The project "Stress Concentration Factors for Simple Tubular Joints" was a UK Department of Energy (DEn) sponsored project, completed in 1990. The primary objectives of this project were to assess current methods for deriving SCFs in simple tubular joints, to discuss the applicability of the more commonly used parametric equations, and to derive a new set of parametric equations that would reduce some of the anomalies in the existing formulae. In this paper the principal conclusions of this project are summarised and the new Lloyd's Register parametric equations estimating SCFs for simple tubular joints are presented.
In the 1970's with the increasing development of the hot-spot stress S-N approach to nodal joint fatigue life estimation, it became clear that the determination of reliable SCFs for tubular joints was fundamental to this concept. The first parametric SCF equations covering simple tubular joints were derived by Toprac and Beale (1967) using a limited steel joint database. The prohibitive cost of testing scaled steel models led Reber (1972), Visser (1974) and Kuang et al. (1975) to use finite element (FE) analyses based on analytical models of cylindrical shells. Subsequent equations by Wordsworth and Smedley (1978) using acrylic model specimens and by Efthymiou and Durkin (1985) employing 3-D shell FE analyses, have made considerable advances both in the accuracy of parametric equations and in the range of joints covered. Over this period, differences arose between the experimental procedures used to derive stress concentration factors for simple tubular joints. These differences led to inconsistencies both in the measured SCFs themselves, and also in the SCF parametric formulae based on these measured SCF values. These inconsistencies in SCF derivation are reflected in the hot-spot S-N curves used to estimate fatigue lives for simple tubular joints. In the UKOSRP II project (DEn, 1987), a limited programme of work investigated anomalies between the existing simple joint parametric SCF equations. A test programme on unstiffened tubular joints performed by Lloyd's Register, (Smedley and Fisher, 1990), has further emphasised inconsistencies between test results and the more commonly used parametric equations. While current guidance merely states that ‘the appropriate SCF’ should be used, it is the intention of future guidance (Reynolds and Sharp, 1990) to give more specific directions on which parametric SCFs may be employed in fatigue design, and the procedures that should be considered prior to performing either experimental tests, FE analysis, or in deriving parametric equations.
