D 



_^ 



D2 

 Ptb 



D 



ptb /b 



D 1,D 



p_tb /bV 



D 



2 \D 



The last term, (pt^j/D^XB/D)^, produced the best correlation with 

 '^b^^ib ^"-"^ ^'^^ FY-71 laboratory tests. A plot of this relationship is presented 

 in Figure 4. The dimensions of this normalized time term are time-force per 

 length to the fourth power. This term can be nondimensionalized by multi- 

 plying by a permeability and dividing by the unit weight of water. Such a 

 nondimensional term can be rationalized as discussed in Appendix C. 



The field and FY-68 laboratory test data were also plotted in the 

 form of Fj,/Fj|3 versus (pt^^/D^){B/D)■^ , as shown in Figure 5. Here the corre- 

 lation is somewhat weaker and the points tend to lie to the left of the data 

 points for the FY-71 laboratory tests ( Figure 4). The shift in the data points 

 appears attributable to a variation in some time-related soil property (possibly 

 permeability) among the different soils tested. It is felt that if this property 

 were measured accurately it could be incorporated into the normalized time 

 term (as discussed in Appendix C) to improve the quality of the correlation. 

 The permeability of the FY-71 laboratory test soil was measured (0.9 x 

 10"^ cm/sec), as was the vertical permeability of core samples obtained at 

 the Gulf of Mexico test site (1.1 x 10"' cm/sec). The latter measurements, 

 however, are unrepresentative of the overall permeability characteristics of 

 that site because the horizontal permeability was not determined. Because 

 sand seams were observed in the cores, horizontal seepage would assuredly 

 dominate the long-term soil response. Therefore, a more accurate normali- 

 zation of the data is not possible at present. 



22 



