Volume 76, Issue 3

In order to perform safety oriented transient analyses, the entire nuclear power plant has to be simulated in most cases. However, a high level of resolution is only needed where the safety parameters may reach the safety criteria. In this paper, three scale resolutions are defined (NPP level, core level, fuel pin level). Thermal-hydraulic and neutron kinetic codes are already available for each scale. By coupling them, it is possible to develop a complete simulation tool giving access to the required level of precision at each scale level. For this purpose the Gesellschaft fuer Anlagen- und Reaktorsicherheit (GRS) mbH has developed a number of coupled thermal-hydraulics/neutron-kinetics code systems which can solve the defined goal on assembly-wise or pin-by-pin level. Three types of couplings are applied depending on the utilized codes and application purposes: internal, external and parallel. An overview of the existing and future coupled systems available at GRS is given in this paper.

The reactor dynamics code DYN3D has been coupled with the CFD code ANSYS CFX and the 3D thermal hydraulic core model FLICA4. In the coupling with ANSYS CFX, DYN3D calculates the neutron kinetics and the fuel behavior including the heat transfer to the coolant. The physical data interface between the codes is the volumetric heat release rate into the coolant. In the coupling with FLICA4 only the neutron kinetics module of DYN3D is used. Fluid dynamics and related transport phenomena in the reactor's coolant and fuel behavior is calculated by FLICA4. The correctness of the coupling of DYN3D with both thermal hydraulic codes was verified by the calculation of different test problems. These test problems were set-up in such a way that comparison with the DYN3D stand-alone code was possible. This included steady-state and transient calculations of a mini-core consisting of nine real-size PWR fuel assemblies with ANSYS CFX/DYN3D as well as mini-core and a full core steady-state calculation using FLICA4/DYN3D.

The Light Water Reactor (LWR) dynamics code DYN3D is extended and adopted for the application to block-type High Temperature Gas-Cooled Reactor (HTGR). A procedure for the cross section generation for the HTGR core calculations was developed. The modified Reactivity-Equivalent Physical Transformation (RPT) approach is applied in order to eliminate the double-heterogeneity of HTGR fuel elements in the deterministic lattice calculations. A full core analysis of the reference simplified HTGR core is performed with DYN3D using macroscopic nodal cross sections provided by HELIOS. The SP 3 transport approximation is integrated into the multi-group DYN3D code to take anisotropy of the neutron flux and heterogeneity of the core more precisely into account. The SP 3 method was developed for hexagonal geometry of the graphite blocks, where the hexagons are subdivided into triangular nodes. A 3D heat conduction module coupled with a channel-type coolant flow model is implemented into the code. It is shown that there is significant redistribution of the produced heat by heat conduction between the graphite blocks.

To investigate the influence of nuclear data uncertainties on reactor core calculations systematically, the sampling based uncertainty and sensitivity software package SUSA developed at GRS was extended for the use with nuclear covariance data. Varied nuclear data are generated randomly corresponding to the uncertainty information from the covariance matrices. After performing a large number of calculations with these data, the results are statistically evaluated; this can be done not only for integral, but also for local output quantities like the assembly power distribution. The method is applied to multi-group Monte Carlo calculations stationary states of the PWR MOX/UO 2 core transient benchmark, and to corresponding nodal diffusion calculations. Unexpectedly large uncertainties result for the radial power distribution. The uncertainties in the nodal results agree very well with those in the Monte Carlo reference results; thus, it is possible to apply the random sampling method to determine the influence of nuclear data uncertainties on transient core calculations.

