Optimal partition of power distribution network service areas based on improved K-Means algorithm

In power distribution network planning, geographic information and administrative boundaries have great influence on the power distribution network service areas partition. In this paper the mathematical model of optimal partitioning for distribution network service area is established and the geographic information and administrative boundaries is fully considered. In the proposed model, Objective function and evaluation function is defined respectively to evaluate the efficiency and rationality of each planning solution. Penalty function is used if the constrains of geographic information and administrative boundaries are violated and the improved K-Means algorithm is adopted to solve the mathematic model. Based on the load types of each node, the weight of the loads and capacity of clusters are considered, and the distance between the substation and each load node is redefined. Finally, a practical distribution network planning case is studied and the optimal service area partition plan is presented. The test result shows that the model and solution method is effective and feasible.