The recent ABC Four Corners program “Power Failure” by Michael Brissenden began inauspiciously:

“By any measure Australia is an energy superpower. We’re the world’s largest exporter of coal, we have abundant sources of renewable energy and by 2020 Australia will be the largest exporter of liquefied natural gas. And yet, this apparently lucky country is now in the grip of an energy crisis.”

When I tell you that our roads are used by cyclists, cars and skateboarders, you probably have a good idea of the relative share of the traffic of each. And everybody would pick quickly that I missed trucks entirely. But almost nobody has any idea of the relative share of the energy system of the sources named; and the rest of the program failed to enlighten them. And I wonder, did anybody spot the energy export more than twice as big as liquefied natural gas that didn’t rate a mention?

According to the Australian Energy Update 2016, our total energy consumption in 2014-15 from renewables was 343 petajoules. A petajoule (PJ) is a large unit of energy. Most people are at best familiar with kilowatt-hours (kwh) from their electricity bills. A PJ is a rather large energy unit being equal to some 277 million kwh.

For comparison, our oil, coal and gas energy consumption was 2237, 1907, and 1431 petajoules respectively; and our energy exports were 13,088 PJs (mostly coal). Of the 343 petajoules from renewables, 186 were from biomass, and 48 were from hydro. Biomass might be renewable, but it isn’t necessarily low-carbon, and wood smoke has all the same toxic pollutants as cigarettes; it’s filthy stuff. Of the 343 PJs, wind came in at 41 and solar photovoltaic (PV) at just 21. People see solar panels on roof tops and don’t understand what a small piece of the energy puzzle they are looking at. Speaking very roughly, household electricity consumption is one-quarter of all electricity consumption, and electricity is one-quarter of all energy use and almost all energy use is from fossil fuels. Electricity certainly generates a disproportionate amount of CO2, but we have to solve the whole climate problem, not just the easy bits.

The other thing Australians, in particular, need to keep firmly in mind is that energy is not the be-all and end-all of our decarbonisation needs. We have more cattle than people in Australia and their methane will generate more warming over the next 20 years than all of our coal-fired power stations. And then there’s the land cleared, or kept cleared, to graze them.

In summary, wind and solar PV featured heavily in Brissenden’s story, despite only being worth 41 and 21 PJs respectively. And what of uranium? Remember uranium? Four Corners forgot it entirely.

The elephant in the room didn’t get a mention

We only produced 6110 tonnes of uranium in 2014-5, rather less than than the 67 million tonnes of coal we burned or the 392 million tonnes we exported. But what’s 6110 tonnes of uranium oxide in PJs? About 2592 PJ according to the Energy Update. That’s right, 123 times more energy than solar PV. But not all PJs are equal. Burning stuff produces around two PJs of heat for every one PJ of electricity. That’s not because engineers are incompetent, it’s a consequence of the laws of thermodynamics. Smart people use the heat for heating (fancy that!), desalination, or other industrial processes.

So if you used the uranium for electricity, rather than heat, then you can cut that 123 down to about 40. Meaning that our 6110 tonnes of uranium exports would produce about 40 times more electricity than our solar PV — plus a lot of heat.

A solid investigation of our energy problems would at least have have asked the obvious question: “Why don’t we use our uranium instead of just exporting it?”

So the Four Corners program began with a glaring omission.

Lakeland solar farm and other toys

Considerable time was devoted to the Lakeland solar farm without pointing out how tiny it is; it’s 10.8 megawatts (MW). The fact that it looked massive on the program is just because solar energy harvesting uses large amounts of land to harvest very little energy.

Brissenden interviewed Lakeland’s Christopher West who claimed, “What we’re creating is a base load power generator”.

A five-second look at the Lakeland specs (as reported to the Australian Renewable Energy Agency) shows that his claim is false; not just exaggerated, but false. Lakeland is 10.8 MW solar farm with a 1.4 MW/5.4 MWh battery. A 10.8 MW base-load power station would provide 10.8 MW continuously for every hour in the day; maintenance notwithstanding. When the sun stops shining on the Lakeland solar panels, it won’t provide 10.8 MW for any hours — not one. But if the battery happens to be full, then it could provide 1.4 MW for about four hours.

