BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//7.2.3.1//EN
TZID:Asia/Kolkata
X-WR-TIMEZONE:Asia/Kolkata
BEGIN:VEVENT
UID:157@cds.iisc.ac.in
DTSTART;TZID=Asia/Kolkata:20251107T113000
DTEND;TZID=Asia/Kolkata:20251107T123000
DTSTAMP:20251104T085638Z
URL:https://cds.iisc.ac.in/events/ph-d-thesis-defense-online-mode-cds-07-n
 ovember-2025-structure-preserving-physics-informed-neural-networks-for-ani
 sotropic-porous-media-with-pressure-dependent-viscosity/
SUMMARY:Ph.D: Thesis Defense: ONLINE MODE: CDS: 07\, November 2025 "Structu
 re-Preserving Physics-Informed Neural Networks for Anisotropic Porous Medi
 a with Pressure-Dependent Viscosity"
DESCRIPTION:DEPARTMENT OF COMPUTATIONAL AND DATA SCIENCES\nPh.D. Thesis Def
 ense\n\n\n\nSpeaker : Mr. NISCHAL KARTHIK MAPAKSHI\nS.R. Number : 06-18-00
 -10-12-19-1-17456\nTitle : "Structure-Preserving Physics-Informed Neural N
 etworks for\nAnisotropic Porous Media with Pressure-Dependent Viscosity"\n
 Research Supervisor : Prof. Soumyendu Raha\nThesis Examiner : Prof. Rajend
 ra K. Ray\, Indian Institute of Technology Mandi.\nDate &amp\; Time : Nove
 mber 07\, 2025 (Friday)\, 11.30 A.M.\nVenue : The Thesis Defense will be h
 eld on MICROSOFT TEAMS\nPlease click on the following link to join the The
 sis Defense:\nMS Teams link\n\n\n\nABSTRACT\nModeling flow through porous 
 media with realistic physical constraints remains a longstanding challenge
  in subsurface engineering. Anisotropy in permeability\, pressure-dependen
 t viscosity\, and non-negativity requirements on pressure fields introduce
  mathematical complexity and numerical instability\, especially in mesh-fr
 ee learning frameworks. This thesis presents a structure-preserving Physic
 s-Informed Neural Network (PINN) formulation for simulating nonlinear Darc
 y flow in anisotropic porous domains governed by Barus-type viscosity laws
 . To enforce discrete maximum principles (DMP) and ensure physically admis
 sible pressure fields\, two constraint strategies are developed. A hard en
 forcement mechanism is implemented via output transformations that restric
 t predictions to within prescribed bounds. In parallel\, a soft enforcemen
 t strategy augments the loss function with penalization terms that discour
 age DMP violations. These approaches are systematically evaluated within b
 oth strong-form PINNs and variational PINNs\, the latter based on Galerkin
  and Variational Multiscale (VMS) formulations. A series of numerical stud
 ies demonstrates the performance of the proposed methods across several se
 ttings. A one-dimensional benchmark using manufactured solutions validates
  convergence. In a square reservoir with a central borehole\, the effect o
 f permeability anisotropy is analyzed by sweeping the directional contrast
  ratio. It is observed that hard constraints are essential to maintain DMP
  adherence under strong anisotropy. In a separate case involving localized
  central forcing\, the impact of nonlinear viscosity is assessed by varyin
 g the Barus coefficient. Increasing nonlinearity results in larger DMP vio
 lations unless physically motivated constraints are imposed. Sensitivity s
 tudies also reveal the influence of boundary condition density\, penalty w
 eights\, and network depth on stability and accuracy. The results indicate
  that while both soft and hard constraints improve physical fidelity\, har
 d enforcement consistently outperforms in preserving maximum principles. A
 mong all tested configurations\, the VMS PINN with hard constraints yields
  the most robust performance\, maintaining zero violations across anisotro
 py sweeps and producing stable velocity and pressure fields. All models ar
 e implemented using the DeepXDE library and use physically meaningful para
 meter ranges relevant to oil recovery and geologic carbon storage. This wo
 rk demonstrates that constraint-aware PINNs can serve as scalable\, reliab
 le solvers for complex porous media problems without sacrificing physical 
 realism.\n\n\n\nALL ARE WELCOME
CATEGORIES:Events,Thesis Defense
END:VEVENT
BEGIN:VTIMEZONE
TZID:Asia/Kolkata
X-LIC-LOCATION:Asia/Kolkata
BEGIN:STANDARD
DTSTART:20241107T113000
TZOFFSETFROM:+0530
TZOFFSETTO:+0530
TZNAME:IST
END:STANDARD
END:VTIMEZONE
END:VCALENDAR