Renewable Generation Diversity Investigation

NZWEA engaged Concept Consulting to investigate the opportunities and challenges that might arise for the New Zealand electricity system due to increasing amounts of wind and solar generation. In particular,  how diversity, both in technology type and location, might assist with reducing the inherent intermittency of these generation types.

It is recognised that the variability of wind and solar, energy will create some significant challenges as New Zealand transitions to an electricity system with very high levels of renewables.

The study looked at how diversity, both in location and technology type, will affect incorporating wind and solar into the electricity system over the next 15 to 30 years.

The approach taken was to look at the correlation between generation from different locations and technology types which were then combined in hypothetical build scenarios and then how these interact with electricity market demand.

Key correlation conclusions include:

  • Wind correlation to Tararua drops with geographical distance with Northland, Dunedin and Southland having a low correlation. Hourly is most diverse with weather systems increasing the correlation on a weekly and monthly basis. Diverse wind results in a higher GWAP/TWAP ratio.
  • Solar correlation is high on a monthly and quarterly level due to seasonal effects. Correlation is lower in winter supporting diversity of siting albeit with lower generation levels and locational diversity does not address the intra-day and seasonal nature of solar.
  • Wind is anti-correlated with solar, particularly in winter
  • Three build scenarios were analysed (balanced wind and solar, solar only, wind only) with a further three scenarios for geographical diversity (concentrated, more diverse, very diverse)

Key build conclusions are:

  • Solar output varies drastically depending on time of day and time of year. Wind exhibits these characteristics but to lesser extent.
  • Wind is very flat but with slightly more generation in the afternoon and in spring with less in late summer.
  • The generation duration curves are very different for wind and solar with solar not generating for 50% of the time during the night. Wind has a roughly linear curve with less variation than solar.
  • Locational diversity has a noticeable difference for wind with a mix of wind and solar flattening the duration curve particularly for daily and weekly generation

How generation scenarios interact with the wider electricity market assessed in relation to peak supply, seasonal swings and dry years are as follows:

  • Residual peak hourly demand is significantly higher for the solar-only series since there is negligible generation during system peak. Residual demand reduces with other scenarios.
  • On a weekly basis solar and wind alone are worse than a combination of both.
  • On  a seasonal basis solar generates significantly less in winter
  • Wind generates slightly more in winter, but slightly less in dry winters
  • Increasing locational diversity can help most aspects

 In summary:

  • Having a diversity between different types of variable renewables (i.e. having a mix of wind and solar) materially reduces the extremes of low and high generation.
  • Solar’s relative lack of generation in winter and early mornings and evenings make it poorly suited for meeting the need for seasonal and peaking generation.  However it has less variability than wind over week-to-week timeframes.
  • Wind’s average output is well matched to daily and seasonal demand shapes, meaning that, on average, there will be reduced need for additional balancing energy for daily peaks and over winter in a wind-heavy future compared to a solar heavy future. 
  • Wind’s positive correlation with winter hydro inflows means it will often exacerbate the need for dry-year firming energy.
  • Having a diversity in the geographical spread of renewables will further reduce the extent of extremes of both high and low generation.  
View the full report and a summary presentation below: