uses electricity, being added to a system that is designed from the ground up to add new things without breaking it.

There is quite a bit to unpack with capacity problems and how they’re managed, so we’ll start at the house level.

Say you buy a beautiful second hand 2014 Leaf and want to install a charger at home. You will call an electrician and ask them to install the charger. The electrician’s full time job is connecting new electricity things and making sure they don’t break anything when they do it. So when they install the charger at your house they will make sure the sum of the potential loads in your switchboard does not exceed the capacity of the switchboard. And if something goes wrong, someone adds 3 heaters to a single power point, then the circuit that is drawing too much current will trip, disconnecting the demand and protecting the switchboard.

But what if everyone on a street installs a charger, every single house with enough capacity to fully use their connection? This can become a problem at the local distribution transformer, which is kind of what the tweet in question is talking about. The difference is though, that in the absence of any protections, yes installing too many chargers can wreck a transformer, but for the same reason turning on too many heaters doesn’t wreck your house switchboard this won’t happen.

The transformer has fuses and circuit breakers too. If everything goes wrong and too much is demanded from the transformer, with or without EV chargers, the circuit breaker will trip or the fuse will blow. This happens only very rarely, because the electricians are all working within limits, and those limits sum to the transformer. But if someone makes a mistake only the fuses in the transformer will blow, there’s no way that the lifetime will be decreased by orders of magnitude.

Play this forward and you can see it being a problem if everyone installs a charger and all try to charge at the same time. Transformers rely on two things to stay safe; “diversity” of loads, and thermal protection of something big and heavy. Transformers are actually quite resilient and can run at much greater than their rated load for short periods of time. A 500kVA transformer can flow 500kVA constant, and peak to 650kVA for an hour or more, some could even get to 1000kVA for a short period. It’s a heat problem. When it’s overused it gets hot. Heat builds until something fails. But they’re big, heavy objects, usually with oil cooling, so a sudden burst in heat can be absorbed and dissipated. Even more so when the transformer is up a pole and it’s a cold day.

Diversity in loads is the degree to which they don’t coincide. Say you have 100A of potential loads at your house, but you never operate them all at the same time, maybe the highest you’ve ever demanded is 50A. You could say you have diversity of 50%. Local distribution networks work on the same theory, that it’s very unlikely that everyone will use enough load at the same time to test the transformer. And if they accidentally do, the fuses fail, not the transformer.

This is an unsatisfactory management strategy, relying on fuses to save transformers, so new technology is coming to manage this. There is a risk every house in the street could plug in at the same time and cook the transformer. How can we manage that better?

With Smart Charging, and this can solve almost all capacity problems in EV charging I’ve seen. The problem of too many chargers on a circuit is already happening in some commercial buildings and is analogous to too many houses on a street turning on at the same time, but it’s the building switchboard under threat, rather than the local transformer. Smart charging relies on the fact that capacity problems are typically very short lived, so if you can defer a load for a little while, chances are you can avoid the congestion. On a suburb scale, that means deferring EV charging until after the 7pm peak, which is easy. Cars are usually stationary for 12 hours and only need a few hours to fully charge.

Using EV chargers with internet connections, it’s possible to orchestrate all of the chargers together and make sure they don’t all draw at the same time. Chargers have a language protocol that allows their power to be turned up and down with standardised commands. So a smart charging system will measure the demand at the switchboard, and then turn chargers up or down to make sure the total doesn’t exceed the design limit. Then another layer of smarts can be applied over this to make sure the cars charge the way you want. Maybe car 3 needs to be full by midday, but car 4 can wait until 4pm. The system can handle all of this. These systems are already in use now, using connected residential chargers and railway commuter car parks.

That covers all the small chargers, what about the big ones? The massive inter-city charge sites that can charge a car in minutes rather than hours?

All of these same protections continue to apply at larger scale, it’s just different people giving approval. Rather than your local electrician giving permission to connect, it’s the network operator making sure we don’t add too much load to the circuit. Rather than your local distribution transformer getting overloaded, the 11,000V network could get overloaded. But it doesn’t get overloaded, for the same reasons; there is a strong process controlling new connections; there is diversity in the loads allowing more to be connected than can run at the same time; if by chance they do demand too much at once, fuses and circuit breakers protect the equipment; and if we find we’re hitting those safeguards regularly we can apply smart charging and control point metering.

This is a long way of saying that yes, there are some technical challenges with adding EVs to the grid, but they are surmountable problems, that can be solved with existing technology and systems. EV chargers are just another load, being added to a system designed from scratch to add new loads. EVs are coming and the grid is cleaning up in parallel. I feel like I’ve said this a bit recently but:

It’s working. Keep going.

What are batteries?

Batteries are coming to a grid near you, if they’re not there already. Little ones in households and big ones like in Hornsdale, South Australia are being deployed in support of the clean energy revolution. But what are they doing there? What can batteries do?

For starters, they’re not really “batteries”, more accurately they’re a battery energy storage system; we have a few at work and we call them “BESS” when we’re being formal, but normally just batteries. The battery part is just where the energy is stored, and it is usually thousands of cells connected in series and parallel to achieve the desired voltage and current. This is the DC part of the system, an energy storage reservoir hidden away inside the BESS. In modern systems these cells are almost always based on a lithium chemistry, normally Lithium-ion or Lithium Iron-Phosphate, with a few more exotic chemistries in niches.

The cells are protected by the Battery Management System, often called the BMS. A BESS is a bit of an onion, layers of system wrapped around each other. At the centre is the cells, the chemical batteries where the energy is stored. Wrapped around the cells is the BMS, a software and control system that monitors the vital statistics of the cells and controls the things that make sure the cells stay in specification. Not all BMS are created equal, some just measure voltage at the DC terminals of the whole array and the temperature. Others measure the temperature, voltage, state of charge and current flow out of each individual cell and can tune their protection accordingly. A good BMS is one way that smart companies can differentiate the quality of their BESS.

Next layer out is the inverter, which is really a bi-directional inverter-rectifier/charger. Inverter takes DC power and turns it into AC for the grid, a rectifier takes AC and turns it into DC, a charger controls how DC goes into the cells. Despite this complexity, we’ll keep calling it the inverter.

In solar and BESS systems, the inverter is the bit that interfaces the electricity production with the grid. I would argue it’s the most important part of the system, once you’re beyond a minimum quality standard in solar panels and battery cells. Inverters are where the magic happens.

In our household system, the inverter is like a traffic controller, directing electricity to where it’s needed. If the sun is shining and there are loads in the house, power gets directed there. If our solar was export limited it would ramp down the power output if there aren’t household loads. The inverter also controls battery charge and discharge, which in a house is quite simple; if the sun is shining and battery has capacity, charge battery. If the sun is not shining, there are loads on the house and the battery has charge, then discharge the battery. That’s most of the algorithm in our house. Additional complexity is possible with off-peak charging and similar time of use tweaking, and this also comes from the inverter.

The outer layer of the onion is the control system beyond your site that orchestrates your battery as part of a larger system. The Virtual Power Plant in South Australia, or a third-party controller like Reposit are good examples of this.

What can batteries do?

The capabilities of a battery are almost entirely governed by the inverter. The cells play a role, they need to be able to give or take DC power quickly enough to keep up with the inverter, but they’re basically just a bucket of energy to keep the inverter happy.

The battery and inverter might be different components, which we call a DC coupled system, or the same box, an AC coupled system, but they perform the same, the inverter governs the battery capabilities.

At a minimum, the inverter will be able to make AC power from DC and put it into the grid safely, take power from solar panels or the grid to charge the cells and control the charging of the cells to keep them healthy. They will monitor solar production, household power and the grid status to control charging and discharging the battery, depending on the operating mode. Most batteries are just capturing solar export and discharging it at night, and they turn off if the grid drops. It’s not intuitive, but having a battery doesn’t mean you’re immune to blackouts, and this requires additional hardware and smarts.

