The CBR-PNN approach to real-time voltage security assessment of power systems
This paper presents the Case-based Reasoning (CBR) technique, a novel approach to voltage security assessment of power system in a real-time environment. The Probabilistic Neural Network (PNN) is used for determining the solution to a new case, using the old cases from the case base of the CBR system. A feature selection technique based on the information gained has been employed to identify the relevant input features for PNN. The CBR technique extracts the information of voltage security from large sets of possible operating conditions of the power system, generated by computer simulation. Also the ranking of contingencies is done by CBR as accurately as the one done by the Continuation Power Flow (CPF) method. Security classes are defined by the threshold value of maximum loadability margins, calculated by the CPF method. The effectiveness of the proposed approach is tested on IEEE test systems. The proposed method gives fast and accurate voltage security information for unknown patterns and it is found to be suitable for real-time monitoring of voltage security in energy management systems.
Keywords: case-based reasoning, CBR, probabilistic neural networks, PNNs, maximum loadability margins, voltage security assessment, electric power systems, simulation, energy management systems
Customer comments
No comments were found for The CBR-PNN approach to real-time voltage security assessment of power systems. Be the first to comment!