A gas lead recovery from Covid-19 – Pipe dream or Hot Air?

First apologies for the two dad jokes in the one headline – but how often do you get a chance like that😊!

The past few weeks have seen a flurry of headlines regarding a gas lead recovery. Including a proposal for new pipelines, a new gas hub styled on the “henry hub” for natural gas in the US and even including potentially a new 1,000 MW gas fired power station in the Hunter to replace the capacity lost when the Liddel coal fired plant is closed by AGL.

Like much of Australia’s electricity policy – these proposals owe more to ideology than good policy.

In this context, it is useful to delve into the potential role of gas fired generation within Australia’s electricity supply mix. You can think of electricity in two components – bulk energy (aka baseload) and dispatchable energy. Bulk energy is the majority of the electricity used the majority of the time. The key feature of bulk energy is that it needs to be cheap. Dispatchable energy is sources of supply that can be ramped up or down to respond to peaks in demand (or for that matter shortfalls in supply – such as when the wind isn’t blowing, or sun isn’t shining).

Gas is not, and never will be, an economically competitive source of bulk energy. The cost hierarchy for bulk energy is renewables, then coal and then gas. The emissions hierarchy is renewables, then gas with around half the emissions of coal, and then coal. Variable Renewable Energy (VRE) simply dominates. Where potentially gas had a role 10 years ago as a transitional fuel. Gas potentially had a role 10 years ago as a transitional fuel (when emissions could have been reduced by reducing coal usage in the 2000s) prior to the scaling up of VRE – that opportunity has now passed. If you want cheap bulk energy – VRE is both the cheapest source and zero emissions – new build gas-fired power stations have no role for bulk energy.

To illustrate this, the following table shows the levelised cost of energy for a combined cycle gas plant based on a $6/GJ gas cost.

At $6/GJ gas, a gas CCGT has a levelised cost of $0.06-$0.08/kWh. The compares with $0.05-0.06/kWh for VRE. For example, the ACT recently signed a PPA with Neoen at $0.045/kWh (fixed nominal) for its latest SA project.

The key problem for gas plants is that gas is just too expensive. It is well over 50% of the levelised cost. For a gas plant to operate at 0.05/KWh – that is to be price competitive with renewables – it would need a gas price of $1.26-$4.01 per GJ (using the scenarios above). While gas prices might get down to that in the middle of the Covid-19 energy price slump – it just isn’t realistic to assume prices will be that low for the 25 year life of a new build gas plant (particularly including the impact of future carbon taxes/emissions restrictions).

That leaves the potential opportunity for gas as a source of dispatchable energy. That is, using gas fired peaking plants to support electricity use in the evening peaks and for instances where VRE supply falls short.

In this role, gas faces two main competitors[1]:

· Flexibility from legacy coal plants. Over the next decade or so, electricity supply in Australia will continue to be dominated by coal. These plants are not as flexible as gas peakers. That is, they find it difficult to ramp up or down quickly (and increased ramping may result in more frequent breakdowns and outages), but they are not completely inflexible. This means the sunk cost we already have in coal plants is likely to be a low marginal cost source of dispatchable capacity over the next decade or so. In particular, where a new gas fired plant needs to have a strategy for recovering its capital cost – an existing coal plant does not.

· Battery/pump hydro storage. These are technologies that can take excess VRE energy during periods of high production and deliver it during periods of peak demand/low VRE production. Batteries are likely to be most competitive for short durations – ie shifting daytime solar into the evening peak. Pump hydro looks more competitive for long-durations – that is dealing with seasonal demand peaks as well as extended periods of low wind. The key challenge for anyone investing in a new gas fired plant today is that batteries are likely to get cheaper. For the investment to make sense, it needs to be competitive versus batteries not just today, but over the next 20-30 years.

To illustrate this the following tables compares the economics of an open cycle gas peaker with a 2 hour battery as a source of 2 hours of daily dispatchable capacity. To keep the tables straight‑forward, the analysis of the gas peak assumes a gas price of $6 per GJ. However, it is important to recognise that gas only represents around a third of the cost of a gas fired plant. The majority of the cost is driven by the cost of capital. In particular, even if gas was free, a gas fired plant needs to earn $0.10/KWh to $0.14/KWh to cover its capital cost.

The battery analysis considers two examples.One based on today’s battery prices.The second is based on a battery price of $200/KWh (which is in line with broadly accepted forecasts of where automotive battery prices are likely to get to over the next decade).While a 50% or so fall in battery prices may seem ambitious – it is actually quite achievable in the context of renewable energy scale economies.Specifically - in the case of batteries – Infradebt has seen the capital cost of utility scale batteries fall by around 50% over the last two years.And for a longer-term example see the history of solar cell costs in the chart below.

In simple terms, the cost of dispatchable power from a battery, at $0.23/kWh in this example, is currently higher than the 25 year levelised cost of a gas peaker ($0.18-$0.21/kWh). However, the cost advantage is reasonably small – at 10-20%. And importantly, this analysis focuses solely on peaking capacity/energy arbitrage (supply of energy at peak demand), it completely ignores revenues derived from Frequency Control and Ancillary Support (FCAS) services (grid stability services). As the Hornsdale Power Reserve shows, FCAS revenue can be meaningful – in the first two years revenues have exceeded the cost of the Hornsdale battery (https://ieefa.org/big-battery-in-australia-proves-profitable-as-neoen-recovers-capital-costs-in-just-two-years/ ). If the forecast battery price of $200/KWh is achieved, then the battery would smash the peaker with around half the levelised costs. And interesting enough, the price of vehicle batteries (a good proxy for utility scale storgare) is circa AUD230 kwh today according to Bloomberg New Energy Finance (https://about.bnef.com/blog/battery-pack-prices-fall-as-market-ramps-up-with-market-average-at-156-kwh-in-2019/).

This underpins my expectation that a gas peaker built today won’t get to operate for its full operating life. This is important – if you reduce the operating life assumption for the gas plant to 10 years – its levelised cost increases $0.25 to $0.31 (or $0.17 to $0.24 on a capital cost only basis). That is, if you acknowledge the likelihood that the gas plant will have a short operating life, then batteries are already cheaper as a source of dispatchable power.

And it is for this reason that the Government’s pro gas agenda is likely to receive a lukewarm reception from investors. In particular, a new gas plant is likely to require a subsidy from government, where new batteries are being developed on a subsidy free basis.

If the government wanted to kick start the delivery of dispatchable power (and get ahead of the inevitable coal plant retirements over the decades ahead) they would be better to:

· Invest effort in getting the market design right to reward those who deliver sustainable dispatchable capacity to the market. Under the current energy only market it is difficult for market participants to be appropriately rewarded for dispatchability. Importantly, any market mechanism needs to focus on appropriate signals for sustainable dispatchable capacity. That is distinguish between subsidising existing fossil fuel plant (ie rewarding capital that is already sunk) versus providing appropriate price signals for new dispatchable capacity.

· Run availability payment auctions for new dispatchable capacity that are aimed at bringing forward new investment in dispatchable capacity. These could involve a range of technologies from utility scale batteries, pump hydro, fleets of household level batteries or even demand response. That is, spend government money on bringing forward the future, rather than handouts to bring back the past – perhaps David Rowe’s cartoon from the AFR illustrates it best.

[1] A third source of dispatchable energy is demand response which is beyond the scope of this article because it is difficult to estimate the marginal cost of adjusting demand.

© 2018 Infradebt