Wind, solar to supply 15% plus of NEM, 64% of South Australia, in 2019

Print Friendly

We expect wind & solar PV supply to represent at least 15 per cent of National Electricity Market demand by 2019 based on committed supply announcements. Even more remarkably, it looks like wind and solar will account for 64 per cent of demand in South Australia – an unparalleled percentage anywhere in the world.

Figure 1 below summarises the data we have put together.

Figure 1 Wind & PV by State. Source: NEM Review, ITKe

Figure 1 Wind & PV by State. Source: NEM Review, ITKe

We expect more and perhaps significantly more supply announcements over the next 12 months. Specifically, even at Origin Energy’s glacial pace, we expect Stockyard Hill wind (500MW) and Darling Downs solar PV (100MW) to be confirmed.

There is over 6 GW of planned utility scale solar PV in NSW alone, most of which can be developed quickly if developers want to.

Also, these developments are in front of the Victorian and Queensland reverse auction processes. Some of the development though could be bid into those auctions.

It’s the subject of a separate note, but we think that the way in which Victoria will invite bids  for electricity supply in the name of the Crown, but passed on to retailers, is an important, unresolved issue. However it’s done it will likely result in more supply than shown in the above table.


image: pv-tech

The figure shows that the small South Australian market is likely to get to roughly 2/3 wind & solar PV supply –  surely the highest penetration rate in the world for a proper grid. That will surely require good transmission connections – and storage – to deal with the fluctations in wind output. It will also require careful management.

One of the main messages that SHOULD emerge from the supply build up is that the AEMC and AEMO need to get their act together in regard to transmission investment. The existing regulatory investment test and the time it takes are inappropriate in the new NEM.

The new supply numbers, although an indication that perhaps prices and policy do work after all is still a modest number. Even in 201o we will still be below where Europe, Texas and California are today. But we are catching up.

Demand is the 12 months to April 5, 2017 total as reported in NEM Review + rooftop PV  and grown by 1%.  Supply in terms of annual energy delivered is built up from the MW tables shown below multiplied by ITK’s estimated capacity factors.

LCOE now in the mid $60- $70s MWh for as available power

ARENA reported at this week’s INFORMA/RE large scale solar conference that they estimated single axis tracking PV to have an LCOE (required electricity price over life of project to recover all costs) in the mid $70s MWh.

The Silverton wind farm, partly financed by QIC, one of Australia’s largest and most experienced infrastructure financiers, has a relatively short term PPA price of $65 MWh real (increases with inflation).

These prices are now broadly comparable with those in the USA where wind PPAs are routinely reported at around US$40 = A$60 MWh (US $20 MWh paid by the customer and US $20 MWh of  production tax credit paid by US Federal Govt).

The reverse auction process of the Victorian and QLD Govts has the potential to produce even lower LCOEs because of the longer term PPAs and the excellent credit quality of the State Govt. However this may be partly offset by the fact that supply has to come from within the relevant State reducing competition.

Price discovery required for power firming

The missing link in seeing much greater expansion of wind and PV is a better understanding of the price required to make that power “on demand” rather than as generated. The LCOE of wind and PV is not the full cost of delivering dispatchable power. We don’t yet know what that dispatchable price is. But we are going to find out.

In the industry this is known as “firming” the supply. We are about to go through a process of technology and price discovery to work out the most economic way to firm up over all supply.

Diversity will only be of modest help to PV (PV will always be a middle of the day producer), but of much greater assistance to wind (wind in NSW won’t always blow at the same time as in South Australia).

Other candidates for firming include  gas, hydro, pumped storage hydro, solar thermal and batteries. Batteries have one key advantage over all other technologies in that they can be distributed. If we had proper policy and analysis in Australia we could properly allow for the savings of distributed power.

In addition demand management and flexible coal can also assist. Its worth remembering that demand for electricity has always varied widely over the day and the existing thermal fleet manages this without complaint.

12 GW of wind & PV

Its interesting that each of the States has a similar amount of combined wind & PV MW although of course in terms of the significance of that to the overall market it matters far more to South Australia.

