Reservoirs Optimized Real Time Operation System, Focusing in Flood Control (PAP014824)

Código

PAP014824

Autores

Rosana de Fátima Colaço Gibertoni, Akemi Kan, Dirk Schwanenberg, ALBERTO ASSIS DOS REIS, José Luís da Silva Pinho

Tema

Flood forecasting and early warning systems

Resumo

Efficient floods management can benefit from advanced information technologies which may include data on hydrological and hydrodynamics characteristics of vulnerable regions, i.e. inundated areas, water levels and flow rates, collected by telemetric gauge stations and meteorological models forecasts. Prediction of flood events can accurately be achieved by applying robust mathematical models for describing rainfall-runoff phenomena as well as surface waters hydrodynamics. CEMIG is a Brazilian energy supplier operating about 55 hydro power plants with an installed capacity of more than 6000 MW. The company currently executes a R&D project of duration of 3 years with the Brazilian institute LACTEC. It aims at building up tools in the field of flow forecasting. This serves several objectives such as the mitigation of flood damage in downstream river reaches of dams and the optimization of hydropower benefits. The huge number of information complicates the operation of forecast system in an integrated way. The operational forecasting system Delft-FEWS, developed by DELTARES, is used to allow integration of all prediction steps and to facilitate the insertion of any types of data and models. This system allows the implementation of decision support tools with high flexibility, security, accuracy and agility in real time environments. It has the data processing as the central focus. Basically, it offers a set of tools and software that can be configured to implement forecasts systems attending the specific requirements of the operational centres. Thus, Delft-FEWS provides proper tools for database management; embed different models (including hydrological and hydrodynamic models) that can be configured to run automatically using configurable workflows and display data and results or warning messages. We present the work developed concerning the set-up of distinct models and data related to the forecasting process within the Delft-FEWS system. The study includes the development of a methodology that integrates all phases of a reservoir operation, using real time data and different types of hydraulic and hydrological models. Methodologies employed in this research were applied in 11 hydropower plants operated by Cemig that are located in 8 different watersheds. Different basins are considered in order to assess the performance of the system related to the distinct regions characteristics and the influence of their drainage area (it ranges from 543 to 71,487 km²). A sound analysis of the obtained results will allow deciding about the future application of the system to other reservoirs and basins. Behind their inherent tools the system is based on a models database that are constituted by: SOBEK-Sacramento, MGB-IPH and HEC-HMS hydrological models; 1D2D SOBEK and HEC-HAS hydrodynamic models; weather forecasting model with ensemble predictions; a method for evaluation of flow restrictions by fuzzy logic; and RTC TOOLS, a package for Real Time Control and modelling that was implemented in Delft - FEWS and can be linked to other hydraulic modelling packages.