values exceeded the range established for the El prediction equations,
the predicted El values seem to be reasonable except for five test sec-
tions. For these sections (I-10A, I-10B, SR 15A and SR 715) the pre-
dicted asphalt concrete modulus tends to be high, considering the mean
pavement temperature and corresponding viscosity of the asphalt binder
during NDT testing. The reliability or accuracy of the predicted EI
values, neglecting estimation errors of E2, are discussed in Section
6.4.
The base course modulus was computed for only three sites (SR 12,
and SR 80 Sections 1 and 2, using an estimate of E,. Two of the pre-
dicted E2 values exceeded the limits originally established for the
prediction equations (Case 4 Equations 4.8 to 4.10). The E3 predic-
tions seem to be reasonable and of the order of magnitude expected in
practice. The accuracy of these values will be determined when the
deflections are modeled using BISAR in Section 6.5.
Table 6.3 shows that the three E4 predictions are in close agree-
ment, especially for Equations 4.16 and 4.35. The good agreement
occurred for Eq values from 5.0 to about 40.0 ksi. However, beyond E4
values of 40.0 ksi, the agreement is good between Equations 4.17 and
4.35. A typical example is the I-10A test pavement in which the 108.4
and 108.0 ksi predictions from Equations 4.17 and 4.35, respectively,
are far closer than that of 113.7 ksi from Equation 4.16.
6.3.3 FWD Prediction of Layer Moduli
The FWD prediction equations and the deflection data listed in
Table 6.2 were used to compute each layer modulus for the various test
pavement sections. Equation 4.20 was used to make most of the E