Difference-of-Convex approach to chance-constrained Optimal Power Flow modelling the DSO power modulation lever for distribution networks
Abstract
The increasing expansion of renewable energy sources leads to the growth of uncertainty in the distribution
network operation. Short-term operational planning performed by distribution system operators should evolve
to address those new operating conditions, in particular to allow the efficient utilization of different flexibility
levers. In this work, the use of a chance-constrained Alternating Current Optimal Power Flow (AC-OPF) is
proposed to model the operational planning problem, considering the activation of several levers such as power
modulation and power curtailment. The correlation between the renewable generation profiles and loads is
considered via a joint probability constraint approximated with scenarios. The main novelty of the present
manuscript is the adoption of a Difference-of-Convex approach that allows to solve the obtained optimization
problem without convexification or linearization of the core OPF equations. Furthermore, the approach yields
a natural and embarrassingly parallelizable scenario decomposition. The method starts with a reformulation
of the model as a Difference-of-Convex optimization problem, and then a proximal bundle method algorithm
is applied to solve it. The proposed methodology is tested in a 33 bus distribution network with 11 different
values for the safety level defining the probability constraint, ranging from 0.75 to 1.
Origin : Files produced by the author(s)