Clearly, talk of this being base load is a load of another sort entirely.

And then there’s the blackout

Think about a person riddled with metastatic tumours and weakened both by the tumours and large doses of chemotherapeutic drugs. When influenza finishes them off, how do you describe what killed them? The immediate cause of South Australia’s September 2016 blackout was the automatic shutting-down of the Heywood Interconnector when it needed to cover the loss of 456 megawatts of wind. The Australian Energy Market Operator (AEMO) has been worried about the stresses to the interconnector caused by renewables in SA for years. A 2014 renewable integration study wrote:

“High variability in non-synchronous generation in SA presents challenges in managing flows within required limits on the Heywood Interconnector. This can occur when non-synchronous generation varies by large amounts over short time frames, and the necessary balancing of the variation occurs via the interconnector.”

They came up with two strategies, firstly to upgrade the Heywood Interconnector to handle 650 MW, and secondly to limit the load on that interconnector to 250 MW. The second interconnector, Murraylink, is much smaller with capacity at just 220 MW.

At the time of the blackout, the two interconnectors were supplying 613 MW of the 1826 MW demand. The wind farms weren’t delivering; they’d slowed down because it was too windy. One hour before the blackout, they’d been supplying almost 1200 MW and within the space of 20 minutes, this slowing down had dropped the power supply by 340 MW to just 860 MW causing the interconnectors to make up the shortfall. This meant it was loaded well above the 250 MW safety limit. AEMO responded by dispatching more gas to try to reduce the interconnector flow. What does this mean? It means they were trying their best to implement their policy of leaving enough capacity on the interconnector to cater for the unforeseen.

This is similar to not travelling too close to the car in front because something might cause that car to brake suddenly. But the unforeseen happened before the interconnector had enough capacity to handle it. The graph below shows the kind of large wind variability that the AEMO had to deal with on that day.

Unfortunately, Brissenden’s coverage of this was limited. He didn’t ask anybody why the interconnector capacity between SA and the rest of the grid had been recently upgraded. He didn’t ask anybody why it was running so close to capacity on the day in question. He certainly didn’t ask anybody why it was so far above the 250 MW limit mentioned in the AEMO renewable integration report.

A European study in 2014 found that 100% renewable systems need interconnectivity to be increased by a factor of five to 10 to optimally shunt power from where it is to where it is needed. This is precisely because of the huge swings that must be stabilised by allowing increased flows.

The devil is in the detail

The appendices in the final AEMO report are particularly fascinating. They detail the conditions under which future blackouts may occur. Here’s a key result:

“These simulation case studies highlight that it is paramount for all SA wind farms to ride through a sufficient number of voltage disturbances in quick succession, to allow the Heywood Interconnector to stably and securely operate at an import level of 550 MW.”

Read this carefully: a robust system is one where multiple components can fail without causing a problem. The internet is a great example. A fragile system is one where a failure in any subsystem can bring the whole house of cards crashing down. The preceding quote makes it clear that when Heywood is heavily loaded, a single wind farm not operating according to specification could bring the system down.

Keep reading and the identified risks grow. If the interconnectors fail or are disconnected (as in September) what are the chances of a blackout.

“These calculations indicate that with the loss of the Heywood Interconnector there is at least 10% likelihood of an unexpected response of protective relays that could result in cascaded tripping across the system. During islanding conditions, the number of synchronous machines online would have a marginal impact on the available fault currents.”

Of ladders and renewable electricity grids

The standout feature of the appendices needs some explanation. Think about a company making five-metre-long ladders. The ladder has a number of rungs and each rung is in a different spot on the ladder. Ideally you want to get the rung-maker to deliver rungs within a certain tolerance and you randomly test a sample of rungs from each batch to ensure they meet specifications. You don’t want to have to test each rung in place on a ladder. You don’t want it to matter where in the ladder this particular rung might end up to do your testing.

