ECS-Mathematical Sciences Energy Seminars: Measurement-based identification of power system dynamic model & Numerical Methods for Optimal Control
16 May 2012 14:00 in E240
Janusz Bialek - Measurement-based identification of power system dynamic model (or inverse-engineering approach to eigenanalysis).
Eigenanalysis is a standard tool to analyse power system dynamics whereby the response of a high-order dynamic system is represented as superposition of responses of first- and second-order systems (so-called modes) defined by eigenvalues of the system state matrix. Obviously to determine eigenvalues it is necessary to know the full system model (i.e. the state matrix). This talk will describe an attempt at inverse-engineering eigenanalysis when the unknown system model is extracted from measurements of system modes and mode shapes. Professional help from linear algebra mathematicians is required to explain unexpected results when it was possible to extract the system model from an incomplete set of measurements.
Max Jensen - Numerical Methods for Optimal Control
I begin this short talk reviewing classical results from optimal control theory to place the approach by Bellman into a wider context. I then describe why the numerical solution of the Bellman equation remains a challenging problem. I conclude with some examples to show how this approach has been used to problems in energy production and finance.
Mathematical modelling underlies much of energy engineering. At Durham, relevant engineering research ranges from power network reliability, economics and planning, through reliability analysis of generation units, to computational fluid dynamics models of wind and steam turbines. For this reason, ECS and Mathematical Sciences are organising a joint seminar series to explore opportunities for future collaborative research and grant proposals.
Each seminar will consist of a 20-30 minute talk from each discipline, followed by an extended discussion. While these seminars are open to any Durham researcher, the series is tightly focused on discovering topics for future external proposals between the Mathematical Sciences and ECS.