Optimal daily operation of reservoirs subject to probabilistic flood constraints

This paper addresses the problem of determining the optimal daily operating policy of a river system subject to probabilistic flood constraints. This optimization problem is stochastic because streamflows are random. One of the difficulties is to take into account the streamflow persistence and, more specifically, the fact that there may be several consecutive days of high flow. This is important because floods usually occur during those periods. Unfortunately, the techniques for solving stochastic optimization problems cannot generally take into account long persistence. This paper presents a method that solves the problem in two steps. The first step determines flood control rule curves for the reservoirs that take into account streamflow persistence. Any reservoir operating policy that respects these rule curves satisfies the probabilistic flood constraints. The second step is used to determine the operating policy, among those that respect the flood control rule curves, that maximizes the generation of hydroelectric power. This is done without taking into consideration streamflow persistence.