Figure 2 Wind & PV supply. Source: APVI, CEC, ITKe

Figure 2 Wind & PV supply. Source: APVI, CEC, ITKe

7.4 GW PV

Figure 3 PV supply. Excludes W.A., TAS. Source: APVI, ITKe

Figure 3 PV supply. Excludes W.A., TAS. Source: APVI, ITKe

Note that we have assumed all the utility PV is single axis tracking but that is not universally the case. As compared to our previously published estimates we have found following this week’s conference another 220 MW of PV in Qld (4 separate projects) that is definitely going ahead.

5.4 GW Wind

The wind numbers are straightforward.

Figure 4 Wind power. Source CEC, ITKe

Figure 4 Wind power. Source CEC, ITKe

Firming supply can be measured on several criteria

Although LCOE is an important driver of pumped hydro, peaking gas, batteries and thermal PV its not the only important factor. Other things to consider are speed to market and response time.

If we look at QLD and South Australia we see that over the past year average spot prices over the day have been quite similar despite remarkable differences in the composition of supply.

Figure 5 12 month to April 4 average spot prices by time of day. Source: NEM Review

Figure 5 12 month to April 4 average spot prices by time of day. Source: NEM Review

The days of $20 MWh overnight prices have gone for the time been.

Despite the similarity in average prices the volatility in South Australia is far higher. The following chart shows the coefficient of variation (Std deviation/mean) of spot prices is far higher in South Australia particularly outside of peak hours. This reflects wind supply and smaller quantity of average demand in South Australia. Never the less even in QLD the price at peak time of day is unpredictable with the standard deviation 4X the average.

Figure 6 Coefficient of variation of spot prices. Source: NEM Review

Figure 6 Coefficient of variation of spot prices. Source: NEM Review

This data suggests that prices in South Australia are unpredictable without knowledge of say wind data. There is therefore likely to be a premium for fast start plant and this is where batteries may come into their own. Certainly we think that new technology plant needs to be able to respond faster than gas if its to capture sudden shifts in price. This doesn’t just mean fast start it also means fast stop.

How big are the hourly price swings?

We looked at the absolute half hourly price change in South Australia and Queensland over the past year. 90% of the time the half hourly price change is less than $50 MWh (up or down)

Figure 7 Percentiles of absolute change in half hourly prices. Source: NEM Review

Figure 7 Percentiles of absolute change in half hourly prices. Source: NEM Review

Looked at in this way there is less difference between the States about 2% of the time you can expect a half hourly price change of over $300 MWh. Qld had 132 half hours with a price change from the previous half hour greater than $1000 MWh and South Australia 148. Being fast enough to capture those movements will be important to generators providing firming capacity. Essentially you need to be faster than gas right now.  If the half hourly settlement is changed to 5 minute settlement we might see some movement in these numbers.  But it will take a bigger spreadsheet.

David Leitch is principal of ITK. He was formerly a Utility Analyst for leading investment banks over the past 30 years. The views expressed are his own. Please note our new section, Energy Markets, which will include analysis from Leitch on the energy markets and broader energy issues. And also note our live generation widget, and the APVI solar contribution.


RenewEconomy Free Daily Newsletter

Share this:

  • WR

    PV with single-axis tracking provides a more or less constant output on a sunny day from shortly after sunrise until shortly before sunset.

    Cloud cover has a larger impact on output than time-of-day (during daylight hours, of course) for tracking systems.

    See the output from the Moree solar farm on the Aneroid Energy website for an illustrations of this.

    • Rod

      Thanks for the link.
      So Moree is the only one on that graph with tracking?
      Gives over 1 1/2 hours more solar when it is needed most. You would think the value of that extra solar and the time it is available would justify the extra cost of tracking.

  • Just_Chris

    My personal opinion is that something pretty substantial will need to change to maintain those capacity factors in SA in 2019. Don’t the new rules after the blackout state that 2 gas generators must be running at all times in the state? I might be wrong on the gas rules but my gut feel is that you will have turn down the wind farms at times. It’s also going to be interesting to see how Victoria responds to SA becoming a net energy exporter, which is what I think will need to happen to accommodate SA’s new generation capacity. Especially since SA have made it a requirement that 34% of “base-load” power has to be purchased in the South Australia.

