Backcalculation of FWD data often results in unrealistic layer moduli. The modulus of the subgrade may be two to three times the expected value, and the modulus of an intermediate granular material may be lower than the subgrade modulus. If stresses or strains measured in the pavement are compared with theoretical values, the agreement is often poor.

All theoretical models for calculating pavement response are based on a number of simplifications with respect to reality and must be verified experimentally. Most models assume that all pavement layers consist of linear elastic materials. This paper demonstrates, that treating the subgrade as a nonlinear elastic material, can result in much more realistic moduli and a much better agreement between measured and calculated stresses and strains.

The response of nonlinear elastic materials can be calculated using the Finite Element Method (FEM). A much simpler approach is to use the Method of Equivalent Thicknesses (MET), modified for a nonlinear subgrade. The paper includes an example where moduli backcalculated using FEM, linear elastic theory and MET are compared. Stresses and strains predicted by the three methods are also compared to measured values. For this example it is found that the best agreement between measured and calculated stresses and strains is obtained with MET. FEM results in stresses and strains within the range of measured values, whereas all the stresses and strains calculated with the linear elastic theory are outside of the range of measured stresses or strains.