E&CSE Research Seminar, Wednesday 23 - 05 - 2007

Abstract:

The load modeling of power system has always been a challenging and equally interesting task for power industry as the load models describe the dynamics of power system and have significant effects on system stability analysis and control. Measurement based load models give a closer representation of non linear nature of power system loads. Artificial intelligence techniques have been used so far for load modeling. Stochastic time series techniques in the form of ARMAX mathematical models are used in the present research for dynamic load modeling using the on-line measurement data taken from five substations in Victoria . This project extends the robustness of the ARMAX model developed with a disturbance data at a location, to predict the disturbance responses at different locations and times. The research aims to explore possible methods to develop general ARMAX models for Victorian power system. It proposes to integrate the model with PSS/E package and establish the compatibility for the on-line transient analysis. The performance of the heuristic models will be compared with the existing composite models used by the energy utility. It proposes new strategies to understand the physical interpretation of the model response in terms of network elements.