Model Predictive Control of Induced Draft Cooling Towers in a Large Scale Cooling Plant

We study the problem of designing a model predictive control strategy for multiple induced draft cooling towers. The common objective towers is to regulate outlet water temperature collection basin desired set point. Each tower can be operated in different modes (bypass, showering, ventilation) meet heat rejection requirement. To deal with interacting dynamics and logic, we propose based on mixed integer programming (MIP) which accounts changing weather conditions active load simultaneously determine best operational mode optimal fan speed each tower. Dwell time constraints are added avoid excessive switching between modes. Although proposed formulation account efficiencies towers, an adaptive penalty designed balance run among identical inherent over-actuated structure process exploited provide robustness against unavailability efficacy approach demonstrated experimentally validated models Simulation results show significant performance improvements energy savings over conventional heuristic solutions.