Case 1. For t, = 3.0 in., E, = 100.0 ksi, 10.0 < E < 85.0 ksi,
 1 1 2

 10.0 < E < 200.0 ksi, and with E3 between 6.0 and 35.0 ksi,
 4

 -K
 E = K (D D ) 2 Eqn. 4.11
 3 1 4 10



 -0.3562 9 -0.7185
 K = 22.74(E ) + 3.503 x 10 (E )(E ) Eqn. 4.12
 1 2 4 2




 0.1528
 -0 10183(E )
 K = [3.4455 + 0.00841(E )](E ) 2 Eqn. 4.13
 2 2 4


 The accuracy of the E3 prediction equation presented above appeared

 good within the stipulated range of variables listed above. However, it

 was not simplified enough to allow the development of a more comprehen-

 sive equation to include varying E1 and t values. Multiple linear

 regression analyses (39) were performed using various combinations of

 variables and transformations in an attempt to develop a relatively

 simple E3 prediction equation for t, values of 3.0, 4.5, and 6.0 in.

 The best results from these analyses produced a complex equation

 containing 13 variables.



 Case 2. For 3.0 < t < 6.0 in., 100.0 < E < 1000.0 ksi,
 1 1

 10.0 < E < 85.0 ksi, and 0.35 < E < 200.0 ksi,
 2 4