Working through the implications of 50% renewables
We regard Queensland’s 50% renewable target by 2030 as potentially by far the largest driver of new renewable energy supply in Australia over the next 20 years. Queensland is the only State where electricity demand is presently growing, it is Australia’s second largest source of electricity demand the current renewable supply is tiny relative to the target. Only NSW has the potential to match Queensland and for now the NSW Govt is more preoccupied with public transport.
For these reasons we think that a large part of the utility industry’s cognitive surplus should be devoted to this topic.
What follows is ITK’s first take on the Queensland Renewable Energy Expert Panel Draft Report [Muggelstone Report].
Our conclusion is that as proposed the system won’t work. It will result in far too much PV in the middle of the day and require the remaining thermal generation to ramp up and down continuously in a way it was never designed for.
Proposed PPAs are too short, and the recommendation to have all the new generation sourced from Qld is too parochial and will increase costs to consumers in several ways. The report is successful politically in that the key conclusions are that prices won’t increase, no thermal generation has to close and system reliability isn’t compromised. These are the key objectives to move forward. But they are less than the full story.
Today’s note shows why the plan probably won’t work as currently outlined. The second part of this note will show, based on work being done in the USA, how it can be made to work very successfully, but it will require far more underlying change to the rules and regulations than is contemplated so far.
30 TWh by 2030 means 22 new TWh over next 13 years
The most important number is 22 TWh. That’s roughly the amount of new renewable energy that will need to be produced in Queensland over the next 14 years to get to a 50% target of 30 TWh by 2030. If we take the numbers in the Mugglestone report, based on Jacobs modelling, and use our estimate of capacity factors we get to:
The composition of supply is that assumed by the Mugglestone report. It certainly doesn’t have to be that way. In our view the most questionable assumption is in fact that there will be another 3400 MW of rooftop PV. Even with just 1600 MW of existing PV, QLD already has 30% of houses with solar on the roof (see APVI map). To get another 3400 MW in just 17 years seems a stretch even assuming a lot comes from the commercial sector. If it does work out that way, its going to have a lot of implications for networks, which we can’t get into here.
This duck has big teeth
The shape of thermal electricity demand during the day when there is significant PV penetration is referred to in the industry as a “duck curve”. I’m no artist but its fairly conceptual. Never mind here’s the past 12 months in Qld. The duck is only starting to emerge.
We don’t have data on Qld wind generation, but we can model the PV impact
In mid May 2016 the AEMO started providing an estimate of rooftop PV supply. The following chart shows the estimate by time of day for QLD. The shape of the chart is entirely as expected but its interesting to note that the maximum ouput as estimated by the AEMO never exceeds 800 MW that is just about 50% of capacity and less than modelled by the AEMO. As we get Summer data these numbers may increase a bit.
Projecting the duck curve forward
In terms of energy, most of the new supply in Qld is projected by the Mugglestone Report, based on Jacobs Modelling to come from Utility scale PV. We assume that this will mostly be single axis tracking. Single axis tracking PV has a much higher capacity factor, on an AC basis, than rooftop PV. We have used the Moree PV farm for the single axis capacity use data and based it on the 56 MW AC capacity rather than the 70 MW PV DC..
The daily thermal ramp up and down is the problem
Figure 5 below is the most important figure in this note. It has been derived as follows.
- We calculated 2016 average daily half hourly demand and then increased it by 10% to allow for demand growth to 2030.
- We forecast the PV output using the same production curve for QLD rooftop PV as already supplied by the AEMO but using the Mugglestone report estimate of rooftop PV capacity installed by 2030
- We used the Moree single axis solar plant as the daily production profile for the single axis utility solar expected to be built by 2030 in the Mugglestone report
- We subtracted the total PV output from the daily demand and what is left is the average daily energy that has to be supplied by thermal (gas and coal) and wind.
- Note that this is a daily average. The variation around this will be enormous, both from hot or cold weather increasing or reducing demand but also from storms or whatever impacting PV output.
In our view its (extremely) unlikely that coal plant will be able to ramp up from 1 GW to 7 GW and down again every single day without suffering absolutely significant economic and physical impacts. The plant wasn’t designed that way. Even gas generators don’t like being cycled too often.
At an absolute minimum the average efficiency of the plant will be way worse. Coal and gas plants are designed to run at one temperature. Almost certainly there will be a substantial increase in the forced outage rate, which in turn will impact daily prices. The figure below shows how little variation there is in average coal and gas generation in Qld over the past 12 months.
To be clear, ITK fully supports 50% and even 100% renewable energy but we just don’t think the Mugglestone Report has had a deep enough think about it.
