|M.Sc Student||Ben-Yehuda Tomer|
|Subject||Evaluating Complicatedness in Mechanical Design|
|Department||Department of Mechanical Engineering||Supervisors||Professor Reuven Katz|
|Dr. A. Zonnenshain|
|Full Thesis text|
Complexity and complicatedness in mechanical design have become the subjects of interest and research in the past few decades. With that, the needs to quantify and measure these properties have emerged. Few researches recognize the differences between complexity and complicatedness, and typically, only the term “complexity” is regarded. In this research, the differences between complexity and complicatedness are explored. It is concluded that complexity is a system property which can describe several aspects of the design, and isn’t necessarily a negative trait. Complicatedness, on the other hand, results from the complexity involved in these design aspects, and can be avoided at the design stage. Several existing methods for evaluating complexity are investigated, and the most relevant ones are used to analyze equivalent systems. The advantages and disadvantages of these methods are noted. We conclude that the complicatedness of mechanical design depends on the complexity of design properties such as number of parts and interfaces, manufacturing processes and assembly. A model for evaluating complicatedness is derived based on these parameters, using both previous methods and new methods that surface in this research.
We analyze three sets of functionally-equivalent systems to verify the complicatedness model’s performance. We then validate the model with an experiment wherein thirty experienced engineers review the mechanical designs of the three sets of systems, and grade the complicatedness of each design on a Likert scale. They then comment on parameters that affect the complicatedness of mechanical systems. The results of the experiment are analyzed in two statistical methods. First, direct statistics are used, and presented with box-diagrams. Second, Kendall’s Correlation coefficient are used to examine the fit between complicatedness rankings based on the results of the model and the rankings of the same systems based on the experiment results. The analysis yields a perfect fit for two of the sets. The mechanical designs in these two sets are: 1. Exposed design, where there is no industrial design involved, and 2. Enveloped design, where the mechanical design is enveloped in an industrial shell. For the third set, wherein the mechanical design is embedded within the industrial design, the findings are inconclusive, as it is impossible for the participants to only assess the complicatedness of the mechanical design and ignore the industrial design. Suggestions are made for improving the model in future research, such as exploring design parameters that are mentioned by the experienced engineers and not used in the model. Finally, to improve the applicability of the model in the design process, it is suggested to explore automatization of the complicatedness model by using software that can be integrated into CAD systems. This can be the basis for future research as well.