A Reassessment of PWR Pressure Vessel Integrity During Overcooling Accidents, Considering 3-D Flaws

Abstract

A continuing analysis of the pressurized-thermal-shock problem associated with PWR postulated overcooling accidents indicates that the previously accepted degree of conservatism in the fracture-mechanics model needs to be more closely evaluated, and, if excessive, reduced. One feature that was believed to be conservative was the use of two-dimensional as opposed to finite-length (three-dimensional) flaws. The degree of conservatism could not be adequately investigated because of computational limitations and a lack of knowledge regarding flaw behavior; however, that situation has changed to the extent that some cases involving finite-length flaws can be studied. A flaw of particular interest is one that is located in an axial weld of a plate-type vessel. For those vessels that suffer relatively high-radiation damage in the welds, the length of the flaw will be no greater than the length of the weld, and recent calculations indicate that a deep flaw of that length (∼2 m) is not effectively infinitely long, contrary to previous thinking. The benefit to be derived from consideration of the 2-m flaw and also a semi-elliptical flaw with length-to-depth ratio of 6/1 was investigated by analyzing several postulated transients. In doing so the sensitivity of the benefit to a specified maximum crack-arrest toughness and to the duration of the transient was investigated. Results of the analysis indicate that for some conditions the benefit in using the 2-m flaw is substantial, but it decreases with increasing pressure, and above a certain pressure there may be no benefit, depending on the duration of the transient and the limit on crack-arrest toughness.