Adaptively constrained stochastic model predictive control applied to security constrained optimal power flow

The number of renewable energy sources in electricity grids is growing and with it the associated uncertainty. For power system operation a good tradeoff between security and cost becomes increasingly important. We consider the problem of including uncertainty from renewable in-feed in a DC security constrained optimal power flow (SCOPF) by formulating chance constraints. The chance-constrained SCOPF is solved repeatedly and can be understood as a chance-constrained Model Predictive Control (MPC) problem, where the time coupling comes from storage in the system and the associated dynamics. The main focus of the paper is to reformulate the problem in a non-conservative way, i.e., to exploit the predefined violation level of the chance constraints based on their empirical closed-loop violations in order to achieve a good tradeoff between operational security and cost. We use and extend a recently proposed adaptive stochastic MPC scheme that starts with a standard conservative chance-constrained reformulation and then adapts the formulation of the constraints online based on the empirical violation level. The proposed approach is demonstrated on a 5-bus network.