Table 11. Regression coeffi-

cients k1 and k2 for (Mr θ)

model  (150-mm-diameter

samples).

%

R2

crushed

k1

k2

100

29.396

0.417

0.70

75

9.8145

0.628

0.93

50

12.752

0.576

0.93

25

7.5508

0.711

0.90

0

25.237

0.468

0.86

groups, k1 and k2 are 13.647, 0.591 (R2 = 0.79), and

15.438, 0.540 (R2 = 0.84), respectively.

Compared with the results from the large-scale tests,

the data from the 150-mm samples show that the resil-

ient modulus is highest with the 0 and 25% crushed

aggregate mixtures. A possible reason for the higher

resilient modulus for the 100% natural aggregates could

be the lower void ratio. It is interesting to note that the

50% crushed aggregates had the lowest resilient modu-

Figure 16. 150-mm-diameter test

lus at the same void ratio. The 100% crushed aggregate

specimen at the end of testing.

had a lower resilient modulus than the 100% natural

material.

sis. Other missing data in Table 10 indicate that the strain

The effect of specimen size on resilient modulus is

measurements were noisy and were not used in the cal-

shown in Figure 18 for three levels of crushed aggre-

culations for resilient modulus.

gates. In general, resilient modulus from the smaller

The change in average resilient modulus as a func-

diameter samples was higher than that from the larger

tion of bulk stress is shown in Figure 17. Individual

samples, even though void ratios for the 300-mm sam-

regression power equations were fitted to the data and

ples were lower. The results agree with Sweere's (1990)

the coefficients (k1 and k2) are presented in Table 11.