Contour refinement by enhanced query-based learning

An enhanced query-based learning neural network is proposed to refine the contour in this paper. The proposed learning approach provides a classifier at a lower computational cost comparing with the other method. The ambiguous regions in the classification process can be thus easier clarified. The proposed approach has been tested on the refinement of security contours in the assessment of power system operations. Results demonstrate the effectiveness and feasibility of the proposed approach for the applications.