No abstract

The results of an experimental stress analysis of the intersection region of two straight cylindrical shells are presented. Two models were used; in the first model the two axes were inclined at 30 degrees while for the second case, this angle was 60 degrees. In each case, the main shell was 6.625 in. in diameter and 0.198 in. thick, while the attached shell was 3.5 in. in diameter and 0.226 in. thick. The intersection region was subjected to in-plane and out-plane moments applied to the attached shell and the measurements were made using foil resistance rosette gauges. These measurements demonstrate that the local stress concentration in the intersection region of the main shell increases with the increase of the acute angle between the axes of the two shells; thus for a given moment loading on the attached shell, the stress concentration will be the largest when the two axes are normal to each other. IntroductionIntersecting cylindrical shell configurations are common in structural components for power and petrochemical plants as well as in off-shore structures. Piping tees and nozzles in cylindrical vessels are some specific examples. Despite their common occurrence, however, proven elastic stress analysis methods have not been generally available and consequently accurate design information for such configurations has also been lacking. This is especially true for configurations consisting of two intersecting cylindrical shells where the two axes are not perpendicular to each other, even for idealized cases with no transactions, reinforcements, or fillets in the junction region. These configurations are often used to provide restraints on piping systems and pressure vessels in power and petrochemical plants, as well as in off-shore drilling structures.A comparatively simple configuration where the two cylindrical shells intersect normally has been studied (1)-(19)t by several investigators, but with limited results. The theoretical analysis (2x4) provided rather limited results. Bijlaard (2) converted the moment loading on the attached shell into a pressure loading on the main shell. These calculations were based on a set of equations similar to the well known Fliigge equations for a cylindrical shell. The calculation assumed a thin cylindrical shell, simply supported at the ends, loaded by an external pressure in a finite region of the shell. The considered external pressure loading in a region was assumed to result from a thrust or external moment load on the attached shell. The solution to the problem was obtained by representing the displacements and loadings by double Fourier series. Naghdi and Erigen (4) obtained stresses in a shell with a circular cut-out due to internal

No abstract

No abstract

In this experiment, the method of phase spectral analysis was used to monitor ultrasonic moduli changes in a two part epoxy adhesive bonded between steel adherends during cure as well as during cyclic loading. Ultrasonic signals were generated using a piezoelectric transducer operated in the pulse-echo mode. These signals were digitized using a high-speed transient digitizer (100 MHz sampling rate) and stored for post-test analysis in the memory of a minicomputer. Based on this information, ultrasonic attenuation and phase velocity measurements (as a function of frequency) were obtained for both longitudinal and shear waves throughout the cure cycle. This technique was also used to monitor the cure of adhesives with different mix ratios in order to evaluate the utility of this approach for quality control purposes. The potential use of the method to study fatigue damage development was also considered.

Traditional remaining strength assessment methods such as B31G [1] and API RP-579 [2] are inherently conservative to account for uncertainties and thus assure safety. With the advent of reliability based assessment methods, built-in conservatisms at the model level are undesirable. Instead, the statistical model descriptions are used to account for the model uncertainties. To date, several research studies have compared the accuracy of these assessment methods as burst pressure predictors with newer limit-state models [3]. In these comparisons, the assessment methods such as B31G do not perform well as burst predictors because their original formulation was used (i.e., safety factor was not removed), leading some to reject B31G for use in a reliability based assessment. In this paper the effectiveness of the burst capacity prediction of these assessment models are compared by reformulating these assessment methods into appropriate limit-state models such that they are compared on the same basis. The comparison shows that the B31G based limit state model is one of the most accurate burst capacity predictors, well suited for use in a reliability based framework.

No abstract

No abstract