Localized Buckling of Buried Flexible Pipelines

ABSTRACT Buckling of submarine pipelines due to axial compressive forces caused by the constrained expansions set up by thermal and internal pressure actions has been extensively studied, but very little attention has been paid to the transition from an assumed pseudo-periodic mode of deformation to a loca1i zed buckl1ing pattern observed in practice. This localizing is of considerable importance for the proper determination of the vertical deflections of a buried flexible pipeline. The paper examines a numerical formulation and checking of a 600 m experimental trench. INTRODUCTION A large amount of literature investigating the risk of buckling of submarine pipelines subjected to constrained expansions, resulting from thermal effects and internal pressure, is available. However, previous studies have made assumptions that do not always seem appropriate. The same tendency can be noted in studies of the deformation of continuously welded railway track, with the except of the paper by Tvergaard and Needleman, (1), Which represented a real advance. Previous studies have formulated both the railway track and pipeline buckling problems in terms of the isolated imperfection pattern, which implies localized buckling. This excludes, a priori, all possibility of the redistribution of extensions among the adjacent imperfections. The experience acquired by COFLEXIP and IFP, through full-size tests and numerical simulations, shows that, at least as far as flexible pipelines are concerned, this approach can be very limiting. We have investigated the real behavior of flexible pipelines with numerical models using a set of representative configurations. After analysis of the results and comparison with observations made on-site and test results, the importance of the role of the principal initial imperfection was confirmed. The assumption that this imperfection is repeated along the pipeline is more or less realistic, and its repetition is related to the concept of influence length, which is defined later. The principal imperfections are, ina11 probability, separated by smaller amplitude waves, and it can be supposed that these secondary imperfections have the same wave1ength as the principa1 ones. The number of secondary imperfections per period defines the equidistance of the principal imperfections. The three main results are as follows:In the unfavorable case, in which the extensions are highly localized at the principal imperfections, the equidistance of the imperfections is the deciding factor in the estimation of the deflections. However, this is valid only for moderate principal imperfection equidistances (for a 6-inch flexible pipeline it is valid for equidistances of the order of 100 m maximum).If the equidistance referred to in (a) is greater than the threshold value indicated, secondary localizations appear. These deformation concentrations illustrate the concept of periodic relaxation, as demonstrated by Courbon (2) for the case of a flexible pipeline without any initial imperfections.