Now compare this with our power system as described in the appendices. AEMO has a couple of computer simulation models and it can simulate the result of various types of failures of any component of the system. Except that the components aren’t like rungs in a ladder. Every single one is different and their position on the grid is critical to the simulation. So AEMO can’t just simulate a wind farm failure, they have to separately simulate the failure of each and every wind farm at it’s exact location and each and every combination of wind farm failures and each and every failure of the many connections between them; and combinations of failure. This is a mathematical and technical nightmare.

The difference between a traditional grid with a small number of fairly large turbines and this extraordinarily complex tangled ball of twine is vast. This complexity is a direct consequence of the distributed nature of renewable energy systems. Weatherill’s claim that renewable energy had no roll in the blackout showed not only that he doesn’t know what he’s talking about, but also that he doesn’t know that he doesn’t know.

In summary, you can get away with driving too close to the car in front for years. And then one day, a tree falls in front of the car in front and you smash into the back. Of course, you can always blame the damn tree.

If we are to appreciate problems of scale and complexity, we must explore the Adelaide blackout in great technical detail, and take into account the miniature size of the Lakeland farm.

Scaling problems

To appreciate why the size of Lakeland is so important, you need to think about what happens when you scale it up.

It’s wrong to assume that things that work at small scales will work at large scales. Bicycles are wonderful things for commuting and just plain having fun, but they aren’t much use transporting a 30 million-tonne wheat harvest.

What does a solar farm look like at the kind of scale that matters?

Here’s a few numbers using a South Korean APR1400 nuclear reactor as a benchmark. The APR1400 is, unsurprisingly, a 1400 megawatt nuclear reactor. You could fit the main buildings on the MCG. There is always a safety zone around a large nuclear plant, but it’s typically used by wildlife so it’s really a benefit, rather than a cost.

Now compare this with the Nyngan solar plant, Australia’s largest and about 10 times bigger than Lakeland. You’d need 47 Nyngans occupying an area equal to about 6030 Melbourne Cricket Grounds to generate the same amount of electricity during a year as an APR1400. And none of that construction area would be wildlife friendly; it’d just be a sea of steel frames and solar panels. Lay those MCGs end to end and you’d have a 41-lane highway stretching 1032 kilometers.

Next we need to add in storage. Lakeland has a Lithium-ion battery, but that’s too expensive at scale, so molten salt is far more realistic. This isn’t sea salt but a mix of potassium and sodium nitrate made from stuff dug up by miners and transformed in large chemical plants. For 12 hours of storage, you’d need 1.6 million tonnes of the stuff and the current global output of these is about 3 million tonnes (see here and here). And that’s just for 12 hours of storage. If you have a few cloudy days in a row, you may still be buggered.

Infinite sunshine but finite harvesting tool

Many people confuse the renewability of sunshine and wind with the resources used to harvest them. These resources are non-renewable and quite normal. They include habitat, steel, aluminium, concrete, trucks, and rare earth metals, not to mention millions of tonnes of battery chemicals. Note that I said habitat instead of land. There’s no such thing as just land, all land is habitat for something and making this explicit helps ensure it gets thought about. We can’t avoid appropriating habitat, but we should be minimising our appropriation, not maximising.

Scaling that works

Uranium as an energy source scales incredibly well because of its energy density. Meaning that you need very little mining for a whole lot of energy. That should be obvious from earlier parts of this article, but here are some more numbers. You need about 280 tonnes of uranium oxide to power a 1400 MW nuclear plant for a year, but the same sized coal plant will need about 1540 kilometers of coal train cars carrying 4.3 million tonnes of coal. We are talking 11,000 to 12,000 tonnes every single day.

So in contrast to the vast coal mines producing our local and export coal, and the vast network of gas wells and pipelines producing gas, our uranium comes from a few small mines. The biggest producer is Olympic dam in South Australia; and this is really just a copper mine. It produces about 220,000 tonnes of copper per year and about 4000 tonnes of uranium oxide.

