The conventional classification of pavement design procedures into "empirical" and "mechanistic" is reexamined. It is submitted thatthe term "mechanistic" originally denoted a data interpretation methodology based on the laws of engineering mechanics, proposed as an alternative to the statistical interpretation techniques used extensively in the interpretation of aasho road test data. Current mechanistic design procedures retain statistical/empirical correlations, but this is a matter of practical expediency, pending improvements in analytical capabilities. On the other hand, incorporating mechanistic results into a statistical/empirical framework rarely leads to reliable conclusions. A case in point is provided by aashto 1986 nondestructive testing method 2 (ndtm2), which combines the mechanistic layered elastic theory with the purely statistical/empirical structural number concept. The derivation of the ndtm2 equations is traced and simplified, and a number of theoretical shortcomings are highlighted. It is recommended that use of ndtm2 be discouraged and that efforts for the gradual elimination of statistical/empirical constructs (e.G., Structural number, equivalents single-axle load, and miner's fatigue concepts) be intensified. Attempts to define statistical/empirical parameters (e.G., Layer coefficients, present serviceability index, and load equivalency factors) using mechanistic theoretical tools should be abandoned. This paper appears in transportation research record no. 1307, Pavement analysis, design, rehabilitation, and environmental factors 1991.
Abstract