Centralized control of load-tap-changing transforms using neural networks

Presents a neural network based centralized control scheme for load tap-changing transformers. To implement the coordinated tap adjustment, the developed scheme employs successive linearization techniques to evaluate the interactions among different transformers represented by a sensitivity matrix. Through updating the matrix using neural network methods, only the local information associated with the participating transformers is desired to perform the centralized tap control. The developed scheme has been verified to be superior to conventional decentralized methods in terms of avoiding unnecessary dynamics and enhancing voltage stability of power systems