More baseload would actually be bad for Australia

Even though Australia is demonstrably extremely bad at public debate, we are having a public debate right now about the future of our electricity grid. It’s a tough one, because few of the participants seem to really grasp the fundamentals and aren’t using the terms that experts all agree have an actual meaning.

The crux of the debate, for want of a better word, hinges on whether or not our grid needs baseload. This would almost be a valid argument if we were using the agreed definition of baseload, but we aren’t, so the argument is bad. Let’s start with the invalid view.

In many renderings of this discussion, Baseload is this special type of electricity that is available 24/7. Like this huge dam of potential electricity sitting behind the plug in the wall, waiting for you to flick the switch and let the baseload issue forth. Baseload is the good electricity, ready to go whenever you are, as happy and enthusiastic to be called into action at 3am as at 3pm. Baseload is the loyal Golden Retriever of electricity, wagging its tail, waiting for your bidding.

But both parts of this view are wrong. There is no such thing as extra electricity, waiting in the system to be used at a moment’s notice. If you demand more electricity an actual electricity generator must increase generation to meet that demand. Supply and demand in the network must match all the time. If they don’t stuff breaks, and we don’t want that. Better an hour’s inconvenience than 6 months of replacing all the generators on the east coast.

Knowing that supply must meet demand, ramping up and down during our daily cycle, coal plants are extremely not suited for making sure the lights stay on. In times of rapidly changing demand, such as the South Australia storm and subsequent blackout, more baseload generators would not have helped the situation. They can’t change generation quickly and don’t want to. No number of coal plants in South Australia would have prevented that blackout.

This is the real definition of baseload: plants that don’t want to turn off, up or down, ever. They want to output at exactly the same power all the time, and any deviation from this costs money. This is due to a genuine mechanical engineering limit; if you have a turbine you have a best efficiency point and any deviation away from this point is inefficiency.

So the old school, centralised-generator way of running the grid was this: run the coal plants flat out, 24/7, that’s your baseload, at something like 15GW on the Australian east coast. Then when people wake up in the morning, make breakfast and start their business for the day, and demand rises from 15GW to 25GW, bring in the easily dispatchable, variable generators. Snowy and other hydro has performed the bulk of this role, gas turbines have picked up more in recent years. Ideally the baseload production matches the minimum demand. Raise the minimum demand and you can run more baseload, hence off-peak hot water heaters.

Those two points are key to understanding why coal (and nuclear because it’s also a steam cycle) are utterly incompatible with a modern grid.

1. Supply must change to meet demand.
2. Baseload generators do not want to change output (and have a technically limited ability to do so).

Coming back to the public debate, we have a “problem” or even a “crisis” in the Australian electricity network, that Needs Fixing. I agree there is a problem, or more a challenge to be solved, but that’s a matter of perspective I guess.

Trying for an unbiased appraisal of the state of affairs, I’ll define our challenge as “maintaining grid reliability, while keeping prices as low as possible”. I’ll also add an additional caveat the luddites will disagree with “while finding lowest-cost CO2 reductions”. This is the trilemma; cost, reliability and emissions.

We need to think about this and actually make decisions, because the current infrastructure is old and failing in places. Like every machine, it is wearing out and bits need replacing. Coal generators come in big lumps of generation, at 500MW and up. Loy Yang A in the La Trobe is 2200MW. Replacing that capacity with 5MW wind turbines would require about 440. But Loy Yang is about 30 years old and burns horrible brown coal that is going to kill us all one way or another. When a big plant fails, what are we going to replace it with?

A big power station will be expected to last 40 years and more, so decisions about future profitability need to be based on a forecast over a long time. The cost of building any power plant is expected to be recovered through the electricity they make and sell. More electricity means the cost of construction is divided out among more units, so they make more money. Coal plants are expected to run with a capacity factor of 0.9, where 1 is running flat-out, perfectly, all the time.

One of the strongest and most remarkable trends in the Australian energy market is the uptake of rooftop solar on houses. Since 2010 the installed capacity has just kept going up and I don’t think it’s ever going to stop. Solar is in a virtuous cycle of development, where each iteration makes the next cheaper. Someone buys a solar panel. More manufacturing capacity is added to meet that need and the cost comes down. As a result of the lower price, someone else buys solar, more capacity is added and the cost comes down. Tony Seba is excellent on this topic.

Australians absolutely love solar and it would have to be made completely illegal to stop it now. And even if someone went mad and tried that it would just accelerate households’ departure from the grid. In some places not joining the grid is cost effective right now. As solar and batteries get cheaper, there will be more and more of these places.

