On the burst of pipes under internal pressure and elevated temperature

This dissertation presents the results of a study of how burst of steel tubes is influenced by elevated temperature. The study involves experiments and modeling of burst. Burst tests were conducted on SAF 2507 super duplex stainless steel and X-52 carbon steel tubes. The tubes have diameter-to-thickness ratios of approximately 34 and lengths of approximately five tube diameters. Burst tests were conducted at constant temperature between 75°F and 350°F (24°C and 177°C). The test specimens were axially restrained and the pressurization was conducted under volume control. The pressure-circumferential strain response has an initial stiff elastic regime which terminates into a pressure plateau caused by yielding. Wall thinning eventually leads to a pressure maximum. The pressure maximum is considered to be the burst capacity of the tube. In the neighborhood of the pressure maximum, the deformation starts to localize in the form of a bulge. After the maximum pressure, wall thinning localizes on one side of the bulge and eventually a linear shear rupture develops.; Uniaxial tension tests were performed at various temperatures and strain rates for both steels. The material is modeled as a finitely deforming elastic-plastic or elastic-powerlaw viscoplastic solid calibrated to the measured stress-strain responses. The rate dependence of SAF 2507 was found to be well represented by powerlaw viscoplasticity. X-52 exhibits Lüders bands during initial yielding. This part of the stress-strain response is approximated as linear with a small positive tangent modulus. This approximation was found to result in pressure-displacement responses which are representative of those recorded in the burst tests.; A sequence of models with increasing complexity was used to simulate the burst experiments. In the simplest model, the tube was assumed to deform cylindrically. In the second model, the tube is represented as an axisymmetric membrane which can bulge. At the next level, the tube is modeled as a finitely deforming shell. The first version is axisymmetric and, in the second version, the restriction is removed allowing a bulge to grow in the unsymmetric manner seen in the final stages of the experiments. This last model enables study of the effect of local thickness imperfections.; All four models produce similar pressure-radial displacement responses up to the pressure maximum provided the length of the tube is longer than five tube diameters. Thus, even the cylindrical expansion model predicts the burst pressure accurately. The shell models allow simulation of the final stages of burst. The rate dependent constitutive model results in good predictions if the volume rate at which the tube is inflated in the physical test is matched in the simulations. The rate independent constitutive model results in good predictions if the experiment is conducted at a slow inflation rate.