Development of a Fast Response Combustion Performance Monitoring, Prediction, and Optimization Tool for Power Plants

Combustion performance monitoring is a challenging task due to insufficient post combustion data and insights at critical furnace areas. Post combustion insights are valuable to reflect the plant’s efficiency, reliability, and used in boiler tuning programmes. Boiler tuning, which is scheduled after all the preventive maintenance suggested by boiler manufacturer, could not addressed the operational issues face by plant operators. A system-level digital twin incorporating both computational fluid dynamics (CFD) and machine learning modules is proposed in the current study. The proposed tool could act as a combustion monitoring system to diagnose, pinpoints boiler problems, and troubleshoots to reduce maintenance time and optimize operations. The system recommends operating parameters for different coal types and furnace conditions. The tool can be used as a guideline in daily operation monitoring/optimization and in risk assessments of new coals. The current paper discusses the general architecture of the proposed tool and some of the preliminary results based on the plant’s historical data.