Optimization under uncertainty for the management of hydroelectric resources

Loading...
Thumbnail Image
Date
2023
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Hydroelectric generation management is considered a fundamental problem in the operation of power systems, where operations research has contributed significantly to decision making. One of the most relevant difficulties is that the decisions of storage and use of water resources in the present affect future decisions. Additionally, this problem presents another important challenge associated with the uncertain nature of hydro inflows, which is becoming even more complex today due to the consequences of climate change. Thus, this thesis proposes the development of optimization under uncertainty models to support medium- and long-term planning of hydrothermal system operations. First, this thesis proposes to develop a statistical model that represents the random behavior of hydro inflows in multinodal and multireservoir hydrothermal systems. The purpose is to capture statistical aspects such as trend, seasonality, spatio-temporal correlation, asymmetry and multiple modes present in historical hydrological data through the use of autoregressive models and variational inference algorithms. In a second part, it is proposed the development of a risk-averse optimization under uncertainty model that incorporates the previously proposed statistical methodologies for the medium-term operation of hydrothermal systems. The main objective is to determine an optimal generation policy capable of operating efficiently under the least favorable realizations of future hydrological trajectories. Finally, the third stage consists of the development of optimization under uncertainty models that support long-term energy planning and operation processes. The idea is to analyze the role of water in the interaction with other productive sectors such as agriculture, thus improving the efficiency of the use of their resources. This tool will also make it possible to anticipate future phenomena such as the penetration of new technologies, regulatory changes in the use of water resources or their low availability, and what would be the impacts on the resource management for different users.
Description
Tesis (Doctor in Engineering Sciences)--Pontificia Universidad Católica de Chile, 2023
Keywords
Optimización bajo incertidumbre, Aversión al riesgo, Procesos estocásticos, Gestión de recursos hídricos, Generación hidroeléctrica, Planificación energética
Citation