Path-Based Distribution Feeder Reconfiguration for Optimization of Losses and Reliability

This paper presents a path-based modeling framework for the distribution feeder reconfiguration (DFR) problem. The framework maps the decision variables-on/off status indicators for the paths-into linear expressions for the network flows and reliability indices. These linear expressions are suitably deployed for DFR optimization, where reliability can feature either as an objective or as a constraint. Additionally, the usual objective of loss minimization can also be incorporated, leading to a multi-objective DFR optimization framework. In its most general form, the proposed framework is a mixed-integer quadratic programming formulation, and can be solved using existing solvers. The paper makes three other important contributions: first, a probabilistic approach is adopted to validate the approximation of load point failure rates as the sum of failure rates of upstream components; second, a graph algorithm is proposed for identifying topologies with best reliability; and third, simulation studies with representative loading patterns are used to identify the appropriate choice of loading conditions for optimizing overall active power losses. Application of the proposed framework to standard test systems demonstrates its ability to optimize network losses or reliability or both. Numerical results also show the effect of topology and loading conditions on the performance of optimized topologies.