Valleys

Thin layers of asphalt and sprayed seals (10-40mm) perform well, and thick asphalt pavements (>100mm) perform well. For a given design, though, there will be a range of thicknesses between the two extremes that just doesn’t work. 40-80mm is the usual dead zone.

This happens due to the properties of the underlying granular material. The strength of a granular base course increases proportional to the stress it is subjected to (sort of like the non-Newtonian corn starch mixture that you get to punch in high school science class). A thicker, less flexible asphalt layer will reduce the stress levels in the base course below, reducing the level of support provided to it (sort of like the asphalt is too proud to burden the gravel with its problems, so it just bottles it all up instead).

Asphalt achieves the greatest fatigue life when it contributes very little to the pavement stiffness and relies entirely on the support of the granular base, or contributes the majority of the stiffness to compensate for the lack of support.

Doesn’t look like a valley in this graph, but please imagine it’s upside down. (From AGPT02)

There are two solutions – one on either side of the valley – but going half-way means increased strain and a reduced lifespan.

I’ve been trying to think of other scenarios where half-way input results in a worse result than either of the extremes. The uncanny valley is an easy pick. Work effort is a potential candidate too – much like an asphalt layer, I know I can work just enough to feel strain but not enough to get a result. I think it’s a good shape of graph to keep in mind when evaluating a problem – perhaps sometimes we need to push through and increase the input when it’s intuitive to pull back, or vice versa.