Prediction of External Corrosion for Steel Cylinders--2002 Report

Schmoyer, RLS

Oak Ridge National Laboratory

关键词: Corrosion;    Management;    36 Materials Science;    Monitoring;    Portsmouth Gaseous Diffusion Plant;   

DOI  :  10.2172/814650
RP-ID  :  ORNL/TM-2002/143
RP-ID  :  AC05-00OR22725
RP-ID  :  814650

美国|英语

来源: UNT Digital Library

【 摘 要 】

The United States Department of Energy (DOE) manages the UF{sub 6} Cylinder Project. The project was formed to maintain and safely manage the depleted uranium hexafluoride (UF{sub 6}) stored in approximately 50,000 carbon steel cylinders. The cylinders are located at three DOE sites: the East Tennessee Technology Park (ETTP) site in Oak Ridge, Tennessee; the Paducah Gaseous Diffusion Plant (PGDP) in Paducah, Kentucky, and the Portsmouth Gaseous Diffusion Plant (PORTS) in Portsmouth, Ohio. The System Requirements Document (SRD) (LMES 1997a) delineates the requirements of the project, and the actions needed to fulfill these requirements are specified in the System Engineering Management Plan (SEMP) (LMES 1997b). This report documents activities that in whole or part satisfy specific requirements and actions stated in the UF{sub 6} Cylinder Project SRD and SEMP with respect to forecasting cylinder conditions. The results presented here supercede those presented by Lyon (1995, 1996, 1997, 1998, 2000), and Schmoyer and Lyon (2001). Many of the wall thickness projections made in this report are conservative, because they are based on the assumption that corrosion trends will continue, despite activities such as improved monitoring, relocations to better storage, painting, and other improvements in storage conditions relative to the conditions at the times most of the wall thickness measurements were made. For thin-wall cylinders (design nominal wall thickness 312.5 mils), the critical minimum wall thicknesses criteria used in this report are 0 (breach), 62.5 mils, and 250 mils (1 mil = 0.001 in.). For thick-wall cylinders (design nominal wall thickness 625 mils), the thickness criteria used in this report are 0, 62.5 mils, and 500 mils. The criteria triples are preliminary boundaries identified within the project that indicate (1) loss of material (UF{sub 6}), (2) safe handling and stacking operations, and (3) standards for off-site transport and contents transfer criteria, respectively. In general, these criteria are based on an area of wall thinning. However, the minimum thickness predicted in this report is essentially for a point--an area of about 0.01 square inches--because the thickness measurements on which the predictions are based are essentially for points. For thicknesses criteria greater than zero, conclusions based on minimum point thicknesses are conservative. Because of the interaction of UF{sub 6}, with atmospheric moisture and steel, a point breach would deteriorate in a year to one-inch diameter hole (DNFSB 1995), however, and so small area approximations should be close for the breach criteria. The most recently collected data, entered into the corrosion model database and not available for the previous report (Schmoyer and Lyon 2001), consists of evaluations of wall loss of 48 inch thin-wall cylinders: 301 cylinders at Paducah, 101 at ETTP, and 139 at Portsmouth; 14 thick-wall cylinders at Portsmouth; and 99 model 30A cylinders at Paducah. However, because of missing values, repeated measures on the same cylinders, outliers, and other data problems, however, not all of these measurements are necessarily used in the corrosion analysis. In several cases, difficulty with the data is also due to a mathematical approach to cylinder corrosion modeling that is used in this report, in Schmoyer and Lyon (2001), and in earlier reports by Lyon. Therefore, an alternative approach is also considered in this report. In previous reports, minimum wall thicknesses have been modeled indirectly through separate models of initial thickness and maximum pit depth. In order to estimate minimum wall thicknesses, the initial thickness and maximum pit depth models are combined using mathematics that assumes independence of the statistical distributions of the initial thicknesses and maximum pit depths. Initial thicknesses are modeled from wall thickness maxima measured at relatively uncorroded wall areas of each cylinder. Maximum pit depths for each cylinder are estimated as differences between the initial thickness estimates and measured minimum wall thicknesses. The pit depth maxima estimates are modeled as a function of age and a cylinder grouping based on location.