Ultimate Strength Design of Tubular Joints

Abstract

ABSTRACT This paper presents a method for designing thick-walled, non-gusseted tubular joints for static loads. The method is applicable to T-, Y-, and K-joints having branch-to-chord diameter ratios of from 0.25 to 0.75. Given the axial load in a branch, the designer can use this method to choose a chord wall thickness which satisfies static strength requirements. Conversely, given the chord wall thickness, the designer can predict the axial load in the branch which would cause the chord to fail. Comparisons are presented between predicted ultimate loads and measured ultimate loads. The success of this design procedure depends upon the ability to 1) predict elastic stresses in the chord, and 2) correlate the branch load causing first yielding to the load causing the joint to fail. As an aid in predicting stresses, mathematical expressions are presented summarizing the results of numerous parameter studies. These studies were performed using two computer programs, one based on classical shell theory and the other based on finite element theory. The empirical correlations between first-yielding loads and ultimate, loads were determined from joint tests wherein the joints were loaded statically to failure. The correlations for T- and Y-joints are based on earlier tests performed at the University of Texas and Southern Methodist University while the K-joint correlations are based on a series of tests sponsored by Humble. INTRODUCTION The importance of providing adequate tubular joints in offshore platforms is well recognized by the offshore oil industry. To be considered adequate, a joint must be able to satisfy certain minimum requirements concerning static strength, cyclic loading, ductility, and weldability. This paper deals exclusively with the most fundamental of these requirements, that of adequate static strength. Presented herein is a method for designing non-gusseted tubular joints for static loads. As used here, the term "static loads" refers to loads resulting from extreme and rare conditions as opposed to high-cycle or fatigue loads. Such "static" loads could well be the result of the dynamic response of a platform to some extreme forcing function. The basic philosophy incorporated in this procedure is to provide the most fatigue- resistant joint known--in our opinion, a thick walled, non-gusseted joint--designed to withstand the maximum expected loads applied in a static manner. We refer to it as an ultimate strength method because it presupposes the ability to calculate the ultimate capacity of joints. This procedure is currently being used by the platform design groups of Humble and its affiliate, Esso Production Research Company. The success of this approach depends upon the ability to 1) predict elastic stresses in the chord (see Fig. 1), in particular the maximum or ''hot-spot'' stress, and 2) correlate the branch load causing first yielding at the hot spot to the load causing the chord to fail. If we have the ability to perform these two tasks, our job is straight forward; namely, to select a chord wall thickness such that the hot-spot stress is sufficiently low to provide an acceptable factor of safety against joint failure.