A Modelling and Experimental Study for Long-Term Prediction of Localised Corrosion in Carbon Steel Overpacks for High-Le

  • PDF / 566,731 Bytes
  • 8 Pages / 414.72 x 648 pts Page_size
  • 38 Downloads / 161 Views

DOWNLOAD

REPORT


A MODELLING AND EXPERIMENTAL STUDY FOR LONG-TERM PREDICTION OF LOCALISED CORROSION IN CARBON STEEL OVERPACKS FOR HIGHLEVEL RADIOACTIVE WASTE. A. Hoch*, A. Honda**, H. Ishikawa**, F. Porter*, S. Sharland* and N. Taniguchi** *AEA Technology, Harwell, Oxfordshire, UNITED KINGDOM **Power Reactor & Nuclear Fuel Development Corporation, Tokai Works, Ibaraki-ken, JAPAN

ABSTRACT This paper describes a joint modelling and experimental study for investigation of pit growth in carbon steel High-Level Radioactive Waste overpacks under consideration in Japan. A mathematical model of the growth of corrosion pits in metals has been developed. This model is implemented in the computer program CAMLE, and includes representation of the chemical, electrochemical and migration processes that control pit-growth rates. Experiments to provide key input data for the model are described, in addition to experiments to measure 'short-term' pit growth. Predictions from the model are compared with these data. Overall, the comparisons are encouraging and the model shows good potential as a tool for assessment of the long-term corrosion behaviour of overpacks under repository conditions. Future developments of the model to improve agreement are di:K~ussed.

1. INTRODUCTION Current plans for the disposal of High-Level Radioactive Waste (HLW) in Japan involve emplacement in a deep underground repository within a multi-barrier repository system, which includes metal overpacks and a bentonite buffer. Carbon steel is one of the materials currently being considered for the overpacks.

As part of the multi-barrier system, overpacks may be required to remain intact for several hundred years, in order to provide a significant contribution to the repository performance. The main factor limiting the performance of the overpacks on these timescales is considered to be corrosion. It is anticipated that carbon steel will corrode under the repository conditions. The method for attaining the required lifetime is to design the overpacks with sufficient metal thickness to allow for the corrosion attack. The estimation of this thickness requires knowledge of the modes of corrosion likely to offect the overpacks in the repository and of the rates of metal wastage. Mathematical modelling, based on an understanding of the key chemical and physical processes, provides a means of making predictions on the required long timescales. Validation of such models is made by comparing the short-term predictions with the results of laboratory experiments that arr. ,lesigned to simulate repository conditions. Corrosion of carbon steel under repository conditions can take a number of forms including both uniform and localised corrosion, depending on the physical and chemical environment surrounding the metal. Pitting corrosion is one possible form, where relatively rapid corrosion occurs at isolated points on the metal surface. A mathematical model of the growth of pits in metals has been developed with the aim of providing a long-term predictive capability. This model is implem