This is the fourth and final in a series of Contractor Reports on buried flexible pipe design. The previous three reports, CR 228, 229 and 230 (see IRRD 842875, 842876, 842877) covered the design methods presently in use in Britain, the development of a new analytical approach by Gumbel for TRL and its application. This report assesses the relative merits of the various methods of design. The assessment considers the validity of both the theoretical and the practical assumptions made in the methods by comparison with the results of model and full-scale tests on buried flexible pipes. The model tests were carried out by Gumbel (see IRRD 800757) and the full-scale tests at TRL were reported by Crabb and Carder (see IRRD 287399). Both Gumbel's method and the widely-used method described in CIRIA Report 78 (see IRRD 240909) are examined in detail in the context of the TRL test results. These tests encompassed pipes with a range of plane-strain flexural stiffnesses from 0.05 to 12.4 kPa and two different soil types. Back-calculation has been used on the test data, at both 5% deflection and buckling failure, to derive appropriate soil moduli. This assessment shows that the derived soil moduli at 5% pipe deflection are of the same order of magnitude for both methods and within the range of expected values. For buckling failure, which may occur at greater or lesser deflections than 5%, the back calculated values from Gumbel's method are more realistic than those from the CIRIA method. The applicability of the results of the tests presented is limited by the shallow burial and strip loading constraints of the test pit and loading system; however published data from other research and Gumbel's model tests suggest that his method may also be sound for other geometries and loading configurations. The most difficult problem remaining, which equally affects both the CIRIA and Gumbel methods, is that of determining the appropriate value of the lateral earth pressure ratio K. It is concluded that Gumbel's method is potentially more realistic than design methods in current use. However, as there is little difference between them in the analysis of deflection, the case for retaining the Spangler-based CIRIA method is strong. But there is clear evidence that the Gumbel method can give better predictions of buckling behaviour than the Meyerhof-based CIRIA method.

Want to know more about this project?