BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20210916T132451Z
LOCATION:Louis Favre
DTSTART;TZID=Europe/Stockholm:20210707T120000
DTEND;TZID=Europe/Stockholm:20210707T123000
UID:submissions.pasc-conference.org_PASC21_sess141_msa384@linklings.com
SUMMARY:A large scale phase-field fracture simulations
DESCRIPTION:Minisymposium\n\nA large scale phase-field fracture simulation
 s\n\nKopanicakova, Krause\n\nThe phase-field approach to fracture models c
 rack paths with a possibly complex topology by means of a diffusive damage
  variable, which is coupled to elasticity. The numerical solution of the u
 nderlying model is computationally challenging as it requires a solution o
 f large-scale, non- convex, bound-constrained minimization problems. Moreo
 ver, the arising linear systems are severely ill-conditioned due to local 
 changes in the damage variable. Here, we propose a recur- sive multilevel 
 trust-region (RMTR) method to efficiently solve such minimization problems
 . The RMTR method combines the global convergence property of the trust-re
 gion method and the optimality of the multilevel method. The solution proc
 ess is accelerated by employing level- dependent objective functions, mini
 mization of which can yield good coarse level corrections for the fine lev
 el problem. In the context of the phase-field fracture approach, it is cha
 llenging to design efficient level-dependent objective functions as the un
 derlying mathematical model relies on the mesh-dependent parameters. We in
 troduce solution-dependent objective functions that combine a fine-level d
 escription of the crack paths with the coarse level discretization. The ov
 erall performance of our solution strategy will be demonstrated using seve
 ral numerical examples, including fracture propagation of complex stochast
 ic fracture networks with up to 1'000 fractures. In the end, we also inves
 tigate the strong and weak scaling properties of our simulation framework 
 up to 9'216 MPI processes and 1.9e8 degrees-of-freedom.\n\nDomain: CS and 
 Math, Solid Earth Dynamics, Engineering
END:VEVENT
END:VCALENDAR