Stress-rupture data obtained from tubular specimens stressed with internal pressure are compared with data from standard tension-bar specimens from the same heats of material. Agreement between the data of the two types of specimens is poor for thick-wall tubes when the hoop stress in the tube wall is calculated on an average stress basis. Better agreement is obtained with thin-wall tubes. These results show that any design formula that neglects the variation of stress through the thickness of a tube wall is unsatisfactory for thicker tubes. A dimensionless parameter method is presented which provides satisfactory correlation of the data for three materials, four test temperatures, and a wide range of wall thickness. Several areas in which additional research would be of value are discussed.

This content is only available via PDF.
You do not currently have access to this content.