There is 6.5GW of solar installed on almost 1.8 million houses, adding about 80MW per month for the last 3 years, but the trend looks up. I don’t think we’ll see a year of below 1GW installations ever again. If it takes 4-years to build a 2GW coal plant, the capacity will be replaced twice just during construction, by something with free fuel.

Obviously the sun doesn’t shine at night, but no one cares. Power generated on your roof doesn’t need to be transported through the network, and you don’t have to do anything at all to get electricity, so it is much cheaper than grid power. Every day the sun comes up and on a good day 6GW pours into the grid from households. Next year it will be 7GW, then 8GW. Cost drops to the point where every warehouse gets covered in panels. 2GW per year becomes normal, and in 5 years we have 15GW of solar on the east coast. In the 40-year lifetime of a coal plant of 2,000MW output, we might add 60,000MW of solar, which the coal can’t compete with because it needs to use a network to transport the power there.

This hollows out demand for electricity in the middle of the day, to the point where electricity has zero, or even negative cost. Already we have seen mild autumn days in Brisbane where the market price has gone below zero. More electricity than we know what to do with. We call this the Duck Curve.

In this market a coal plant then has to sell their power at a low or negative price, or turn off and then turn back on again, which takes about 36 hours. Either way, that is less money for the coal plant, and it will only get worse as more solar keeps going in. All of this is repeated for nuclear, which is based on the same basic thermal process. Baseload generators will get absolutely thrashed in a market with increasing penetration of solar. If your coal plant is designed to run at capacity factor of 0.9, but actually runs at 0.45, then the electricity costs twice as much.

In parallel with the development of solar, electricity generated by the wind keeps getting cheaper, creating another competitor for coal, with zero cost of production. Wind is less time-of-day predictable than solar, but the economics are the same. The wind blows and electricity comes out. There are no fuel costs, so the marginal cost of production is zero. A sunny day followed by a windy evening can already lead to 16 hours of very low electricity prices.

What’s left is a competition between generators to fill in the gaps, which will be fought by hydro, pumped and dammed, open-cycle gas turbines and batteries on the grid and in houses. There might be some exotics in energy storage, but I doubt we’ll see much more, at scale, in the next decade. There is a limited suite of generators that can meet this changing load.

Australia has a bit of hydro already, in the Snowy Mountains, Tasmania and a couple of other spots along the Great Dividing Range. We have lots of opportunity for more, and Andrew Blakers at ANU has some great work on this. But hydro power needs to be transported through the network like coal power, so a system behind a customer’s meter will be cheaper. Pumped hydro is a big project too, meaning long planning and approval times, and a very long lived asset at the end.

Open-cycle gas turbines are inefficient and polluting, and their cost is largely driven by the cost of gas, which is going through the roof. Gas has been declining recently and I don’t see much changing that soon. There will be some big dramas in the Australian gas market in the next few years.

That leaves batteries, at grid scale and in homes and eventually businesses. Grid and home batteries are solving two different problems, using the same technology. On the grid the goal is charging and discharging to stabilise the grid. The battery can turn on and off much faster than any other generator, so it can manage the changes in voltage and frequency that happen faster than other generators can respond to.

In the home, batteries are about moving solar consumption later in the day. Rather than exporting unused power in the middle of the day, it is stored for use at night, cooking dinner and watching tele, when demand rises again. I fully expect battery installations in homes to follow the curve of solar installations, as very similar economics are at play, and the installation structures that already exist for solar can be used for batteries as well, giving the industry a head start. Much like residential solar, resi batteries are very easy to install, and despite the technical advantages of pumped hydro, I expect residential batteries will eat the pie that pumped hydro wants to eat. Consider the effort required to build a billion dollar pumped hydro facility. No number of household decisions to invest will get one built. But if a household wants to invest in a battery, it is very easy and fast. Call a solar company and ask for a battery, then they install a battery.

A million households with a 5kW battery is the equivalent of a 5GW power plant that can run for two hours or so, every day. The solar curve will get flattened and stretched out. Rather than negative demand during the day, there will be zero demand, as batteries charge and smart devices watch the market. The sun goes down and the battery discharges. Batteries get bigger and cheaper. They last for two hours, then three, zero demand stretches until 5pm, then 6 and 7, and before long they last overnight. Already at our house we have gone days without buying power from the grid during summer, with a 6kW solar system and 6.5kWh of storage.

Demand on the grid dwindles until there is basically nothing left. Just one freak day a year when the sun hasn’t shone for days and the wind has stopped blowing. And the coal plant missed its chance to sell some electricity, because they forgot to start the plant two days ago.