    Always interesting to use TWh in these conversations as I am sure someone will bring up the battery options – 500 MWh of batteries cycled once a day gives you 0.2 TWh per year, which is significant but it won’t dramatically change the numbers either way for what is a very big battery – the average power requirement for SA is 1470 MW. At 500 MWh I assume that the batteries would be dictating the spot price on the SA grid.

    • MikeH

      It aims for 36% of all local generation to be “synchronous” but I don’t think that is spinning – it just means that it can be fired up if necessary. And 36% is about where they are now with gas – it tightens to 50% by 2025 but according to this article from Giles, it can include gas, solar thermal, pumped hydro as well as being “open for inverter-based technologies such as wind and solar with battery storage,and batteries as long as they could provide a similar service.”

      And given the scare campaign being run by the Victorian Liberals about the impact of Hazelwood’s closure, there should be no reason for Victoria to object to SA imports.

      • Malcolm M

        The requirement for 2 gas generators is imposed by AEMO, and adds ~200 MW of power into South Australia that is necessary only for stability purposes. From what their new CEO has said of distributed energy, I expect that this requirement would be relaxed once tests have shown that batteries can provide similar services. A gas generator would be equivalent ~100 MW/100 MWh of battery. So at times of high wind or solar, gas generators could switch off, allowing another ~200 MW of renewable energy to be exported from the SA network. Now that Hazelwood has closed, there doesn’t seem to be much negative price impact on SA power exports on Vic prices.

        • Ian Fordham

          The most common scenario that meets the synchronous requirement in SA currently is 2 Torrens island ‘B’ units, which at minimum generation are 40MW each, just like now, they are the only ones running in SA. Rest is wind and SA is exporting >450MW. There are other scenarios that meet the minimum requirement but looks like those scenarios would all require a higher level of generation. Edit – I posted this just after Osborne ramped down, the export is limited now a lot further due to the RoCoF constraint. I expect some wind is curtailed or ‘wasted’ at the moment. Batteries or storage would be useful here!

  • George Darroch

    If we add western Victoria to the mix (and we should, since this is the NEM and electrons don’t recognise state borders) the percentage becomes even higher.

    The wind/solar border is somewhere near Ballarat.

  • Alastair Leith

    Thanks for this analysis made free David!

    When you say “looks like wind and solar will account for 64 per cent of demand in South Australia – an unparalleled percentage anywhere in the world.”

    Are you forgetting two states in Germany that are already over 100% RE (net), Scotland, and Denmark?

  • Peter F

    Thanks again for the analysis David.
    Two quibles.
    1. Storage is probably a lot cheaper than new interconnects so I don’t believe new Transmission infrastructure to SA will be needed, maybe more NSW/QLD and perhaps NSW/Victoria but that is a fair way down the track.
    2. Latest figures suggest rooftop solar could well exceed your projections

    Adding 500-600MW of renewables in Tasmania, upgrading the peak capacity of the Tassie hydro and and duplicating Basslink might be an interesting exercise. quite possible cheaper than Snowy 2 per MWhr delivered

    A serious crack at power to heat in the form of grid controlled hot water and ice banks is probably the cheapest way of all to lower peak demand and power prices at the same time

    • David leitch

      Peter I agree that the debate about storage v transmission is well worth having. One that some comprehensive modelling could help with. Many commentators, and I include myself in that number, think that some extra transmission will help the development of utility scale renewables. It did wonders in Tx for instance. Tasmania is getting left well behind at the moment despite its excellent wind and hydro resource. At first blush that seems a bit nutty.

      That does not preclude development of storage all around the grid. Batteries out at the street level and on the fringes of the grid can do a tremendous amount to lower electricity costs and increase reliability. The USA studies to demonstrate this already exist. We just need to get one done in Australia to help with appropriate policy.

      • Peter F

        I like the batteries at street level idea. I often look at pole mount transformers and think that we could mount a bettery on the same pole and collect all the excess local solar, reduce the transformer size, reduce iron losses, reduce peak transmission losses etc. I haven’t done the calculations I will one day

  • Peter Campbell

    15% in 2019? That doesn’t seem like much. Wasn’t the Renewable Energy Target, even when the LNP reduced it, over 20% in 2020? That leaves a lot to do in one year.

    • this just refers to wind and solar which is the new build – it does not include hydro, which will take it well beyond 20%.