The root of the problem

Seeing the failed towers and other wind turbine destruction in India (Windpower Monthly, September 1998) certainly gave a dramatic picture of how strong the forces of nature can be. While it is not practical or economical to build turbines strong enough to survive all possible storms, I believe there is a powerful method to limit the damage, which our industry does not as yet appear to employ: selecting the point of failure.

It is standard practice in electrical design to use current limiting fuses to protect various parts of a circuit. Wind turbines can do a similar thing by choosing a bolted root design wherein the bolts will yield and fail before the tower or foundation does. Once relieved of the aerodynamic drag of its blades, the wind turbine tower and nacelle can withstand a great deal more wind than if the blades stay attached. Trees shed leaves and limbs in excessive winds as a matter of course, and thereby greatly limit the number that are uprooted. Whenever we see a wind turbine on the ground with its blades still attached, we should recognise that we may not be as clever as nature has already shown us how to be.

This comment is not made casually. In my work with the wind industry over the past 20 years, I've designed many successful wind turbine blade roots. Ductile yield in metal is a quite consistent and predictable property, and can be used to assure that the bolts will fail and the blades will be shed before the turbine is torn down to the ground. Of course one must take care to assure that the nacelle is not twisted from the tower top when only some of the blades are gone. While this is not trivial, it does appear achievable.

The biggest barrier is perhaps psychological. It is difficult for a designer or manufacturer to specify weaker bolts, when for very little more cost, stronger ones can be used. It is the natural tendency of all engineers to design their subsystem to not fail, including root designers. But if we do not decide where the structure will fail if overloaded, then it is not likely to fail in the least dangerous and expensive way. We must choose how it fails, because we cannot absolutely prevent out of design extreme conditions.