Date of Award

5-2018

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Mechanical Engineering

Committee Member

Joshua D. Summers, Committee Co-Chair

Committee Member

Marissa L. Shuffler, Co-Chair

Committee Member

Cameron J. Turner

Abstract

The systematic engineering design process equips designers with tools and methods necessary to understand and solve a given design problem. Function decomposition is one such tool that allows designers to decompose the given problem into sub-problems which may be easier to address. Research on Function modeling, specifically Function Structure models, has focused on improving model construction techniques and using the Function Structure models to support concept generation. Additionally, Function Structure models have also been traditionally used as individual design tools; however, most other conceptual design tools are used in a collaborative setting (e.g. gallery sketching, method 3-6-5, etc.). This research investigates the use of Function Structure models as a collaborative tool by using seed models constructed using three different chaining methods (forward chaining, backward chaining, and nucleation) identified in a pilot protocol study. These seed models were intended to represent a partially completed model created by one designer, which was then delivered to the next designer for completion. A designer study and a protocol study were conducted to identify differences between the final Function Structure models generated using different seed models, based on the percent increase in the number of functions and flows, change in model complexity, and a rubric based evaluation of the model. Results show that using a nucleation seed model yield a higher increase in function and flows, as well as a larger change in model complexity. Analysis of the rubric based model evaluation shows that the presence of the seed model improves the evaluation scores, however, the type of chaining method used does not impact the final score. These results suggest that teaching of Function Structure models should include explicit identification of the different chaining methods, and recommends nucleation as the chaining method of choice. Moreover, future research areas are identified with respect to further comparison of chaining methods, as well as investigation of behavioral patterns in the modeling activity.

Share

COinS