| description abstract | Some of the materials problems associated with the use of mild steels in large gas-cooled reactor pressure vessels are discussed. Tests to failure of 5-ft-dia 0.36 percent carbon-steel vessels with through-thickness longitudinal slots, supported by tests on 7-ft-wide centrally slotted flat plates, have indicated that rapid failure at working-stress levels can only initiate from very long cracks, feet rather than inches in length. Of the mechanisms whereby realistic defects can grow to these sizes, brittle-crack propagation is considered the most important and this can be prevented by the maintenance of a minimum pressurization temperature, based on the crack-arrest temperature. The tests used to assess the crack arrest temperature of plates up to 4 in. thick are described; compared with tests on thinner specimens the thick plate gives arrest temperatures higher by approximately 10 deg C per in. of test-specimen thickness. A comparison is made of crack-arrest temperature and data given by small-scale tests, particularly the Charpy V-notch test. Mechanical limitations of creep deformation in some current designs have been more restrictive on design stress than the values allowed by the existing BS.1500. The test data quoted for stress-rupture and fatigue indicate that these modes of crack extension are not important in current designs. Possible magnitudes and effects of stress concentrations are quoted but, other than a large body of satisfactory service operation, there is little direct evidence of the effect of operating in the creep range on these stress concentrations. The importance of work of this type in justifying higher design stresses and more economic use of material is emphasized. | |