Like solar inverters, battery inverters must turn off when there is a grid blackout. This sounds like a conspiracy, but there’s a genuine safety reason that must be followed; if the grid is down, and you connect a generator to it and put power in, that could injure someone working on those lines. I have heard this is a genuine problem after cyclones. Lines go down all over the place, old mate remembers he’s got a generator out the back, and re-energises the house. It also puts power back into unknown parts of the remaining network.

To get power during a blackout, the house needs to disconnect from the grid first. In a blackout the inverter detects no voltage from the grid and opens a contactor which breaks the connection from the grid to the house. The inverter then starts its own grid just for the house. A good inverter will be able to do this very quickly, enough to not even make the TV flicker. Check out the Inertia post for more on these grid forming inverters.

That’s pretty much the limit of residential inverters, grid forming capability for some, generally just shifting solar, some might be fast enough to be part of orchestrated systems like the Virtual Power Plant.

At commercial scale inverters become more sophisticated and don’t just plod along absorbing solar to discharge it at dinner time, they’re fast and sophisticated enough to modify the quality of electricity in the area in real time.

At this scale, battery systems become the most sophisticated power supply in the grid and by some margin. Good inverters have what they call symmetrical 4-quadrant power which basically means they can charge and discharge at full power, from power factor of 1 to zero. They can make any type of AC electricity possible, in a matter of milliseconds.

As an example, I’ve seen specs on a 1MW battery that can go from charging at 1MW, to discharging at 1MW in 50ms, then back again in another 50ms. For one this is way faster than the fastest conventional generators in the electricity network. A fast-start diesel generator might make it from off to on in 30s if things go well, then ramp from 30% to 100% over 30s. I’ve heard a hydro turbine can get down to 10s or but I’m having trouble believing they could synchronise in that time. A battery could go from off to on about 100 times in 10 seconds.

What makes them peerless though is that batteries can act as a negative power supply as well. Sure they can charge up and store power, that’s what we mostly think of them doing, but pumped hydro can do that. Apply a constant load to the grid and pump water up hill. A battery can take power from the grid too, but it can vary that demand as quickly as it can vary its output.

This is easiest to think about in microgrid applications, with a limited number of loads and power supplies. These grids must keep the balance between supply and demand very tight to maintain frequency, which means the supplies need to be able to respond as quickly as the loads. Lots of little diesel grids around the world will be adding solar and batteries now, because it’s cheaper, but also because it works better.

Imagine a small school grid running on a diesel generator. It’s fine when the lights and computers are working, but when the water pump comes on the generator lags and the lights and computers flicker. That’s because the generator can’t change output quickly enough. A little battery in parallel with that system would take out those spikes, in both directions, and make better quality power.

So what can batteries do? They can charge and discharge, they can monitor and respond to key grid parameters or be controlled as part of larger systems. They can do anything a conventional generator can do but much faster and with more accuracy, and they can do it *negatively* as well.

How do we use them then?

The obvious opportunity for batteries is residential solar, and this is probably the area that is getting the most attention at the moment. It makes sense because residential solar is all around us so of course you’d want a battery. I think residential batteries will be a contributor for sure, but the dark horse here, and the main reason I’ve written this long post is that batteries are most cost effective right now in capacity problems, and that has some incredible outcomes.

Capacity Problems

A capacity problem is when a part of the network, say a power supply to a building, or an interconnector between states or even just your household switchboard demands more power *instantaneously* than it can handle. It might literally be 15 minutes a year, but this is a hard limit and if you exceed a capacity constraint something trips or catches on fire. If the interconnector is running at its rated potential during a heatwave, then that’s it. Additional demand will go unserved and there will be blackouts. To get more electricity down the wire requires a new wire and often that is an expensive job.

There are capacity problems everywhere; businesses, networks, households trying to decide if they need to upgrade wires so they can use more electricity. The thing about capacity problems though, is they are only a problem at the very limits of the system, because electricity demand varies over time. I see this obviously in our household use data; we use about 20kWh/day, suggesting an average demand of a touch under 1kW.

You can see the intermittent things running through the night, fridges, modems that sort of thing. The coffee machine turns on at 0700 and demand peaks at 1700 when cooking dinner and charging the car. Our record peak is 7kW. If we had a capacity constraint of 3kW, this would be a problem for about half an hour out of 24. If we wanted to upgrade the capacity we have a choice between installing a new wire to be used half an hour a day, or installing a battery which will also deliver other services.

This sort of profile appears at all sorts of scales; the peak demand of a system is way higher than the average demand, and so capacity constraints are for a limited time only. As another example, here’s the averaged daily load profile for a 24-hour service station.

This sort of average hides some things though, the hottest days will peak higher than that, the mild days less throughout. If they have a capacity constraint, how often is it realised? We answer that with a load-duration curve, one of my favourite graphs.

This curve takes a year worth of 15-minute interval electricity data and sorts it from largest to smallest. Plot that against percentage time and you get an idea of how often a system reaches different demand points. Notice how steep the right hand end of the curve is; average demand for this site is about 20kW, peaking at 50kW. But the bit from 30kW to 50kW is vanishingly small, the final 40% of demand occurs less than 1% of the time! These are 15 minute intervals, of which there are 35,040 each year. 1% of that means there are 350 15-minute intervals each year where demand is above 30kW. If we applied a 20kW battery to this it would only have to discharge for 15 minutes each day, to reduce the demand by 40%. If I need 20kW for 15 minutes a day, is it cheaper to apply a battery or upgrade the wire? These are easy questions to answer with quotes from suppliers, so it’s easy to make a sensible decision here.

Because the capacity limit is typically short, batteries for these problems only need an hour or so of discharge, even less in some cases. The new ESCRI battery in South Australia is 30MW and just 8MWh, or 24 minute flat out discharge.

Implications

It’s a long road to here I’m sorry, but this background lays the ground work for where batteries will be used in Australia over the next decade.

The intuitive response is that households with solar will install a battery, and I agree with this. But they’re not really cost effective just yet and my model says they won’t make a huge contribution in the next decade. Practically speaking these will mostly be low-sophistication batteries shifting solar export into the night, and will have the affect of reducing residential demand.

We’re starting to see grid scale battery deployments, mostly in grid support functions, either arbitrating through the wholesale market or providing FCAS or inertia services. These big deployments show us what’s possible, but I think they’re mostly going in because they’re the easiest ones for the regulatory framework to consider. A big battery is pretty much a generator and the utilities can cope with that.

But it’s the next phase that I think will do the heavy lifting and surprise many in the sector; batteries deployed to solve capacity problems. There are capacity problems that are cost effective right now against upgrading wires. And as I’ve spent 2000 words showing, capacity problems don’t typically fully utilise the battery.

I’ve seen projects that have a 15-minute PER WEEK capacity problem, that will cost $300,000 to fix. A battery that does the same job costs $330,000, but the battery makes you money in reduced demand charges and possibly solar shifting. And when it’s not doing that one of the orchestrators can bid that battery into the grid and make money from frequency services.

The big idea then is this; many batteries will be deployed to solve capacity problems. They will be cost effective in this role and achieve very low utilisation. So fully paid for assets with extraordinary capabilities will be sitting around just waiting for someone to make money with them. And the potential scale of this is staggering. How many capacity bottlenecks are there in Australia? Thousands? Tens of thousands? What about in future? Every new building or suburb, is it cheaper to build a bigger connection or add a battery?

Take this idea to it’s engineering conclusion and even interstate interconnectors are in scope, as covered in this excellent podcast with battery developer Fluence. An interconnector might extend hundreds of kilometres, something like Murray Link is 180km long. If that cable reaches capacity, 180km of cable needs to be upgraded. That’s a big job. At $1million a kilometre that’s a lot of money to spend on something else.