Scaling that doesn’t work: batteries

Scaling problems are also particularly relevant to the various battery proposals Brissenden looked at in the program.

Tesla got a mention, as did zinc-bromide flow batteries. Let’s think about the battery scaling problems.

Elon Musk is building the world’s biggest battery factory at present in the US. When it is complete in 2020 it will have as much Li-Ion production capacity as the entire global Li-Ion industry had in 2013. It will then be able to supply half a million car batteries a year. And how many car batteries do we need? We produced 95 million motor vehicles in 2016; up from 91 million in 2015. Can you see the problem? … and the giga factory will have taken five years to build when it is finished.

Now think what will happen if we start putting batteries into houses? We will have less battery capacity to put in cars. We desperately need batteries for cars because we have so few other clean ways of powering them. It’s a really tough problem and we need to solve it. Putting batteries in houses just makes that incredibly tough problem even tougher. It’s like trying to keep our antibiotics for really sick people while patients with colds keep demanding prescriptions.

Brissenden presented Tesla driving software guru Simon Hackett with his great zinc-bromide flow batteries and allowed him spin wondrous yarns of a city full of these things, all connected by intelligent software. The world produces plenty of zinc, but bromide is produced in quite small quantities and zinc-bromide is produced in even smaller quantities. Zinc-bromide is a Class 8 corrosive marine pollutant and it’s dangerous enough that you can’t legally carry more than 1 litre of it at a time on a passenger plane. Which makes it fairly innocuous in the toxic world of battery chemistry, but it should be enough to warn you that scaling it up to solve our global energy problem will require more than just smiling nicely and saying how cheap it is.

Case studies in energy planning that have worked

Brissenden was right when he said that we have had a total failure of leadership on energy in Australia. Not because nobody has done any leading, but they have been leading us up the garden path.

Let’s look at a few examples of what intelligent leadership coupled with planning and foresight could have delivered.

  1. Aside from a little residual coal use, France decarbonised its electricity system in about 15 years using nuclear power between about 1975 and 1990. No other significant industrial country has come even close to this. The International Energy Agency measures this somewhat indirectly in tonnes of CO2 per terajoule of primary energy (tCO2/TPES). Between 1975 and 1990 France dropped its tCO2/TPES from 61.2 to 36.8. This shows that they’ve basically cleaned up half of their energy problem: electricity. They still need to deal with transport and direct industrial use of fossil fuels. In contrast, Germany during the first 14 years since its Renewable Energy Act in 2000 reduced her tCO2/TPES from 57.7 to 56.4. The numbers tell a truth that all the hype about the German renewable energy revolution can’t hide.

  2. In the same year that I switched from being anti-nuclear to pro-nuclear, 2009, the United Arab Emirates started on their nuclear program. Meaning they started to build the regulatory structures. The first reactor build was begun in July 2012. It will come on line soon this year. Remember that Lakeland solar plant that featured in the Four Corners report? Well the South Korean APR1400 nuclear plant will produce about 580 times as much energy each year (assuming a capacity factor of about 20% based on this story). In 2018, the second reactor will start up. That’s another 580 Lakelands worth of electricity. And again in 2019 and 2020.

And lastly we need to look at China.

  1. For the past 17 years, China has been rolling out what are called super-critical coal plants. These run at a higher temperature and are more efficient than older plants. But the increase in efficiency is a minor matter. In 2006, China announced that work on a high-temperature gas-cooled nuclear (HTR-PM) reactor was a priority. The demonstration plants, 2×105 MW reactors, are due to start operation later this year.

A coal plant is a boiler hooked up to a turbine and generator, which are hooked up to a grid connection. Unremarkably, the Chinese have designed the HTR-PM to be plug-compatible with the super-critical coal boilers. The designs have identical steam production characteristics. Disconnect the coal boiler and connect the nuclear boiler. You can keep all the rest of the infrastructure. This provides a rapid and fairly cheap way to decarbonise almost all of China’s electricity.

Sadly, the state of reporting on energy in Australia is almost as dismal as the state of policy.