Tests on variable cross-section thin-walled pressure vessels

Abstract

The object of this project is twofold. The primary object is to evaluate and analyse the strain set up in a variable cross-section (frustum shaped), thin walled pressure vessel, under variable hydraulic pressure. Then, from the results obtained from the first tests, other vessels were designed and constructed with the same outside dimensions but of variable thickness, such that the hoop stress was kept constant along the slanting face of the frustum. For the first purpose, a mild steel pressure vessel was constructed and subjected to several hydraulic tests, each time with a smaller wall thickness. From these tests, the hoop strain was measured at various radii for different thicknesses and pressures. Also, theoretical and experimental values of strain were compared and discrepancies analysed. From a combination of the results obtained from the first tests and theoretical considerations, it was possible to design other vessels of the same outside and end dimensions, but of variable wall thickness, so as to give a constant hoop stress along the slanting face when subjected to pressure. Unfortunately, the designed vessels could not be constructed with a parabolic surface, as is required theoretically, due to machining difficulties. Therefore, two vessels were constructed with the best approximate linear surface. Then, from the results obtained, the error in the approximation was analysed. Besides these vessels, two other tool-steel vessels were machined; both had the same outside end dimensions, but differed in the slanting face length. From these vessels, besides studying the effect on strain of using a different material, size and end effects were analysed. Again, theoretical and experimental values of strain were compared and discrepancies analysed. Finally, it should be noted that the S.I. system of units was used in all the calculations carried out in this work, even though some of the apparatus was calibrated in British Units. However, it was found to be easier and neater to use the centimetre as the unit of length in drawing certain graphs.