Dynamic modeling and optimization of a coal-fired utility boiler to forecast and minimize NOx and CO emissions simultaneously

Increasing penetration of renewable energy sources to the power grid has prompted new ramping scenarios to dispatchable thermal power plants to balance the variability caused by intermittent renewable supplies. With many thermal power plants designed to be base-loaded, ramping of the power output results in increased emission of pollutants. This study develops a dynamic data-driven model of a coal-fired utility boiler that estimates NOx and CO emissions simultaneously. Given a production schedule of a power plant, estimation of NOx and CO emissions for 3 h into the future is performed that can be further utilized in a dynamic optimization algorithm to minimize the emissions over a horizon. It is observed that a dynamic model always has a higher prediction accuracy than a static model, when training and forecasting of the models are concerned. Application of dynamic and steady-state optimization also results in reduced emissions as compared to historical plant emissions. (C) 2019 Elsevier Ltd. All rights reserved.