A stochastic approach to analyze trade-offs and risks associated with large-scale water resources systems

摘要

Water resources development projects often involve multiple and conflicting objectives as well as stochastic hydrologic inputs. Multiobjective optimization techniques can be used to identify noninferior solutions and to construct a trade-off relationship between conflicting objectives. This paper presents a methodology for analyzing trade-offs and risks associated with large-scale water resource projects under hydrologic uncertainty. The proposed methodology relies on the stochastic dual dynamic programming (SDDP) model to derive monthly or weekly operating rules for multipurpose multireservoir systems taking into account the stochasticity of the inflows, irrigation water withdrawals, minimum/maximum flow requirements for navigation, fishing, and/or for ecological purposes. In SDDP, release decisions are chosen so as to minimize the operating costs of a hydrothermal electrical system. Irrigation water demands and other operating constraints are imposed on the system through the SDDP model. The proposed methodology is illustrated with the Southeastern Anatolia Development project, commonly called GAP, in Turkey. The GAP is a multidimensional-development project involving primarily the production of hydroelectricity and irrigation. Simulation results using 50 hydrologic scenarios show that the complete development of the irrigation projects would reduce the total energy output by 6.5% and will increase the risk of not meeting minimum outflow at the Syrian border from 5% to 25%.