Abstract
This thesis investigates the H2-optimal model order reduction problem within finite time and frequency intervals. Several oblique projection-based algorithms are proposed that construct near-optimal reduced-order models in a computationally efficient way. The deviations in the optimality conditions decay quickly as the order of reduced models is increased. A class of proposed algorithms has interesting properties like stability preservation, monotonic decay in error, and pole placement. The algorithms proposed in the thesis are effective design tools in reduced-order controller design, reduced-order filter design, and fast power system simulation.
Original language | English |
---|---|
Qualification | Doctor of Philosophy |
Awarding Institution |
|
Supervisors/Advisors |
|
Thesis sponsors | |
Award date | 26 Jun 2021 |
DOIs | |
Publication status | Unpublished - 2021 |