Stochastic integration of demand response and reconfiguration in distribution network expansion planning

Planning the distribution network of the future involves forecasting the most likely scenario to make appropriate investment decisions. Many uncertainties concerning, e.g. the evolution of conventional loads, renewable production and electric vehicles (EVs) make it difficult to predict the location of the distribution network's weaknesses (overvoltages, undervoltages and overcurrents) and their occurrence. In some cases, alternative solutions such as demand response (DR) and reconfiguration can remove the constraints and prevent expensive network investment. This study proposes a two-stage algorithm that is able to give the probability that no technical constraints will appear as a function of the reinforcement cost with and without using DR and/or reconfiguration. The first stage of the algorithm consists in running Monte Carlo simulations based on realistic scenarios for loads, EVs and renewable production development provided by French governmental roadmaps. The cost of reinforcement per line and per hour of constraints enables selection of the feeders, where DR (solved with linear programming) and/or reconfiguration (exhaustive research) will be implemented in the second stage of the algorithm to remove these constraints. The methodology is applied to a real part of a French distribution network.