moduli and thickness of layers 1 and 2. Figures 4.14, 4.15, and 4.16
show that for a fixed t the relationships between E and D D4 are
not significantly affected by E3 and E4. Also, the plots suggest that
the effect of E2 becomes negligible as the thickness of the asphalt
concrete increases. This relationship was therefore used to develop
power law equations to predict E1 with an estimate of E2 for different
layer combinations using the Dynaflect modified system.
The sequential development of layer moduli prediction equations
using the BISAR generated Dynaflect and FWD deflections is presented in
Section 4.3.2 for both NDT devices. Equations for the FWD used theore-
tical deflections from a 9-kip FWD load. The methodology employed
involved the use of simple power law relationships and multiple linear
regression analysis (39) procedures.
4.3.2 Development of Dynaflect Prediction Equations
4.3.2.1 Prediction Equations for E Initial analysis of theo-
retical Dynaflect deflection basins had indicated that E1 and E2 were
essentially independent of E3 and E4 using D1 D4 as shown in Figures
4.14 through 4.16. Thus E1 could be expressed as a function of E2, t1l
t2, and D, D where D1 and Dq are deflections at sensors 1 and 4,
respectively, in the modified Dynaflect sensor array (see Figure 4.2).
The use of constant base course thickness (Table 4.1) removes the
influence of t2.
Power law relationships between E1 and DI D4 for several combina-
tions of t and E2 were developed. Regression analyses were then per-
formed to establish intercepts and slopes for the linear trends. The
basic form of the regression equation used in the analyses was