Theses and Dissertations at Montana State University (MSU)
Permanent URI for this collectionhttps://scholarworks.montana.edu/handle/1/733
Browse
1 results
Search Results
Item Manufacturing reliability for C-channel composite beams(Montana State University - Bozeman, College of Engineering, 2014) Bauer, Michael Wayne; Chairperson, Graduate Committee: Douglas S. CairnsA new manufacturing method has been developed for fabricating c-channel composite beams. The beams are to be used as test articles in four point bending tests. The motivation behind this thesis is to study the effects that specific manufacturing parameters have on the resulting amounts of porosity and fiber volume in these three-dimensional sub-scale structures. The parameters considered are number of layers of flow media, fabric architecture, flow rate of the resin, temperature of the resin, and level of vacuum pressure used. The manufacturing parameters were varied in a 1/2 factorial design of experiments where sixteen beams were manufactured, all with varying values for each parameter. A taguchi design of experiments was also formed to provide a comparison. The resulting average porosity percentages and fiber volume percentages were then determined for every beam. In addition, compression and tension tests were conducted to find the average maximum stresses for each. Once all the data had been gathered an Analysis of Variance (ANOVA) study was conducted to determine the effects and their levels of significance. It was found that the level of vacuum pressure has the most significant effect on the porosity while the fabric architecture has the most significant effect on the fiber volume. Overall, every parameter has some sort of quantifiable effect on porosity and fiber volume. There are also significant two and three way interaction effects present for each. Additionally, the 1/2 factorial design seemed to provide more accurate results compared with the taguchi design, which was inherently not comprised of data with a normal distribution and does not include interaction effects. Regression models were developed for the output levels of porosity and fiber volume. This allows manufacturers to create these beams with predetermined output levels for each and can improve testing capabilities. Also, using two layers of flow media greatly improved the consistency of the beams, while reducing porosity and slightly reducing fiber volume percentage. It is recommended to further implement the use of two layers of flow media into large sub-scale structures and potentially full scale turbine blades.