A population based approach to design multi-loop medium voltage distribution network

• A population based approach is developed.
• For mv/lv classification, MSVM and K-means algorithm is used.
• ACS–GA based approach for MV network design is developed.
• Applicable to large size mv network design.

In this paper, a method is presented to design a multi-loop medium voltage network which supplies each mv/lv substation with an alternate feeding point from hv/mv substation to assure the reliability of the system while the total cost of installation, cost of energy losses and cost of unserved energy is minimized subject to operating constraints. A generalized technique for multi-loop mv network design has been developed. In the proposed technique, at first step, mv/lv substations are clustered corresponding to geographical coordinates using MSVM and K-means method. Then multi-loop are formed using population based approaches Ant Colony System and Genetic Algorithm (ACS–GA) in combination. The developed method can explore and exploit search space efficiently. In the next step, using economical criteria, conductor size is selected for each loop under worst operating condition (feeder segment outage near to source end) subject to constraint on voltage magnitude at each node and admissible current carrying capacity of the conductor. Finally, for radial operation of multi-loop, a heuristic approach is presented. The developed technique has been applied effectively on 45-bus test system.