Date of Award

12-2007

Document Type

Thesis

Degree Name

Master of Science (MS)

Legacy Department

Mechanical Engineering

Advisor

Summers, Joshua D

Committee Member

Kurfess , Thomas R

Abstract

This thesis presents a case study on the implementation of a rule based design (RBD) process for an engineer-to-order (ETO) company. The time taken for programming and challenges associated with this process are documented in order to understand the benefits and limitations of RBD. These times are obtained while developing RBD programs for grid assemblies of bottle packaging machines that are designed and manufactured by Hartness International (HI). In this project, commercially available computer-aided design (CAD) and RBD software are integrated to capture the design and manufacturing knowledge used to automate the grid design process of HI. The stages involved in RBD automation are identified as CAD modeling, knowledge acquisition, capturing parameters, RBD programming, debugging, and testing, and production deployment. The stages and associated times in RBD program development process are recorded for eighteen different grid products. Empirical models are developed to predict development times of RBD program, specifically enabling HI to estimate their return on investment. The models are demonstrated for an additional grid product where the predicted time is compared to actual RBD program time, falling within 20% of each other. This builds confidence in the accuracy of the models. Modeling guidelines for preparing CAD models are also presented to help in RBD program development. An important observation from this case study is that a majority of the time is spent capturing information about product during the knowledge acquisition stage, where the programmer's development of a RBD program is dependent upon the designer's product knowledge. Finally, refining these models to include other factors such as time for building CAD models, programmers experience with the RBD software (learning curve), and finally extending these models to other product domains are identified possible areas of future work.

Share

COinS