Calculation of Characteristic Attributes of Consumer Aggregations Based on Multi-objective Clustering

The concept of consumer aggregation and power sale market analysis dimensions are proposed.After characteristic variables of consumer aggregation are selected,the hybrid visual cluster technology with self-organizing map(SOM) and k-means is chosen for division of power sale market under user-defined target.Considering load and economic indices,the attribute measurement algorithm based on entropy weight improvement is presented and employed to perform the elaborating and integrated calculation of consumer aggregations with different division targets to compare the relative properties.Real data of 317 consumers are used as sample of analysis.The results show that different from the traditional electric load characteristics analysis methods,the proposed method can find out more effective information of power sale market.The method also enables more effective data mining and detailed multi-objective consumer characteristic analysis.