Summing up

Capacity problems occur when parts of an electrical system are near the limit of how much power they can shift. Batteries are amazing and can act as a near perfect power supply or load, with incredible flexibility. Because of this they are suitable to substitute into power systems as if they were additional grid capacity. Most capacity problems are short lived so high power batteries can solve them fairly easily.

This suggests that the focus on “long-duration storage” is misplaced. What problem needs the same power output for a week? I don’t think there are any. I can see a lot of capacity problems, but not many requiring a week or even a day of storage. I’ve looked at a lot of solar data for Australian houses over the last decade and never seen a zero generation day, even in the worst weather. There will be solar production every day somewhere in the NEM. We just need enough storage to get through the night.

And like solar, I expect that battery capacity in Australia will look like it is massively overbuilt. We’ll end up with 80GW of solar in a 30GW peak market, just because the marginal cost of adding more solar is nothing. Similarly I expect we’ll end up with 80-100GW of 1-2 hour batteries, and basically no longer duration storage. Because the problem, the lack of power, only lasts for short periods in between the sun and wind. And whenever we add more of either, the gap gets smaller again.

So don’t get too carried away with this idea that we need storage for a week or more, as David Osmond’s excellent study demonstrates the need isn’t there. But there are loads of capacity problems just waiting to be solved with fast lithium, with the added bonus of supporting the clean energy transition at the same time!

Electricity Grid Inertia – What is it?

There’s been a lot of talk in Australia this week about ‘inertia’, how much we need of it in bits of the electricity grid and where we can get more. This discussion has been kicked off by this little report from AEMO, on levels of inertia required in South Australia.

I’m interested in this because I’ve been working on the battery side of this equation for a couple of years, and have learned a bit about what they can do, and how it compares to the old ways of doing things. This post is mostly a way for me to arrange my thoughts and see if I actually understand this stuff. Constructive contributions always welcome.

Inertia

Physical inertia, Newtonian inertia, is a tendency to not change in response to external forces. Newton’s first law states that “a body in motion [or at rest] will tend to stay in motion, unless acted on by external forces”. In mechanics it’s mass times velocity, so if you have more mass you have more inertia; more velocity gives you more inertia. Consider a rugby league tackle situation. You have a choice of either a 120kg prop running at you or an 80kg halfback. Incredibly, they both run at *exactly* the same speed. Which one will be easier to stop? I’ll take the halfback every day.

Rotational inertia is similar, except velocity doesn’t apply to rotating things, it’s angular velocity instead. The time to complete revolutions, rather than the time required to move in a straight line. I find children’s play equipment often has some great angular momentum problems, things like a merry go round is easy to stop and start with just one kid on it, rather than the whole extended family.

Power station inertia

Rotational inertia is the one that impacts grid inertia. This is because we (have for a long time) made electricity using spinning machines. In a coal fired powerstation, the workhorses of the Australia grid for a long time, coal is burned to heat water. Water becomes steam and drives a turbine. When the turbine is being driven, providing constant power to the grid, the heat-flow rate of the steam is constant. Something like 750MW of coal becomes 300MW of steam, which becomes 250MW of electrical output. To increase electrical output, first increase the coal burning rate, then the steam rate will rise, and more power will come out.

The turbine, shown above is a Doosan steam turbine, is a big metal thing on a shaft, that spins when it’s making electricity. There are tonnes of material there, spinning at 50Hz. Consider how much inertia is there. Even if the steam system disappeared instantly, the spinning turbine would keep spinning, and slow down as energy is withdrawn from it. This characteristic drives grid inertia; if something surprises the electricity system, this spinning machinery has a very predictable response.

Frequency and Active Power

If you want to hear an expert talk about it, I strongly recommend this talk by Kate Summers. Kate has a lot of experience in how generators respond to external shocks, and there is loads of data in there from Australia. I’ll try and summarise as best I can.

In any electrical system, be it a national grid or a small off-grid house, power being supplied must equal power being used. If there is 1GW of electricity being used, about 1GW of electricity, or “active power” as the grid likes to call it, needs to be supplied [plus a little bit more to cover system losses]. In this steady state, the generator and load are both operating at exactly 50Hz, right where the grid wants it to be.

During steady state operation, another 100MW load connects, and now we have 1.1GW being powered by 1GW. When the power being supplied is less than the load, grid frequency will start to fall. This is where grid inertia comes in.

Mechanical Inertia

In a system with zero inertia, at the moment the 100MW connects, grid frequency will move *instantly* to the new frequency. The grid will go from 50Hz to 45Hz, everything connected with frequency sensing will disconnect, and that little system that was in balance is now hopelessly out of balance and we have a blackout. You can demonstrate this at home with a little petrol generator and a large resistive load. Connect a light bulb and an electric heater to a 1kW petrol generator. It’s happy pottering along powering the lightglobe. Turn on the heater though and the generator will suddenly start labouring. Make it labour enough and it will stall completely.

In a system with generator inertia, this incident plays out differently. The 100MW load connects, but instead of the grid instantly moving to a new frequency, the inertia in the rotating machines prevents it moving that quickly. A system with higher inertia is much harder to change the frequency of. This is because the inertia of the shaft can be another source of active power for the grid.

Inertia acts as a brake on changes in frequency. It’s okay if it wobbles around a little bit, but it’s bad if frequency moves suddenly. In grid circles they often call this “rate of change of frequency” or RoCoF, and it’s a measure of how fast frequency can change.

System inertia gives power systems engineers time to react to problems. In the instant after a new load connects inertia from the turbine will make sure that the frequency doesn’t change too quickly, then its governor will make sure that the power ramps up to meet the new load. Inertia gives us cover between when something happens and when we can respond to it.

What provides system inertia? Spinning things, both generators *and* loads. There is inertia on both sides of the equation. For the generators it’s anything with a spinning shaft; coal plants, gas plants, both CCGT and OCGT, hydro plants, little biomass units. For loads, any AC motor provides inertia to the grid; it stores rotational energy that can return electrical energy to the system. This excellent podcast with Dean Sharafi, head of System Management in Western Australia covers this in some depth. Synchronous condensers are an extreme example of load-side inertia; basically an electric motor, added as a load to the system, to provide load-side system inertia.

What doesn’t provide inertia? Solar inverters are the main ones, but anything that makes electricity without rotating machinery will ordinarily not deliver any system inertia. But there are some exceptions coming down the pipeline. Why is this a problem? As discussed way up in the AEMO report, as the proportion of electricity from non-synchronous generators increases, our stability when responding to frequency changes is reduced. Reduce it enough and the system stops working.

Synthetic Inertia

The exception to my rule above about “spinning things provide inertia” is wind power, which despite using a massive spinning blade to make electricity, doesn’t usually link that rotational inertia to electricity production. This is an outcome of how wind turbines make electricity; the wind blows at whatever speed, spinning the blades and making AC power. But because the wind could be any speed the output isn’t constant frequency power, so they convert that AC to DC, then back to AC for the grid.

Synthetic inertia can be extracted from wind farms, by reprogramming the inverter system. This is a good paper from IEEE on the topic. As far as I can tell, synthetic inertia is basically an inverter program; rather than providing constant output, if there’s a disturbance in frequency the inverter takes a bit more energy from the turbine and moves between frequency points slower than if there was no program change, which gives the system a few more seconds to respond to the disturbance.

Virtual Inertia

Virtual inertia goes a step further and makes inertia-like responses, even though there is no rotating machinery involved. The batteries we’ve been installing for work can do this and it’s mind blowing.

This post from Watt Clarity has some great info on virtual inertia, which is like physical inertia, in that it relies on inherent system hardware, but different because it doesn’t involve spinning machinery.

