Degree Type


Date of Award


Degree Name

Doctor of Philosophy


Nuclear Engineering

First Advisor

R. A. Danofsky


Since 1977, a program has been underway at experimental breeder reactor-II (EBR-II) to evaluate the performance of metal and mixed-oxide liquid metal reactor (LMR) fuel elements after clad failure (breach). The motivation for this activity, called run beyond cladding breach (RBCB) testing, is to continue safe operation of EBR-II after occurrence of a single or multiple clad failures. Principal safety concerns are excessive release of fission gas (FG) and mixed-oxide fuel/sodium interaction, which result in release of by-products to the sodium and may result in blockage of coolant flow. Excessive release of FG is controlled by the cover gas cleanup system (CGCS) which purifies the cover gas by cryogenic means. The probability of coolant blockage due to accumulation of the by-products of the fuel/sodium interaction is small but is considered a safety issue. To comply with safety regulations, the location and growth of the breach must be known during RBCB operation. An expert system (ES) has been designed to advise the reactor operator under normal and/or RBCB operation. It contains as a knowledge base (KB) (a) the already developed and validated analytical models (i.e., tag gas (TC), delayed neutron (DN), and FG analyses); (b) rules obtained from inductive learning (i.e., learned by experience); (c) heuristic rules; and (d) rules obtained from new data. The TG analysis is used to identify the breached element. The DN analysis is primarily used for indicating a new breach and monitoring the extent of old breaches. FG analysis is also used for indicating a new breach, and gaining other information such as the location of the breach (i.e., fuel column or plenum breach), type of breach (i.e., fuel clad mechanical interaction (FCMI) or creep-rupture, only in mixed-oxide fuel), and fission product release to the primary coolant. The ES uses the information provided by the analytical models and supplies information to the reactor operator, thus, ultimately automating diagnostics during RBCB operation.



Digital Repository @ Iowa State University,

Copyright Owner

Ramin Mikaili



Proquest ID


File Format


File Size

185 pages