A detailed assessment of the JEFF 3.1 and ENDF/B VII.0 nuclear data libraries with the KANEXT code system has been performed on each level of the reactor simulation steps. This allowed for the sensitivity estimation of the lattice and core parameters to changes in nuclide and specific reaction cross section. The direct calculation of cross section data influence, in particular for specific reaction types, is possible due to the unique substitution feature of the group constant module GRUCAL. Further on, the different impact of several reaction types on the lattice versus core calculation points out the “engineering uncertainty” using the two major libraries for the identical problem. MCNP(X) and KANEXT cell models can be mutually adjusted such that differences between stochastic and deterministic data processing can be identified. The current analysis emphasizes, that the two major nuclear data libraries JEFF 3.1 and ENDF/B VII.0 exhibit to some extent compensation of deviations rather than reliable integral data calculation. On a special type of a fast reactor application we show that exchange of JEFF 3.1 and ENDF/B VII.0 cross sections introduce large discrepancies. A major part is attributed to the inelastic cross section. Further investigated discrepancies concerning the capture and the fission cross section in the epithermal energy range might be also a concern for LWR which was beyond the scope of the current study.

The Boltzmann Transport equation is the governing formalism upon which simulations of nuclear reactors are performed, in particular when strong absorption or anisotropic scattering are significant. On the left (loss) hand side of the balance equation one finds the absorption and the scattering cross section Σ a (E′), Σ s (E′) respectively. Those cross sections are energy and temperature dependent i.e. Doppler Broadened. The scattering cross section appears explicitly again on the right (production) hand side of the equation in its differential form ∫∫ 0 ∞ Σ(E → E′); Ω → Ω′)dE dΩ. However, this term is commonly evaluated at 0 K and it does not account for the existing resonances which are the underlying characteristic for Doppler Broadening. Evidently one gets an inherent inconsistency between the integral and differential scattering cross section within the transport solver codes. In this study this missing Doppler Broadened formalism for the differential scattering cross section is introduced in its stochastic and deterministic form. The impact on core criticality is shown to be up to 600 pcm and the change in the nuclides' inventory significant, in particular the 239 PU content can be changed by several percents.

The paper describes the implementation of S3R (core neutronics) and RELAP5-3D (RCS thermal-hydraulics) in a PWR training simulator for the Grohnde NPP located at the Kraftwerks-Simulator-Gesellschaft (KSG) in Essen, Germany. The models are briefly described as well as the coupling between these two codes and the interface with the rest of the simulator. The paper also describes the procedure that will be used to update the simulator core data after future core reloads. Results from the integrated simulator are presented.

This paper describes the coupled code system S3K-RELAP5 and the application of the code system to BWR transients without scram. In addition short descriptions are given for application of the coupled system to fast transients and rod drop accidents. The models and methods of the code system are presented. Validation results are shown for a recorded transient event in the Forsmark 3 reactor. It is concluded that S3K-RELAP5 adequately captures the complicated interaction between physical processes in the reactor as well as the essential reactor protection and control systems, which qualifies it for applications to fast and slow transients, with or without scram, and Reactivity Initiated Accidents (RIA).

For the nuclear power plant Lungmen (two blocks of Advanced Boiling Water Reactor (ABWR)) in Taiwan a plant model for the thermal hydraulic program TRACE (TRAC/RELAP Advanced Computational Engine) was developed. This model is and will be used for the simulation of normal and anomalous plant behaviour as well as for the simulation of incident scenarios. The presentation of this model is done in three steps: The first step is the development of a TRACE model of Lungmen nuclear power plant (NPP) which includes the vessel, the main steam lines and important control systems (such as the feedwater control system, recirculation flow control system, etc.). Key parameters were identified to refine the model further in the frame of a steady state analysis. The second step is the performance of TRACE transient analyses, such as MSIV closure direct scram (MSIVCD, MSIV = Main Steamline Isolation Valve) and loss of feedwater flow (LOFW). The above transient data of Final Safety Analysis Report (FSAR) are used to verify the Lungmen NPP TRACE model. The trends of their analysis results are roughly similar. It indicates that TRACE model is satisfying for the purpose of Lungmen NPP safety analyses. The third step is the prediction analysis of Lungmen NPP startup tests by using the TRACE model. The prediction analysis results of TRACE comply with the startup tests procedure criteria.