The Buckling Design of Stringer-Stiffened Shells Subject to Combined Pressure and Axial Compression

Abstract

ABSTRACT The paper describes the procedure of, and results from, a test programme to determine the buckling characteristics of stringer stiffened shells subject to various combinations of pressure and axial loading. The tests were carried out on small scale models manufactured from sheet steel by welding. The material characteristics and fabrication imperfections were carefully measured and are reported in the paper. A simplified analysis model has been developed to permit the design of practical shells, and the correlation of the predictions of this analysis and the corresponding test results is shown to be good. INTRODUCTION The advent of floating and tethered production platforms for offshore oil or gas exploitation has emphasised the requirement for a design analysis relevant to thin-walled stiffened shell components. However, there is little experimental evidence with which to test the efficacy of any proposed analysis. The few available tests usually relate to aero-space practice in which the shell material is made from an aluminium alloy and the stiffeners are very closely spaced. The test programme reported here was designed to eliminate part of this gap in design information. A total of twelve shells were tested. Manufacture of the models I measurement of material and imperfection characteristics and the testing was carried out in the SERC London Centre for Marine Technology. A method of design analysis for this type of shell component has been developed by personnel at Det norske Veritas and in the paper predictions are compared with corresponding test results. MODEL FABRICATION The models were manufactured from thin steel sheet having material properties similar to structural steel using TIG welding. The special fabrication and stress-relieving techniques l provide small-scale shells with geometric imperfections within design tolerances applied in practice. A total of twelve models were made and Table 1 gives the geometrical details. Subsequent to the fabrication, the ends of the shells were lapped flat and parallel and end rings were fitted. The rings were designed to achieve a fully clamped boundary condition and hence provide a known boundary restraint for comparison with analysis. The detail of the end ring is shown in Figure 1. A sand-Araldite mixture was used to cast the shell into the groove. INITIAL IMPERFECTIONS Each shell was mounted in a measuring device and 160 readings were taken on the shell surface and 7 to 13 readings were taken along each stiffener. The number and location of the readings were dependant on the length of the shell. The apparatus is described fully in Ref L The datum for the imperfection measurements was provided by an accurately machined solid cylinder, accurate to ± 0.0025 mm of a perfect shape. It is estimated that the readings relative to the test cylinder surface were accurate to within ± 0.05mm. Readings were processed to determine deviations from a best fit cylinder and Rourrier coefficients of a series representing the imperfect surface were completed.