Optimization of Natural Frequencies for Fabrication-Aware Shape Modeling

Christian Hafner, Przemyslaw Musialski, Thomas Auzinger, Michael Wimmer, Leif Kobbelt
Optimization of Natural Frequencies for Fabrication-Aware Shape Modeling
Poster shown at ACM SIGGRAPH 2015 (2015-08-09-2015-08-13) In Proceedings of ACM SIGGRAPH 2015 Posters. August 2015.
[abstract] [eposter] [thumbnail] [video]

Information

Abstract

Given a target shape and a target frequency, we automatically synthesize a shape that exhibits this frequency as part of its natural spectrum, while resembling the target shape as closely as possible. We employ finite element modal analysis with thin-shell elements to accurately predict the acoustic behavior of 3d solids. Our optimization pipeline uses an input surface and automatically calculates an inner offset surface to describe a volumetric solid. The solid exhibits a sound with the desired pitch if fabricated from the targeted material. In order to validate our framework, we optimize the shape of a tin bell to exhibit a sound at 1760 Hz. We fabricate the bell by casting it from a mold and measure the frequency peaks in its natural ringing sound. The measured pitch agrees with our simulation to an accuracy of 2.5%. In contrast to previous method, we only use reference material parameters and require no manual tuning.

Additional Files and Images

Additional images and videos

image: image
thumbnail: thumbnail

Additional files

abstract: abstract
eposter: eposter
video: video

Weblinks

No further information available.

BibTeX

@misc{hafner-2015-onff,
  title =      "Optimization of Natural Frequencies for Fabrication-Aware
               Shape Modeling",
  author =     "Christian Hafner and Przemyslaw Musialski and Thomas
               Auzinger and Michael Wimmer and Leif Kobbelt",
  year =       "2015",
  abstract =   "Given a target shape and a target frequency, we
               automatically synthesize a shape that exhibits this
               frequency as part of its natural spectrum, while resembling
               the target shape as closely as possible. We employ finite
               element modal analysis with thin-shell elements to
               accurately predict the acoustic behavior of 3d solids. Our
               optimization pipeline uses an input surface and
               automatically calculates an inner offset surface to describe
               a volumetric solid. The solid exhibits a sound with the
               desired pitch if fabricated from the targeted material. In
               order to validate our framework, we optimize the shape of a
               tin bell to exhibit a sound at 1760 Hz. We fabricate the
               bell by casting it from a mold and measure the frequency
               peaks in its natural ringing sound. The measured pitch
               agrees with our simulation to an accuracy of 2.5%. In
               contrast to previous method, we only use reference material
               parameters and require no manual tuning.",
  month =      aug,
  booktitle =  "Proceedings of ACM SIGGRAPH 2015 Posters",
  event =      "ACM SIGGRAPH 2015",
  location =   "Los Angeles, CA, USA",
  note =       "Lecturer: P. Musialski",
  publisher =  "ACM",
  note =       "Poster presented at ACM SIGGRAPH 2015
               (2015-08-09--2015-08-13)",
  pages =      "--",
  keywords =   "natural frequencies, modal analysis, shape optimization",
  URL =        "https://www.cg.tuwien.ac.at/research/publications/2015/hafner-2015-onff/",
}