CORRECTED VERSION 3:36pm. For correction details relating to wholesale electricity price estimates see bottom.
Last year the Australian Energy Market Operator was commissioned by the federal government to examine the feasibility of operating the eastern states’ National Electricity Market using entirely renewable energy sources for the period of 2030 and 2050. On Friday a draft version of its modelling results was released.
It finds that it is indeed possible to operate the NEM with 100 per cent renewable energy while meeting the current NEM reliability requirement. This means that even with 100 per cent renewable energy it was possible to meet the energy needs of the NEM 99.998 per cent of the time.
There are two key reasons why this modelling represents a major advance in assessing the hard physics involved in utilising renewable energy:
1. AEMO used real-world hour-by-hour weather data across 42 regions within the NEM states. This allows them to determine how wind and solar power output is likely to vary, and the ability for other sources of power to ramp up and down to cope with this while meeting likely electricity demand levels.
2. AEMO took into account constraints in existing transmission power lines to deliver electricity to where it’s required, and examined what new transmission line infrastructure would be economically optimal to manage weather variability, and take advantage of rich but remote renewable energy resources.
It’s important to note that while AEMO found it was possible for 100 per cent renewable energy to reliably supply our electricity needs, it doesn’t mean it would be easy and straightforward. There would be far greater variability in supply than occurs presently and the need for substantially larger amounts of spare capacity.
Also it would represent a major step-up in cost compared to our current high polluting system. However it’s hard to say whether it would be noticeably more expensive than an alternative low carbon or moderately carbon-intensive system that made greater use of fossil fuels or nuclear, because such an alternative wasn’t modelled.
This is rather unfortunate because it means we’re left with a huge scary number on the capital cost of the 100 per cent renewable energy system at $219 to $332 billion without a valid point of comparison that would also provide some rather frightening dollar values.
If you dig further the study finds that the wholesale electricity price required to support the 100 per cent renewable energy system would be $111 to $133 per megawatt-hour. This is actually about the same price as what Treasury’s carbon price modelling projected under government policy for 2030, and lower than what it projected for 2050. Yet Treasury’s electricity system had much higher use of fossil fuels and much higher emissions.
In the end though economic projections are almost invariably a very rough guess. This doesn’t mean they should be ignored, but the real value from this study is its assessment of the physical feasibility of maintaining high reliability with large amounts of renewable energy.
The key to how they manage to meet reliability is through a combination of:
1. Using a range of energy sources and locations that help to smooth out the variability of weather-dependent energy sources;
2. Using a significant amount of capacity from biogas-fuelled turbines and solar thermal with nine hours of thermal energy storage to balance out wind and solar PV output; and
3. Assuming that between 5 per cent and 10 per cent of demand in any hour could be shifted in time to better match electrical output.
Interestingly, AEMO’s modelling assumes no use of battery storage that would export to the grid (although electric vehicles are assumed to import energy for battery charging), as it was deemed too expensive relative to using biogas turbines and solar thermal with storage.
The fuel mix the modelling landed-on is illustrated below for the two scenarios examined. Scenario 1 (‘S1’ in chart below) could be considered the optimistic scenario which makes 100 per cent renewables easier to achieve because electricity demand growth is slower, demand is easier to shift in time, and geothermal in particular realises major cost improvements. Scenario 2 assumes a slower rate of technological improvement thereby favouring the existing lowest cost renewable technology of wind power.
Annual electricity generation by fuel source for 2030 and 2050 period
Source: AEMO (2013)
The location of this alternative generation mix is far more geographically dispersed than the current NEM’s structure around coal deposits. It involves substantial amounts of generation in the north and inland regions which would necessitate substantial new transmission infrastructure.
The chart below provides a good illustration of how shifting demand in time, as well as biogas, solar thermal with storage and hydro help address the inability to control wind and solar PV output. As found in prior studies, winter becomes the most challenging period for managing reliability because there is a demand peak in the evening as solar PV output drops away.
Demand shifting (illustrated by the dashed black line) becomes very important in moving some demand from the late afternoon and evening into the middle of the day. In addition solar thermal (orange), biogas (light green) and hydro (blue) are pivotal in filling in the late afternoon and early evening period of demand when solar PV output fades. It’s also worth noting that there’s some excess energy spilled in the middle of the day on Tuesday and Saturday due to an inability to completely align demand and supply.
Example of supply and demand in a winter week (scenario 2 in 2050)
Source: AEMO (2013)
This study represents a major breakthrough in moving us beyond the rule of thumb approach to assessing the potential of renewable energy personified by such put downs as, ‘did you know the sun doesn’t shine at night, and sometimes the wind doesn’t blow?’.
The next step is to use the model to explore alternatives that use fossil fuels and also nuclear power in combination with renewable energy.
CORRECTION: In the original version of this story it incorrectly stated that AEMO's estimates of the wholesale electricity price did not include financing costs and thus could not be readily compared to estimates by the Australian Government Treasury's carbon price modelling. AEMO has since confirmed that their estimates of the wholesale electricity price do include a weighted average cost of capital (the "financing costs") of 10%.