Structural Number of a Flexible Pavement

The structural number is an index. It is not a stiffness, not a strength, and not a modulus — it is a single dimensionless figure standing in for the whole load-carrying capacity of a flexible pavement, built by multiplying each layer's thickness by a coefficient describing what it is made of, and by a further coefficient describing how well water leaves it.

Two things follow from "index". First, the same $SN$ can be reached many ways and they are not equivalent in service. Thick asphalt over a thin base and thin asphalt over a thick base can share a structural number and behave quite differently — differently enough that the standard design procedure does not stop at the total. It requires a staged check: the surface must satisfy the structural number required above the base, the surface plus base must satisfy the number required above the subbase, and so on down. That check is where the design is actually made. The equation on this page is the total, and the total alone will let you build something silly.

Second, the layer coefficients are regression coefficients, not material properties. They were fitted to one road test and have been recalibrated ever since, agency by agency, against local materials, local climate and local construction practice. Two neighbouring states will assign different coefficients to the same crushed stone and both will be right for their own conditions. The tables that hold them live in the copyrighted 1993 AASHTO Guide and in state design manuals, so they are inputs here and always will be. Use your agency's values, and if you do not have any, that is the thing to go and get rather than a number to guess.

A word on the units, because this is where calculators quietly go wrong. A layer coefficient is a per inch quantity: $a_1 D_1$ has to come out dimensionless, so if $D_1$ is a length then $a_1$ is a reciprocal length. This site types it that way — a real $\mathrm{in}^{-1}$ — which means the product is correct whatever units each factor is entered in. Type a coefficient in $\mathrm{in}^{-1}$ and a thickness in millimetres, as a metric drawing would give it, and the answer is still right. A calculator that treats $a_1$ as a bare number returns an answer 25.4 times too large in that situation, and it looks entirely plausible while doing it.

The drainage coefficients deserve their own paragraph, because they are the most subjective numbers in the design and they are pointing at the most destructive thing that happens to a pavement. Water in an unbound layer softens it, and repeated loading then pumps the fines out of it — each axle pass squeezing water sideways through the aggregate, carrying material with it, so the layer loses both stiffness and support progressively. Under a rigid slab the same mechanism erodes support at the joints and produces faulting. The $m$ factors are the method's admission that drainage is structure. They depend on how fast a layer drains AND on what fraction of the year it sits near saturation, which are two different questions: a base that drains in two hours but is soaked four months of the year is not a well-drained base. Where the answer matters, money spent on edge drains and a daylighted base beats money spent arguing about the coefficient.

Finally, the honest framing. This is empirical design, and it is regionally calibrated by necessity. The relations behind it came from one test on one soil in one climate over two years, and they have been stretched over sixty-five years, every soil in North America, and traffic the Road Test never imagined. Mechanistic-empirical design — computing strains under a modelled load and relating them to distress through transfer functions — has largely superseded this approach in agency practice, and the two do not always agree. The 1993 method nevertheless remains in wide use, because it is quick, because generations of engineers have calibrated their judgement to it, and because the mechanistic route needs material characterisation that many projects cannot justify. Knowing which one you are using, and what it can and cannot see, is more important than which one you pick.