Analytical introduction of uncertainty into long term distribution systems decision-making

Distribution system planning consists in imagining the evolution of the design and operation of distribution systems over a horizon going from several years to decades. There is a lack of standard methodologies that integrate the growing number of uncertainties. In this article, our aim is to provide a framework for integrating uncertainty, from the diagnosis of network constraints and the setup of solutions to their economic evaluations. To do so, we start by modeling the network under load uncertainty. This allows us to use probabilistic power flow calculations for constraint estimations. We use these to determine the best strategy, between line reinforcement and demand response. Finally, we use a compound option model to assess the economic validity of undertaking a unique action, or series of actions, when uncertainty is taken into account. This framework was successfully applied to the IEEE 70 bus network with a discussion on DSO's options: "waiting for more information", "investing" or "activating demand response". Results show that demand response is not optimal on the lower part of the network but should be used on the upper part since only some nodes would see under-voltage constraints during 0.2% of the year. Also, a sensitivity analysis on the cost of demand response enables drawing the DSO's willingness to pay considering two scenarios (expected load evolution and worst case).