002856549 001__ 2856549
002856549 003__ SzGeCERN
002856549 005__ 20230505171429.0
002856549 0247_ $$2DOI$$a10.1016/j.ifacol.2022.07.438
002856549 0248_ $$aoai:cds.cern.ch:2856549$$pcerncds:FULLTEXT$$pcerncds:CERN:FULLTEXT$$pcerncds:CERN
002856549 035__ $$9https://inspirehep.net/api/oai2d$$aoai:inspirehep.net:2640238$$d2023-04-19T13:14:45Z$$h2023-04-20T04:00:03Z$$mmarcxml
002856549 035__ $$9Inspire$$a2640238
002856549 041__ $$aeng
002856549 100__ $$aGhawash, Faiq$$uNorwegian U. Sci. Tech.
002856549 245__ $$9submitter$$aModel Predictive Control of Induced Draft Cooling Towers in a Large Scale Cooling Plant
002856549 260__ $$c2022
002856549 300__ $$a7 p
002856549 520__ $$9submitter$$aWe study the problem of designing a model predictive control strategy for multiple induced draft cooling towers. The common objective of the cooling towers is to regulate the outlet water temperature of the common collection basin to the desired set point. Each cooling tower can be operated in different modes (bypass, showering, ventilation) to meet the heat rejection requirement. To deal with the interacting dynamics and logic, we propose a model predictive control strategy based on mixed integer programming (MIP) which accounts for the changing weather conditions and active heat load to simultaneously determine the best operational mode and optimal fan speed for each cooling tower. Dwell time constraints are added to avoid excessive switching between different operational modes. Although the proposed formulation can account for different operational efficiencies of the cooling towers, an adaptive penalty is designed to balance the run time among identical cooling towers. The inherent over-actuated structure of the cooling process is exploited to provide robustness against the unavailability of an operational mode of a cooling tower. The efficacy of the proposed approach is demonstrated on experimentally validated models of induced draft cooling towers. Simulation results show significant performance improvements and energy savings over conventional heuristic solutions.
002856549 540__ $$aCC-BY-NC-ND-4.0$$uhttps://creativecommons.org/licenses/by-nc-nd/4.0/
002856549 542__ $$dThe Authors$$g2022
002856549 65017 $$2INSPIRE$$aEngineering
002856549 690C_ $$aARTICLE
002856549 690C_ $$aCERN
002856549 700__ $$aHovd, Morten$$uNorwegian U. Sci. Tech.
002856549 700__ $$aSchofield, Brad$$uCERN
002856549 773__ $$c161-167$$n7$$pIFAC-PapersOnLine$$v55$$y2022
002856549 8564_ $$82449237$$s564455$$uhttp://cds.cern.ch/record/2856549/files/1-s2.0-S2405896322008394-main.pdf$$yFulltext
002856549 960__ $$a13
002856549 980__ $$aARTICLE