Efficient frontiers: Finding the right balance between wind and solar

It’s one of the big questions for market operators, networks, planners and investors as we migrate to a grid dominated by variable renewable energy rather than centralised fossil fuel generation. What exactly is the right mix of wind and solar?

There’s a lot to think about. We’ll focus here on the right mix for efficient investment, and we’ll assume – for the purposes of this exercise – that there was no rooftop solar, no interstate transmission, and we’ll ignore storage and demand too.

If ITK’s assumptions about capital costs are correct, then having about three quarters of your NSW variable renewable energy portfolio made up of wind and about one quarter from solar produces an efficient mix of output relative to cost.

As NSW is embarking on a 12GW supply procurement path – to replace the bulk of its coal generators that are retiring – such considerations seem worth talking about to me.

Of course the general conclusion shown in the graph below is not really news, it’s just been presented  a little differently. Being at the efficient point likely means minimizing the cost of the system once the cost of storage is added in.

Figure 1 Source: ITK using NEM Review data

Figure 3 Source: NEM Review, ITK, data to Nov 9, 2021

Figure 3 Source: NEM Review, ITK, data to Nov 9, 2021

And for the winds farms it’s been a good year, with two reporting capacity factors of more than 40% for the past 12 months.

Note how much the overall variability declines when the wind farms are considered as a portfolio. There is much better diversification benefit than solar, and it’s why from a system point of view some planning is preferred.

Figure 4 Source: NEM Review and ITK

The half yearly output for the past 12 months is shown in the following graph, which isn’t very useful.

Figure 5 Source: NEM review

So the way we proceeded was to express the wind and solar as an index relative to its average for the year.

So the average output was 1 for both solar and wind but because of capacity factors it takes 4.1MW of solar to produce a through-the-year hourly average of 1MWh of solar and 2.9MW of wind to get an average 1MWh of wind.

Next we assign costs of $1.2m/MW to solar and $1.9m/MW to wind. So the total cost is the MW of wind and solar * the cost of each. As you add more wind relative to solar the total output goes up so to adjust for that we get a metric total cost/average total output.

That’s one axis of the efficient frontier. The other axis is the variance of the combined portfolio, or actually the mean/std deviation.

After that its just a mix of experimenting, or in reality using excel to create a table using various ratios of solar to wind and then plotting an efficient portfolio.

As others have done almost identical work previously we expected to find that about 20% solar and 80% wind was the “sweet spot” and so it proved.