In this work, we studied, in a two-dimensional context, the coupling between Biot’s equations, which describe the behavior of a poroelastic medium, and a transport equation governing the distribution of a solute concentration within the pore fluid. Biot’s equations were introduced in both the two-field formulation, where the unknowns are the displacements of the solid and the fluid, and in the three-field formulation, which includes the fluid pressure as a third variable. The two formulations were compared, deriving the discrete form and highlighting the main differences. After introducing the weak formulation of the coupled problem, spatial approximation was performed using the PolyDG method, while time discretization was carried out using the beta-Newmark method. The transport problem was also discretized in space with the PolyDG method and, for time discretization, the Crank-Nicolson method was used. The PolyDG method allows greater flexibility in mesh and polynomial degree selection, thus enabling easy coupling. The choice of the beta-Newmark and Crank-Nicolson methods is justified by the goal of maintaining second-order time accuracy for both problems. The numerical analysis of the two-field coupled problem was carried out in MATLAB using the Lymph library, while the three-field coupled problem was studied using the multiphenics library in Python. Moreover, Mandel’s model for poroelasticity was considered in an inertial context, highlighting differences with results from the literature for short time intervals due to the introduction of an inertial term. These effects diminish over longer time intervals. The analytical solutions of Mandel’s problem coincide with the numerical results obtained using the three-field poroelastic model. Finally, a section of the human brain was analyzed by simulating physiological conditions to study the behavior of the three-field coupled model and compare it with results from the literature. By imposing pressure at the domain boundaries and a stress in the ventricular region, a strong agreement with the literature results was observed, obtaining a pressure range between 6 mmHg and 8.5 mmHg in the brain parenchyma and maximum solid displacement of 1 mm in the brainstem, third and fourth ventricle.
In questo lavoro abbiamo studiato, in un contesto bidimensionale, l’accoppiamento tra le equazioni di Biot, che descrivono il comportamento di un mezzo poroelastico, e un’equazione di trasporto che governa la distribuzione della concentrazione di un soluto nel fluido presente nei pori. Le equazioni di Biot sono state introdotte sia nella formulazione a due campi, in cui le incognite sono lo spostamento del solido e del fluido, sia nella formulazione a tre campi, che considera come terza variabile la pressione del fluido. Sono state confrontate le due formulazioni, ricavandone per entrambe la forma discreta ed evidenziandone le differenze. Dopo aver introdotto la formulazione debole del problema accoppiato, si è proceduto con la discretizzazione spaziale mediante PolyDG e la discretizzazione temporale attraverso il metodo beta-Newmark per le equazioni di Biot e il metodo di Crank-Nicolson per il problema di trasporto. Grazie al metodo PolyDG è possibile una flessibilità nella scelta della mesh e del grado polinomiale, consentendo così la possibilità di un facile accoppiamento mentre la scelta di adottare il metodo beta-Newmark e il metodo di Crank-Nicolson è giustificata per mantenere un’approssimazione in tempo di secondo ordine per entrambi i problemi. L’analisi numerica del problema accoppiato a due campi è stata svolta su MATLAB sfruttando la libreria Lymph mentre il problema accoppiato a tre campi è stato studiato mediante la libreria Multiphenics di Python. Inoltre, è stato considerato il modello di Mandel per la poroelasticità, evidenziando le differenze con i risultati della letteratura per intervalli di tempo brevi a causa dell’introduzione di un termine inerziale. Tali effetti risultano attenuati per intervalli di tempo più lunghi. Le soluzioni analitiche del problema di Mandel coincidono con i risultati dell’analisi numerica ottenuti mediante il modello poroelastico a tre campi. Infine è stato analizzata una sezione del cervello umano, simulando le condizioni fisiologiche, per studiare il comportamento del modello accoppiato a tre campi mettendolo in relazione con i risultati della letteratura. Imponendo la pressione ai bordi del dominio e uno sforzo nella porzione ventricolare è stata osservata ottima analogia con i risultati della letteratura, ottenendo una pressione variabile tra 6 mmHg e 8.5 mmHg nel parenchima cerebrale e uno spostamento massimo del solido di 1 mm nel tronco encefalico, nel terzo e nel quarto ventricolo.
A polyDG-discretized coupled model for scalar transport in a poroelastic medium
MORASCA, ALESSANDRO
2024/2025
Abstract
In this work, we studied, in a two-dimensional context, the coupling between Biot’s equations, which describe the behavior of a poroelastic medium, and a transport equation governing the distribution of a solute concentration within the pore fluid. Biot’s equations were introduced in both the two-field formulation, where the unknowns are the displacements of the solid and the fluid, and in the three-field formulation, which includes the fluid pressure as a third variable. The two formulations were compared, deriving the discrete form and highlighting the main differences. After introducing the weak formulation of the coupled problem, spatial approximation was performed using the PolyDG method, while time discretization was carried out using the beta-Newmark method. The transport problem was also discretized in space with the PolyDG method and, for time discretization, the Crank-Nicolson method was used. The PolyDG method allows greater flexibility in mesh and polynomial degree selection, thus enabling easy coupling. The choice of the beta-Newmark and Crank-Nicolson methods is justified by the goal of maintaining second-order time accuracy for both problems. The numerical analysis of the two-field coupled problem was carried out in MATLAB using the Lymph library, while the three-field coupled problem was studied using the multiphenics library in Python. Moreover, Mandel’s model for poroelasticity was considered in an inertial context, highlighting differences with results from the literature for short time intervals due to the introduction of an inertial term. These effects diminish over longer time intervals. The analytical solutions of Mandel’s problem coincide with the numerical results obtained using the three-field poroelastic model. Finally, a section of the human brain was analyzed by simulating physiological conditions to study the behavior of the three-field coupled model and compare it with results from the literature. By imposing pressure at the domain boundaries and a stress in the ventricular region, a strong agreement with the literature results was observed, obtaining a pressure range between 6 mmHg and 8.5 mmHg in the brain parenchyma and maximum solid displacement of 1 mm in the brainstem, third and fourth ventricle.| File | Dimensione | Formato | |
|---|---|---|---|
|
2025_07_Morasca_Tesi_01.pdf
accessibile in internet per tutti
Dimensione
7.19 MB
Formato
Adobe PDF
|
7.19 MB | Adobe PDF | Visualizza/Apri |
|
2025_07_Morasca_Executive Summary_02.pdf
accessibile in internet per tutti
Dimensione
2.98 MB
Formato
Adobe PDF
|
2.98 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/10589/240257