An analyze of slow voltage variations from the electric distribution systems with a clustering based approach

The paper is focused on slow voltage variations in the nodes of an electric distribution system. A clustering based approach to classify the nodes (substations) in categories described by linguistic terms from viewpoint of slow voltage variations is proposed. The linguistic terms are defined in function by the daily average load of nodes and the voltage quality indices (average value and mean square deviation of voltage). The approach was applied to a zone of a distribution system with 117 substations (20/0.4 kV) that supplies mostly the residential consumers, but it can also be applied to substations operating at other voltage levels and supplies another consumers' categories.