AEMO exposes truth on wind
The Australian Energy Market Operator has released a report today providing a variety of interesting data illustrating characteristics of wind power in South Australia. Data covered includes:
-- Monthly wind farm output from 2006 to June 2012;
-- Contribution of wind farm capacity to peak demand;
-- The relationship between wind farm output and electricity prices with a particular focus on negative price events;
-- How wind farm output tends to vary by time of day and by region; and
-- The extent and speed of variation in wind farm output.
Of particular interest from this report is that AEMO has analysed how geographic spacing of wind farms might moderate the variability of wind farm output. The rate of change in the output of wind farms (and indeed all power sources) as well as the predictability of that change is a major concern for maintaining power quality and reliability. This is because it takes time for other power stations to adjust their output.
The degree of inflexibility in power systems is often overplayed by climate and wind sceptics who seem to be unaware that there is already large underlying variability in the power system without wind (due to variation in demand and unplanned break-downs of conventional power plants). Nonetheless, as the amount of wind becomes quite large as a proportion of the overall supply, the variability of wind becomes more important.
One of the potential ways this variability might be reduced is through spacing wind farms over larger geographical areas. The logic being that weather tends to vary across large distances, so while the wind speed might be growing rapidly in one area, it could be slowing down in another. This then acts to balance out, or at least moderate, the speed of change in wind power output.
AEMO's analysis broke down wind farm output in SA into three separate regions illustrated in the picture below.
What it found was that variability does reduce when you aggregate wind farms across the three regions. The two charts below illustrate the extent of change in the output of wind farms within a five-minute period (the top chart) and within an hour (the bottom chart).
The rate of change within five minutes is especially important because the capability for other power producers to respond is far more constrained and more costly than what they might be able to achieve if given an hour.
Five-minute interval - Frequency and extent of change in wind farm output
Red – aggregate of all regions; Blue – Coastal peninsula; Orange – Mid-North; Yellow – South-East
Hourly interval - Frequency and extent of change in wind farm output
Red – aggregate of all regions; Blue – Coastal peninsula; Orange – Mid-North; Yellow – South-East
The top chart shows that even within a single region the rate of change in wind output from one five minute period to another as almost always below 5 per cent of installed capacity. Furthermore when you aggregate across all the regions (the red line) you see that frequency of no change in output from one time period to another becomes even greater and the frequency that change exceeds 5 per cent of output becomes even smaller. So geographic dispersion in SA has helped to reduce variation within short time periods.
In terms of hour to hour variation, while the likelihood of there being no change in output from one hour to the next is reduced with greater geographic dispersion, the more extreme changes in output above 10 per cent of installed capacity are noticeable reduced. Overall this is also good news for wind. That's because when given an hour's lead time other power stations should have little difficulty covering a change of 10 per cent in the output of wind, even at high penetrations, but very big changes beyond 20 per cent are more challenging.
It's worth noting that this study was constrained to SA and this is hardly the limits of geographic dispersion that we might be able to achieve in the interconnected National Electricity Market. Another thing worth noting is that the AEMO has undertaken a cost benefit analysis that suggests it would be appropriate to upgrade the transmission interconnector between SA and Victoria. This will further assist in managing the growth of wind in SA as well as Victoria.
One further additional point is that while I was in Western Australia last week I caught up with their Independent Market Operator (IMO) about the potential costs of managing the variability of wind in their main grid in the South West. Prior modelling work undertaken by ROAM consulting had estimated that the costs of additional quick-response generation to manage wind would substantially undermine the economic viability of new wind farms. However according to discussions with the head of the IMO, these costs have been significantly overstated due to problems Western Power had in attributing the sources of unplanned variation in the electricity system (primarily demand). The IMO will be releasing updated analysis on this issue which will provide a greatly improved assessment of the likely costs of managing variation in wind farm output in the South-West Interconnected System.