Design of secondary circuits of distribution networks using clustering and shortest path algorithms

A solid and practical approach for designing an optimal secondary distribution network is proposed. The methodology starts by optimally locating and sizing medium voltage/low voltage transformers then finding the optimal path for secondary circuits. The optimal number of transformers and their location are determined by k-means clustering algorithm, and validated using Davies-Bouldin index. In addition, the obtained centroids represent the load centers where transformers should be located. Once the transformers are located, the Dijsktra's algorithm, for shortest path problem, is used to find the optimal number and path of transformer's secondary circuits. This algorithm helps finding the minimum distance from transformers to consumers. Moreover, as a part of the database, electric poles are considered to find the path of secondary circuits. This methodology is implemented in a real distribution network in Guano-Riobamba, Ecuador.