The simulation of Enhanced Oil Recovery (EOR) processes such as the injection of miscible gas in a reservoir requires the use of a fully compositional multiphase model for flow in porous media with detailed physical laws based on equations of state. From a mathematical point of view, this problem consists in solving a large system of nonlinear equations coupling the conservation laws for the components with the thermodynamic equilibrium constraints. Traditionally, the reservoir simulation industry relies on the formulation introduced by Coats, based on natural variables (pressures, saturations and component molar fractions) and on flash calculations to detect phase appearance and disappearance. Although this formulation is stable with respect to phase transitions, it adds complexity because it requires to manage the set of present phases and the associated unknowns and equations at each point of the time-space domain. Recently, new formulations with nonlinear complementarity conditions have been proposed for modeling systems which have to switch often between different states (single-phase, two-phase, three-phase). In particular, the formulation introduced by Lauser et al., that uses phase pressures, saturations and component fugacities as main unknowns and complementarity conditions for handling phase transitions, leads to a fixed set of unknowns and a fixed set of equations regardless of the actual present phases. Therefore it allows to avoid the flash calculation, otherwise necessary to determine the set of present phases and their composition. However this scheme requires to solve, for every cell of the time-space computational domain, a system of nonlinear equations to compute the molar fractions and their derivatives as a function of pressures, temperature and fugacities. In this work, we analyzed and compared the natural variable and the complementarity conditions formulations for solving compositional multiphase flow problem with cubic equations of state. We implemented each formulation in a Fortran90 program and we validated the complementarity condition formulation via several test cases. Moreover, in order to evaluate the performance of our programs, we compared our results with the ones given by the C++ software DuMux.
La simulazione di processi di Enhanced Oil Recovery (EOR) come l'iniezione di un gas miscibile in un reservoir richiede l'uso di un modello multifase e multicomponente per il flusso in mezzi porosi con leggi fisiche basate sulle equazioni di stato. Da un punto di vista matematico, questo problema consiste nel risolvere un ampio sistema di equazioni non lineari in cui le leggi di conservazione per i componenti sono accoppiate ai vincoli di equilibrio termodinamico. Tradizionalmente, l'industria della reservoir simulation utilizza la formulazione introdotta da Coats, basata sulle variabili naturali (pressioni, saturazioni e componenti molari) e sulla flash calculation per individuare la comparsa e la scomparsa delle fasi. Sebbene questa formulazione sia stabile rispetto ai cambiamenti di fase, essa aggiunge complessità al sistema perché comporta la gestione dell'insieme delle fasi presenti e delle relative incognite ed equazioni in ogni punto del dominio spazio-tempo. Recentemente sono state proposte nuove formulazioni con condizioni di complementarità non lineari per modellizzare sistemi che devono cambiare spesso stato (monofase, bifase, trifase). In particolare, ci siamo concentrati sulla formulazione introdotta da Lauser et al., che utilizza pressioni, saturazioni e fugacità dei componenti come principali incognite e condizioni di complementarietà per gestire i cambiamenti di fase. Questa formulazione porta ad un insieme fisso di incognite e di equazioni indipendente dall'insieme delle fasi presenti e perciò permette di evitare la flash caculation, altrimenti necessaria per determinare l'insieme delle fasi presenti e la loro composizione. Tuttavia questa formulazione richiede di risolvere, per ogni cella del dominio computazionale spazio-temporale, un sistema di equazioni non lineari per calcolare le frazioni molari e le loro derivate in funzione di pressioni, temperatura e fugacità. In questo lavoro di tesi, abbiamo analizzato e confrontato la formulazione con le variabili naturali e la formulazione con le condizioni di complementarietà per la modellizzazione e la simulazione di un flusso multifase multicomponente con equazioni di stato cubiche. Abbiamo implementato ciascuna formulazione in un programma in Fortran90 e abbiamo convalidato la formulazione con le condizioni di complementarità tramite diversi casi di test. Inoltre, al fine di valutare le prestazioni dei nostri programmi, abbiamo confrontato i nostri risultati con quelli forniti dal software C++ DuMux.
Numerical methods for compositional multiphase flows in porous media
LUSETTI, IRENE
2017/2018
Abstract
The simulation of Enhanced Oil Recovery (EOR) processes such as the injection of miscible gas in a reservoir requires the use of a fully compositional multiphase model for flow in porous media with detailed physical laws based on equations of state. From a mathematical point of view, this problem consists in solving a large system of nonlinear equations coupling the conservation laws for the components with the thermodynamic equilibrium constraints. Traditionally, the reservoir simulation industry relies on the formulation introduced by Coats, based on natural variables (pressures, saturations and component molar fractions) and on flash calculations to detect phase appearance and disappearance. Although this formulation is stable with respect to phase transitions, it adds complexity because it requires to manage the set of present phases and the associated unknowns and equations at each point of the time-space domain. Recently, new formulations with nonlinear complementarity conditions have been proposed for modeling systems which have to switch often between different states (single-phase, two-phase, three-phase). In particular, the formulation introduced by Lauser et al., that uses phase pressures, saturations and component fugacities as main unknowns and complementarity conditions for handling phase transitions, leads to a fixed set of unknowns and a fixed set of equations regardless of the actual present phases. Therefore it allows to avoid the flash calculation, otherwise necessary to determine the set of present phases and their composition. However this scheme requires to solve, for every cell of the time-space computational domain, a system of nonlinear equations to compute the molar fractions and their derivatives as a function of pressures, temperature and fugacities. In this work, we analyzed and compared the natural variable and the complementarity conditions formulations for solving compositional multiphase flow problem with cubic equations of state. We implemented each formulation in a Fortran90 program and we validated the complementarity condition formulation via several test cases. Moreover, in order to evaluate the performance of our programs, we compared our results with the ones given by the C++ software DuMux.File | Dimensione | Formato | |
---|---|---|---|
2018_7_Lusetti.pdf
non accessibile
Descrizione: Testo della tesi
Dimensione
18.3 MB
Formato
Adobe PDF
|
18.3 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/10589/141688