Stochastic MPC For Optimal Operation of Hydropower Station Under Uncertainty

Abstract

The operational condition at the Dalsfoss power station is complicated due to many requirements related to environmental regulations and safety constraints such as the seasonally varying water level requirement at the reservoir. However, the operation becomes more difficult due to uncertainties in the system. In this paper, at first a certainty equivalent MPC is applied to the uncertain hydro power system and it has been shown that its robustness property is poor. Secondly, to prevent the constraint violations due to the uncertainties in the system, two measures are taken. One measure is to introduce a safety margin for the constraints and further design a certainty equivalent MPC. The other measure is to implement a multi-stage MPC for robust constraint satisfaction. Two types of multi-stage MPC are considered in this paper. The first employs all of the possible 50 scenarios of the uncertainty of an input disturbance variable, and the latter generates and uses three synthetic scenarios to approximate all of the possible 50 scenarios. All of the simulation results are compared for their robust performances and computational time.