Direct numerical simulation (DNS) of turbulent flows is nowadays not affordable for computers due to high memory request, but in very small computational domains, where dimensions are in the order of centimetres. The simplifying approach of large eddy simulation (LES) is becoming more and more used and studied in academical environment, but still is difficultly applied to industrial large scale simulations with reasonable costs. Reynolds averaged Navier-Stokes (RANS) equations are nowadays the most applied solution to the problem of turbulence simulation in industry. Here rises the issue of understanding the interaction between turbulence and chemical kinetics. Much effort was spent towards gaseous reactions (oxidations in particular) and many successful models were developed. Unfortunately, these models cannot be applied properly to liquid reactions as they are. An important branch of computational fluid-dynamics (CFD) in industry is dedicated to reactor design and safety. In particular, prediction of thermal behaviour is crucial in exothermic reactions, where runaway may occur. Prediction of reaction pathway is also essential when selectivity problems are analysed. CFD can be used in order to reduce (or even avoid) experiments with these kind of reactions if a good turbulence-kinetics interaction (TKI) model is adopted. Moreover, precise design requires knowledge of detailed fluid dynamics, which is difficultly predicted without CFD. The aim of this work is then the development of a new TKI approach to liquid reactions, its implementation in the commercial software ANSYS Fluent and its application to reactor design and safety problems.
Allo stato dell'arte, la simulazione diretta (DNS) dei flussi turbolenti non è fattibile, se non in domini molto piccoli dell'ordine del centimetro, a causa dell'elevato costo computazionale richiesto. La semplificazione introdotta nelle large eddy simulation (LES) non è a tutt'oggi sufficiente per applicazioni industriali con costi ragionevoli, mentre desta sempre più interesse in ambito accademico. La soluzione più adottata a livello industriale per la simulazione di flussi turbolenti è l'uso delle equazioni mediate nel tempo (Reynolds averaged Navier-Stokes, RANS). Nasce di conseguenza il problema di modellazione dell'interazione cinetica-turbolenza (turbulence-kinetics interaction, TKI), largamente studiata per reazioni gassose (maggiormente ossidazioni) e con discreto successo. Questi modelli, per come sono ora, sono risultati però inadatti alle applicazioni in fase liquida. La fluidodinamica computazionale (computational fluid-dynamics, CFD) trova applicazione industriale nel progetto e nella verifica di sicurezza dei reattori. La stima del profilo termico del sistema è essenziale per esempio per prevenire le conseguenze di runaway in reazioni esotermiche. Al contempo, è importante caratterizzare il profilo di reazione quando si hanno problemi di selettività. La CFD può essere adottata per ridurre (o addirittura evitare) la fase sperimentale per queste reazioni. Ciò è possibile solo se un buon modello di accoppiamento cinetica-turbolenza viene usato. Lo stesso modello è anche necessario per caratterizzare completamente la fluidodinamica del sistema, con lo scopo di progettare il reattore; senza modellazione CFD questa procedura può risultare assai ardua. Lo scopo di questo lavoro è quindi lo sviluppo di un modello TKI per le reazioni liquide, la sua implementazione del software commerciale ANSYS Fluent e infine la sua applicazione ai problemi di progetto e verifica di sicurezza dei reattori.
A new approach to micromixing modelling for liquid reactions and its application to reactor design and safety
FLORIT, FEDERICO
2016/2017
Abstract
Direct numerical simulation (DNS) of turbulent flows is nowadays not affordable for computers due to high memory request, but in very small computational domains, where dimensions are in the order of centimetres. The simplifying approach of large eddy simulation (LES) is becoming more and more used and studied in academical environment, but still is difficultly applied to industrial large scale simulations with reasonable costs. Reynolds averaged Navier-Stokes (RANS) equations are nowadays the most applied solution to the problem of turbulence simulation in industry. Here rises the issue of understanding the interaction between turbulence and chemical kinetics. Much effort was spent towards gaseous reactions (oxidations in particular) and many successful models were developed. Unfortunately, these models cannot be applied properly to liquid reactions as they are. An important branch of computational fluid-dynamics (CFD) in industry is dedicated to reactor design and safety. In particular, prediction of thermal behaviour is crucial in exothermic reactions, where runaway may occur. Prediction of reaction pathway is also essential when selectivity problems are analysed. CFD can be used in order to reduce (or even avoid) experiments with these kind of reactions if a good turbulence-kinetics interaction (TKI) model is adopted. Moreover, precise design requires knowledge of detailed fluid dynamics, which is difficultly predicted without CFD. The aim of this work is then the development of a new TKI approach to liquid reactions, its implementation in the commercial software ANSYS Fluent and its application to reactor design and safety problems.File | Dimensione | Formato | |
---|---|---|---|
2017_07_Florit.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: Testo della tesi
Dimensione
5.48 MB
Formato
Adobe PDF
|
5.48 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/134980