作者: S. Galelli , A. Castelletti , A. Goedbloed
DOI: 10.1002/2015WR017595
关键词: Water quality 、 Cluster analysis 、 Reduction (complexity) 、 Water supply 、 Control theory 、 Environmental engineering 、 Environmental science 、 Process (computing) 、 Flood control 、 Model predictive control
摘要: This paper contributes a novel High-Performance Integrated Control framework to support the real-time operation of urban water supply storages affected by quality problems. We use 3-D, high-fidelity simulation model predict main dynamics and inform controller based on Model Predictive Control. The integration into control scheme is performed reduction process that identifies low-order, dynamic emulator running 4 orders magnitude faster. reduction, which relies semiautomatic procedural approach integrating time series clustering variable selection algorithms, generates compact physically meaningful can be coupled with controller. used design hourly Marina Reservoir, 3.2 Mm3 storm-water-fed reservoir located in center Singapore, operated for drinking flood control. Because its recent formation from former estuary, suffers high salinity levels, whose behavior modeled Delft3D-FLOW. Results show our reduces minimum levels nearly 40% cuts average annual deficit about 2 times active storage (about 4% total demand).