Search

Experimental results obtained with 31 nickel circular cylindrical shell specimens are evaluated in the light of the theory. The test results have already been reported in SUDAAR No. 415. The experimental creep buckling times obtained in axial compression at a temperature of 650F were found to be in reasonable agreement with the theoretical formulas. (Author).

Test equipment suitable for the study of the creep buckling of axially compressed circular cylindrical shells was developed and built. With the aid of this equipment, thirty-one electroformed nickel cylinders of radius-to-thickness ratios ranging from 30.6 to 96.4 were tested at a temperature of 650F. The loading of each specimen was interrupted usually once, and in some cases twice, to permit an exact measurement of the creep deformations produced by the axial compression. Diagrams showing the deformed shapes of eight generators of each specimen are presented at two or three stages of the creep buckling process. Inspection of the figures reveals that specimens of this kind either buckle axisymmetrically, or begin the creep buckling process in a axisymmetric manner but change over to a multilobed pattern in later stages of the deformations. (Author).

Two additional experiments using the single specimen approach described in detail in this report supplement and extend the authors previous research by considering circular cylinders with higher R/t ratios which buckled at a lower percentage of the classical critical load. The shells had R/t ratios of 946 and 1419 and buckled at 43.9 and 24.6 percent of classical, respectively. In all four experiments, 204 tests were conducted. Ranges in test machine extensional stiffness from 589,000 lb/in. to 2400 lb/in. were considered; R/t values ranged from 226-1419 and Pcr /Pcl values from 0.249 to 0.77. L/D ratios were approximately consistant in all cases . The results provide overwhelming evidence that the Tsien criterion is inapplicable in all problems considered. This may be due to the invalidity of the criterion itself or to the inadequacy of the large displacement analysis. The consequences, however, are the same; scatter in experimental results is not due to the influence of test machine rigidity and a new look at the large displacement analysis and the appropriate criterion is needed.

Liquid-gas interfaces are studied from two viewpoints. The first method uses the classical surface tension model of the interface. The differential equation defining the equilibrium shape of the interface relates the surface tension with the pressure difference across the interface. The boundary conditions for the equation assume that the contact angles (the angles the liquid forms with the container waalls) are constant and independent of the body forces. An iterative method for finding the equilibrium shape is developed, for the case of small body forces. The method is used in two cases and compared with results of other methods. The stability of the interface is investigated for a two-dimensional layer spread over a segment of a rigid cylinder with rigid radial walls. The eigenfunctions of the vibrations of layers of nonuniform depth are expanded in terms of the eigenfunctions for the layers of uniform depth. Approximations are made by truncating the series. The coefficients are obtained by minimizing the mean square error along the interface. The results are given in the form of a relationship between a dimensionless stability parameter and the angle between the radial walls for the case of limiting stability. A comparison of the approximate results with known solutions shows good agreement.