Page 118 - Buried Pipe Design
P. 118

Design of Gravity Flow Pipes  93

         the hyperbolic soil model 6,42,43  provides a nonlinear soil model that has
         been used successfully in finite element analyses of buried pipe. Thus,
         the hyperbolic model is incorporated in most finite element programs
         that are used in buried pipe analysis. Examples are CANDE and
         PIPE5.
           The Iowa formula, as proposed by Spangler, predicted the change
         in the horizontal diameter of the pipe due to soil placed over the top
                                          56
         of the pipe. Watkins and Spangler proposed the use of the modulus
         of soil reaction  E  with units of force per length squared. Later
         Watkins, Spangler, and others showed that the vertical and horizon-
         tal deflections were about equal for small deflection. They also
         showed that the vertical deflection was the better predictor relating
         to pipe performance. While the Iowa formula has been criticized by
         some, it remains the best known simplified method for computing
         deflections.
           Howard’s  E  values (Table 3.4), back-calculated from measured
         vertical deflections of many flexible pipe installations, are conserv-
         ative. For the back-calculation, he had to assume the bedding factor
         and the lag factor. Some have proposed an increasing soil modulus
         with depth of cover, but Howard found no correlation between E
         and depth of fill. His data were limited to 50 ft of cover, so he stated
         that his proposed values of  E  may not be valid for cover greater
         than 50 ft.
           As noted, many researchers have attempted to correlate the modulus
         of soil reaction E  with other true soil properties that can be evaluated
         by test. The most common parameter used in these efforts is the con-
         strained modulus M s which is the soil stiffness under three-dimensional
         strain, where strain is assumed to be zero in two of the dimensions
         because of restraint (Fig. 3.11).




                                s z
                s z
                                               M
                          P V                   s
                                               M s
                                                                    P
                          P V                                        v
                                               M s
                                        M s                        P v

                                              Œ z
         Figure 3.11 Constrained compression test schematic.
   113   114   115   116   117   118   119   120   121   122   123