The IAEA ICSP on Numerical Benchmarks for Multiphysics Simulation of Pressurized Heavy Water Reactor Transients was initiated in 2016 to facilitate the development of a set of open access, standardized, numerical test problems for postulated accident scenarios in a CANDU styled Reactor. The test problems include a loss of coolant accident resulting from an inlet header break, a loss of flow accident caused by a single pump trip, and a loss of regulation accident due to inadvertently withdrawn adjusters. The Benchmark was split into phases, which included stand-alone physics and thermal-hydraulics transients, coupled steady state simulations, and coupled transients. This thesis documents the results that were generated through an original TRACE/PARCS coupling methodology that was developed specifically for this work. There is a strong emphasis on development methods and step by step verification throughout the thesis, to provide a framework for future research in this area. In addition to the Benchmark results, additional studies on propagation of fundamental nuclear data uncertainty, and sensitivity analysis of coupled transients are reported in this thesis. Two Phenomena and Key Parameter Identification and Ranking Tables were generated for the loss of coolant accident scenario, to provide feedback to the Benchmark Team, and to add to the body of work on uncertainty/sensitivity analysis of CANDU style reactors. Some important results from the uncertainty analysis work relate to changes in the uncertainty of figures of merit such as integrated core power, and peak core power magnitude and time, between small and large break loss of coolant accidents. The analysis shows that the mean and standard deviation of the integrated core power and maximum integrated channel power, are very close between a 30% header break and a 60% header break, despite the peak core power being much larger in the 60% break case. Furthermore, it shows that there is a trade off between the uncertainty in the time of the peak core power, and the magnitude of the peak core power, with smaller breaks showing a smaller standard deviation in the magnitude of the peak core power, but a larger standard deviation in when this power is reached during the transient, and vice versa for larger breaks. From the results of the sensitivity analysis study, this thesis concludes that parameters related to coolant void reactivity and shutoff rod timing and effectiveness have the largest impact on loss of coolant accident progressions, while parameters that can have a large impact in other transients or reactor designs, such as fuel temperature reactivity feedback and control device incremental cross sections, are less important. / Thesis / Master of Science (MSc) / This thesis documents McMaster’s contribution to an International Atomic Energy Agency Benchmark on Pressurized Heavy Water Reactors that closely resemble the CANDU design. The Benchmark focus is on coupling of thermal-hydraulics and neutron physics codes, and simulation of postulated accident scenarios. This thesis contains some select results from the Benchmark, comparing the results generated by McMaster to other participants. This thesis also documents additional work that was performed to propagate fundamental nuclear data uncertainty through the coupled transient calculations and obtain an estimate of the uncertainty in key figures of merit. This work was beyond the scope of the Benchmark and is a unique contribution to the open literature. Finally, sensitivity studies were performed on one of the accident scenarios defined in the Benchmark, the loss of coolant accident, to determine which input parameters have the largest contribution to the variability of key figures of merit.
| Identifer | oai:union.ndltd.org:mcmaster.ca/oai:macsphere.mcmaster.ca:11375/25927 |
| Date | January 2020 |
| Creators | Groves, Kai |
| Contributors | Novog, David, Engineering Physics |
| Source Sets | McMaster University |
| Language | English |
| Detected Language | English |
| Type | Thesis |
Page generated in 0.0043 seconds