Towards an Online Prognostic System for Predicting the Axial Shrinkage of AGR Cores

Abstract


 
 
 In the UK, there is the desire to extend the operation of the Advanced Gas-cooled Reactor (AGR) power plants beyond their initial design lifetimes of 35 years. As part of the justification of extended operation, an increased understanding of the current and future health of the graphite reactor cores is required. One measure of the health of the AGR power plants is the axial height of the graphite core, which can be determined through measurements undertaken during statutory outages. These measurements are currently used to manually make predictions about the future height of the core, through identifying the relevant data sources, extracting the relevant parameters and generating the predictions is time- consuming and onerous. This paper explores an online prognostic approach to support these manual predictions, which provides benefits in terms of rapid, updated predictions as soon as new data becomes available. Though the approach is described with reference to a case study of the UK’s AGR design of power plant, similar challenges of predicting passive structure health also exist in other designs of power plant with planned license extensions.