This equation is similar to Equation 4.20 except that the
regression coefficients and range of parameters applicable are
different. Two pavements had prediction errors of -21.8 and 16.8
percent. These were SR 24 and US 301, with asphalt concrete thick-
nesses of 2.5 and 4.5 in., respectively. All other pavements had
El predictions of the order of 10 percent error. Therefore, Equa-
tion 6.11 may be preferred over Equations 4.19 and 4.20, since it
covers a broader range of variables and also developed from tuned
test data.
6.7.3.2.2 Base course modulus, E2. Analysis of the tuned
data using E2 as the dependent variable resulted to an equation
similar to Equation 4.21 which was obtained from the theoretical
analysis.
log E2 = 3.280 0.03326(t ) 0.1179 log (D )
+ 3.3562 log (D1 02) 9.0167 log (D1 D4)
4.8423 log (D1 Ds) Eqn. 6.12
(R2 = 0.959 and N = 22)
Error analysis indicated that only two pavements (SR 15A M.P.
6.549 and 6.546) had -15.6 and 15.8 percent prediction errors.
Prediction errors for the others were 10 percent or less. Equation
6.12 should be used in place of Equation 4.21 unless the applicable
range of the former is excessively exceeded.
6.7.3.2.3 Stabilized subgrade modulus, E3. Multiple regres-
sion analysis of the data set resulted in the following equation