The most comprehensive data base available to date was published in 2004 on behalf of the Netherlands Agency for Energy and the Environment. This was a study that looked at the experience of 43,000 turbine-years of experience in Germany, Denmark and the Netherlands up to 2002. In that time, they analysed “over 200 severe accidents and incidents” (their words) and found that 62 of these appeared to be relevant for the safety of nearby objects (or persons, should they be nearby). They found the following “recommended values” for failure rates: (Statistically risk analysis increases the values found slightly when the data of failures is not large to be statistically significant.)
– Loss of blade 8.4 x 10-4 (failures per year per turbine)
– Collapse of entire tower 3.2 x 10-4
– Collapse of rotor or nacelle 1.3 x 10-4
You can say this means for a single turbine, the severe accidents that will cause loss of life (loss of blade that can travel up to 500 metres), collapse of the tower (limited to pretty well the total height of the tower plus a bit for impact bits – say about 150 to 200 metres ), or collapse of the nacelle or rotor (generally limited to about 50 metres) the total failure rate would be about 1.3 x 10-3, which would that if one had 1000 turbines one might expect 1.3 of these failures per year.
But, remember that the body of data was based on turbines up to 2002, and a lot of those turbines were smaller than today’s.
Looking at relevant differences, we find looking at information from the US Department of Energy Renewable Energy Laboratory, who reported in a paper published in 2002 that new taller wind turbines might be subject to a new failure rate because of the difference in wind speeds across the rotor (wind speeds at the top of the rotor are higher than wind speeds at the bottom resulting in the blade flexing as it goes around the circle.)
Ontario experience to the end of Sept 2008 comprises about 615 turbine years of experience. In that time we’ve seen 2 blade failures, one at the Port Burwell wind farm in April 2007, and one at the Prince Farm wind farm in January 2008. 2 failures in 615 years is a blade failure rate about 5 times greater than the failure rate seen in the Netherlands study. It is true that we do not have a great deal of data yet, but still it is troubling given that many municipalities are using setbacks of 50 metres from turbines to public roads and the lot lines of neighbours.