Rotational stiffness governs my wind turbine foundation — is that normal?

Daniel ParkConstruction Methods Engineering · 12 yrs

Verified expertise in Construction Methods Engineering

You have correctly identified your weak link, and for a stiffness-governed design it is the binding one. SPT-derived small-strain stiffness carries a scatter band you should think of as roughly a factor of two, not a percentage, and the correlation was developed for particular soils. Across ground varying from weathered rock to soft alluvium you are almost certainly applying a correlation outside its calibration range at several positions. When the governing requirement is a stiffness, that scatter propagates straight into base diameter, and you are sizing concrete against a number with a factor-of-two uncertainty.

The economic argument for direct measurement is usually easy to make at this point. A seismic cone or a cross-hole survey at three or four representative positions, covering the range from weathered rock to alluvium, costs a small fraction of the concrete that a factor-of-two stiffness uncertainty forces you to pour across fourteen foundations. It also lets you defend the design properly, because a stiffness-governed foundation sized from an SPT correlation is difficult to justify to a reviewer who asks the obvious question about where the stiffness came from. Measure it and the whole design chain firms up.

Sofia AdeyemiMine Automation Engineering · 12 yrs

One addition on the position-by-position approach, which is the right call: check the dynamic stiffness consistently as well as the static. The supplier requirement is usually a dynamic rotational stiffness at small strain, and if you are deriving it from a static bearing-based model you may be comparing the wrong quantity. The two differ substantially in soft alluvium, and mixing them across fourteen positions can make some foundations appear to fail and others appear to pass for reasons that are entirely about which stiffness definition was used where.

Also worth confirming how your soil stiffness varies with depth in the model rather than using a single equivalent value. Rotational stiffness of a shallow gravity base is dominated by the soil within roughly half a base diameter of the edge, so a profile that stiffens with depth gives a different answer to an averaged value, and the direction of the error is not the same in weathered rock as in alluvium. Across your variable ground that could well explain part of why nine positions land on the stiffness constraint and five do not.