Automated design and sketching of mechanisms based on specified design requirements by employing expert system methodologies

Yih-Min Wu, University of Rhode Island

Abstract

In the early stages of mechanism design, type synthesis requires intuition and engineering experience to achieve the desired task. This can only be accomplished by an experienced design engineer.^ Current type synthesis design techniques are not fully automated. This project utilizes the computer to aid type synthesis of mechanisms by employing symbolic manipulation techniques and frame base knowledge representation in InterLisp-D environment. The study involves the design and creation of a LISP based program the purpose of which is to perform type synthesis operations which are now performed by designers with an expert knowledge of type synthesis methods. The objective of this work is two-fold: First, schematic representations of mechanisms can be automatically developed from graphs. Second, functional evaluation of mechanisms is accomplished. Computer code was developed to draw mechanisms automatically with various revolute, prismatic, and gear joints for different mechanism inversions. A designer can visually inspect the schematics to determine if the mechanisms satisfy the designer's requirements before a dimensional analysis is performed.^ Functional evaluation is based on features and joint connectivity and it includes five functional and three structural constraints. The automated process can examine large number of mechanisms in a fraction of the time traditionally required. These techniques were applied to a robot gripper design and two optimum mechanisms were found in which the functional and structural criteria are satisfied. ^

Subject Area

Engineering, Mechanical|Artificial Intelligence

Recommended Citation

Yih-Min Wu, "Automated design and sketching of mechanisms based on specified design requirements by employing expert system methodologies" (1987). Dissertations and Master's Theses (Campus Access). Paper AAI8811573.
http://digitalcommons.uri.edu/dissertations/AAI8811573

Share

COinS