In many parts of the world, reservoir operations are based on empirical rule curves to manage reservoir operation for different seasons or months. These rule curves are based on historical data and are expected to maintain a pre-defined water level in the reservoirs. However, predefined operation curves often fail to operate the reservoir optimally in unexpected scenarios due to climate change, increase in pollution, and change in land cover. In cascading reservoir systems, operations are more complex because they must consider different operating objectives and reservoir characteristics/hydrology. Advances such as forecasted reservoir inflows can be used to update storage and reservoir elevations, allowing optimization of reservoir system performance by reducing unnecessary spills, avoiding hazardous flood conditions, and maximizing generation. We develop an integrated multi-objective forecast-decision operation approach for the large reservoir system and demonstrate its realistic application to Apalachicola-Chattahoochee-Flint (ACF) basin’s four cascade reservoirs - Lake Lanier, WestPoint, George, and Woodruff in the southeast U.S. The approach is designed to quantify and maximize water storage while satisfying all other system demands (e.g., supply) and constraints (e.g., flood control, minimum release policies) through an adaptive forecast-based real-time optimal operation. The Historical Analog Extended Streamflow Prediction (Analog ESP) method is adopted to forecast the inflow into the reservoir system. To model the reservoir operation in real-time, Orthogonal Function Dynamic Programming (OFDP) is used with the heuristic water management rule curves developed by the US Army Corps of Engineers (USACE). Scenarios of various hydrologic conditions are used to demonstrate the improvement in system performance through a marginal benefit assessment of forecasting using retrospective simulations and comparing benefits of cooperated operation versus individual operation. The results demonstrate the efficacy of the adaptive forecast-based optimization approach in improving systematic operation and its ability to improve reservoir system adaptability given future environmental changes.
.jpg "Carly Hansen, PhD photo")
