Please use this identifier to cite or link to this item: https://scholarbank.nus.edu.sg/handle/10635/128686
Title: Assessing the effectiveness of a real-time control method for Marina Reservoir management
Authors: Goedbloed, A.
Galelli, S. 
Schwanenberg, D.
Keywords: Inflow predictions
Model predictive control
Multi-objective optimization
Multi-purpose reservoir operation
Issue Date: 2011
Citation: Goedbloed, A.,Galelli, S.,Schwanenberg, D. (2011). Assessing the effectiveness of a real-time control method for Marina Reservoir management. MODSIM 2011 - 19th International Congress on Modelling and Simulation - Sustaining Our Future: Understanding and Living with Uncertainty : 4036-4042. ScholarBank@NUS Repository.
Abstract: Marina Reservoir is one of the largest fresh water body in Singapore, recently constructed with the purpose of increasing the drinking water supply and Singapore's self-sufficiency. Besides this strategic role, the reservoir also serves for floods control and lifestyle attraction. The largest portion of the inflow volumes comes from five main uncontrolled catchments, which show fast-varying discharge variations due to the strong storms affecting this region. Moreover, highly paved areas cause rapid runoff processes, with a concentration time lower than one hour. Such hydrological context creates difficulties in satisfying the operational objectives and provides a challenging environment for the application of real-time control methods. With the purpose of evaluating the potential of real-time control for the optimal operation of Marina Reservoir, this work adopts Model Predictive Control (MPC), a form of deterministic control that employs the current state of the system, the future inflows trajectory furnished by a predictive model and a further model describing the internal dynamics of the controlled sub-system to determine an optimal control sequence over a finite prediction horizon. The rationale behind the choice of MPC is that this method is characterized by reduced computational requests and the capability of exploiting any exogenous information, as precipitation or evaporation measurements, that allow for an accurate prediction of the inflow events, which is crucial for an effective management of fast-varying hydrological systems. In the present application, Marina Reservoir catchment is described with a data-driven rainfall-runoff model (in the form of M5 model trees) that fully exploits the hydro-meteorological information available in real-time, thus enhancing the accuracy of the inflow prediction. The model is combined with a non-linear MPC scheme that optimizes the barrage operation according to the different operational objectives. Preliminary results show the effectiveness of the proposed control method, which outperforms the currently-used operating rules.
Source Title: MODSIM 2011 - 19th International Congress on Modelling and Simulation - Sustaining Our Future: Understanding and Living with Uncertainty
URI: http://scholarbank.nus.edu.sg/handle/10635/128686
ISBN: 9780987214317
Appears in Collections:Staff Publications

Show full item record
Files in This Item:
There are no files associated with this item.

Page view(s)

45
checked on Oct 14, 2021

Google ScholarTM

Check

Altmetric


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.