Real time optimal operation of reservoir-river system under flooding conditions

Flood control reservoirs are placed upstream of river reaches for the sake of flood control and for other additional purposes such as water supply during draft period, power generation, etc. The function of the flood control reservoir is to store that portion of flood waters to protect downstream areas that would otherwise be damaged. However, accumulation of the flooding water will result in increasing water surface elevations of the reservoir itself, as well as increasing the water surface elevations of the most upstream extremities causing a potential damage source for upstream locations. Starting flood operations late may make it possible to flood the down stream area or, in the most extreme case, cause a dam failure.; The objective of this research was to develop an efficient methodology and the computer software model for evaluating optimal reservoir releases policies under flooding conditions that minimize the total damages for both the upstream and downstream reaches of reservoirs. The problem is formulated as a discrete-time optimal control problem in which reservoir releases represent the control variables, and water surface elevations and discharges represent the state variables. Constraints imposed on the reservoir's water surface elevations and reservoirs releases to the downstream reaches into objective function using a penalty function method. The optimal control model consists of two primary interfaced components. Full EQuation (FEQ) routing model is to simulate the dynamics of the river-reservoir system and an optimization technique, Simulated annealing (SA) that updates deterministic operation policies and solves the augmented control problem.; Numerical experiments are conducted to evaluate parameter sensitivity and demonstrate the capabilities of the optimization technique in determining the near global optima. The new reservoir operation model was applied to a hypothetical river-reservoir system as well as to the real river-reservoir system of Lake Travis on the Lower Colorado River in Texas. The model application to Lake Travis revealed the usefulness of the model in improving a given operation policy to the near global optimal operation policies. Regardless the type of the objective function i.e. linear or non-linear the model located the near global optimal operation policies. The methodology and the operation model developed here are unique since they can be applied to any river-reservoir system, do not require simplification of nonlinear problem, and guarantee the determination of the near global optima.