Developing optimal operating reservoir rule-curves in drought periods

Abstract

In this study a simulation-optimization model is developed for deriving operation rule-curves in drought ‎periods. To each reservoir, two rule-curves with adjustable monthly levels are introduced dividing the ‎reservoir capacity into three zones between the normal water level and minimum operation level. To each ‎zone of a reservoir and for each month of the year a hedging coefficient is introduced that determines the ‎release from the reservoir. Accordingly, an optimization problem is developed in which the objective is ‎the minimization of water demands deficits in drought and the decision variables are the rule-curves ‎levels and hedging coefficients. For optimization of the problem, a genetic algorithm equipped with a ‎self-adaptive constraint handling strategy is used. To evaluate the objective function and constraints ‎violations, the flexible and widely-used WEAP simulation model is exploited and coupled to the ‎optimization solver. The model is then applied to the Zohreh three-reservoir system in the southwest of ‎Iran and compared to the Standard Operation Policy (SOP). According to the sustainability indices for ‎the system operated in drought, the obtained operating rule-curves are found significantly superior to the ‎SOP. As a result of applying the rule-curves, the modified shortage index (MSI) and vulnerability ‎‎(extent) of the system are respectively improved by 22% and 28% compared to the SOP. Consequently, ‎the developed policy application resulted in longer periods of deficit (but less severe) as shown by ‎decrease in reliability (5%) and resilience (40%) indices. ‎