The Woodruff School

SUMMARY

SPaRC, Stochastic Particle Response Calculator, is a new stochastic neutron transport code that has been developed and optimized for the computation of response functions. SPaRC transports neutrons from a specified fixed source distribution and computes responses as neutrons stream through and then exit regions of interest. The code makes use of both multi-group and continuous energy nuclear data and takes advantage of parallel computing through the message passing interface (MPI). In order to test the neutron transport routine, various small benchmark problems were solved with SPaRC and compared to results generated with MCNP. Results show very good agreement between the solutions generated by these codes for both multi-group and continuous energy calculations. The responses generated by SPaRC have been tailored for use in the coarse mesh transport (COMET) method. COMET is a hybrid stochastic/deterministic method shown to compute fast and accurate solutions for a variety of nuclear systems. In order to obtain these solutions, COMET makes use of stochastically pre-computed response functions aggregated into a library for use in a deterministic iteration scheme. SPaRC generates these response functions for use with the COMET method, with the added capability of performing these calculations as needed. This on-the-fly capability for response generation enables the use of the COMET method for calculations where the state of a problem changes with time. SPaRC’s ability to generate responses during a calculation eliminates the need for a fully pre-computed response library to cover the entire possible solution space, extending the capability of COMET to thermal-hydraulic and depletion calculations. Sample calculations on the reactor assembly level were performed in order to test the accuracy of the SPaRC generated response functions. First, responses were generated for uncontrolled, controlled, and gadded assemblies with both MCNP and SPaRC. Next, COMET calculations were performed using these two sets of responses for the different assembly types in order to generate eigenvalue and pin fission density. The results generated from the MCNP and SPaRC responses agreed within 0.05% for the core eigenvalue and within 0.002% for pin powers. SPaRC is a newly developed fixed source radiation transport code. The neutron transport method has been benchmarked against the stochastic transport code MCNP with good agreement and new database management and creation routines have been developed to aid response generation. SPaRC introduces response function flexibility to the COMET method that facilitates thermal hydraulic and depletion calculations.

SUBJECT:	Ph.D. Dissertation Defense

BY:	Ryan Hon

TIME:	Thursday, March 31, 2016, 10:00 a.m.

PLACE:	Boggs, 3-39

TITLE:	Development of Methods and Tools for On-the-Fly Response Function Generation for Criticality Calculations

COMMITTEE:	Dr. Farzad Rahnema, Co-Chair (NRE) Dr. Bojan Petrovic, Co-Chair (NRE) Dr. Dingkang Zhang (NRE) Dr. Edmond Chow (CSE) Dr. Alireza Haghighat (VT) Dr. Luiz Leal (IRSN)

SUMMARY SPaRC, Stochastic Particle Response Calculator, is a new stochastic neutron transport code that has been developed and optimized for the computation of response functions. SPaRC transports neutrons from a specified fixed source distribution and computes responses as neutrons stream through and then exit regions of interest. The code makes use of both multi-group and continuous energy nuclear data and takes advantage of parallel computing through the message passing interface (MPI). In order to test the neutron transport routine, various small benchmark problems were solved with SPaRC and compared to results generated with MCNP. Results show very good agreement between the solutions generated by these codes for both multi-group and continuous energy calculations. The responses generated by SPaRC have been tailored for use in the coarse mesh transport (COMET) method. COMET is a hybrid stochastic/deterministic method shown to compute fast and accurate solutions for a variety of nuclear systems. In order to obtain these solutions, COMET makes use of stochastically pre-computed response functions aggregated into a library for use in a deterministic iteration scheme. SPaRC generates these response functions for use with the COMET method, with the added capability of performing these calculations as needed. This on-the-fly capability for response generation enables the use of the COMET method for calculations where the state of a problem changes with time. SPaRC’s ability to generate responses during a calculation eliminates the need for a fully pre-computed response library to cover the entire possible solution space, extending the capability of COMET to thermal-hydraulic and depletion calculations. Sample calculations on the reactor assembly level were performed in order to test the accuracy of the SPaRC generated response functions. First, responses were generated for uncontrolled, controlled, and gadded assemblies with both MCNP and SPaRC. Next, COMET calculations were performed using these two sets of responses for the different assembly types in order to generate eigenvalue and pin fission density. The results generated from the MCNP and SPaRC responses agreed within 0.05% for the core eigenvalue and within 0.002% for pin powers. SPaRC is a newly developed fixed source radiation transport code. The neutron transport method has been benchmarked against the stochastic transport code MCNP with good agreement and new database management and creation routines have been developed to aid response generation. SPaRC introduces response function flexibility to the COMET method that facilitates thermal hydraulic and depletion calculations.