BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Design and Control of Structures that Adapt to Loads through Large
  Shape Changes - Genaro Senatore\, Swiss Federal Institute of Technology (
 EPFL)\, Applied Computing and Mechanics Laboratory\, Lausanne\, Switzerlan
 d
DTSTART:20201009T140000Z
DTEND:20201009T150000Z
UID:TALK152716@talks.cam.ac.uk
CONTACT:Dr Maria Marques de Carvalho
DESCRIPTION:Well-designed adaptive structures are able to meet typical str
 ength and deflection requirements using significantly less material and en
 ergy resources compared with passive structures. Structural adaptation thr
 ough sensing and actuation is employed to counteract the effects of loadin
 g through control of the internal load-path and the structure shape (exter
 nal geometry). The design criterion is minimization of the structure whole
 -life energy comprising an embodied part in the material and an operationa
 l part for adaptation to loading. \nRecent work has investigated the effic
 acy of structural adaptation through large shape reconfigurations that inv
 olve geometrically non-linear effects. A new design strategy that includes
  a combination of geometry and actuator placement optimization with non-li
 near shape control has been formulated\, whereby the structure is designed
  to ‘morph’ into optimal shapes as the load changes.  \nStructures pro
 duced by this strategy can meet must stricter deflection limits with respe
 ct to passive solutions\, which enables new design such as super slender h
 igh-rise buildings\, bridges and self-supporting roof systems. A near-full
  scale prototype has been successfully tested validating key assumptions a
 nd numerical predictions. A new mechanics-based control algorithm has been
  developed based on a linear-sequential formulation which combines shape o
 ptimization and non-linear force method. This formulation enables accurate
  shape control using minimum computational cost\, which makes it suitable 
 for real-time applications.\nClassification based on supervised learning h
 as been employed to infer position and magnitude of the applied load as we
 ll as to improve control accuracy. Experimental results are in good accord
 ance with numerical predictions showing that effective stress homogenizati
 on can be achieved through controlled large shape changes\, which results 
 in 39% embodied energy savings with respect to an equivalent weight-optimi
 zed passive structure.\n
LOCATION:Zoom (email structures-admin@eng.cam.ac.uk for link)
END:VEVENT
END:VCALENDAR