Adaptive Management of the Storage Function for a Large Reservoir Using Learned Fuzzy Models
2021
Tomas, Kozel | Milos, Stary
A single study investigates the possibility of using adaptive control of the reservoir Vranov Reservoir situated on the Dyje River. The control algorithm uses a fuzzy model that approximates the I/O relationships contained in the behaviour matrix of the target reservoir, constructed using the differential evolution optimisation method. Recurring predictions of water inflows into the reservoir are constructed using a fuzzy model, which is based on the idea of similarity of the course of a real series of average monthly flows during the year. After calibration of the control and predictive models, the whole control is tested for the period 2004–2018. The results obtained by the discussed models are compared with the results obtained by dispatching graphs. The results of adaptive control show that the method is very suitable for driving during long low-water periods. In periods with plentiful water, the results barely differ.
显示更多 [+] 显示较少 [-]