Computational Intelligence Based Power Systems Operation

The power system disturbances related to power quality have been important issues to electric utilities and

their customers. Upon malfunction in power systems everything goes disarray impacting industrial units,

business, and a large population.

To prevent power system failures an appropriate corrective action must be taken well before beginning

of cascaded failures. Computational Intelligence (CI) based techniques; Discrete Wavelet Transform (DWT),

Hidden Markov Models, Recurrent Neural Networks, Bayesian networks, mixture of Gaussians and

independent component analysis are proposed for real time identification/classification of the power

system disturbances. For the real time control of power system disturbances, DSpace system implements

intelligent control algorithms for Power Quality Converter to correct a selected set of individual disturbances

as well as overlapped disturbances.

The CI based techniques will perform time series analysis to forecast future disturbances based on disturbance log/history. The real time detection and control techniques will also be used for system probing to predict the load and develop the model of the load in a downstream power system.

The electric power engineering research and education needs new ideas from computational sciences for an

intelligent control of electric grid system. The power system disturbances, caused by natural factors, over loading, power quality events, and malfunction of power system gears, are growing concerns to electric utilities. Therefore, this proposal also strives to develop a new interdisciplinary course titled 'Computational Intelligence in Power System'.

It is expected that the proposed new course will impart training to future power engineers with broader perspectives on interdisciplinary research and education.