There are two families of inverters; grid following and grid forming. Grid following is most of our solar inverters; they watch the frequency of voltage on the grid and follow it. All good when the grid is running at 50Hz, but if it drops the grid following inverters will drop too and make it worse.

Grid forming inverters have their own clock and opinion on what the frequency should be. They constantly work to push frequency to where it thinks it should be. In this way, when there is a disturbance in grid frequency, grid forming inverters work against that disturbance automatically, because they’re already trying to maintain frequency. Thus they resist changes in frequency, which is the same as what inertia provides.

Inertia Summary

We get inertia in the grid from spinning machinery, both loads and generators. This is good because it makes it hard for the grid frequency to change quickly in response to a disturbance. Solar inverters typically provide no inertia, and so as the amount of solar in the grid increases, the amount of inertia in the grid for the same power level decreases, which makes frequency less stable. There are solutions available to increase system inertia, such as synchronous condensers, synthetic inertia from wind farms and virtual inertia from grid forming inverters. It is a real problem that requires management, but there are tools available and the system managers seem to be on top of it.

FCAS

Any discussion of grid frequency usually involves FCAS at some point; or Frequency Control Ancillary Services. This is the power market services that maintain frequency and there’s a good summary of these here from AEMO. FCAS is a set of instructions that generators get paid for; either raise or lower your output in a given time and they will get paid for changing output, rather than the energy delivered.

FCAS goes up or down, and in different duration; 6 seconds, 60 seconds or 5 minutes. Different generators are able to participate in different markets, say a coal plant can’t raise quickly enough for the 6 second market, but hydro can so hydro competes in 6s, coal focuses on 5 minutes. There are multiple providers of the different services and they compete to provide at the cheapest price. This is how the grid responds to changes in demand.

The shortest time there is 6 seconds, which gives you an indicator of how quickly grid operators think they need to respond to network changes. That’s a legacy figure based on the amount of inertia we had previously. Lots of coal plants running, something changes, all that inertia means there are a few seconds to respond.

As noted above, as system inertia decreases, the time to move between points gets shorter; as more solar comes into the grid, inertia has decreased, so the grid responds more quickly to disturbance. As inertia decreases, we need to respond faster.

That’s where the proposed Fast Frequency Response market comes in. There is a proposal to add this market with the regulator at the moment, and it doesn’t appear materially different to the normal FCAS market, just a 1s response, rather than 6s.

It’s a control strategy outcome more than anything, and brings us back to Kate’s talk up the top. Yes, if we have less inertia we need to respond more quickly to system changes. Fortunately we can respond more quickly now! Modern inverters can do so much more than spinning machinery, and they can do it *very quickly*. Our battery in Goulburn can go from charging at full power to discharging at full power in 50ms, then back again in another 50ms. To go from off, to full power, to back off again would take about 16 hours in a coal plant, depending on how safely you want to shut it down. Their performance is scarcely comparable, which is why this is a problem now.

There are loads of technologies we can build a clean grid from scratch with, but what we actually have is a massive grid that has grown organically over 70 years or more. Inertia is a transition problem, not an end-state problem. We need to work with the generators that we have for the moment, but in the future, we’ll be able to do without them. Just need a bit of planning and management between here and there.

2020-2030 Renewable Energy Forecast – Australia

What better time than at the start of a decade to nail one’s colours to the mast and make some bold predictions. What follows is my forecast for renewable energy uptake in Australia for the next ten years, and what I think that means for the electricity sector as a whole.

This isn’t going to be a very sophisticated analysis, more just applying rates of change and multiplying them out to the end of the decade. I’m also going to work in capacities rather than any sort of fancy electricity market model. I’ll point out some outcomes of the capacity as we go through.

Start of 2020

As we wobble into the 2020s Australia’s electricity grid is definitely in transition, but we’ve got loads of work to do. Here’s the summary of the National Electricity Market, the NEM, which is just the east coast. For what it’s worth I think Western Australia will follow roughly the same route in the next decade, scaled proportionally.

Generator type	Number of Facilities	Capacity (MW)
Coal	16 (3 brown)	23,049 (4,690)
Gas	51 (8 reciprocating) (20 open-cycle) (11 combined-cycle) (2 steam) (10 coal-mine waste)	10,400 (304) (5,612) (2,428) (1,700) (277)
Distillate (diesel)	19 (includes the 2 temporary installations in SA)	1,379
Bioenergy	37 (6 biomass, mostly sugar cane bagasse) (31 biogas, mostly sewage methane)	433 (246) (187)
Hydro	59	7,712
Wind	69 (nice)	6,887
Solar	49	3,327

Snapshot of NEM connected generators at the start of 2020. Does not include behind the meter generators like solar on buildings and things like biogas generators. I know of at least 8MW of biogas on Sydney water sites for example that operate behind the meter and aren’t included above.

In addition to the above, we have a lot of solar on houses and businesses that don’t participate in the market, but still reduce electricity use. APVI is the best source for solar data, which hasn’t been updated from September yet, where they have total solar capacity across Australia at 13.9GW. Take out 15% for WA and that’s about 11.9GW in the NEM.

Batteries are just starting to be added to the grid and they’re hard to get data on. Solar data is quite good because there’s a registration process for each system to get the subsidy. There’s no parallel process for batteries so we don’t know how many systems are installed behind electricity meters around the country.

There are 215MW of grid scale batteries across five installations, which probably has about 300MWh of storage behind it. The vast bulk of the cells included are lithium-ion which are typically 1-2 hour duration chemistry.

Distributed across houses and businesses are a bunch of battery installations. There’s no solid data on this, but chatting with industry people the number is probably about 70,000 installations at about 3kW/9kWh average.

Solar

Australians love solar, and I don’t think this will ever change, because the fundamentals driving the solar market won’t change. As unlikely as it sounds, I am forecasting constant growth in solar for the next ten years, and because of the power of compound interest, this leads to some staggering outcomes.

Why do Australians love solar? I’ve been involved in some decent sized surveys of solar and battery customers, and the results were slightly surprising for me. At residential scale, and I can’t link to this data, the top three reasons offered for wanting solar were:

1. To save money
2. To reduce our carbon footprint
3. Because I hate my retailer

Solar on your rooftop is cost-effective against purchasing electricity from the grid, with paybacks available in the order of 5-years. There are two parts to this; the cost of solar, and the cost of grid power.

https://www.solarchoice.net.au/blog/solar-power-system-prices

Solar keeps getting cheaper, and I don’t see any reason for that to change on average. Grid power is getting more expensive, and even if it was sitting still the constant reduction in solar prices means there are more locations that solar works. Something drastic would have to happen to the solar industry and Australian electricity prices to break that trend.

On the other two reasons, it will be a long time before the grid is cleaner than a home solar system, and who thinks 2020 is the year that electricity retailers will suddenly become nice businesses to deal with?

The fundamental drivers for solar are here to stay. So I’m extrapolating growth rates of previous years out to 2030.

APVI has great data on the solar market in Australia, including installations by month all the way back to 2001. I’ve analysed these numbers here, and summarised below.

This data goes all the way back to April 2001, 19 years of monthly data. In the 224 monthly records I have the *lowest* month to month growth was 0.86% (I’m ignoring September because the latest entry tends to get revised up as more paperwork comes in). Monthly data is a bit lumpy to take seriously, so I’ve calculated the annual growth rate in the table. The lowest ever annual growth was 17.5% back in 2016.

If solar continues growing at the lowest annual rate recorded in almost 20 years, we’ll be at 70GW of solar in 2030, and a couple of other potential growth rates there for comparison. I need to pick a number though and:

I’m forecasting 80GW of solar in the NEM by 2030.

This will mostly be behind the meter, with large, sporadic additions of grid solar. Big numbers I know.

