Multi-objective modeling of boiler combustion based on feature fusion and Bayesian optimization

The physical field (temperature, gas concentration, etc.) inside the furnace is closely related to the boiler combustion optimization. A novel multi-objective prediction framework based on feature fusion is proposed to provide the basis for the online combustion optimization of coal-fired boilers. Firstly, the physical field information is obtained through the CFD, which presented a strong correlation between thermal efficiency and NOx generation. Then the eXtreme Gradient Boosting and Bayesian Optimization are used to construct the model according to the changes of the real-time physical field and operating conditions. The modeling results demonstrated that the prediction accuracy of thermal efficiency from the model with the fusion information can be improved by 1.49% compared with the model using the operational data. The prediction accuracy of thermal efficiency and NOx generation is improved by 2.57% and 0.13%, respectively, which indicated that the expression ability of the model improved by combing the typical real-time physical field information.

Automated summary

A novel multi-objective prediction framework based on feature fusion is proposed for online combustion optimization of coal-fired boilers. By combining real-time physical field information, the model demonstrates improved prediction accuracy for thermal efficiency and NOx generation.