Thinking about our winters of discontent in the 2050s

The debate around the cost, viability and regulatory framework that decides on new transmission investment is hitting a nerve. And that is because it touches upon much deeper issues for the future of the electricity system.

The decision on how much and whether to proceed with a new 800MW transmission line between South Australia and NSW is not really just about the specifics of this project. It will start the complete re-design of how the grid will work in the future. And the biggest questions of all are is this the right design and what are the credible alternatives?

The cost of the project has now blown out to $2.4 billion, 20 per cent higher than the numbers used for its cost-benefit analysis and well above the regulatory sign off cost of $1.5 billion.

These seem big numbers, until you look forward at what underpins where this thinking is going.

AEMO’s 2020 Integrated System Plan (ISP) modestly outlined the scale of new investments needed for a renewables based electricity system through to 2040: 26,000MW of new renewables, a doubling or tripling of rooftop solar, up to 19,000MW of new storage or dispatch able generation and $25 billion of additional transmission to do two things: bring the renewables to the main urban markets and move much larger licks of energy between states.

To understand this NEM Risk Bulletin has looked further forward, to the 2050s when to meet the proposed net zero emissions target the electricity system will need to be pure renewables. To get an idea of how much generation is needed, we have taken the generation and demand data from the winter of 2019 for the four mainland states – NSW, Queensland, Victoria and South Australia.

For simplicity we have assumed demand remains constant in each state until 2050, which means continued energy efficiency cancels out continued population and other demand growth. We assumed that gas generation would be phased out, and that nuclear remains either too expensive, too illegal or both. Hydrogen is part of the mix of storage technologies, along with pumped hydro and batteries. There may be some new unicorn technology that arrives over the next 30 years, but as we don’t know what it looks like, we can’t include it in our simple model.

Using the fluctuations in renewable generation and demand as a reasonable proxy for what will be needed over the winters of the future, we then extrapolated how much more renewables would be needed to run the NEM during the winters of the 2050s, when all the energy would come from wind and solar.

All renewables NSW, Qld, Vic, SA June-August 2019 Source: NEM Review

The aggregate renewables chart for winter 2050 is a busy chart. Winter is typically the time of lowest generation for renewables systems, as solar output is reduced by shorter and cloudier days and wind speeds drop as winter settles in. It is the time of the year that would most constrain a renewables-only grid.

Renewables capacity 2019

We then solved for solve for the same demand pattern in 2050, using only wind, large scale solar and distributed solar as the sources of energy. Assuming a constant build rate of each technology in each jurisdiction, we estimated this would require roughly six times more renewables: increasing total capacity from 18,096MW to around 108,000MW. That is a lot.

Projected renewables build for net zero emissions by 2050

This is, of course, a relatively simplistic projection. It is possible that some state build more or less wind or solar PV over the next 30 years. We have assumed Tasmania will be like a giant battery: a net exporter at whatever the size of the transmission capacity it has to the mainland. There are likely to be geographic constraints: Victoria may struggle to build that much wind, and we have assumed six times more DER which means very large rooftop solar PV on literally every available roof in Australia, which is possible but extreme.

The point is if we are going to talk about net zero by 2050 and a renewables grid, then it’s useful to see what it ends up looking like. There are a number of important trends which appear likely to hold regardless of any of the possible changes in location or generation mix.

South Australia net exports and imports by 2050

1. South Australia becomes a mega- exporter

South Australia has smaller demand but more space, so if we stay on current trend it is likely to be the biggest exporter of electricity by 2050. By a country mile. Under this “six times larger” scenario SA hardly ever needs to import or firm its generation because it has so much wind that even at low levels it can mostly cover its demand. There are times like windy, sunny days when it will need to be finding a home for around 10,000MW of surplus generation. This will require very large transmission connections into both Victoria and NSW.

On the scale of the numbers coming from this exercise the size of these interconnections could easily be in excess of 5,000MW each. They will be very inefficient and be poorly utilised for most of the year, but this is the efficiency price for relying on highly intermittent generation technologies.

2. Storage will need to be huge and long duration

Even will all this renewable generation this system will require storage with a maximum output of around 27,000MW, and periods of 15-20 hours of storage with an output of around 15,000MW. To give somer sense of scale the Snowy 2.0 pumped hydro project is 2,000MW. The proposed Tasmanian Battery of the Nation is around 2,500MW. They will be swamped by demand for their electrons on dark and still nights and awash on windy, sunny weekends. Balancing this storage volume will  the main role for hydrogen to fill, but storing hydrogen at scale for long periods of time may be much harder than making it.

3. There will be periods of massive, massive over supply

Because on current maths it is cheaper to overbuild renewables rather than try to match it perfectly with storage, this means there will be regular periods of massive over generation. This 2050 grid has regular periods of oversupply in excess of 30,000MW, reaching to 46,000MW! This model does not factor in the demand from electrification of transport which will help so long as it is configured correctly. It also doesn’t factor in the increased demand peaks from a decline in the use of residential and commercial gas and increased electrification of these and other industries.

4. This will be much more expensive than $50 per MWh

The levelised cost of producing and consuming renewable electricity has been falling, and can be as cheap as $50 per MWh, which is great for decarbonisation. But the system needed to support this will add significant cost. The extra costs come in the additional transmission costs, over build or renewables and cost of large additional storage, which doesn’t generate, but actually loses 20-40 per cent of energy every time it is recharged. So it has two costs – the capital and operating costs and the efficiency losses.

A back of the envelope estimate comes in at around $70 billion on wind, $32 billion on large scale solar, $91 billion on DER, $60 billion on transmission and $65 billion on storage.

Back to EnergyConnect

In this broader context, the argument about EnergyConnect is pretty trivial. We are likely to need another six or seven energy connects, their costs part of a multi-billion dollar spend on all the kit needed to run a renewable-based electricity system.