الفهرس 
chapter 1Only 14 pages are availabe for public view
 |  | from | 230 |  |  |
| | | from | 230 | | |
Abstract
This thesis ,studies the applicability of new fast learning algorithms
to solve power system instability problems. Transient stability problem of power system is briefly discussed, and it is shown there will be a need for more modifications on the existing methods. New methods of transient stability assessment were investigated. Advantages and disadvantages of those methods are discussed. Adoption of o new fast learning on-line prediction algorithm, that has been used in robotics area is designed and applied to test systems. The ’ prediction algorithm does not require any prior knowledge of system parameters or configurations except the measurements of the predicted variable. This algorithm proves its validity to be applicable in prediction of generator’s rotor angle instability for a simple machine and infinite bus system with an acceptable error. Generalization of the method to be a.pplieah1e to real life networks bas been investigated and proven applicable of transient rotor angle instability prediction of most disturbed generators for the IEEE test systems (Reduced Iowa, WSCC 29 generators, and IEEE 50 generators Test Systems). More investigation was done in the area of voltage instability. Transient voltage instability was predicted for IEEE 50 generators system and conclusions are made. Recommendation of using this pioneering method for on-line power system parameters prediction have been proven and concluded,