Search

Hydropower is one of the world's primary renewable energy sources whose usage has profound economic, environmental, and social impacts. We focus on the dispatch of generating units and the storage policy of hydro resources. In this context, an accurate assessment of the water opportunity-cost is crucial for driving the sustainable use of this scarce resource. Nevertheless, traditional computational tools use stochastic dual dynamic programming under the Hazard-Decision (HD) modeling simplification, where dispatch decisions are determined assuming perfect information about the current inflows. In practice, however, some dispatch decisions are made before water inflows are observed, i.e., under a Decision-Hazard (DH) scheme. This inconsistency generates an optimistic assessment of the opportunity cost of the water, inducing a sub-optimal use of energy resources. Thus, our objectives are: (i) to raise awareness of the HD issue bridging research and implementation with a clear recommendation to system operators and regulators; (ii) to incorporate the DH scheme into the long-term hydrothermal scheduling problem; (iii) to assess the economic impacts and other market distortions due to the actual energy usage policy based on the HD simplification. For a representative case-study with realistic data from the Brazilian power system, results show that the HD simplification introduces a regret cost of 12%. Furthermore, we show that incorporating the DH scheme into the hydrothermal planning model, the energy supply cost can be reduced up to 5.3%. We also show that spot-price distortions and expensive thermal generation could be mitigated in the DH scheme. • SDDP applied to hydrothermal dispatch relies on Hazard-Decision (HD) approximation. • The applied HD approximation implies in one-step-ahead anticipative process. • Proposed Decision-Hazard (DH) model considers here-and-now decisions in each stage. • Numerical simulations based on the Brazilian case reveals an overcost of 12% due to HD. • Incorporating the DH model, simulations show a reduced overcost up to 5.3%. [ABSTRACT FROM AUTHOR]

Real-time operation of electric power systems under high penetration of renewable energy generation has to incorporate strategies for managing the uncertainty associated with this type of power sources. In the case of the unit commitment problem, models based on robust optimization have been widely used for the day-ahead computation of power dispatch and reserves schedule. Typical approaches use uncertainty sets with fixed levels of conservativeness, without considering the real-time performance of the solutions. This paper proposes a solution scheme for the adaptive robust unit commitment problem using online adjusted data-driven uncertainty sets. Conservativeness parameters of the uncertainty sets are dynamically calculated as a function of previous operating results and incoming data via reinforcement learning, resulting in an online learning-optimization framework. The paper also develops an experimental framework to simulate real-time operation under the proposed methodology. Out-of-sample experiments illustrate the effectiveness of the proposed scheme against well-known benchmarks with fixed robustness levels. Results show improvements in power generation costs, voltage violations, and use of reserves. Two systems of different sizes are analyzed, illustrating the scalability of the proposed approach. • Q-learning algorithms for data-driven robust uncertainty sets adaptation. • Uncertainty sets dynamical adaptation based on real-time system indicators. • Novel framework for real-time power systems operation simulation. • Results comparison against two-stage robust and stochastic optimization approaches. • Statistically significant lower generation costs and use of system reserves. [ABSTRACT FROM AUTHOR]