9419458 Girgis The mail objectives of this research are to develop new methods for on-line voltage stability and transient stability. Load forecasting and the optimal estimation of voltage phasors, generator frequency deviations, and their rate of change will be the basic information for the development. Based on these values, a static index and a transient index will be developed for voltage instability prediction and transient stability detection, respectively. During normal operating conditions, the static index for voltage instability is calculated, and its future values are predicted. Once the system enters a transient state, the transient index, which is for transient instability detection, will be computed. Load flow solutions are considered as the intersections between a space and a space curve. The angle, between the gradient vector of a space surface and the tangent vector of a space curve at an operating point, may be considered as the static index for voltage instability. Its closeness to 90 degrees may indicate the proximity of the operating point to the voltage collapse point. The directional derivatives of the bus injection power in the space surface equation, in the direction of the tangent vectors of the space curves, are possible candidates for weak bus identifications. The proposed indices are available for the voltage phasors and the load flow Jacobian. To predict voltage instability, the voltage instability index may be computed based on the future load and its corresponding load flow solutions. By comparing the voltage instability index for the current operating condition with the index for the predicted load, one may conclude if the system will become close to or away from the voltage collapse point. If the system is in transient state, the unstable mode will be detected primarily. The predicted generator angle difference in the post disturbance period is proposed for this purpose. Once the unstable mode is pred icted, the high dimensional power system model may be aggregated into a low order model. The kinetic energy of this aggregated low order system is proposed as the transient index for transient in stability detection. ***
