A transmission expansion model for dynamic operation of flexible demand

Abstract

This paper proposes a model to include investments in demand flexibility into traditional transmission expansion problems under uncertainty. To do so, a dynamic power flow model is proposed. The model is solved via applying a value function approximation in form of a neural network on the operational problem, allowing to yield a result for the non-convex investment problem. Additionally, robust sets are applied and linearized to deal with uncertainty and decrease computational complexity. In similar manner, Karush Kuhn Tucker conditions are used to transform a tri-level into a bi-level problem. Case studies for systems of varying complexity show the convergence of the algorithm as well as that flexible resources can be used as a cost-effective substitute for transmission lines in grid expansion.