predicted using Dynaflect and FWD tuned moduli are comparable for SR

268, US 441, and SR 15C test sites. The similarity in response predic-

tions for the two NDT devices may be attributed to the high stiffness of

layers 1, 2, and 3 relative to the underlying layer support values.

However, Table 8.2 shows that significant differences between Dynaflect

and FWD occurred on SR 24 because of large differences in base course

and subbase moduli even though the asphalt concrete and subgrade moduli

were identical for both NDT devices.

 Tensile stresses were predicted in the base course and subbase

layers of SR 26B and SR 15C. However, the magnitude of these stresses

are too low (2.8 psi maximum at SR 15C) to be of concern. Also, verti-

cal stresses on top of the base course layer were higher for thin pave-

ments (SR 24, US 441, and SR 80) than pavements with thick AC layers (SR

26B and SR 15C). The vertical subgrade stresses were generally low,

with a maximum value of 4.8 psi obtained for US 441. This value consti-

tutes 4.0 percent of the total vertical stress of 120 psi applied by

each wheel. This value is less than the limiting 10 percent stress

level conventionally used with the classical Boussinesq's solution

(59,133). Thus, the relatively high stiffnesses of the pavement layers

(i.e., the asphalt concrete, base, and subbase) result in a reduction in

the stresses and strains on top of the subgrade.

 Table 8.6 lists a summary of the stress analyses for each pavement.

The maximum surface deflection ranges from 8.1 to 38.5 mils. The major-

ity of these maximum deflections occurred in the subgrade layer, as

indicated by the corresponding percent compression values. In general,

high deflections were associated with pavements with low subgrade

moduli. However, the high deflections obtained in most of these