RUcore Resource Object
RUcore Resource Object
PDF-1(689.14 KB)
TitleApplication of dynamic global sensitivity analysis in complex systems
NameNiotis, Vasilios (author), Ierapetritou, Marianthi G. (chair), Androulakis, Ioannis P. (co-chair), Berthiaume, Francois (internal member), Rutgers University, Graduate School - New Brunswick,
Degree Date2012-01
Date Created2012
SubjectBiomedical Engineering,
Global analysis (Mathematics),
Production management ,
Tablets (Medicine),
Pharmaceutical industry--Production control
DescriptionOne of the major problems of complex mathematical models that are used to approximate systems and processes is the lack of precise parameter values. This often leads to a high degree of uncertainty in the simulated processes, which in most cases is an undesirable constraint. The uncertainty in parameter values can be addressed using sensitivity analysis, which is the study of how output variations can be apportioned to different sources of variation in the input parameters. The first part of this work consists of the application of time-varying global sensitivity analysis techniques in a mathematical model of human endotoxemia. In general, biological systems contain a large number of components that interact with each other, making the application of sensitivity analysis a valuable tool to decipher the most critical dynamics of the system. Through sensitivity analysis the parameters or components that have little effect on the model but are experimentally observed to be significant for the system, are identified. The results imply the need for better parameter estimation, after further experimentation, or model modifications that will capture the experimentally observed system dynamics. In the second part of this work, the complexity of how interactions between the different unit operations of a continuous tablet manufacturing flowsheet simulation affect the overall product quality is studied. Both quantitative and qualitative results reveal how different uncertain variables of a process dynamically affect an output through the use of time-varying global sensitivity indices. Thus the most important and critical parameters for a certain output are identified at different time points. Such an approach of global sensitivity analysis is not only used to draw significant conclusions about the interactions between specific uncertain inputs to outputs, but also points out necessary correlations that the model fails to capture. Through this work it is shown that sensitivity analysis should have an important part during the development and validation of a computational model in any scientific field. It allows the quantitative and qualitative investigation of variation and perturbation effects on the system behavior and correlation with experimental data.
NoteM.S.
NoteIncludes bibliographical references
Noteby Vasilios Niotis
Genretheses
Persistent URLhttp://hdl.rutgers.edu/1782.1/rucore10001600001.ETD.000064154
Languageeng
CollectionGraduate School - New Brunswick Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.