Based on the thermal network and the MATLartificial intelligence toolkit,
a combustion optimization hybrid modelling of a 300 MW coal-fired power
station boiler is carried out. The boiler is optimized for combustion, and
the weight co?efficient method is used to convert the multi-objective
optimization problem into a single-objective optimization problem. The
results show that the relative error average absolute value of the boiler
thermal efficiency and NOx emission mass concentration calibration samples
are 0.142% and 1.790%, the model has good accuracy and generalization
ability. The weight coefficient method can select the corresponding weight
coefficient according to the actual situation, with the boiler thermal
efficiency or NOx emission mass concentration as the optimization focus,
which has certain guiding significance for combustion optimization.
Research on Prediction of Vertical Water Wall Temperature of Power Station Boiler Based on Deep Learning and Expert Experience