Probabilistic Minimal Loss Reconfiguration for Electric Power Distribution Control

Although it is well-known that electric demand is a random variable, the reconfiguration theory still approaches this problem from a deterministic point of view; i.e., most of the network losses optimization problems solutions are built primarily for a specific profile of demand. Even if there are many random behaviors involved in this subject, the main proposal of this paper is a didactic line of attack for a probabilistic minimal loss reconfiguration problem, that takes into account the randomness of load variations of electric distribution systems. The method is based on the Monte Carlo (MC) technique to assign a random load level in each node of the system. The statistical behavior of the loads is modeled using uniform probability density functions. The model defines a useful stochastic index for medium and long-term operation planning reconfiguration. The optimization problem of reconfiguration is solved using a Genetic Algorithm (GA). Principal contributions of this paper are: a) a didactic way to understand a complex electric distribution control theme b) the idea of “expected or characteristic topologies” controlling the influence of wide random load variations on the system configuration and c) a modal topology (in statistical sense) corresponding to the reconfiguration solution for the average demand values. Preceding outcomes are very important: 1) from the realistic modeling point of view, 2) the decision-making strategies and 3) a simple understanding of the topic.