A report such as this is largely a political document. The real work is done once the policy is adopted. What we are signaling is that much more radical change is required. This is not a bad thing, it should and will be fully embraced. The technical answers are clear, they are just not being spelled out. We will get to them.
Next consider the implications for price.
We already see with only modest PV penetration, relative to what’s coming much higher prices in the early evening compared to midday.
Other issues even ignoring daily profiles
In ITK’s view some of the recommendations in the Queensland Renewable Energy Expert Panel Draft Report [Muggelstone Report] might be debated in the interests of lowering the price of renewable electricity to both Queenslanders and Australians. Two areas that immediately spring to mind are:
- The report suggests the tenor [length] of PPA’s to be 7-15 years. However its likely that longer PPAs would produce even lower unit prices as the longer the PPA the lower the residual risk to the supplier. The Govt is the party best placed to carry risk because its cost of capital is the lowest. In our view the QLD Govt should be considering a PPA life of 15-25 years and focusing on a 20 year PPA.
- The report suggests that all the renewable electricity be supplied from QLD. In ITK’s view this will artificially reduce potential supply and likely increase costs.
Why should all Queensland’s renewable demand and supply be in Queensland?
One of the draft recommendations in the Mugglestone report is that all of Queensland’s new renewable energy to meet the 50% by 2030 target be located in Queensland.
Its easy to understand the appeal of this to Queenslanders, who along with West Australians always seem to feel “State first, Australia second” but is it good policy?
A feature of the ACT’s small but effective 100% renewable scheme was that the generation could be located anywhere. For sure this creates the lowest price for consumers by encouraging the lowest cost supply.
Additionally it likely leads to a more balanced and less distorted NEM.
About electricity generation in Queensland
The distinguishing feature of Qld generation is that very little electricity gets into Qld that wasn’t produced there.
The figure below shows most (95%) of Qld electricity generation sorted by total output over the past 12 months. Qld is a net exporter of electricity, all of it to NSW, but those exports have declined in the past 12 months making an important contribution to the significant increase in NSW futures prices.
The chart shows that the gas generators, Darling Downs, Braemar, Braemar 2 and Yarwun are already reducing output due to rising gas prices, and in some cases because the available gas has physically been sold to an LNG plant. Our data runs just to the end of September so this reduction in gas may have further to run. That’s the case even though the pool price data shows prices over $100 MWh available in the late afternoon in Qld.
Over the past couple of years units at Stanwell Power Station and Tarong were shut in as Qld demand reduced and gas generation and PV increased. However except for PV those dynamics have reversed and the units are all running. If all the major generation units in Qld ran at 90% capacity utilization output could be as much as 16 TWh higher. One mothballed State Govt. owned gas generator, Swanbank is still listed by the AEMO as restarting in 2017.
However in ITK’s opinion Gladstone’s ability to significantly lift ouput over an extended period is limited and we see the main prospects for supply increase from Stanwell, Tarong and Kogan Creek. These stations collectively may well be able to produce about 6 TWh a year more than in the past 12 months and this would be enough, from NSW’s perspective to offset the closure of Hazelwood. Don’t forget if “The Age” speculation is to be believed the Board decision on Hazelwood is this month.
Transmission is the issue
Over the past seven years, with rare exceptions, far more electricity has flowed out of Qld to NSW than in the other direction. The chart below shows the total and other work we have done makes it clear that for the most part it doesn’t matter how much higher the pool price in Qld is relative to NSW its still difficult for NSW generators to access the Qld market. We think there are three reasons for this.
- Some of the Qld generators are located close to the NSW border. Whenever pool prices rise in Qld those generators, specifically Kogan Creek tend to “hog” the transmission up to the South East of Qld.
- There are physical constraints on the NSW transmission line around the New England region that make it harder for power to flow North
- The main interconnector has been designed to allow for higher capacity from Qld than to Qld.
The combined Qld to NSW capacity is presently about 1.3 GW. If this was to run at 100% capacity all year round, then Qld could export 11 TWh to NSW or about 8 TWh more than at present. So again in theory this could partly compensate for Hazelwood closure.
There are proposals around to lift transmission from QLD not just to NSW but all the way to South Australia (1600 km). This could be done by a DC link where costs are falling rapidly but would likely cost about $1.3 bn according to “Trans-Australian HVDC Interconnection Investigation” by Prof Simon Bartlett of the University of Queensland. DC links would seem to be way cheaper looking forward considering that the PV and wind is DC by nature.
By contrast the proposed NSW SA interconnect is expected to be half that cost, but a standard AC line. It allows more renewable development along its corridor but almost by definition wind built in that corridor will be more correlated with wind in South Australia, than wind built further North.
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.
