1. The relative force level appears to be a significant quantity. It will 

 be identified in later sections as Fjj/Fit,, with F,^ equal to the portion of the 

 line force, Fg^,, carried by the soil (assumed equal to F^^^ - W^, + W^). 



2. The complex process of crack formation and propagation appears 

 to play a major role in determining the breakout response. At least three soil 

 mechanics concepts could enter into this problem: 



a. Undrained creep of the soil surrounding the object 



b. Drained creep with water flowing into the soil 

 with time 



c. Water flowing through the soil into an expanding 

 cavity beneath the object 



Concepts (a) and (b) seem more probable, because they would lead 

 to an expansion of the crack system (as observed), whereas (c) probably would 

 not. An analysis of concepts (b) and (c) would lead to the conclusion that the 

 hydraulic gradient, p/D, is an important quantity (p is the gage pressure in the 

 pore water beneath the object and was assumed equal to Ft,/A). An analysis 

 of concept (a) would lead to the conclusion that the relative force level by 

 itself is the important quantity in effecting long-term breakout. The impor- 

 tance of these terms can be determined through empirical data correlation. 



3. The object embedment depth appears to be very significant in 

 determining breakout. 



The significant parameters which were varied among the tests included: 



Original embedment depth of object, D 



Width of object, B 



Hydraulic gradient, p/D 



Relative force level, F^/Fit,, with Fjt, estimated from Equation 5 



The dependent variable of all of the long-term breakout tests was the 

 breakout time, t^,. It was felt that the most important independent variable 

 was the relative force level, F^/Fib. Therefore, the technique followed for 

 determining the best method for correlating data was to plot F^/Fi^, versus 

 t^, normalized by the geometry parameters D and B and the hydraulic gra- 

 dient p/D. 



Several normalizing procedures were followed. The terms plotted 

 versus Fk/F.^ included: 



21 



