Files
Abstract
The purpose of this paper is to report the results of a capstone project for the Milwaukee School of Engineering’s [MSOE] Master of Science in Engineering [MSE] program. The contents of this paper discuss the overall purpose of the project and its justification in its field of research. In addition, the relevant published literature and final project results are investigated in detail. The general purpose of this project was to perform an analysis of various kinds of plastic hole mounts in multiple mounting situations that mimic common scenarios found in real-world applications. These plastic mounts or fasteners are associated with numerous applications in larger product assemblies such as automobiles. In the project investigation, the fastener’s material, size, and moisture content were varied along with the size and geometry of the hole it was mounted in. These variables were tested in combination to measure the force required to mount the fastener [push-in force] and extract it [pull-out force] from the same hole. The collected data were then exported for statistical analysis within Minitab. Applying a Multiway Analysis of Variance [ANOVA] general linear model approach, a residual analysis was conducted to determine if the data met the initial requirements for a Multiway ANOVA. A multitude of data outliers—the result in part of manufacturing defects—contributed to a dubious analytical result [i.e., extremely high push-in force values but negligible pull-out force values]. All outliers were removed, and data were again tested for compliance with general linear model assumptions. Furthermore, Ryan-Joiner normality testing—with its powerful predictive ability in association with large data sets and long-tailed distributions—was employed and indicated normally distributed data. Regression models were then developed for push-in and pull-out forces, and a Box-Cox transformation was applied to each model, resulting in adjusted R-square values of 84% and 67% for the push-in and pull-out models, respectively. This result indicates that the models account for most—but not all—the behavior of the forces. Recommendations are also offered to improve the accuracy of the models, including the use of automated testing equipment, further standardization and conditioning of all tested parts, and the use of enhanced part identification methods, along with the need to investigate part geometric deformation with respect to part resin, as well as statistical noise associated with manufacturing inconsistencies. The primary goal of this research is to inform current and future manufacturing and design of similar products in this area. This knowledge will help mitigate unexpected failure of these products as they are used in the field and reduce waste via the manufacturing process.