each Dynaflect loading wheel), the potential for eccentric loading and
its subsequent effect on D1 is reduced with the use of an average value
from the two sensor deflections. In the case of the FWD, the prediction
equations had a high degree of accuracy. Therefore, the suspect E1
values using FWD predictions could be affected by other factors which
are discussed in the subsequent section.
6.5 Modeling of Test Pavements
6.5.1 General
The theory of a linear elastic model generally implies that defor-
mations (or strains) are proportional to the loads applied to the medium
or media. For flexible pavements, recoverable deformations are con-
sidered elastic even though they are not necessarily proportional to
stress nor instantaneous. In accordance with the terminology first
introduced by Hveem (48), recoverable deformations are referred to as
resilient deformations and the corresponding moduli as resilient moduli.
Analysis of the load-deflection response of the FWD measurements
had indicated that a linear elastic model could be used to analyze most
of the test pavements. Therefore, the BISAR elastic layer computer pro-
gram was used to determine the moduli of the pavement layers from the
Dynaflect and FWD deflection basins. The subgrade was characterized as
a composite value, as conventionally done in multi-layer analyses.
The layer moduli determined from the prediction equations and
summarized in Tables 6.3 and 6.4 were used as input into the BISAR
computer program to compute Dynaflect and FWD deflections, respec-
tively. These modulus values plus layer thicknesses (Table 5.1) and
Poisson's ratio (Table 4.1) served as input data for BISAR. The