Economic dispatch under uncertainty: The probabilistic envelopes approach

With fast paced renewable energy integration, stochastic methods are emerging as a viable alternative to traditional power system operations planning under uncertainty. The value of stochastic planning lies in its effectiveness at reducing operational costs, while allowing higher penetration levels of renewable sources, and without much sacrifice to system security. This is under the premise that the decision maker may not need to fully hedge against more extreme and costlier events that exhibit lower probabilities of occurrence. On the other hand, stochastic methods do suffer from the curse of dimensionality, which renders them computationally intractable for practical application. Here, we propose a novel hybrid robust-stochastic approach based on the flexibility envelopes concept. It circumvents the curse of dimensionality by using probability-weighted envelopes to enclose the evolution of the net load uncertainty over the planning horizon. The new approach is illustrated by a receding-horizon economic dispatch example, to compare its effectiveness to robust planning and stochastic planning with a scenario tree.