LAUSR

Abandoned underground quarries or mines may be rehabilitated as natural reservoirs for underground pumped hydro energy storage (UPHES). In addition to the inherent modeling inaccuracies of the traditional PHES that arise from, e.g., approximating the nonlinear pump/turbine head-dependent performance curves, the optimal operation of these underground plants is also affected by endogenous model uncertainties. The latter typically arise from a limited knowledge of the physical characteristics of the system such as the geometry and hydraulic properties of the underground cavity. In this paper, chance-constrained programming is leveraged to immunize the day-ahead scheduling of an UPHES owner against both these model uncertainties and the modeling approximations. The proposed method is tested on a fictitious UPHES system using an existing underground quarry as lower reservoir. Results demonstrate that the methodology allows finding a compromise between conservativeness and economic performance, while being computationally efficient. This model may thus be integrated in the daily scheduling routine of UPHES owners, or may help regulators and system operators to better estimate the available flexibility of such resources.