Wind

We’re at about 6.9GW of wind now. I’m less familiar with the drivers behind wind deployment, but let’s apply the same brute force technique and see what comes out. Using Open-NEM generation data this time

The early years can be safely ignored, and there were some dark times when Abbott et al were fighting culture wars about wind turbines, but somewhere around 10-15% annual growth feels about right, with some high and low years. Scale that out for ten years and here’s what we get.

That is an implied capacity factor across the fleet of 28% which seems reasonable too. I don’t know as much about wind, and it’s more impacted by policy uncertainty than residential solar, so I’m taking the 10% number.

At the end of 2030 we’ll have almost 18GW of wind in the NEM, making about 44TWh/year.

Energy Storage

We’ll need a lot of storage of various types to smooth all this variable generation, and how this comes to market over the decade will make a big difference to our clean energy goals.

I wrote my engineering thesis on energy storage way back in 2007, a technology assessment of the best storage to match with a wind turbine in a small island grid. The answer was batteries. Back then, the best choice was sodium-sulphur, which sadly never really got commercialised because people were scared of storing sodium at 300 degrees C. Cowards

But that analysis included cost too. Excluding cost, lithium-ion has the best performance characteristics for grid support, mostly in cycling frequency and power quality. Since writing my thesis we’ve had the mobile computing revolution and subsequent decline in lithium battery prices, to the point that lithium is almost always the cheapest option. Some of the more exotic chemistries, like the Toshiba titanates can handle higher power, but they’ll be niche applications only. Lithium will eat >95% of the storage market.

How fast will it come? In my experience, residential storage is not being sold on economics alone. I know how many Powerwalls were sold at the start and I know what the economics on that unit was, and people clearly weren’t making purely economic decisions. They were buying batteries because they like them and like the independence they provide. Again, that driver isn’t going away.

The available and likely market for residential storage is existing solar customers. They have solar, it’s a sunk cost, they’re exporting power and seeing it in their bills, and want to fix that with a battery.

Residential batteries are sold and installed by the same people selling and installing solar. They use the same electricians, sales and service teams. Batteries are just a new product, introduced to a sales channel with existing capabilities. There will be little to stop battery sales ramping up as new models become available.

For these reasons, my model assumes that battery sales will follow the *installation* rate of solar, from whenever the equivalent period was. At the end of 2017 I would have guessed 10,000 battery installations in Australia, which puts us at about the start of 2008 for solar installations.

These systems will tend towards 5kW/10kWh, with some bigger commercial systems dragging up the average. I’ll model them all as 5kW/10kWh though.

Using solar installation rates as a proxy for battery installations, starting from the 2008 data, with batteries starting in 2018 seems a bit fast now that I’ve modelled it. 10,000 installations becomes 100,000 in two years! I’ve decided to halve that instead. So battery installations are modelled as the same installation rate as solar in Australia, starting in 2008, applied to the start of 2018, but divided by two as a fudge. Even that feels high, with 100,000 battery installations forecast for Australia in 2020. I’m sticking with it anyway.

Multiply that out and in 2029 we end up with 5.6GW of battery storage, able to deliver that for 2-hours, for a total of 11,265MWh.

AEMO thinks we need 5GW by 2030, I think we’ll have that! Okay that number doesn’t feel too crazy.

Snowy 2.0

I haven’t included this in the model, because I honestly can’t see it being completed by 2030. It’s a big project! Lots of ways it can be delayed or made more expensive. I’ve also excluded it because I’m not sure what the impact will be. 2GW with a week of storage will probably be used to firm wind for some seasonal variation. Solar really drives daily charge-discharge cycles, which would be a bad use of Snowy.

I think the most interesting aspect of Snowy at the moment is the potential market power it might have. 2GW and a week of storage, flexing its muscles in a 5GW storage market is a significant degree of control. It will be very interesting to see how the government manages that.

What does all this mean?

If this much capacity does come on, what does that mean for the fossil fuels still in the grid?

I haven’t done a full time series model because it’s too hard and I don’t have the skills. But, I do have some rules of thumb I can apply to the grid with rigid modelling and crank out some outcomes.

Rule 1: Solar closes coal plants. This seems counter-intuitive to many because solar is intermittent; how does something intermittent close BASELOAD plant? By destroying its business case.

Once solar is installed, electricity comes out, that’s it. It doesn’t matter if the market is low, people aren’t home or there’s little demand, when the sun shines electricity comes out, and that peak output matches the sun’s peak output. That is, in the middle of every day. Solar pushes daytime demand (on the grid) down, lowering minimum grid demand. That minimum demand gives us an idea of what baseload plant can service. If the average daily minimum is 10GW, then there’s probably an economic case for about 10GW of coal (or steam-gas, like Torrens Island). Baseload plant does not want to turn off every day, or run at negative prices for 4 hours a day.

The minimum demand in the NEM is around 20GW, probably less, but I’ll leave it at 20GW to give the coalies a chance. The minimum demand served by coal is closer to 12GW. So I’m assuming that any coal above the minimum demand is uneconomic.

How much does solar lower minimum demand? Minimum demand is 20GW, with 14GW of solar in the NEM. I’ll assume that half of any additional solar lowers minimum demand. So if we ADD 6GW of solar, that will lower minimum demand by 3GW, and another 3GW of coal becomes uneconomic.

Again, the numbers are a bit crazy. That minimum demand isn’t really what I think the minimum grid demand is, more like a metric indicating the sort of demand we *could expect*. That’s relevant because baseload plant want regular output. If for say, 30% of days of the year they’re sure there’s zero demand, then that’s 30% of their possible generation days gone.

This is really just putting numbers to something I’ve believed for a while; coal-fired electricity will be uneconomic across the NEM by 2030. It probably won’t all be closed, but it will require special payments to remain open, either in electricity market distortions or straight-up government handouts.

I haven’t attributed wind here either, because its impact on daily minima is less predictable. Broadly, I think wind takes energy sales from coal and that could be argued using the generation data.

Black and brown coal are getting smaller over time, gas is fairly steady, wind and solar getting bigger.

A note on nuclear; we don’t have any in Australia, and I say we never will. Why? Cold hard economics. Nuclear is like an expensive coal plant. It uses broadly the same thermodynamic cycle, and is best deployed as a baseload generator, which means that any reduction in capacity factor impacts profits. There is no chance of a nuclear generator operating in Australia before 2030, and I think we’ll have 80GW of solar by then. Maybe it’s going to be less than that, but no one is going to bet $20b to find out in ten years if they were right about solar installation rates falling. 

Rule 2: Batteries close gas. They’re solving the same problem, filling in the gaps between variable renewables. Difference is that the batteries will be filled with free electricity, while gas needs $300/MWh to cover their fuel costs. Because the fuel is expensive, gas is driven by operational costs, not paying down capital. So they don’t run much, in the order of 10% capacity factor, waiting for the right market conditions to generate.

OpenNEM data again

Gas in the NEM is fairly stable over time, wobbling around 19TWh for the last 14 years. I’ll assume their potential market is 20TWh and work backwards from there.

Most batteries will be deployed shifting solar into the evening. Houses that showed no demand or export during the day with solar without a battery, will move to zero demand during the day, export in the afternoon once the battery is full, then the battery covers the evening peak until it runs out. This ramp is typically covered by gas, illustrated by this grab from Queensland yesterday:

Sorry the axis starts at 1 in the afternoon.

Moving left to right along the graph, solar is pumping at 1500, then tails off to zero by 1930. Combined-cycle gas is rising from 1500 as solar declines, then the open-cycle rises once solar is gone. Coincidentally, this is the time that prices spiked, shown in the second axis.

Batteries will dampen the rise of both of those generators, and push the need for them later and later into the night. The degree to which solar over-supply is shifted into the evening will determine the future of gas generators.

