BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:EXPERIMENTAL VALIDATION AND UNCERTAINTY QUANTIFICATION OF PARTITIO
 NED MODELS
DTSTART:20150505T103000
DTEND:20150505T113000
DTSTAMP:20260407T081511Z
UID:f2c36b49ecd84854c41adcbe15b5515806d8f8368d4ff33b4c9627ad
CATEGORIES:Conferences - Seminars
DESCRIPTION:Garrison Stevens\, Glenn Department of Civil Engineering\,Clem
 son University\, Clemson\, South Carolina\, USA\nMulti-scale and multi-ph
 ysics modeling involves the simultaneous use of multiple constituent model
 s that resolve different scales or physics by iteratively communicating ou
 tputs of these models with each other\, referred to as partitioned analysi
 s. Such coupled models aim to achieve a compromise between accuracy and ef
 ficiency when a single scale or physics is not a sufficiently accurate rep
 resentation of the system. In strongly coupled models\, the iterative natu
 re of coupling operations causes uncertainties and errors in constituent m
 odels to be passed back and forth through the interface\, which may lead t
 o the accumulation of these uncertainties and errors in the coupled model.
  Thus\, the predictive ability of a coupled model depends upon the accurac
 y and precision of its constituent models. A distinct advantage of partiti
 oned analysis\, that has not been recognized previously\, is the transpare
 ncy it offers for the quantification of uncertainties and errors in model 
 predictions through the decomposition of a physical system into multiple s
 cales. The predictive ability of constituent models developed independentl
 y can be improved in an isolated manner by rigorously calibrating model pr
 edictions against experimental measurements conducted in the corresponding
  domains. Furthermore\, understanding and quantifying the sources of uncer
 tainties and errors in the coupled models can allow model developers to st
 rategically improve the model’s predictive capability. Herein\, a partit
 ioned approach for experiment-based validation and uncertainty quantificat
 ion is presented. Bias-corrected partitioned analysis utilizes separate-ef
 fect experiments conducted within each constituent’s domain to test the 
 validity of the independent constituents and integral-effect experiments e
 xecuted within the coupled domain to validate the entire coupled system. T
 his approach for calibration and validation utilizing both separate- and i
 ntegral-effect experiments is demonstrated on a multi-scale plasticity mod
 el coupling a finite element code at the macro-scale and visco-plastic sel
 f-consistent code at the meso-scale. Results demonstrate that improvements
  in predictive capability can be achieved by utilizing separate-effect exp
 eriments as a means of appropriately bias-correcting constituent model pre
 dictions during coupling iterations.
LOCATION:GC G1 515 http://plan.epfl.ch/?lang=fr&zoom=19&recenter_y=5864267
 .48334&recenter_x=730958.60537&layerNodes=fonds\,batiments\,labels\,inform
 ation\,parkings_publics\,arrets_metro&floor=1&q=GC_G1%20515
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR