Journal
WATER RESOURCES MANAGEMENT
Volume 28, Issue 1, Pages 131-147Publisher
SPRINGER
DOI: 10.1007/s11269-013-0476-z
Keywords
Optimal operation rules; Multi-objective optimization; Performance criteria of reservoir; Real-time operation; Inflow forecasting; NSGA-II; Pareto curve
Categories
Ask authors/readers for more resources
To extract optimal reservoir operation policies, it is important to consider different objectives simultaneously. In this study, by applying a meta-heuristic, multi-objective optimization approach, real-time optimal operation rules of Karoon4 dam are extracted as two-objectives by considering performance criteria of the reservoir as objective functions. The rules are extracted by relating water release to storage volume and inflow with simple linear equations for two states of real-time operation, which are called dependent on forecast state and independent of forecast state. In the dependent on forecast state, inflow volume is considered during the current period and in the independent of forecast state, inflow volume is considered during the period before the operation. In fact, by associating water release in each period to inflow during a past period, inflow forecasting is employed. Multi-objective optimization results of conflicting objectives of reliability and vulnerability in hydropower generation of Karoon4 are exhibited as a Pareto curve by employing the non-dominated sorting genetic algorithm II (NSGA-II). Each point on the Pareto curve represents an optimal operation policy. Actually, based on the priority and desired criterion of the reservoir operator, for the value of any criterion, the optimal value of another criterion and its optimal operation policy can be extracted by using a Pareto curve. Maximum reliabilities of Pareto curves in first and second states of real-time operation are 60.83 and 60.00 %, respectively, with corresponding minimum vulnerabilities of 8.52 and 9.08 %. Although the dependence of reservoir release in each period to inflow during the previous period (i.e., independent of forecast state) improves the values of objective functions compared with the dependent on forecast state, the difference is insignificant. Since the independent of forecast state of real-time operation does not depend on inflow forecasting, the small difference is negligible and so this state seems more efficient. Also, a comparison of results of long-term operation with real-time operation shows that bcomputed real-time operation rules are flexible and accurate.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available