Author(s): Taesoon Kim; Jun-Haeng Heo; Deg-Hyo Bae; Younghun Jung
Linked Author(s):
Keywords: Piecewise-linear operating rule; Single reservoir; NSGA-II
Abstract: Piecewise-linear operating rule for Soyanggang reservoir in Korea Peninsula is developed using multi-objective genetic algorithms (NSGA-II) and time series modelling. Implicit stochastic approach is applied and 100-year synthetic inflow is generated by autoregressive order one (AR-1) time series model. The simulation model is performed to evaluate the performance of the developed operating rule using historical average initial storage at the start of year and historical annual inflow. Regression analysis is not necessary to get the operating rule since each end point of piecewise-linear operating rule is calculated by optimization model directly, and multi-objectives which are to maximize sum of hydroelectric power production and to minimize water shortage index is considered to get a proper operating rule. Although the piecewise-linear operating rule is developed with only annual synthetic inflows which are larger than the sum of designated water supply, it can be used when annual inflow is not enough to supply water resources to downstream. Storage and release pattern also show good results comparing with historical operation.
Year: 2007