Modelling and numerical simulation of combustion and multi-phase flows using finite volume methods on unstructured meshes
- Autores: Jordi Muela Castro
- Directores de la Tesis: Assensi Oliva Llena (dir. tes.), Carlos David Pérez Segarra (codir. tes.)
- Lectura: En la Universitat Politècnica de Catalunya (UPC) ( España ) en 2018
- Idioma: español
- Tribunal Calificador de la Tesis: Antonio Lecuona Neumann (presid.), Jesús Castro González (secret.), Jose Ignacio Nogueira Goriba (voc.)
- Programa de doctorado: Programa de Doctorado en Ingeniería Térmica por la Universidad Politécnica de Catalunya
- Materias:
- Enlaces
- Tesis en acceso abierto en: TDX
- Resumen
The present thesis is devoted to the development and implementation of mathematical models and numerical methods in order to carry out computational simulations of complex heat and mass transfer phenomena. Several areas and topics in the field of Computational Fluid Dynamics (CFD) have been treated and covered during the development of the current thesis, specially combustion and dispersed multi-phase flows. This type of simulations requires the implementation and coupling of different physics. The numerical simulation of multiphysics phenomena is challenging due to the wide range of spatial and temporal scales which can characterize each one of the physics involved in the problem. Moreover, when solving turbulent flows, turbulence itself is a very complex physical phenomenon that can demand a huge computational effort. Hence, in order to make turbulent flow simulations computationally affordable, the turbulence should be modelled. Therefore, throughout this thesis different numerical methods and algorithms have been developed and implemented aiming to perform multiphysics simulations in turbulent flows.
The first topic addressed is turbulent combustion. Chapter 2 presents a combustion model able to notably reduce the computational cost of the simulation. The model, namely the Progress-Variable (PV) model, relies on a separation of the spatio-temporal scales between the flow and the chemistry. Moreover, in order to account for the influence of the sub-grid species concentrations and energy fluctuations, the PV model is coupled to the Presumed Conditional Moment (PCM) model. Chapter 2 also shows the development of a smart load-balancing method for the evaluation of chemical reaction rates in parallel combustion simulations.
Chapter 3 is devoted to dispersed multiphase flows. This type of flows are composed of a continuous phase and a dispersed phase in the form of unconnected particles or droplets. In this thesis, the Eulergian-Lagrangian approach has been selected. This type of model is the best-suited for dispersed multiphase flows with thousands or millions of particles, and with a flow regime ranging from the very dilute up to relatively dense.
In Chapter 4, a new method capable of performing parallel numerical simulations using non-overlapping disconnected mesh domains with adjacent boundaries is presented. The presented algorithm stitches at each iteration independent meshes and solves them as a unique domain.
Finally, Chapter 5 addresses a transversal aspect to the previously covered topics throughout the thesis. In this chapter, a self-adaptive strategy for the maximisation of the time-step for the numerical solution of convection-diffusion equations is discussed. The method is capable of determining dynamically at each iteration which is the maximum allowable time-step which assures a stable time integration. Moreover, the method also smartly modifies the temporal integration scheme in order to maximize its stability region depending on the properties of the system matrix.
