Proof-Test-Based Life Prediction of High-Toughness Pressure Vessels

Abstract

The paper examines the problems associated with applying proof-test-based life prediction to vessels made of high-toughness metals. Two A106 Gr B pipe specimens containing long, through-wall, circumferential flaws were tested. One failed during hydrostatic testing and the other during tension-tension cycling following a hydrostatic test. Quantitative fractography was used to verify experimentally obtained fatigue crack growth rates and a variety of LEFM and EPFM techniques were used to analyze the experimental results. The results show that: plastic collapse analysis provides accurate predictions of screened (initial) crack size when the flow stress is determined experimentally; LEFM analysis underestimates the crack size screened by the proof test and overpredicts the subsequent fatigue life of the vessel when retardation effects are small (i.e., low proof levels); and, at a high proof-test level 2.4 × operating pressure), the large retardation effect on fatigue crack growth due to the overload overwhelmed the deleterious effect on fatigue life from stable tearing during the proof test and alleviated the problem of screening only long cracks due to the high toughness of the metal.