In the last week, renewable generation in SA has varied from 126% of local demand to only 0.6%. During this period of low wind in the state, SA was importing ~600MW from Victoria. At that same time, renewables were thankfully generating 34% in VIC, along with brown coal pretty much running at maximum capacity.
This event highlights the two key elements of our future renewable-based grid, which will require:
- A level of firming capacity within a region; and
- Increased interconnection to other states to rely on generation from diverse weather patterns.
A thesis could be written on this topic. So we will merely look at some elements that may indicate the level of “other generation” that may be required to cover the difference between P50 and P10 wind profiles across SA, VIC and NSW by time of day and by season. To do this analysis, we looked at the dispatch profiles of wind generation in SA, VIC and NSW from 2020 to May 2023.
Wind generation profiles obviously differ significantly by season, with wind generation over summer (November to February) essentially following an inverse solar profile, with its lowest generation in the middle of the day ramping up to its highest time of day generation during the evening peak by about 6pm before beginning to drop off at 6am. This compares to winter (June to August), where the generation profile is much flatter over the course of the day.
The graph below shows the difference between the P50 and P10 time-of-day profiles for wind in NSW, SA, and VIC as a proxy estimate for firming generation that may be needed and scaled up as wind capacity increases.
Let’s take the average winter P50/P10 difference as a percentage of the current installed wind capacity. We find that roughly 20-25% of ‘firming’ may be needed to cover those extremely low wind generation periods. If we use these figures to extrapolate out to 2030 (based on our latest Q2 Benchmark Power Curve expectations of installed wind capacity), we find that NSW will need around 1.9GW, VIC, 2.1GW, and SA 750MW of firming. To put that into context, in the most recent event last week in SA, when wind was providing 0.6% of generation, the in-state firiming generation required was just over 1GW.
The graph’s ‘firming’ value essentially highlights the difference between the generation we expect wind to provide and a potentially significantly lower wind generation period by time of day. Interestingly, this difference over summer for VIC and SA between midnight and 6am is very similar, even though VIC has 75% more wind capacity. However, over the evening peak in summer, we see that this gap increases due to the P50 generation profile increasing over the evening peak while the P10 profiles for VIC and SA remain somewhat similar.
As a result, there is the potential for a larger need for peak firming in VIC due to the larger delta between the P50 and P10 profiles.
There is a much more nuanced view that can be built around the required volumes, including:
- Looking at the number of instances when generation of wind in one state is high versus low generation in interconnected regions – which would provide a historical view of the value potential value of increasing interconnection
- Looking at the duration of those low-generation events would allow for a quantification of the duration of this ‘firming’ capacity.
The increasing need for firming assets as the grid transitions to the much higher level of renewable penetration and understanding the corresponding level of volatility that this brings to wholesale market pricing is key to evaluating future captured prices for renewables and storage alike. For more information on our recently released Benchmark Power Curve, please contact email@example.com.