Technical information

4-7
there is no requirement to press a ‘run’ button to execute the code; in fact, any change
of value in one of the input tables will be automatically reflected in real time in the
results table. This feature is what makes this program extremely easy to use and
appealing for further analyses compared to any other available LET code. Note also that
no units are specified as the computations are done in dimensionless form (see Equation
4.2.3); the user must be consistent with his choice. In the example, units of psi are used
for moduli and stresses; inches are used for thicknesses, radii and coordinates.
4.2.2 Calibration to APT Conditions
It is well recognized that pavement materials do not comply with isotropic LET
assumptions. The resilient response of HMA mixtures is known to be anisotropic and
nonlinear viscoelastic (Shields et al., 1998; Levenberg, 2006; Uzan and Levenberg,
2007). The resilient response of unbound layers is nonlinear elastic and stress-state
sensitive (Uzan, 1985; 1992) and also anisotropic (e.g., Tutumluer and Thompson,
1997). As argued in Section 4.1 use of isotropic LET may be considered appropriate, at
least as a first order approximation, given that as-constructed pavement layers are not
stress-free even without external loads. These result in built-in stresses which diminish
somewhat the inconsistency with actual material behavior. Nevertheless, a systematic
error is introduced into the analysis when isotropic LET is applied. Minimizing this
error can be accomplished by deriving the free model parameters (i.e., elastic moduli)
through a process of inverse analysis (or backcalculation) using the time history of
embedded gauge readings.
Following this approach, subsequent stresses, strains and deflections calculated
with the calibrated model will resemble measured responses even though the model
assumptions are fundamentally incorrect and over-simplified. In this connection it
should be noted that LET cannot inherently simulate certain features that were seen in
the experiment (see Subsection 3.5.2). One example refers to the offset observed
between peak responses and load location which in LET must coincide. Another
example is the non-symmetric response relative to the load location, i.e., the differences
between approaching and receding curves as recorded by the gauges which in LET is
always symmetric (see also Elseifi et al., 2006; Al-Qadi, 2007).