The impact of batteries across the network is complex. Houses that have solar and are exporting now, will add a battery and reduce their export, and reduce their peak in the afternoon, but their solar output will stay the same. In other areas, the battery will be export limited during the day, so the battery will liberate additional kWh from existing generators. These are the most cost effective deployments at the moment, so for this model I’ll assume that the electricity charging the batteries is from curtailed solar.

This is just capacity, there’s an interaction between the amount of deployed capacity and the degree to which it is utilised. Lots of batteries will be very well utilised; fully charged during the day and fully discharged overnight. Others will be oversized and never fully charge or discharge each day. Others still will do nothing at all for days on end as an electrical trip or some other problem means the battery is doing nothing. 1GWh doesn’t mean the grid gets 1GWh of effective storage.

For this model I’ll assume 20% of the potential daily battery capacity gets utilised to reduce gas use. If there is 1GWh of batteries deployed then the GWh they can steal from gas is 1GWh x 365 days x 0.2. I modelled 20% and it was a bit grim, so added 50% as comparison.

Well that one didn’t go as I’d hoped, maybe the next decade is the one for batteries. As modelled, batteries reduce gas use by 5-10% by the end of the decade. Wind will take some gas too but I haven’t modelled that yet.

What will impact this?

Solar

I don’t think anything can stop residential solar, apart from making it illegal. The economics are a run-away freight train and we’re all a lot better on the train than trying to stop it. I’m reluctant to type it, but I don’t think removing the small-scale renewables subsidy will slow the train very much. I’d like to see it stay though, as it gives a lot more control over a distributed power system, and is probably cost-effective anyway. It could be tweaked a little to get more of what we want, like prioritising west-facing panels, or inverters with smart control or voltage response, but I’m fine with leaving the subsidy.

Technical limits on the grid might slow uptake a little, but in my experience residential solar customers aren’t doing the maths in great detail. Export limits just reduce the economic case, but don’t impact the other reasons people like solar, so they’ll keep buying it.

Many will argue that “solar eats its own lunch”, and this will limit uptake. This view holds that because all solar produces power at the same time, the value of the electricity in the market reduces as more solar is added.

This only applies to market-exposed solar, plants that are connected to the grid just to sell electricity into the national market. In Australia we have almost 10GW of solar behind meters now, none of which is exposed to the market. Exposing behind the meter solar to the market requires significant market reforms that have been talked about for a decade and seem no closer to coming. This will have limited impact in Australia.

Further, the Australian electricity network is very expensive, so making electricity on site and avoiding paying for shipping it through the electricity network will always have an advantage. I’ve seen sites in the last year where the cost of solar on site is less than the network component of their electricity supply. That is, the electricity could be provided for free, but paying to transmit it through the network is more expensive than installing solar on the property.

The networks could change their tariffs to make solar completely unattractive, but I think they’re reluctant to push that button just yet. The networks have seemed to me to be walking a bit of a tightrope with residential solar; it’s an existential threat, but if they fight against it the threat will get worse and people will leave the grid. In the last 12 months I’ve seen their views shift more to viewing solar as an opportunity to be worked with, rather than against.

Wind

Network inaction could significantly limit wind installation rates, as grid restrictions make it impossible to build new plants. At some point they need to start building transmission infrastructure to open up new generation regions, but this is held up by Commonwelth of Australian Government (COAG) meetings and the processes behind that. There’s some movement now, but it could be very slow.

Conclusion

This clearly isn’t a perfect method but I’m sticking with it. Here’s where I think we’ll be in 2030.

Solar will keep getting installed at the rates we’ve seen in the last decade. Wind will go in fits and starts, but there’s enough projects in the pipeline that I still think we’ll hit that rate on average. Batteries are new and the economics look good in coming years, but it’s to be seen how much people want to buy them and how the market rules respond to them.

Someone take a pdf of this and sticky tape it to the front of the internet. Here is my prediction for 2030 and I’m not changing it.

Here’s the spreadsheet I did the calculations in.

Thanks for reading this far and let me know what you think in the comments. 

EB

 

 

Is nuclear a good investment?

In the much belated follow up to the eligibility criteria post on nuclear power, I consider the first merit criterion; is nuclear power a good investment? Is building a nuclear power plant a good way to turn a pile of money into a bigger pile? Or are there better ways of doing that?

Imagine you’ve been saving for a while and you have $50 billion burning a hole in your pocket and want to invest it in some clean electricity production. Nuclear looks okay, people keep telling you it’s clean and reliable, and everyone in the neighbourhood seems to have solar now as well, so might look at that too.

There’s a bit to consider; nuclear takes a long time to build, but once it’s running it produces electricity almost all the time. Solar is very quick to build, but only produces power when the sun shines. Which will give you more electricity from your investment?

In the first post I assumed an 8-year construction schedule, the Nuclear Energy Agency allows 4-8 years depending on where you are, and excluding permitting, finance raising and design. As this would be the first ever nuclear power plant built in Australia I think ten years is reasonable. It’s taken Hinkley Point, with full government support, at least two years to even decide if they’re going to build it and they’re expecting electricity in 2022 or so.

$50b is about the expected cost for the Hinkley Point reactor. For that price you get 3200MW, provided by two steam turbines. These will generate power almost all the time, for a modeled capacity factor of 85%. This might be a touch low, but it doesn’t make much difference in the model; the US hit about 91% capacity factor last year, a record, and their long term average is around 90%. France in comparison was closer to 80%, but nuclear makes up a much higher proportion of generation in France, so has to ramp up and down to follow the load. The construction period is about ten years, so construction costs are around $5b each year.

How much solar would that buy? I’ve been using $2/Watt as an estimate for about 18 months, and it’s probably a bit high now for rooftop where the installation market is very competitive. I’ve seen commercial systems go in recently at $1.1 – $1.4/Watt, but they have all been on flat rooftops. If you spend $5b a year on solar it will require more space than some factory roofs, so the cost of installing in a field is more accurate. These tend closer to $2.4/Watt, which is what I’ve modeled.

Solar’s capacity factor is quite easily determined using climate data, driven by maps like this from  the Bureau of Meteorology. Find the insolation, multiply it by the panel efficiency and it’s possible to calculate the average energy output per kW. Like this table. For the model I have assumed 4kWh/kW/day, slightly more than Sydney’s average, but a lot less than Brisbane, Cairns and Alice Springs. 4kWh/kW per day is an implied capacity factor of 17%.

Solar is much easier to build. Australia has been adding about 800MW each year for the last 5 years, without really trying. At $2.40/Watt, $5b will buy a touch over 2GW of capacity. That represents a pretty big increase in current installation rates, so maybe 2GW in the first year is a bit ambitious. If this were an actual 10-year program then your installation capabilities would ramp up and should easily install 2GW/year by the third or fourth. I have modeled a constant program of 2GW per year to keep it simple. Again this doesn’t make much difference to the result.

I also haven’t modeled any change in the price of solar, which is extremely generous to nuclear. Installed prices have roughly halved in the last 5-years, continuing the trend of the last 40-years. It would be quite reasonable to assume that the cost of solar would halve again during this period.

The price of nuclear hopefully doesn’t change during the build, although these plants have a bad record of hitting their construction quotes and timelines. This excellent article in Grist gives a summary of some studies into power projects and how often they run over time and budget. Nuclear is the worst, on average being more than 100% over budget, and almost 100% of projects running over time. Of generating technologies, solar is the most reliable, with average cost overrun of less than 10%, with 40% of projects running over schedule.

First the graph of capacity over time. Solar installs just over 2GW a year, nuclear installs nothing for ten years. This is just peak output though, nuclear can provide that around 90% of the time, solar less than 20%.

.

