Ultra-supercritical boiler closed-loop combustion optimal control method

Abstract

The invention discloses an ultra-supercritical boiler closed-loop combustion optimal control method; based on a combustion system dynamic model built by an untracked Kalman filtering least squares support vector machine, the boiler efficiency and the dynamic characteristics of NOx emission changed along with loads can be accurately reflected; and meanwhile, through introduction of updating mechanisms, the excellent self-adaption capacity and prediction capacity of the dynamic model under different working conditions are also guaranteed. When the load of a 1000 MW fired coal boiler is variableor stable, and the adjusting quantity, the controlled index and related parameters are all located in a reasonable range and stably changed, the boiler efficiency can be kept stable; and meanwhile, after SCR inlet conversion, the NOx concentration is obviously reduced after application.

Description

technical field [0001] The invention relates to the field of thermal automatic control, in particular to a closed-loop combustion optimization control method for an ultra-supercritical boiler. Background technique [0002] Adjustable parameters such as the air distribution mode of the boiler, the oxygen amount and the operation mode of the pulverization system have an important impact on the boiler efficiency and NOx emissions. At present, the DCS control system can generally only realize the control of the total coal volume and the total air volume according to the load demand. For the coal volume and air volume distribution of each layer of burners, the uniform distribution method is adopted, or the operating personnel can adjust it according to experience. Obviously, this mode of operation is relatively "extensive" and cannot adapt to changes in coal types. Therefore, it is necessary to further tap the potential of boiler energy saving and emission reduction through comb...

AI Technical Summary

Problems solved by technology

Obviously, this mode of operation is relatively "extensive" and cannot adapt to changes in coal types. Therefore, it is necessary to further tap the potential of boiler energy saving and emission reduction through combustion optimization.

[0003] At present, the boiler combustion optimization methods mainly include: combining the data-based nonlinear modeling method with the intelligent optimization algorithm, first establishing the boiler combustion characteristic model, and then optimizing the operating parameters with the optimal boiler efficiency and emission as performance indicators. This kind of method is suitable for the combustion optimization of steady-state working conditions, and lacks the adaptability to changing factors such as coal types; and then proposes dynamic modeling of boiler systems based on unscented Kalman filter least squares support vector machine, when the model accuracy does not meet the requirements When , the sample update strategy is designed and the online matrix inversion is avoided, thereby reducing the amount of online calculations. However, one of the biggest shortcomings of the above online algorithm is that the kernel parameter σ, which has an important impact on the model accuracy, can only be obtained through repeated trials. The combined method is determined offline and cannot be updated online

Images