Draft ISP 2024 “Hot take”

Like many of you other nerds I’ve been poring over the recent draft Integrated System Plan (ISP) release and listened to the interesting interview with Merryn York on Energy Insiders. I wish Merryn had more time to talk about the weather modelling – but it probably needs explanation with charts. At the end of the interview, she said AEMO are open for feedback, so here are a few of my notes. I’m not going to focus on generation or transmission which seem pretty uncontroversial at this point. Most people would say “yeah the future will look sort of like that” for those sections. Which is good – it’s helpful to have a rough map (it’s not a mandated plan or forecast) of the general direction we are heading.

However, some of the stuff outside the electricity industry is a bit… unrefined. This could do with some more focus before the next ISP as the assumptions have a strong impact on the buildout in the electricity sector.

It starts with a dissonance in how ISP assumptions are imported into the modelling. AEMO asks different consultants to come up with forecasts for a bunch of different technologies before bringing them together and filling in the majority. This kind of process doesn’t allow an equilibrium between competing options and a winner gets picked. Take residential solar and batteries – the growth assumptions behind the ISP expect a massive reduction in install costs, without harmonisation with large scale solar and battery costs. Small-scale systems are assumed to be cheaper than large-scale systems which is not likely, then a huge amount of small-scale is assumed to be installed, with large-scale solar/battery merely filling in the gaps. The rest of the model is at least harmonised by AEMO after a few weird inputs, and they end up with more rational outcomes, like how they basically dismiss offshore wind as too expensive and it only makes an appearance because Victoria is expected to subsidise some.

The same input collection process is driving the expectation of outcomes in other segments that are not economically competitive. The principal one is hydrogen demand. Let’s look at a few of the input streams.

Here we have the gas fuel assumptions. They come from three sources, ACIL Allen does the fossil methane forecasts, and CSIRO did the other two separately.

The methane forecasts are based on the expectation of a large reduction in global oil demand, and gas price being oil-linked. They used several international sources to estimate the price suppression in such a scenario. Gas generation pays a bit more than industry because they source it less consistently, and sometimes at short notice. That all sounds pretty reasonable.

AEMO doesn’t talk about biomethane in the ISP so I’m guessing they just wrote it off which is totally fair. The assumed reductions in cost by CSIRO are based on public subsidisation of biomethane processing facilities. Biomethane processing is only practised experimentally because the economics are dreadful. Biogas is usually only available in modest quantities and is geographically dispersed. This means that you need an expensive processing facility relative to the volumes processed. Almost all biogas is used without processing - burned directly for process heat or carbon credits, or in a gas engine to produce electricity. Scaling biogas is very challenging. Germany is a good example – they successfully collect and digest most household organic waste at huge scales, as well as feeding in energy crops to bioreactors. The result is just 2% of their gas consumption, nearly all turned into electricity and not used as a replacement for gas. In short – biomethane is not going to be a major energy source.

They CSIRO hydrogen assumptions are that electrolysers get 5x cheaper and massively improve efficiency. The efficiency part is believable if Hysata is successful. I’m more doubtful of the capex reduction – electrolysers are century old technology. Neither assumption is ridiculous, but they are perhaps heroic. Also, why do academics love PEM so much? It’s picked as the winner all the time despite the fact that alkaline electrolysers are cheaper, more efficient, and more flexible.

Despite optimistic assumptions, when you look at the cost of the alternative gases you can immediately tell that they are never economically viable. Isn’t it a bit weird that so much hydrogen gets used so much then?

There are several segments of hydrogen demand we should scrutinise.

Ammonia

The need for green hydrogen for ammonia is a given – there is no alternative if we want to feed billions of people without carbon dioxide emissions.

Iron

Reducing iron ore with hydrogen is the expected pathway to decarbonisation, but direct electrolysis is still in the running.