The interaction between capacity and capacity factor are shown in the annual output graph. Solar’s output increases each year for the first ten as new capacity is added, then decays over time as the panels degrade. I’ve modeled 1% reduction in capacity each year. Nuclear generates nothing for ten years, then 3200MW forever. Degrading at 1% solar’s annual output falls to match nuclear’s some time after year 30. The solar array would have replaced some panels and inverters by this time, but many of the original panels would still be working.

Solar gets a massive headstart through easy construction and modular design, which theoretically generates electricity from day 1. Even with the panels degrading and nuclear power’s famous high capacity factor, it doesn’t generate as much electricity as solar until about 80 years into this scenario.

I’ve just realised that I graphed Year in all three. MWh 

There’s a gentle curve in solar as the panels wear over time, but in anything less than a 50-year investment solar is going to make more electricity per dollar than nuclear, by some margin. Solar provides returns much faster and with much lower risk than nuclear, with the strong chance that construction will get cheaper throughout a program, rather than more expensive. The numbers used above are extremely favourable for nuclear, assuming that the nuclear project will be delivered on time and on budget, which is demonstrably unlikely, and assuming that solar will not experience any cost improvement, despite doing the same thing every year for ten years, with a product demonstrated to lower its manufactured price over time. Nuclear produces nowhere near as much electricity as solar on conservative estimates and it’s only going to get worse.

The absurd complexity in manufacture and design of nuclear power plants is a weakness that leads to severe cost and timeline uncertainty, while the opposite is true of solar. No one will build a nuclear power plant in Australia without significant government intervention. And it won’t just require money, but the political stamina to commit to a project of staggering cost and no discernible output for over a decade. It won’t happen. Nuclear has already lost.

Rational Gold-Plating

Much has been written and spoken about electricity prices, particularly the role of network prices in driving up household bills. I’ll admit I haven’t listened to this Background Briefing, but there is one point I have wanted to discuss for a while:

Federal Treasury estimates that 51 per cent of an average household bill is spent on network costs. Most of that is going towards paying off the $45 billion network companies have spent on updating our poles and wires over the last five years.

The definitive work on electricity pricing is the Australian Energy Market Commission’s 2013 Electricity Price Trends. From page ii onwards gives a good summary of the biggest price drivers in the last 12 to 24 months. Yes the various ‘green’ schemes have an impact, making up 17% of the average bill. These include legacy solar feed-in-tariffs, which look incredibly generous in hindsight. I think there are some houses in the ACT and Queensland still getting 60c/kWh, where at the moment about 20c/kWh will give you a 10-year payback. The carbon price contributed about 10%, the RET a bit less, and then the state efficiency schemes even less.

The real action is in network prices, making up almost exactly half of the average residential bill. This has risen in the order of 40% in the last few years, driven by the network investments. The question for today then is “Are network operators trying to rip us off?”

I have heard in a number of places now the charge that network companies are cynically using the network investment return rules; that when they build new networks they can claim a 10% return on investment through power prices. Network businesses are state owned or regulated, so the amount they can charge is regulated by a central agency, like IPART in NSW. The allegation at the centre of the Gold Plating charge is that knowing these rules and wanting to make money, the networks built unnecessarily large infrastructure. The Background Briefing cited above claims this is demonstrated by unused or oversized infrastructure, like substations that aren’t being used. This doesn’t look great, but I argue this isn’t as sinister as it appears and that network companies are making rational business decisions. Two things drive this; the time it takes to make investment and infrastructure decisions and the fixed costs associated with upgrades.

In 2006-7 I was working for Sydney Water, a massive electricity user in NSW. About 1% of state demand, or 2% if the desalination plant is running. During this time I was in meetings with the network provider about upgrades to the Sydney ring main. The ring main is the circuit of wires under Sydney that powers everything. It’s all underground, under roads and under buildings. Upgrading it is a big job, and it was going to happen one day. Wires can only carry so much current, and at some stage they must be upgraded as demand in the city grows. That time was coming and planning was well underway. So these meetings were going from about 2005 until the project finished in around 2012, about 7-years worth of planning. The result is a $400m or so project that greatly increased the capacity of the main and improved its reliability. If you’re interested I’m pretty sure the solid grey building near the light rail and Bellevue park near Central is the new substation associated with this project. Planning electricity upgrades takes a long time because it is a massive job.

So in 2006 they were nearing capacity of the network and wondering what to do about it. This conversation was being had all over the network; the initial build outs of the 70s and 80s were reaching capacity and needed to be replaced.

Considering the Sydney example again, how are the costs apportioned in a big job like this? I don’t know the specifics, but in smaller projects I’ve worked on, getting staff to site, foundation work, meetings and the big one, closing roads to do the work, can take up as much as 80% of the total cost. These are the fixed costs. It doesn’t matter what capacity you install, these costs are the same. The network infrastructure is 20%. Since it’s such a massive job just showing up, you might as well put the biggest cable you can in the hole. Even if you double the carrying capacity of the infrastructure, that only increases the project cost by 20%. This means there is a very strong technical incentive to oversize the infrastructure as it avoids those fixed costs.

Compounding this, how do you decide how big to make a network anyway? Residents don’t call the network provider and tell them they’re thinking of doubling their load with a new air conditioner and to some degree neither do industrial customers. The rational thing is to observe trends and make a guess. And the trends in electricity use in the National Electricity Market have side swiped virtually everyone.

This graph from The Conversation tells the story. If you were making electricity decisions in 2007 the data says electricity use grows every year, for almost the last 30 years (this trend goes back a long time). I did not see anyone, anywhere, predict the decline in electricity use that happened from 2009. Even two years into the decline the market operator was forecasting growth in the following years. Now though we have just finished the 4th straight year of declining electricity use.

Put all this together; it is a massive job upgrading electricity infrastructure, requiring disruption to supply, holes under roads and co-ordination of dozens of stakeholders. Of the cost of upgrading, a small component is impacted by deciding to go large, so network companies upgrade based on a 20-year forecast. The previous 20 years showed unrelenting growth and no one predicted the slide. What would you do? Exactly what the network companies did. Mitigate fixed costs by upgrading massively when you have to and assume the trend of the last 20 years is going to continue. The greatest risk then is an oversized network and increased cost; the greatest risk of an undersized network is blackouts, political tension and increased costs as the fixes are done urgently.

So I argue; network businesses made rational decisions with the information available to them. I agree there is an incentive to oversize based on the infrastructure return-on-investment rules, but this is dwarfed by the incentive to build big networks when you have to. Given the networks are state regulated there is good argument to reign in those investment rules, but I don’t think it would make much difference. Network businesses are large, unwieldy and rightly conservative organisations. As a result, we have some of the best electricity supply certainty in the world.

What can a consumer do about rising network prices? Leave the network. And the more people leave the greater the incentive for those remaining to go as network costs fall to a smaller customer base. This is something I am *very* interested in at the moment and can’t wait to see what happens.

A letter to the Federal LNP

I got angry about climate change again this week, in particular this stunt by NSW Liberal MPs. With my blood up after my first coffee for the day I wrote the below and sent it off. Maybe there are some points in there you’d like to take to your local members? Politics aside, the ACT’s response to climate change is the most appropriate in the country. It only appears radical because everyone else, particularly federal politicians, are doing so little.

 

Good morning Ms Goward,

I note with disappointment your media event and associated opposition to the ACT’s ambitious renewable energy plan. Two things in particular are disappointing and I would have thought below you. First, the idea that electricity should be produced in the region it is consumed, or any other product for that matter, is absurd and not how we have ever operated. Do the people of Lithgow complain about hosting our power plants? Do you drive out there and apologise from time to time? Do you feel guilt that the cheap power that has driven our economy is killing the residents of Morwell? A functioning economy has always depended on distributing costs and benefits and electricity generation is no different.

