Out came that much-used 2011 picture of a Scottish wind turbine that burst into flames during a storm. Commentators queried the amount of CO2 emitting from a turbine fire, and if the cost of repair was built into the estimated advantages of "these things". RenewableUK questioned the reliability of the source data, but most of the papers ignored this. In fact, the main source of data was anti-wind lobby group Caithness Windfarm Information Forum, and the report's conclusion was based on a weak ten-times extrapolation of data.
After haranguing from RenewableUK and various publications, the study author updated his blog, and the academic body amended its website, pointing to a lack of data and repeating that the ratio of turbine fires has decreased significantly since 2002. Industry publications reported the amends, unlike the popular press.
The original report was an academic study into fire-protection engineering in wind turbines. It acknowledged that an increase in wind farm fires is balanced against a big increase in operating turbines; and it did say that poor statistical records are a concern and hinder research efforts. But a press release from one of the originators of the report ran with the headline that fires are a major cause of wind-farm failure.
Such a shame. Here was some research being carried out because wind is a growing source of renewable energy and any kind of breakdown of turbines hinders energy production and so increases costs. The lack of publicly available data led the researchers to base their calculations on biased, insufficient and sensational information, and then magnify these figures to conclude that previous calculations had been wrong - the industry was hiding something. The author acknowledged a frustration at the limited data, but the report went ahead. Research using poor data does not give good results. The number of turbine fires was not the point of the report, but it was the most dramatic.
Turbines have been operating in large numbers for a long time now. Data has been captured, and reports have been carried out (see p43 for one example), but clearly more needs to be done.
We need facts, not wild guesstimates. And with so much attention being paid by operators to "big data" and control systems that capture, analyse and respond to data, these facts can be produced. The wind industry has to stop hiding behind a cloak of competition-based confidentiality and make its data available, somehow.
The industry would benefit from wider research into many areas, including fire prevention, even though the likelihood, of a wind turbine catching fire, even according to the above report's author, is less than one in 10,000.
Let me put another figure out there. This comes courtesy of wind blogger Mike Barnard, who reported on a 2012 calculation by an Australian energy analyst using data from many sources, from the Caithness anti-wind lobby to GWEC. This analyst calculated that if you were to stand underneath a random wind turbine you'd have to wait around 1,693 years before it burst into flames.
Jacki Buist is editor of Windpower Monthly