Spectral Cascade-Transport Turbulence Model Development for Two-Phase Flow

Turbulence modeling remains a challenging problem in nuclear reactor applications, particularly for the turbulent multiphase flow conditions in nuclear reactor subchannels. Understanding the fundamental physics of turbulent multiphase flows is crucial for the improvement and further development of multiphase flow models used in reactor operation and safety calculations. Reactor calculations with Reynolds-averaged Navier-Stokes (RANS) approach continue to become viable tools for reactor analysis. The on-going increase in available computational resources allows for turbulence models that are more complex than the traditional two-equation models to become practical choices for nuclear reactor computational fluid dynamic (CFD) and multiphase computational fluid dynamic (M-CFD) simulations. Similarly, increased computational capabilities continue to allow for higher Reynolds numbers and more complex geometries to be evaluated using direct numerical simulation (DNS), thus providing more validation and verification data for turbulence model development. Spectral turbulence models are a promising approach to M-CFD simulations. These models resolve mean flow parameters as well as the turbulent kinetic energy spectrum, reproducing more physical details of the turbulence than traditional two-equation type models. Previously, work performed by other researchers on a spectral cascade-transport model has shown that the model behaves well for single and bubbly twophase decay of isotropic turbulence, single and two-phase uniform shear flow, and single-phase flow in a channel without resolving the near-wall boundary layer for relatively low Reynolds number. Spectral models are great candidates for multiphase RANS modeling since bubble source terms can be modeled as contributions to specific turbulence scales.;This work focuses on the improvement and further development of the spectral cascadetransport model (SCTM) to become a three-dimensional (3D) turbulence model for use in M-CFD codes. To aid in SCTM development and validation a spectral analysis of single and two-phase bubbly DNS data in different geometries was performed with investigation of the modulation of the turbulent kinetic energy spectrum slope due to the presence of bubbles. A new spectral analysis technique was developed to show that modifications to the energy spectrum slope are due to the presence of bubble wakes. Spectral analysis results are essential aids in turbulence model development and validation. Further work on the one-dimensional (1D) SCTM formulation was performed to improve model behavior for higher Reynolds number channel flow than previously examined, where the boundary layer close to the solid wall is now resolved and good agreement was achieved between the SCTM and DNS data. The SCTM was then implemented into the 3D MCFD package NPHASE-CMFD and tested for turbulent single-phase, monodispersed bubbly twophase, and polydispersed bubbly two-phase flow in various geometries. The SCTM predictions were compared with the k-a model, experimental data, and DNS data. The objective of the work is to improve and develop the SCTM and subsequently provide the numerical framework for the SCTM to be used in M-CFD predictions of multiphase flow in complex nuclear reactor geometries.