Influence of Outliers on Transformer Power Losses Estimation Using a Statistical Based Data Mining Approach

摘要: Distribution network operators (DNOs) are alarmed for fluctuating behavior of power consumption in the process decision-making. In Smart Meters data, outliers can occur from recordings irregularity, undiscovered demand, illegal user's connection, inadequately installed devices etc. The paper presents an original data mining based approach that acquires basic concept knowledge discovery on databases (KDD), focused to identification recorded by Meters. This information was input losses estimation transformers. order validate proposed approach, a real database with 300 rural substation considered. By applying methodology DNOs know technical and also proper operation planning actually “smart” networks be applied.