Available at: http://digitalcommons.calpoly.edu/theses/350
Date of Award
MS in Electrical Engineering
John Y. Oliver
This paper describes an experimental study offered by Designs of Experiments (DOE) within the defined factor domains to evaluate the factor effects of simultaneous characteristics on the benchmarked hard disk drive performance by proposing well-organized statistical models for optimizations. The numerical relations of the obtained models permit to predict the behaviors of benchmarked disk performances as functions of significant factors to optimize relevant criteria based on the needs.
The experimental data sets were validated to be in satisfying agreement with predicted values by analyzing the response surface plots, contour plots, model equations, and optimization plots. The adequacy of the model equations were verified effectively by a prior generation disk drive within the same model family. The retained solutions for potential industrializations were the concluded response surface models of benchmarked disk performance optimizations.
The comprehensive benchmarked performance modeling procedure for hard disk drives not only saves experimental costs on physical modeling but also leads to hard-to-find quality improvement solutions to manufacturing decisions.