zirconium oxidation and TRACE at Penn State

 Today I found this:

http://d-net.us/pdf/Nuc_E_431W.pdf

 

NucE431W S2013

Spent Fuel Pool Analysis of a BWR-4

Fuel Bundle Under Loss of Coolant

                          Conditions Using TRACE
 
                            For the period of January 7, 2013 through May 1, 2013


The Pennsylvania State UniversityDepartment of Mechanical and Nuclear Engineering
University Park, PA 16802
5/1/2013

ABSTRACT

The purpose of this project was to investigate the ability of TRACE to accurately model a single fuel bundle within a spent fuel pool and to determine the conditions for when a fuel bundle might initiate rapid oxidation of its zirconium cladding. Three models were developed and simulated using the TRACE thermal-hydraulics system code: a full height BWR bundle cooled by air, a partial height test composed of five BWR fuel bundles, and a full-height BWR bundle initially filled with water. The first two models represented experiments recently completed at the Sandia National Laboratory, while the third is a model of a planned experiment. Each model was used to estimate the time at which breakaway (rapid oxidation of the cladding) would begin, which is at approximately 1100 K based on experimental data. It was found from the full-scale air test that TRACE has the ability to approximately model a fuel bundle heat up within a spent fuel pool up to approximately 1073 K. The data prior to 1073 K has been determined to be an accurate representation with regards to the specified input parameters. It was found that fuel bundles will undergo ignition (the point in TRACE where the metal water reaction calculations begin and where the temperature increase jumps to several degrees per second – also known as the breakaway point) within eight hours for power levels above 4 kW for the full-scale water model. Additionally, it was found that there exists a clear direct relationship between the coolant flow rate and collapsed water height. More importantly, the TRACE results provided data indicating the ignition time as a function of bundle power.

EXECUTIVE SUMMERY   (Note by Leyse: The spelling is by Penn State.)
 
Initial tests conducted at the Sandia National Laboratory assumed that the worst case scenario for heat up of a fuel bundle in a spent fuel pool would be the case of complete fuel uncovery. Cooling of the bundle would occur by natural convection to air in this case, with flow entering through the bottom nozzle and exiting through the top of the bundle. This case, while a significant threat to the safety of the fuel bundles, may in fact not be the worst case scenario. The worst case scenario occurs when there exists only a small amount of water at the bottom of the fuel bundles which will cause a blockage in air flow through the bundle. This case would effectively reduce the heat transfer rates from the fuel causing the fuel to heat up at a higher rate than if natural circulation was occurring. Understanding the behavior of spent fuel bundles under beyond-design basis scenarios such as fuel bundle ignition has become important. In order to understand fuel bundle behavior in these situations, computer models and physical tests must be performed and benchmarked against one another before any one program can be considered reliable to produce accurate models.

In order to determine the modeling capabilities of TRACE two models were created and benchmarked against tests performed by Sandia National Laboratories. These models were the Full Length BWR-4 Full Bundle in Air, the 1x4 Partial Length BWR-4 Fuel Bundle Configuration in Air and the Full Length BWR-4 Fuel Bundle in Water. After determining the modeling capabilities with regards to the experimental data a third model, Full Length BWR-4 Fuel Bundle in Water, was generated based upon the original Full Length BWR-4 Full Bundle in Air model. This third model was generated in order to provide a starting point to develop a Full Scale Water ignition test at Sandia National Laboratory. The Full Scale Air test currently uses two “break” components within TRACE to create a pressure differential to drive the flow through the bundle whereas the actual test had a controlled forced flow which gives rise to some variations in the observed data from TRACE and that provided by Sandia National Laboratory.
 
Ultimately it was found that the bundle heat up and ignition behavior of the experiment conducted by Sandia National Laboratory is accurately modeled by TRACE. From here the water model was created by changing the working fluid from air to water and modifying the bottom “break” component in TRACE to a “fill” component to provide a controlled forced flow through the model. Varying bundle powers were examined in addition to three distinct flow rates. Relationships were generated for time to breakaway as a function of bundle power and collapsed water height as a function of flow rate. From these relationships it was found that increased powers reduce the time to breakaway and higher flow rates result in larger collapsed water heights for all examined cases which is to be expected.