Dr. Sridhar Ungarala
 |
Associate Professor
|
|
- Research Group
- Publications
- Teaching
- Profile
Research Interests
The ability to optimally interact and guide a process towards its specified goals depends upon our capacity to observe and understand the process as accurately and completely as possible. Chemical processes are often nonlinear, non-stationary and exhibit intrinsic multiscale structures, where the physico-chemical phenomena occur at various scales or resolutions. Such structures manifest themselves in an aggregate form in our observations and interactions with the process. Traditional methods of process data analysis, modeling and control tacitly assume that these inter linked tasks can be carried out at one aggregate resolution. The multiscale cell-space method is a novel approach to model real processes and offers improved techniques to observe and manipulate them. Our research unifies the wavelet and cell mapping theories to develop mathematical foundations to isolate, examine and exploit the multiscale structures in process dynamics. A volume-discretized cell-space in multiple resolutions offers a natural framework to analyze dynamical systems. Current research is centered on the analysis of qualitative behavior of nonlinear processes, state and parameter estimation, and detection and diagnosis of process abnormalities. The interdisciplinary research is a combination of applied statistics, functional analysis and signal processing which will serve the needs of chemical engineering as well as allied fields such as biochemical and biomedical engineering.