Greenpeace on wind variability

This debate is about variations in wind power, mitigated by averaging over a large geographic region (many wind farms). I will have much more to say about this later, for reasons you will see (later).

For now, a quick half-rebuttal to Greenpeace' latest polemic:

Greenpeace UK | Wind power myths are blown away [press release]

It links among other things to these two studies:

Greenpeace UK | Wind Power: Managing Variability (David Milborrow, 2008)

University of Oxford | Wind power and the UK

The claim is:

So what does it say? First of all, that old chestnut about the wind dropping and the lights going out is just not true. Of course, the wind does fluctuate, but averaged out across the country that fluctuation is much less. (The UK is, of course, the windiest country in Europe.) This means that while the output from one wind farm might dip as the wind subsides, the wind will still be blowing somewhere else, and the larger the nationwide network of wind farms, the smaller the variations in electricity generation.

This is true, and should be quantified. But moving on:

In fact, research by Oxford University's Environmental Change Unit shows that low speed wind events affecting 90 per cent of the country only happen for, on average, one hour every year (pdf).

Very misleading. The claim is true - I quote the Oxford report page 8:

Low wind speed conditions

It is common to refer to low wind speed conditions as "calm" periods but this underestimates the conditions when electricity generation will cease. Large wind turbines do not generate electricity in winds below 4m/s, so all hours where winds are below this speed are included in the definition of low wind speed conditions.

UK Met Office records show that whils low wind speed conditions can be extensive, there was not a single hour during the study period where wind speeds at every location across the UK were below 4m/s. On average there is around one hour per year when over 90% of the Uk experiences low wind speed conditions (Figure 4)

And look how misleading that statement was! That very rare event is this: simultaneous zero-power-generation in 9/10 sites. That's what "low wind event" means - zero power. What did you think it meant - something reasonable, like "10% below average"? No; it means nothing at all.

The obfuscation is even deeper than that. Quantifying the percentage of wind farms that are down is one thing, but what we're really interested in is - what is the total generation put together? The difference is illustrated as: if 90% of wind turbines are barely-moving at 10% of their power capacity, and the other 10% aren't moving at all - well, that's fine: only 10% of turbines are completely down (or "low wind event" as some would call it); 90% are up! But of course it is a grid catastrophe - you're running on 9% capacity!

And (surprise!) this is typical. The total, combined generation of wind turbines across a country is quite often very low, EVEN though are many wind turbines spinning, so their output is not exactly zero (the very rare event Greenpeace reassures us about). Here for example, are excerpts of the SUM generation of all Irish wind farms put together - and there are dozens - in units of megawatts:

(Example is from David MacKay's Sustainable Energy - without the hot air, p. 187. This in turn is compiled from www.EirGrid.com).

These are the meaningful numbers: total electricity generation. And they are awful. Look at the third row - you have a whole continuous week in February were all of Ireland is operating around 5-10% capacity. And this happens far more often than "one hour a year" - it's a whole, nonstop week! And if you look at shorter, day-long outages, well they're everywhere!

But Greenpeace hides this. They obfuscate: they throw useless statistics like chaff.

Likewise, the full Greenpeace report seems (I only skimmed) to ignore this stuff completely. The gist of it is chapter 3. They conclude wind power is very non-variable at all, when it is averaged together. But again it's a red herring: they only consider "intra-hourly" variability - fluctuations on timescales of minutes. But of course the real issue is fluctuations on the scale of days - e.g. the week-long power outage.

And they say, for instance:

An analysis of the wind power fluctuations in Western Denmark in 2007 suggests that for 42% of the year (3700 hours) the intra-hourly fluctuations were within the range plus or minus 25 MW (1% of the wind capacity). Extending the range to plus or minus 50 MW captures another 1800 hrs of fluctuations. At the extremes, fluctuations in excess of plus or minus 375 MW (16% of capacity) only occurred 10 times in the year. The complete histogram of power swings is shown in Figure 3.The standard deviation of the fluctuations is around 3%.

...going on and on about those intra-hourly fluctuations in Western Denmark, when all you need to do is zoom out and see

(source - IEA study, page 212)