This grant provides funding for the development of an innovative research approach to modeling and analysis of complicated engineering reliability problems. The project consists of four research and education tasks - 1) developing new statistical models that can extend reliability analysis to multiple-stress profiled reliability testing data; 2) fusing reliability information from various resources, including mechanistic model, expert opinions, and field failure data, into reliability estimation and prediction using Bayesian methods and their associated computational tools; 3) demonstrating the advantages of using generalized linear models and Bayesian analysis methods for the design and optimization of reliability tests; 4) engaging science and engineering students, particularly Hispanics students, in active learning and using probabilistic and statistical models in real-world applications.
If successful, this project will advance the knowledge of statistical reliability analysis of a complex testing plan and multiple failure modes, develop a data analysis method which will eliminate many unrealistic assumptions and incorporate prior engineering knowledge into reliability analysis, and promote a broader understanding and appreciation of computational statistics methods among quality and reliability researchers and industrial practitioners. The methodology developed in this research can apply to the automobile industry, electronic industry, semiconductor industry, and many other applications in the fields beyond engineering, such as medical and biological sciences. In addition, this project will significantly enhance the basic science and engineering education in a major Hispanics-serving university, foster a collaborative research environment across the disciplines of industrial engineering and mathematical science, and benefit a minority student group who is traditionally underrepresented in science and engineering careers.
