A parametric approach to the optimization-based design of compliant mechanisms

Matthew B. Parkinson, Larry L. Howell, Jordan J. Cox

Research output: Chapter in Book/Report/Conference proceedingConference contribution

26 Scopus citations


Several optimization-based strategies have been proposed for compliant mechanism design that do not rely on the experience or intuition of the designer. This paper demonstrates an optimization-based method wherein compliant mechanisms are modeled parametrically within an optimization and a finite element analysis package. Topological optimization is performed to minimize an objective function representing the fitness of the design. This methodology exploits the nonlinear nature of compliant mechanisms and augments optimization-based methods previously proposed. Using this method, constant-force mechanisms optimized for a displacement from 4% to 25% of the mechanism's total length were predicted to remain within 3.58% of constant force. Results from the testing of fabricated mechanisms are: for 4- 25% displacement, within 7.5% constant force; for 18-65% displacement, within 2.3%. Path generation mechanisms were designed with similarly encouraging results.

Original languageEnglish (US)
Title of host publication23rd Design Automation Conference
PublisherAmerican Society of Mechanical Engineers (ASME)
ISBN (Electronic)9780791880449
StatePublished - 1997
EventASME 1997 Design Engineering Technical Conferences, DETC 1997 - Sacramento, United States
Duration: Sep 14 1997Sep 17 1997

Publication series

NameProceedings of the ASME Design Engineering Technical Conference


ConferenceASME 1997 Design Engineering Technical Conferences, DETC 1997
Country/TerritoryUnited States

All Science Journal Classification (ASJC) codes

  • Mechanical Engineering
  • Computer Graphics and Computer-Aided Design
  • Computer Science Applications
  • Modeling and Simulation


Dive into the research topics of 'A parametric approach to the optimization-based design of compliant mechanisms'. Together they form a unique fingerprint.

Cite this