Ultimate Capacity Equations for Tubular Joints

ABSTRACT A total of 137 ultimate strength tests on simple T, Y, DT, and K tubular joints is used as a basis for development of new ultimate capacity formulas. The data are taken from a variety of sources and only relatively large geometries are considered. Axial tension, axial compression, in-plane bending, and out of- plane bending loads are represented. The failure condition is taken as the minimum of either maximum load, first crack load, or load at an excessive deformation limit. Several formulas are recommended to predict the capacity of the different joint and load types. The accuracy of these formulas is then studied in a statistical manner. Predictions of past and present API RP 2A formulas are compared to the same data base. It is found that the new equations are more consistent in their level of prediction and result in less scatter. The new equations are also relatively simple in format. INTRODUCTION Circular tubular members are used extensively in offshore structures because their drag characteristics minimize wave forces on the structure and their closed cross section provides for buoyancy needed during installation in the ocean environment. Tubular members are also used in many truss type structures which require long slender compression members since the tubular cross section exhibits a high strength-to weight ratio. In most instances the connections of tubular members are formed by full penetration welding of the carefully contoured ends of the branch members to the continuous chord of the truss. Unlike piping connections, the plugs are not removed. Typical simple tubular joints are shown in Fig. 1 along with some of the geometric terms, including ratios ß and ?, used in defining joint behavior. The word "simple" implies no stiffening devices, no grout and, in the case of K joints, non-overlapping branches. The objective of this paper is to establish improved ultimate capacity equations for tubular joints. Design of typical welded and bolted connections generally considers the post-yield behavior because of the large reserve capacity exhibited. Since theoretical prediction of this behavior is difficult, test results have often been used to develop empirical capacity equations. The empirical approach is employed here; however, the equation development benefits from a carefully screened data base, which existed as of early 1979. Some of the data are new, unpublished results from the University of Texas, including data for out-of-plane moment loading on the branch (25). After preliminary discussion of the data base and the equation format, this paper systematically addresses axially loaded T, Y, and DT joints, axially loaded K joints, and moment loaded joints. In each section there is a review of joint performance characteristics, a presentation of the new equation(s), and a statistical analysis of each equation's accuracy relative to the data base. The prediction accuracy of the equations in the 1973 and 1977 editions of the American Petroleum Institute Recommended Practice (API RP 2A) (1) is also given for the same data base.