HIGH ACCURACY MODELING FOR ADVANCED NUCLEAR REACTOR CORE DESIGNS USING MONTE CARLO BASED COUPLED CALCULATIONS
Open Access
- Author:
- Puente Espel, Federico
- Graduate Program:
- Nuclear Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- April 19, 2010
- Committee Members:
- Kostadin Nikolov Ivanov, Dissertation Advisor/Co-Advisor
Kostadin Nikolov Ivanov, Committee Chair/Co-Chair
Maria Nikolova Avramova, Committee Chair/Co-Chair
Dr Stefan Misu, Committee Member
Prof Robert Edwards, Committee Member
Arthur Thompson Motta, Committee Member
Victor Nistor, Committee Member - Keywords:
- MCNP5
COBRA-TF
NJOY
NEM
Coupled Calculations
BWR - Abstract:
- The main objective of this PhD research is to develop a high accuracy modeling tool using a Monte Carlo based coupled system. The presented research comprises the development of models to include the thermal-hydraulic feedback to the Monte Carlo method and speed-up mechanisms to accelerate the Monte Carlo criticality calculation. Presently, deterministic codes based on the diffusion approximation of the Boltzmann transport equation, coupled with channel-based (or sub-channel based) thermal-hydraulic codes, carry out the three-dimensional (3-D) reactor core calculations of the Light Water Reactors (LWRs). These deterministic codes utilize nuclear homogenized data (normally over large spatial zones, consisting of fuel assembly or parts of fuel assembly, and in the best case, over small spatial zones, consisting of pin cell), which is functionalized in terms of thermal-hydraulic feedback parameters (in the form of off-line pre-generated cross-section libraries). High accuracy modeling is required for advanced nuclear reactor core designs that present increased geometry complexity and material heterogeneity. Such high-fidelity methods take advantage of the recent progress in computation technology and coupled neutron transport solutions with thermal-hydraulic feedback models on pin or even on sub-pin level (in terms of spatial scale). The continuous energy Monte Carlo method is well suited for solving such core environments with the detailed representation of the complicated 3-D problem. The major advantages of the Monte Carlo method over the deterministic methods are the continuous energy treatment and the exact 3-D geometry modeling. However, the Monte Carlo method involves vast computational time. The interest in Monte Carlo methods has increased thanks to the improvements of the capabilities of high performance computers. Coupled Monte-Carlo calculations can serve as reference solutions for verifying high-fidelity coupled deterministic neutron transport methods with detailed and accurate thermal-hydraulic models. The development of such reference high-fidelity coupled multi-physics scheme is described in this dissertation on the basis of MCNP5, NEM, NJOY and COBRA-TF (CTF) computer codes. This work presents results from studies performed and implemented at the Pennsylvania State University (PSU) on both accelerating Monte Carlo criticality calculations by using hybrid nodal diffusion Monte Carlo schemes and thermal-hydraulic feedback modeling in Monte Carlo core calculations. The hybrid MCNP5/CTF/NEM/NJOY coupled code system is proposed and developed in this dissertation work. The hybrid coupled code system contains a special interface developed to update the required MCNP5 input changes to account for dimension and density changes provided by the thermal-hydraulics feedback module. The interface has also been developed to extract the flux and reaction rates calculated by MCNP5 to later transform the data into the power feedback needed by CTF (axial and radial peaking factors). The interface is contained in a master program that controls the flow of the calculations. Both feedback modules (thermal-hydraulic and power subroutines) use a common internal interface to further accelerate the data exchange. One of the most important steps to correctly include the thermal hydraulic feedback into MCNP5 calculations begins with temperature dependent cross section libraries. If the cross sections used for the calculations are not at the correct temperature, the temperature feedback cannot be included into MCNP5 (referred to the effect of temperature on cross sections: Doppler boarding of resolve and unresolved resonances, thermal scattering and elastic scattering). The only method of considering the temperature effects on cross sections is through the generation (or as introduced in this dissertation through a novel interpolation mechanism) of continuous energy temperature-dependent cross section libraries. An automated methodology for generation of continuous energy temperature-dependent cross section libraries has been developed as part of the hybrid Monte Carlo-based coupled core studies at PSU. This tool is used together with the automated cross-section temperature interpolation capability for intermediate points. The automated methodology, combined with the interpolation capability, has considerably reduced the cross section generation time. A new methodology for generation and interpolation of temperature-dependent thermal scattering cross section libraries for MCNP5 is introduced as well. Using the interpolation methodology specially designed for thermal scattering cross sections, a thermal scattering grid at the desired temperature was generated. This gives the possibility of performing MCNP5 criticality calculations at the correct moderator temperature and improving the accuracy of the calculations. A cross section update methodology has been included, which efficiently reduces the time of the cross section libraries update. Several acceleration strategies are introduced and implemented in the hybrid coupled code system. The computation process is greatly accelerated by calculating the 3-D distributions of fission source and thermal-hydraulics parameters with the coupled NEM/CTF code and then using coupled MCNP5/CTF code to fine tune the results to obtain an increased accuracy. The PSU NEM code employs cross-sections generated by MCNP5 for pin-cell based nodal compositions. Finally, the hybrid coupled system is automated and enhanced in order to provide the user with an efficient and easy to use high accuracy modeling tool.