For both ammonia and iron, Australia is assuming that hydrogen is our ticket to large export earnings. We need a reality check on how likely that is. State and federal governments have earmarked billions in spending on hydrogen projects that are not likely to lead to viable export opportunities unless there are significant and global prices on carbon. Even if that eventuates, how do we know that Australia will be the one to produce those materials? We have not checked our competitiveness against other countries. Yes, we have large renewable resources relative to our domestic demand, but that does not determine where future industries will locate – price does. I would bet that Texas or Canada would be more attractive places for green industry, heck once Argentina starts using USD it could be a cheaper option than Australia.

Heat

There is a large (55 TWh p.a.) demand for “emerging hydrogen” excluding export industries. I assume most of this is supposed to be for industrial heat.

Here’s the hot take (pun intended) – hydrogen heat is going to decline by 2050. Currently a bunch of hydrogen is burned as a byproduct of oil refining. That oil refining will decline AND their free hydrogen will be used for better purposes in the meantime.

Hydrogen is utterly uncompetitive in heat and the conditions it needs to become competitive also advantage the alternatives (cheap electricity).

The base case for heat decarbonisation should be direct electrification. It is low capex, reliable, more efficient than combustion, and physically capable of delivering any temperature.

For low temperature applications like hot water, heat pumps are a safe bet. They are much more efficient over low temperature deltas, but they cost more than direct electrification.

If you have volatile electricity prices, thermal energy storage is your best option as you can just charge up on cheap offpeak. With flat electricity prices direct electrification is cheaper.

This last point leads me to a conflict in the ISP. The operational demand profiles are expected to flatten out over time, thanks to lots of storage and controlled loads.

However, storage and controlled loads need volatility to be viable. Flat operational demand is not an equilibrium outcome because storage and controlled loads will stop being deployed OR renewables will deploy more because prices are stable.

Transport

The question that drives the fuel cell vehicle assumptions is the wrong one. “How many fuel cell EVs could there be in future?” leads to some large positive and overly optimistic assumption. If you instead ask, “what kind of vehicles will people drive?” you get a different result. I definitely don’t believe that hundreds of FCEVs will arrive in Australia next year, growing more and more over time. There’s nowhere to fill up, no plans for hydrogen stations, and the cars cost more than EVs with fuel that costs more than petrol. You can’t even get the cars. I had a look on the websites of Toyota and Hyundai and while ostensibly “available” you can’t be directed to a dealer or shown a price for a fuel cell car, which is easy to do for every other model.

On the other hand, there is a missing demand for hydrogen in alternative fuels. Currently several pathways to sustainable aviation fuel (SAF) use hydrogenation. This is a more likely success story for hydrogen in the short term because there is mandated demand for SAF in other markets, and the green premium is smaller.

Raw hydrogen

In the Step Change and Green Exports scenarios hydrogen is exported as is in huge volumes. This is much less likely than hydrogen derivatives because of how difficult hydrogen is to transport.

Why does this matter?

We are exaggerating the amount of energy needed to be delivered by the NEM, and therefore exaggerating the amount of infrastructure required to be built. We are talking about 70 TWh p.a. in the Step Change and 2,300 TWh (!) in the Green Exports scenario.

So that’s good news, this is all a lot easier without wasted energy in hydrogen processes.

Picking hydrogen as a winner in the ISP also justifies state and federal governments to continue their infatuation with it. The budgets for hydrogen dwarf all other industrial decarbonisation efforts combined. That’s not to say hydrogen spending should be stopped. We’ve successfully demonstrated that FCEVs are terrible in several states for only a few million dollars. Funding breakthrough tech like Hysata is a small cost with potentially gigantic benefits. I’d love to see a green ammonia or iron project demo. But that leaves a couple of billion dollars that can be saved or diverted to more efficient decarbonisation efforts.

We also should take the chance to do a global competitiveness check. Whilst we have huge renewable resources, we aren't the lowest cost jurisdiction - what can we do to improve that?