Input space decomposition and multilevel classification approach for ANN-based transient security assessment

This paper proposes an ANN-based multilevel classification approach for fast transient stability assessment of large power systems. A two-level classifier incorporating two feedforward ANNs is built to obtain a stability index for security classification using some general abstract post-fault attributes as its inputs. The ANNs are trained by a newly-developed semi-supervised learning algorithm. The proposed approach can not only distinguish whether a power system is stable or unstable based on the specific post-fault attributes but also provide a relative stability quantifier. The numerical results on applications to the 10-unit New England power system demonstrate the validity of the proposed approach for transient security assessment