Study and prediction of damping in large interconnected power systems with applications in monitoring device location, situation awareness and system operation.

摘要

This thesis presented research, analysis and results which confirm the proposal of a novel application for statistical regression analysis in the study of small signal stability on an interconnected power system for accurately calculating and predicting damping on electromechanical modes of oscillation. From a practical applications perspective, the proposed methodology may be applied to the optimizing of monitoring device location such as Dynamic Swing Recorders (DSRs) and Phasor Measurement Units (PMUs) on the power system. Also it may be practically useful in establishing rules by which system operators may effectively react to damping conditions. The availability of real time Wide Area Monitoring Systems (WAMS) creates the possibility of applying the developed expression and utilizing the established rules in real or near-real time.;The bulk of this research was verified on the Manitoba Hydro power system but the methodology may be applicable to any large interconnected power system. Research of the literature supports the conclusion that to date regression analysis has not been applied to the study of damping on inter-area, inter-plant and local electromechanical modes of oscillation in power systems.;Results show that the developed regression expression can produce accurate calculations of damping on the dominant inter-area mode with Root Mean Square Error (RMSE) of 0.11 for 12 measurable power system variables (selected from an initial 64 variables) with high correlation to damping. This suggests that on average, damping can be predicted to within +/- 0.11%. This expression may be used to accurately predict damping without the need for eigenanalysis. The measurable power system variables included line MW flows, generator active power output, voltage magnitudes and angular differences between voltage phasors at identified locations. Sensitivity studies confirmed that, using the regression expression, appropriate rules may be established by which system operators can effectively react to damping conditions.;The path to proposing the regression expression along with its possible practical applications required the development of linearized models of the power system for small signal stability studies. These models were verified on the IEEE New England 39 bus, to generator power system using the Small Signal Analysis Tool (SSAT) and an analytical programme developed in MathLab. Results from post-processing of disturbance recorder data in conjunction with off-line eigenvalue and time domain stability studies were pivotal in establishing damping trends on the power system and eventually in developing variable selection and adjustment rules for input to the regression model. Observability concepts were developed and validated using the NE-39 bus system and results confirmed that this methodology coupled with the regression expression may be used to aid in optimizing the location of recording devices.