What I am most disappointed about though is the argument against the ACT’s principled stand. It appears from the article in the Canberra Times that you said “Ninety per cent by 2020 is really quite outrageous – it’s pandering totally to a green movement. It’s unrealistic, it’s impractical and wind turbines are notoriously unreliable as well.” This amounts to you asking the ACT to stop being ambitious. Their action is making you and the laggards in the LNP look bad. Rather then try and work harder on your own response you try and bring others down. It is disgusting and will be viewed with utter contempt in years to come.

You have an opportunity to do the right thing by your electorate and their children, by being ambitious and taking action. Yet you choose to bring others down instead. Party unity is one thing, doing the right thing something else entirely.

Why not think about the jobs in your region? Or the health benefits that come from reducing coal use and mining? The Goulburn region has a stunning wind resource, that can bring jobs and money to the region. While climate change ruins farms and rain patterns, you choose to do nothing and stand in the way of a new income stream for these properties.

I hope you change your mind and support the most ambitious and appropriate government response to climate change in Australia at the moment. I have not read a report yet that thinks Direct Action will work, and there are no penalties if it doesn’t. If you are serious about addressing climate change you can stand up and say something. If you are not, history will judge you very harshly.

I am available to speak further about this on [mobile number] and happy to discuss more appropriate policy responses to climate change if you desire.

Regards,
Evan Beaver

More like Urgh Hour

I’ve been complaining loudly this week about how ineffectual Earth Hour is and as a result how much I hate it. I despair that this massive, well organised and well funded marketing machine has set their sights on getting people to turn their light off for one hour, once a year. Anyway, this is an old hobby horse of mine, and if you want to hear more, follow this link to my TEDx talk from last year.

What can we do then? What concrete actions can we take, and could Earth Hour promote, that would contribute to de-carbonising our economy and making our lives more energy efficient?

Buy GreenPower. This commits your electricity retailer to purchasing enough renewable energy to cover your use. This is a cost and effectively subsidizes renewable electricity sources. It is additional to the mandatory renewable energy target and audited by a federal government agency. I trust it absolutely, noting GreenPower is a product, and many electricity retailers have “green” supply options, which might not all be renewables. If everyone was on GreenPower Australia’s decarbonising job would be done.

Contribute to a community renewables project or put solar on your roof. Community renewables include: Hepburn Wind, SolarShare who I volunteer with or this excellent project in Freemantle. If you want solar on your roof use an accredited installer sourced through the Clean Energy Council.

Buy more efficient appliances. Chances are if your fridge is more than ten years old you should replace it, and the energy saving will cover the cost of a new fridge in 6-10 years. You can compare the efficiency of all sorts of new appliances here and there are calculators to work out how much you can save. I encourage you to leave the sticker on these appliances, or move it to the back, so you can make comparisons later and understand how much energy different things use. The star system is an indication of how efficient a unit is compared to a standard, which is updated periodically. To get a feel for how much different appliances use, compare the kWh figure, the number printed in the middle of the sticker. As an example, our fridge uses 320kWh a year, while a big plasma TV is closer to 450.

Ride your bike more!

Make your house more efficient. I’ve got a blog post here on how to make a Canberra house more efficient, and here at the Conversation on how to control heat flows in your house.

 

 

Waste heat?

I thought this was a great piece this week by Glenn Platt on the anatomy of the Australian electricity network. I just wanted to expand on one point that I’ve been meaning to write about for a while.

“A typical coal-fired power station loses (or wastes) almost 70% of the energy that goes into it, when converting the energy in coal to electricity, and up to a further 10% is lost during the transmission and distribution stage. An old-fashioned light bulb then loses 98% of this energy to make light.”

These numbers are all of similar magnitude to what I have heard, except no one really uses incandescent lights anymore do they? What many people are surprised by in these figures is the 70% waste out of a coal fired power station. Surely there is something we can do with all that wasted energy?

The range of coal plant efficiencies in Australia is approximately 25% for the oldest plants to about 35% for the newest. The old plants could be improved with new technology, but the cost of retrofitting is so high that it is often cheaper to build a new plant. Or just keep operating the old one.

The theoretical maximum efficiency of any system that uses heat to create electricity is governed by the Carnot relationship, or even better the Chambadal-Novikov efficiency. Leaving out the maths, both of these relationships state that a conversion process is more efficient when the hot part is hotter and the cold part is colder. In a power station the hot part is the water in the coal boiler and is usually somewhere between 500 and 600 degrees Celsius. This temperature is limited by the materials available and their cost. An old unit might use copper or stainless steel, where a newer plant would be more exotic materials like titanium alloys. The material needs to be able to tolerate the 1000 degrees and more in the burning coal stream and not corrode due to the steam passing through it. When a steam tube fails it leaks steam into the boiler and once enough go there’s no point running the boiler any more. Shutting down and restarting a coal boiler is a two day job and so they don’t want tubes to fail.

The cold part is essentially ambient temperature, but power stations do a few cunning things with water to create a cold vacuum on the turbine and maintain efficiency. The huge curved concrete cooling towers we associate with power stations, both nuclear and coal, are one of the solutions to keeping water cool. This technology has generally been surpassed and newer plants favour fan forced cooling.

So the maximum heat in the steam tubes governs maximum efficiency, and materials science governs the maximum temperature.

There are however reasons beyond this that plants do not extract every last joule from their coal. Coal is essentially pure carbon, burning which gives CO2. But buried alongside the coal are other elements and compounds, some of which hinder the combustion process while others combust to give other damaging compounds. Water occurs in quite high concentrations in brown coal, as high as 60% in some places. Water, somewhat obviously, hinders the combustion process because some of the energy that should have been heating the steam pipes is actually heating the water in the coal. There are also sulphur compounds, which combust to give SO2, sulphur dioxide, one of the compounds responsible for acid rain. Acid rain forms when water and SO2 mix, and since there is water in the exhaust stream these can combine and form acid in the smoke stack, which causes big problems. To stop this happening they keep the stack temperature elevated, around 130 and above, to ensure that acid does not precipitate in their stack.

Others have regulatory restrictions placed on them which mandate minimum stack temperatures. I know of one plant whose state regulatory authority mandates an exit gas temperature above 150 degrees. This higher temperature helps the exhaust gas rise faster to be dispersed in the jet stream.

Coal plants aren’t running as efficiently as they theoretically could be, but there are good technical and economic reasons why they aren’t.

There aren’t many waste-heat opportunities in the electricity sector that haven’t been exploited to some degree. The best example of this ruthless efficiency is the combined cycle gas turbine. In this system a gas turbine, essentially a plane engine, burns gas and generates electricity. Carnot does apply here, but since the combustion is happening inside the engine the theoretical maximum efficiency is much higher. Practically though, an ‘open cycle’ gas turbine achieves somewhere around 40% efficiency. Gas turbines are mostly trying to convert kinetic energy, expanding gas, into electricity, so the gas that comes out is roaring hot. To capture this heat combined cycle gas turbines employ a steam boiler in the gas stream which also drives a turbine and creates electricity. This arrangement pushes the efficiency up to 60% and higher.

These efficiencies are possible with coal, but it’s difficult and currently cutting edge technology. In these proposed systems the coal is liquefied and sprayed into an internal combustion engine, conceptually similar to a diesel engine. In this case the hot bit is inside the cylinder, with no need to heat water, so the maximum possible efficiency is much higher, over 50% some have suggested. But this is hard to do; coal has all sorts of things in it that you would not want inside your tractor motor  and removing them costs energy.

It’s getting harder to see why anyone bothers burning the stuff at all.

 

TEDx Talk – a low energy use life is glorious

I did a TEDx Talk in Canberra recently. Here’s the video:

In this talk I wanted to get people excited about an energy efficient lifestyle. Rather than the Earth Hour view that to save the planet we need to go without I think an energy efficient lifestyle is actually pretty nice and this talk focuses on some of the co-benefits that come with energy efficiency, using some examples from my industrial career.

Let me know what you think.