Browsing by Author "Lorca, Alvaro"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Multistage adaptive robust optimization for the hydrothermal scheduling problem
    (PERGAMON-ELSEVIER SCIENCE LTD, 2023) Favereau, Marcel; Lorca, Alvaro; Negrete-Pincetic, Matias
    The current water scarcity faced by many countries increases the need to consider an appropriate representation of future hydro inflows in power system operation and planning models. Hydrothermal scheduling is the problem that seeks to use the water stored in reservoirs throughout time in order to find an optimal dispatch policy between hydro and thermal power plants. Due to both the inherent randomness of water inflows and the intertemporal decision process, this problem has been typically approached through multistage stochastic optimization, minimizing the total expected operational cost over the entire planning horizon. However, this approach has some practical disadvantages. Among the main ones we highlight (i) the complexity of balancing the statistical representativeness of the stochastic processes and the computational efficiency of the optimization model; (ii) the need to employ computationally intensive decomposition methods for its solvability; and (iii) the need to carry out network simplifications to tackle tractability issues arising in large networks. As an alternative, we propose a multistage adaptive robust optimization model for the hydrothermal scheduling problem. Robust optimization is useful to prevent the previous disadvantages because it does not make any distributional assumption and it works with the so-called uncertainty sets instead of carrying out sampling processes. In particular, we propose an efficient formulation based on linear decision rules and vector autoregressive models to represent the uncertainty in hydro inflows. Our experiments, based on the Chilean electric power system with hundreds of hydro nodes and connections, show the proposed model's efficiency for large-scale systems and provide insights into the adequate balance between cost-effectiveness and reliability that robust optimization models guarantee.
  • Thumbnail Image
    Item
    Robust streamflow forecasting: a Student's t-mixture vector autoregressive model
    (SPRINGER, 2022) Favereau, Marcel; Lorca, Alvaro; Negrete-Pincetic, Matias; Vicuña, Sebastian
    Accurate streamflow forecasting is one of the main challenges in the management of reservoirs, where autoregressive models have been commonly used. Typically, the noise of these models is considered Gaussian. However, this assumption can overestimate the presence of outliers, generally presented in water inflow real-world data. Motivated by this, we propose a novel streamflow forecasting method by modeling the noise of a vector autoregressive model as a multivariate Student's t-mixture based on the use of the variational expectation-maximization algorithm. The proposed model is able to capture the trend, seasonality, and spatio-temporal correlations of hydro inflows, along with both asymmetry and multimodal features of the vector autoregressive process' residuals. Based on 12 of the main inflows of the Chilean hydroelectric network, our experiments show the proposed model's efficiency and improvements for forecasting medium to long-term inflows over a classical vector autoregressive model. Results show that the expected forecasts are improved with the proposed model and the predictive distributions present tighter intervals based on standard and state-of-the-art metrics.