- Project Summary
This research presents the analysis of various techniques/parameters in developing a suitable Neural nework for performing static voltage stability analysis and its application. At first numerous functions and methods for training an Artificial Neural Network is studied and the best is found out. Then an alrogithm is developed with the most suitable neural network architecture and tested on different standard power systems such as Ward Hale and IEEE 30 bus systems. The efficiency in learning the variations of system parameters such as load and the ability of identification of the new load changes by the network is assessed.
The research also tests the trained Neural network for the common disturbances such as generator outages and line outages. The apparent comparison in the ability of learning the normal and also the contingency by a single network and with seperate networks was one of the study measures. It was found that the single ANN is capable of learning both conditions even though for larger systems the accuracy is less when compared with trained independent ANN's.
- For further reading
Sukumar Kamalasadan, "A neural network approach to Voltage Stability and Improvement - Masters Thesis Report [ ]
D Thukaram,Sukumar Kamalasadan and Adel A Ghandakly, "A MATLAB based neural network approach to Voltage Stability Monitoring and Assessment -16th International Conference on Compter Appliations in Industry, ISCA-CAINE 2003 []
D Thukaram,Sukumar Kamalasadan and Adel Ghandakly, Artificial Neural Network approach to Voltage Stability Monitoring and Assessment - Manuscript in Progress